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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
80237570-6a3a-4ebf-838c-a9b8672c7374 | first-arrival-picking-using-u-net-with-lovasz | 2104.02805 | null | https://arxiv.org/abs/2104.02805v1 | https://arxiv.org/pdf/2104.02805v1.pdf | First arrival picking using U-net with Lovasz loss and nearest point picking method | We proposed a robust segmentation and picking workflow to solve the first arrival picking problem for seismic signal processing. Unlike traditional classification algorithm, image segmentation method can utilize the location information by outputting a prediction map which has the same size of the input image. A parameter-free nearest point picking algorithm is proposed to further improve the accuracy of the first arrival picking. The algorithm is test on synthetic clean data, synthetic noisy data, synthetic picking-disconnected data and field data. It performs well on all of them and the picking deviation reaches as low as 4.8ms per receiver. The first arrival picking problem is formulated as the contour detection problem. Similar to \cite{wu2019semi}, we use U-net to perform the segmentation as it is proven to be state-of-the-art in many image segmentation tasks. Particularly, a Lovasz loss instead of the traditional cross-entropy loss is used to train the network for a better segmentation performance. Lovasz loss is a surrogate loss for Jaccard index or the so-called intersection-over-union (IoU) score, which is often one of the most used metrics for segmentation tasks. In the picking part, we use a novel nearest point picking (NPP) method to take the advantage of the coherence of the first arrival picking among adjacent receivers. Our model is tested and validated on both synthetic and field data with harmonic noises. The main contributions of this paper are as follows: 1. Used Lovasz loss to directly optimize the IoU for segmentation task. Improvement over the cross-entropy loss with regard to the segmentation accuracy is verified by the test result. 2. Proposed a nearest point picking post processing method to overcome any defects left by the segmentation output. 3. Conducted noise analysis and verified the model with both noisy synthetic and field datasets. | ['Hien Van Nguyen', 'Jiefu Chen', 'Xuqing Wu', 'Wenyi Hu', 'Pengyu Yuan'] | 2021-04-06 | null | null | null | null | ['contour-detection'] | ['computer-vision'] | [ 1.77676409e-01 -3.21267657e-02 4.66065288e-01 -2.76769698e-01
-1.23136234e+00 -4.69800264e-01 2.78832883e-01 2.74094939e-01
-7.22652853e-01 5.15479863e-01 -1.17507584e-01 -4.04913649e-02
-5.33534884e-01 -8.36901784e-01 -7.76874304e-01 -9.80944157e-01
-4.25399601e-01 3.40445846e-01 5.74357510e-01 -2.30388433e-01
5.63381732e-01 5.30671418e-01 -1.49428773e+00 2.07174391e-01
1.02628922e+00 1.20149887e+00 4.28846121e-01 4.49635148e-01
-1.61795206e-02 1.65390730e-01 -5.91020107e-01 6.48168474e-03
5.55550337e-01 -2.74264276e-01 -6.81875348e-01 -3.27442616e-01
8.45468044e-02 -7.67364800e-02 2.10810322e-02 1.09451580e+00
8.72222900e-01 3.27490151e-01 7.86958337e-01 -1.05096292e+00
-5.33882640e-02 7.11765766e-01 -8.48851442e-01 3.18941772e-01
1.93268478e-01 1.19855829e-01 7.78283358e-01 -7.38312483e-01
4.46500540e-01 9.35996115e-01 1.16021180e+00 -6.14679903e-02
-9.35057282e-01 -6.68552518e-01 -2.37628564e-01 2.85196632e-01
-1.42635298e+00 8.01200792e-02 9.43688989e-01 -5.56419730e-01
5.80972552e-01 3.62709790e-01 4.95085478e-01 5.74863076e-01
1.83294848e-01 6.49549067e-01 1.11819208e+00 -4.87913072e-01
2.61656940e-01 -2.87250012e-01 3.43522549e-01 2.87060410e-01
4.40179482e-02 -9.86860842e-02 -2.83962935e-01 -4.53041606e-02
5.25486112e-01 -3.48618209e-01 -3.87289762e-01 2.24885643e-01
-1.11532044e+00 7.63620555e-01 5.68064392e-01 4.84902948e-01
-3.71406645e-01 3.34715769e-02 4.92073506e-01 2.39792801e-02
5.16084552e-01 5.06078184e-01 -2.37782419e-01 -5.22500016e-02
-1.46353269e+00 3.75307649e-01 4.65846211e-01 4.95276451e-01
7.95119524e-01 1.06458366e-01 -3.59850675e-01 8.33637953e-01
2.78052032e-01 7.14338779e-01 5.22683084e-01 -8.88770044e-01
5.34355342e-01 3.75200510e-01 -2.10514620e-01 -1.29797220e+00
-6.62446260e-01 -6.53632104e-01 -9.19104159e-01 3.44328552e-01
4.82757747e-01 -2.85068423e-01 -9.98068094e-01 1.44537485e+00
1.26079574e-01 5.62910736e-01 -5.42467237e-02 9.68775749e-01
6.87006295e-01 7.42795229e-01 -2.69289166e-01 -7.90322870e-02
1.22510195e+00 -5.24658680e-01 -6.93337083e-01 -3.20316013e-03
3.41790348e-01 -8.76865387e-01 8.07459831e-01 5.50257206e-01
-9.19179380e-01 -5.61334312e-01 -1.10127926e+00 4.29833978e-01
-2.66522974e-01 -3.60153727e-02 2.31431141e-01 5.65672874e-01
-8.92366052e-01 9.65712905e-01 -9.31410432e-01 -2.77657248e-03
3.99873137e-01 1.96029559e-01 -1.48111060e-01 2.62592375e-01
-1.28558087e+00 4.62184012e-01 6.18079364e-01 4.07582015e-01
-5.25086880e-01 -8.13664794e-01 -6.03053391e-01 -3.40715870e-02
3.51051241e-01 -1.28193185e-01 7.87476003e-01 -6.96471512e-01
-1.21944964e+00 5.14789999e-01 1.92029372e-01 -6.76970065e-01
8.71312559e-01 -2.12747023e-01 -1.88724950e-01 3.40202957e-01
3.71678114e-01 3.17505509e-01 6.12525642e-01 -1.43106425e+00
-5.16518116e-01 -1.66314632e-01 -4.92382348e-01 -2.00668797e-02
2.06173927e-01 -2.32239574e-01 -3.67331266e-01 -9.47950125e-01
7.27464020e-01 -7.01932311e-01 -1.51939318e-01 -5.73528588e-01
-6.51518166e-01 1.64731756e-01 8.12628329e-01 -1.07613027e+00
1.25365686e+00 -2.25349379e+00 -4.02929842e-01 8.87642205e-01
-7.45051131e-02 2.59001762e-01 1.97540671e-01 3.03387165e-01
-1.67454928e-01 2.86013573e-01 -9.94222224e-01 -1.56951413e-01
-2.57814854e-01 -9.74824354e-02 -1.40238479e-01 5.97924411e-01
7.16760904e-02 3.63212943e-01 -6.98279381e-01 -5.47421098e-01
1.97921097e-01 3.02309632e-01 -4.50315475e-01 2.19170190e-02
1.58806771e-01 4.89862829e-01 -3.03297848e-01 4.50620651e-01
1.09937811e+00 3.92195046e-01 -5.29021323e-01 -5.21445930e-01
-2.36036167e-01 -3.76337796e-01 -1.73820531e+00 1.53373361e+00
-1.44613564e-01 5.15560567e-01 1.45135671e-01 -1.07787824e+00
1.25768566e+00 3.54512066e-01 7.79867947e-01 -6.56898081e-01
-7.48484284e-02 5.23454309e-01 -5.29717319e-02 -7.35996544e-01
3.10421348e-01 -3.47906835e-02 6.50464669e-02 7.62930363e-02
-1.64231598e-01 -2.75149763e-01 1.30921543e-01 -6.25835508e-02
9.34644461e-01 2.23329201e-01 -9.06453729e-02 -5.84350169e-01
6.40535951e-01 1.85310110e-01 7.19594181e-01 9.92514133e-01
-1.10986635e-01 1.28249943e+00 4.46086079e-01 -1.45293936e-01
-8.40240657e-01 -9.71785367e-01 -3.01829666e-01 5.79667449e-01
3.28448743e-01 1.52490526e-01 -1.05305958e+00 -3.81671309e-01
-2.09143773e-01 7.32569277e-01 -3.23067278e-01 1.39404714e-01
-8.66114259e-01 -1.10903192e+00 9.51987267e-01 3.29193264e-01
8.41858745e-01 -1.09877205e+00 -6.85890377e-01 2.75322944e-01
-4.16233957e-01 -8.26266885e-01 -2.61393160e-01 3.58664066e-01
-8.31130385e-01 -1.01823688e+00 -8.01223338e-01 -6.01354957e-01
3.51140708e-01 -1.88181430e-01 7.09852397e-01 2.84958966e-02
-2.42264062e-01 7.22285584e-02 -6.16255462e-01 -3.08181852e-01
-1.90513983e-01 -9.04802233e-03 -2.89187223e-01 2.54376233e-01
-7.66120106e-02 -6.03603125e-01 -7.55233169e-01 2.50196129e-01
-1.06579304e+00 -3.32170755e-01 4.34922427e-01 9.13634777e-01
6.78318083e-01 3.63571256e-01 7.92225420e-01 -6.62752926e-01
7.59399414e-01 -4.12088603e-01 -5.78592300e-01 4.54988331e-02
-4.86195952e-01 4.24994119e-02 5.63147902e-01 -1.57583922e-01
-1.02860034e+00 -1.63146388e-02 -4.75825489e-01 -1.37540221e-01
-2.29107618e-01 6.47291839e-01 -2.95500569e-02 -3.01859882e-02
5.60510039e-01 1.06848352e-01 -5.09899966e-02 -5.87218165e-01
-5.22214584e-02 6.98218763e-01 7.37051308e-01 -5.44078708e-01
6.43174887e-01 5.85037887e-01 2.11039111e-01 -1.02953959e+00
-4.91709143e-01 -7.14197338e-01 -6.51330173e-01 -3.63299429e-01
1.11521792e+00 -3.81261140e-01 -6.74854279e-01 7.57906675e-01
-1.20011532e+00 -9.67109129e-02 -2.16529965e-01 5.67760587e-01
-4.49720144e-01 5.86319864e-01 -4.09493446e-01 -1.01669967e+00
-5.12628913e-01 -1.42716622e+00 9.66985822e-01 3.47855479e-01
2.54510313e-01 -8.33739758e-01 -6.22820668e-02 2.66548395e-01
3.54712486e-01 8.04686189e-01 5.58193982e-01 -9.09945130e-01
-3.19458783e-01 -3.48787576e-01 -1.39121115e-01 4.76902246e-01
-2.93459415e-01 6.25806749e-02 -1.10724998e+00 -7.30290264e-03
3.00968558e-01 1.51002586e-01 1.10394883e+00 8.02519500e-01
1.16265690e+00 4.83228005e-02 -1.67019546e-01 7.14674652e-01
1.65215075e+00 5.60298324e-01 8.98857474e-01 5.54345310e-01
5.48127234e-01 5.82630634e-01 7.20076799e-01 3.35713595e-01
-9.22068432e-02 5.98604202e-01 3.97619218e-01 -2.10508063e-01
-1.15429252e-01 3.06914169e-02 5.77813946e-02 7.85884976e-01
-7.00722188e-02 -1.32028610e-01 -1.31543183e+00 7.16850221e-01
-1.83084822e+00 -9.98555422e-01 -5.62045395e-01 2.30421543e+00
5.92734277e-01 3.64918023e-01 -1.79291800e-01 7.02046931e-01
7.80563712e-01 6.37231693e-02 -1.55245066e-01 -2.34892890e-01
-3.75437707e-01 3.45910817e-01 8.80338788e-01 7.40823507e-01
-1.38090503e+00 5.03391922e-01 4.64846611e+00 1.39432311e+00
-1.16636395e+00 1.19505107e-01 7.83283651e-01 5.27275860e-01
-7.46598542e-02 -4.36919667e-02 -4.87616062e-01 7.53909469e-01
5.13643861e-01 3.34489763e-01 5.44516323e-03 4.48217481e-01
5.88037312e-01 -4.15058166e-01 -5.82305312e-01 9.34221447e-01
-1.15392506e-01 -1.13851953e+00 -3.59872013e-01 -2.54124850e-01
6.91715181e-01 7.44593935e-03 -2.09944606e-01 -1.32778242e-01
-3.19630593e-01 -9.76495862e-01 8.72154832e-01 8.66069376e-01
4.65492368e-01 -7.65691161e-01 1.26716220e+00 4.03499573e-01
-1.30352235e+00 -3.37071121e-02 -7.42805600e-02 1.02006584e-01
4.75480109e-01 9.87753212e-01 -7.75937736e-01 9.98041272e-01
8.99156094e-01 3.42440248e-01 -4.98092413e-01 1.65185678e+00
7.71185309e-02 9.55712497e-01 -7.42736578e-01 2.57524699e-01
5.53869486e-01 -5.12115002e-01 9.19061959e-01 1.47087979e+00
4.71115887e-01 -9.82597768e-02 2.86303252e-01 7.95726895e-01
3.74762088e-01 2.44569421e-01 -4.54155177e-01 6.92308426e-01
5.65366864e-01 9.54468608e-01 -1.26586258e+00 -5.76834902e-02
1.83684722e-01 5.25597632e-01 -5.00630856e-01 3.85594100e-01
-7.35502660e-01 -7.85713255e-01 -6.82396591e-02 2.90822715e-01
2.52358735e-01 -1.07500099e-01 -8.20624769e-01 -4.67720240e-01
2.08074421e-01 -4.94485855e-01 2.53826946e-01 -5.01679480e-01
-1.21721327e+00 7.08305299e-01 3.86782974e-01 -1.43377399e+00
7.78313056e-02 -3.05167973e-01 -9.39049304e-01 9.87367690e-01
-1.37099004e+00 -8.60307038e-01 -4.44468409e-01 2.18988344e-01
5.38320482e-01 -4.13967893e-02 2.51374215e-01 6.12418115e-01
-4.60168123e-01 4.46778923e-01 3.57816219e-01 2.87680715e-01
5.21287441e-01 -1.23796022e+00 1.09181479e-01 1.05741143e+00
-1.58578575e-01 1.42643064e-01 8.56674790e-01 -8.73187244e-01
-6.29652143e-01 -7.98355818e-01 4.16978896e-01 1.26148820e-01
4.87724751e-01 -1.19507149e-01 -1.18306148e+00 1.26995027e-01
1.09763131e-01 -1.89832866e-01 1.28139749e-01 -6.49365306e-01
2.95613825e-01 -3.60266536e-01 -1.37164402e+00 2.60542125e-01
5.51073670e-01 -3.80910710e-02 -5.56682706e-01 2.32769936e-01
6.66285396e-01 -3.47082734e-01 -8.23836923e-01 7.33501792e-01
3.09627086e-01 -1.12402976e+00 1.06729305e+00 1.41143113e-01
3.94573718e-01 -5.33715248e-01 -9.89532769e-02 -1.14073813e+00
6.10476248e-02 -5.43003201e-01 5.69849133e-01 1.48907888e+00
6.05210364e-01 -6.40772462e-01 6.83857679e-01 2.36641139e-01
-4.50391710e-01 -7.81753182e-01 -1.29653406e+00 -8.73816729e-01
3.31835188e-02 -8.44884098e-01 4.62425560e-01 6.46241367e-01
-4.09054279e-01 -2.70850360e-01 -1.00818664e-01 3.36807072e-01
9.64873672e-01 -2.71598339e-01 3.60150307e-01 -1.22468150e+00
-2.00182736e-01 -5.58418751e-01 -4.73054886e-01 -6.54079199e-01
-2.73712128e-01 -9.05103087e-01 3.82391840e-01 -1.55746841e+00
-3.55669469e-01 -7.65107632e-01 -2.24134639e-01 1.46288812e-01
-1.65332422e-01 2.65780956e-01 2.50391454e-01 2.42108777e-01
2.16129422e-02 2.86908060e-01 1.17041326e+00 -1.22717731e-01
-3.77757221e-01 3.73017758e-01 -1.19354241e-01 8.24097157e-01
9.28646743e-01 -6.46904051e-01 -1.87627718e-01 -2.47876719e-01
1.01866603e-01 1.81189075e-01 3.90975773e-01 -1.44242144e+00
3.24367374e-01 1.58023924e-01 -1.07504027e-02 -9.59252417e-01
1.17906921e-01 -9.20668960e-01 1.29963890e-01 4.59557891e-01
-1.65207818e-01 6.68126270e-02 1.07716121e-01 4.12568957e-01
-4.73020405e-01 -7.54900277e-01 8.33854377e-01 -3.30281840e-03
-7.30866790e-01 -2.03101724e-01 -5.05977422e-02 1.58054039e-01
1.06030428e+00 -5.47273934e-01 -2.46857312e-02 -2.52281696e-01
-7.21738338e-01 2.32776970e-01 -3.96498963e-02 -8.78352076e-02
6.73265040e-01 -9.62546408e-01 -8.81425500e-01 2.15454593e-01
-2.95095742e-01 2.73077935e-01 5.46951830e-01 1.17050123e+00
-1.13648891e+00 -1.69582888e-01 -8.42264108e-03 -8.48405182e-01
-9.15104032e-01 1.33538187e-01 5.92484176e-01 -3.38705927e-01
-7.92727888e-01 8.13153267e-01 -1.66753322e-01 -2.97392845e-01
2.89777070e-01 -5.28055131e-01 -3.46397281e-01 2.55491674e-01
2.42846802e-01 8.44150960e-01 2.89041191e-01 -6.37438536e-01
-3.82046014e-01 9.13240433e-01 3.95472109e-01 -4.06143695e-01
1.41673887e+00 -6.73449859e-02 -1.27968252e-01 4.47959870e-01
1.18365443e+00 -5.20373173e-02 -1.18121243e+00 1.53581366e-01
2.71094054e-01 -3.51756960e-01 1.49259761e-01 -8.03950310e-01
-1.22837830e+00 8.70623946e-01 1.03495979e+00 3.04439694e-01
1.20240879e+00 -4.23238516e-01 9.71368730e-01 2.93505825e-02
1.12784207e-01 -1.51068509e+00 -2.67117023e-01 4.95659173e-01
8.71955097e-01 -1.24837911e+00 -9.79290809e-03 -3.22658330e-01
-6.69752479e-01 1.15397954e+00 1.54543206e-01 -3.56172770e-01
1.02148950e+00 5.04978478e-01 1.56171948e-01 -2.19448984e-01
2.85166651e-01 -1.14549741e-01 2.23497748e-01 5.65443814e-01
2.57597297e-01 8.74856189e-02 -6.45494163e-01 7.29535639e-01
-2.41361097e-01 -1.58070788e-01 4.39935207e-01 8.34799945e-01
-5.86509228e-01 -7.06439197e-01 -9.08832133e-01 5.30530870e-01
-6.58521354e-01 3.53825674e-03 2.66094744e-01 6.23867273e-01
5.33150256e-01 1.08955622e+00 -1.49658233e-01 -4.75047350e-01
4.46234792e-01 -2.87725292e-02 -1.40571311e-01 -1.16273873e-01
-9.34073210e-01 3.26291978e-01 -2.29241207e-01 -4.45813596e-01
-4.60121036e-01 -6.99287057e-01 -1.58952963e+00 1.92133725e-01
-4.52536106e-01 3.11785102e-01 8.32919836e-01 1.00470626e+00
-1.20610736e-01 6.81748927e-01 5.72761059e-01 -9.35117841e-01
-4.71305966e-01 -9.58096206e-01 -7.56346107e-01 6.24957800e-01
3.15058440e-01 -6.02258861e-01 -5.46699107e-01 5.64128980e-02] | [7.209409236907959, 2.106745958328247] |
b78a1573-7d97-402b-9357-0bc568388c45 | a-model-oriented-approach-for-lifting | 2208.03095 | null | https://arxiv.org/abs/2208.03095v1 | https://arxiv.org/pdf/2208.03095v1.pdf | A Model-Oriented Approach for Lifting Symmetries in Answer Set Programming | When solving combinatorial problems, pruning symmetric solution candidates from the search space is essential. Most of the existing approaches are instance-specific and focus on the automatic computation of Symmetry Breaking Constraints (SBCs) for each given problem instance. However, the application of such approaches to large-scale instances or advanced problem encodings might be problematic since the computed SBCs are propositional and, therefore, can neither be meaningfully interpreted nor transferred to other instances. As a result, a time-consuming recomputation of SBCs must be done before every invocation of a solver. To overcome these limitations, we introduce a new model-oriented approach for Answer Set Programming that lifts the SBCs of small problem instances into a set of interpretable first-order constraints using a form of machine learning called Inductive Logic Programming. After targeting simple combinatorial problems, we aim to extend our method to be applied also for advanced decision and optimization problems. | ['Alice Tarzariol'] | 2022-08-05 | null | null | null | null | ['inductive-logic-programming'] | ['methodology'] | [ 6.55334234e-01 5.08775592e-01 -3.43558550e-01 -3.38816643e-01
-5.82614660e-01 -6.23076618e-01 4.70386803e-01 7.60748327e-01
-2.53695231e-02 1.04838717e+00 -4.39462483e-01 -7.69454539e-01
-6.45967364e-01 -1.34070659e+00 -5.26682258e-01 -4.03486997e-01
-2.13624999e-01 1.08494914e+00 4.72369403e-01 -1.88392058e-01
4.21224922e-01 6.24718785e-01 -1.74333096e+00 7.79260874e-01
9.74634230e-01 1.09977293e+00 -1.90013602e-01 2.76868105e-01
-5.09325206e-01 7.03354478e-01 -5.07659495e-01 -2.10244343e-01
3.90469939e-01 -4.98899668e-01 -1.11953855e+00 1.52285332e-02
6.97656870e-02 1.35597482e-01 5.82234085e-01 1.17572141e+00
-2.50984967e-01 8.62327889e-02 4.80094582e-01 -1.53661311e+00
1.97578490e-01 5.96708953e-01 -8.37338716e-02 -1.50411278e-01
5.59408605e-01 -2.05201522e-01 1.20809925e+00 -2.04302087e-01
7.77448058e-01 1.09925961e+00 2.80178726e-01 3.77012342e-01
-1.61919200e+00 -2.66162902e-01 4.37241375e-01 7.28228867e-01
-1.18167043e+00 -9.57788676e-02 8.62984061e-01 -2.79976487e-01
1.02618122e+00 9.84893978e-01 8.83143246e-01 4.97511894e-01
-2.24933267e-01 5.44359565e-01 1.36493397e+00 -7.01427400e-01
8.19008708e-01 2.08067611e-01 3.14052373e-01 5.77124894e-01
5.28233171e-01 -1.11768581e-01 -2.29860932e-01 -3.14316154e-01
2.63306558e-01 -4.37081307e-01 -2.12327167e-01 -6.07497990e-01
-8.48121166e-01 1.05924726e+00 1.10599753e-02 3.33081335e-01
-1.29163757e-01 9.62638482e-03 4.76365119e-01 4.07621711e-01
1.94838405e-01 8.45467925e-01 -5.50963104e-01 -2.34029908e-02
-1.11103284e+00 4.90900010e-01 1.24040163e+00 8.22345555e-01
8.34516764e-01 -6.53292537e-01 -4.31337133e-02 1.90595537e-01
-3.27688232e-02 -3.90348323e-02 -3.64725031e-02 -7.69079447e-01
5.59011400e-01 1.04173648e+00 2.14136541e-01 -1.25794101e+00
-4.58115995e-01 -2.48716563e-01 -3.30750495e-01 3.56262743e-01
5.82133412e-01 9.72537994e-02 -4.32623565e-01 1.48733044e+00
4.92768079e-01 -3.23293246e-02 1.63272750e-02 7.56409049e-01
1.87375709e-01 7.70356059e-01 -2.18255103e-01 -6.74913287e-01
1.01119101e+00 -8.13544571e-01 -4.45582956e-01 -1.50029361e-01
9.41364765e-01 -3.32665056e-01 6.12942398e-01 9.71805811e-01
-1.27023566e+00 1.32567078e-01 -1.09461844e+00 1.93030592e-02
-3.57505381e-01 -6.92158788e-02 9.74520326e-01 4.30227399e-01
-6.52848303e-01 5.25146425e-01 -6.14753842e-01 7.83478916e-02
2.01597571e-01 8.32623541e-01 -2.79868454e-01 -2.80156672e-01
-8.56852174e-01 1.06265402e+00 7.57036030e-01 3.83284599e-01
-2.67720103e-01 -5.27493596e-01 -7.42564321e-01 3.49034518e-01
1.13494396e+00 -3.44953448e-01 1.11768496e+00 -1.14191270e+00
-1.42420030e+00 8.52441311e-01 -3.19538563e-01 -4.45867419e-01
5.54520726e-01 3.78798187e-01 -6.79510608e-02 1.29958436e-01
-2.47999296e-01 4.69539489e-04 5.90955734e-01 -1.28437078e+00
-5.52912772e-01 -3.89154613e-01 7.37380624e-01 -1.63371354e-01
1.29201608e-02 -4.61229123e-02 -1.54096588e-01 -1.14700049e-01
5.70679009e-01 -8.48029315e-01 -5.00589252e-01 -2.04205260e-01
-5.46647966e-01 -1.82789966e-01 4.20818508e-01 -3.38323027e-01
1.33881807e+00 -1.61650181e+00 9.22713578e-01 8.01907063e-01
-2.07679328e-02 2.96543390e-01 9.96459499e-02 4.41759437e-01
-2.99117327e-01 2.29254633e-01 -3.62954915e-01 7.10481033e-02
2.59199679e-01 5.67111194e-01 -2.83467919e-01 8.69806334e-02
2.70426989e-01 7.21521795e-01 -8.04593384e-01 -5.01221359e-01
3.51305187e-01 -4.29434448e-01 -1.08248794e+00 -9.35333371e-02
-1.05988538e+00 3.24114472e-01 -5.61611354e-01 3.65332723e-01
8.69977176e-01 3.62968929e-02 7.36837447e-01 8.70646015e-02
-1.71352178e-01 5.43739676e-01 -1.65751517e+00 1.35085034e+00
-7.22572982e-01 1.48675025e-01 1.35905832e-01 -1.54589641e+00
6.95424080e-01 -1.01791047e-01 4.68000740e-01 -7.16865182e-01
1.13540463e-01 4.84902591e-01 2.38155834e-02 -4.59310710e-01
1.28275394e-01 -3.56899291e-01 -1.58391163e-01 4.65079963e-01
-3.92056584e-01 -2.96328843e-01 7.00304449e-01 -1.74271092e-01
1.06211162e+00 5.47567867e-02 3.79785866e-01 -2.64077723e-01
1.17231047e+00 3.82683456e-01 9.10025060e-01 4.11714464e-01
6.85809493e-01 1.47966325e-01 1.28996837e+00 -8.40250611e-01
-7.46162772e-01 -5.51658869e-01 -1.41879305e-01 6.02476001e-01
8.25606510e-02 -8.09844196e-01 -7.48802662e-01 -8.08789313e-01
-1.35469958e-01 8.92627656e-01 -3.61286342e-01 -1.28917009e-01
-7.54178345e-01 -6.64413154e-01 2.15595812e-02 1.38081893e-01
4.72796783e-02 -8.63264084e-01 -9.22305226e-01 3.21869880e-01
-2.44743809e-01 -9.69691396e-01 4.68586206e-01 4.83355314e-01
-1.07637715e+00 -1.47572029e+00 9.99939516e-02 -3.30444872e-01
9.89567399e-01 -3.48587155e-01 9.65544105e-01 4.76517439e-01
-2.50913650e-01 -2.53742248e-01 -4.47013140e-01 -2.94727892e-01
-4.85985368e-01 1.68618694e-01 -2.88042635e-01 6.28688559e-02
1.63379550e-01 -4.83002782e-01 2.36679435e-01 2.47890592e-01
-1.03092492e+00 2.45815530e-01 3.20487976e-01 7.87490368e-01
7.77202368e-01 5.75544000e-01 2.64417738e-01 -1.13176179e+00
4.27137882e-01 -1.89191565e-01 -1.49344766e+00 7.03564703e-01
-4.12008852e-01 6.24511123e-01 9.56833363e-01 -2.98446834e-01
-9.04340982e-01 2.25412786e-01 1.06288493e-01 1.34134576e-01
-1.06580675e-01 9.38092947e-01 -3.15890729e-01 -1.29190460e-01
3.89956027e-01 1.64244715e-02 -4.29185033e-01 -2.68992633e-01
7.75822699e-02 2.87143946e-01 7.73156658e-02 -1.04140556e+00
7.75797069e-01 2.59796321e-01 7.03611553e-01 -3.63629669e-01
-7.01075375e-01 -1.91435009e-01 -4.75068480e-01 7.95606822e-02
4.06539351e-01 -8.39060098e-02 -8.46488953e-01 -1.56277612e-01
-1.25786626e+00 -3.56705457e-01 -5.38932323e-01 1.94669887e-01
-7.36028075e-01 3.59652400e-01 -2.02448703e-02 -1.02530563e+00
2.41489798e-01 -1.36860776e+00 8.06014180e-01 -7.67205730e-02
-4.50658590e-01 -7.71229684e-01 4.66912314e-02 4.71804827e-01
-8.01995397e-03 2.79546559e-01 1.65734184e+00 -7.77216733e-01
-7.81588674e-01 -5.15268922e-01 -1.31083176e-01 6.81963488e-02
-2.41251945e-01 -2.73869559e-02 -5.81901431e-01 2.88511626e-02
1.50762096e-01 -1.65302232e-01 3.14343333e-01 5.43783009e-02
1.46168888e+00 -5.97247422e-01 -3.34176362e-01 3.29426318e-01
1.49344921e+00 8.53303820e-03 6.58380330e-01 4.84590143e-01
-1.62298903e-01 7.87601411e-01 8.22600424e-01 4.02746856e-01
6.17321767e-02 9.66166973e-01 6.19698226e-01 2.70374924e-01
4.91831750e-01 5.56110144e-02 -1.13925755e-01 1.95915967e-01
-4.64429200e-01 -1.23544239e-01 -8.46872270e-01 3.42457026e-01
-1.96159911e+00 -1.09358680e+00 -4.80318815e-01 2.47461605e+00
9.25310493e-01 3.20550710e-01 1.21012619e-02 7.62594461e-01
5.08065104e-01 -3.22383881e-01 -8.97629783e-02 -1.06071877e+00
2.41263494e-01 5.96178114e-01 2.61337072e-01 8.34503353e-01
-7.48441696e-01 6.91721261e-01 5.67183781e+00 5.65014958e-01
-1.03651392e+00 -1.35824993e-01 2.13015676e-01 -2.17136905e-01
-4.47544008e-01 4.80694950e-01 -6.14327073e-01 1.17349684e-01
7.25797176e-01 -1.25177741e-01 7.83081234e-01 9.32117403e-01
5.60960034e-04 -5.13452709e-01 -1.53377759e+00 6.15978360e-01
-2.18038201e-01 -1.37402272e+00 -1.17245466e-01 7.05110952e-02
6.45126760e-01 -8.35515559e-01 -2.53829449e-01 2.73223430e-01
-2.71777928e-01 -1.09904480e+00 6.43001378e-01 1.91938117e-01
2.99965143e-01 -9.07560945e-01 7.65086293e-01 6.22117102e-01
-8.04521143e-01 -5.56667805e-01 -2.84062952e-01 -5.42259216e-01
1.06093086e-01 7.65696824e-01 -7.55583704e-01 7.85149753e-01
2.76676565e-01 1.59601364e-02 -4.60586220e-01 1.15137732e+00
-4.86111253e-01 1.40105158e-01 -6.86251819e-01 -1.71795070e-01
1.62432596e-01 -3.64338487e-01 3.94437790e-01 8.24005306e-01
9.30623263e-02 3.21592987e-01 -2.47276034e-02 1.05115271e+00
4.40234303e-01 8.84311125e-02 -3.71794373e-01 -1.79392360e-02
5.03159463e-02 1.19302619e+00 -1.00832844e+00 -2.49446958e-01
-2.68027008e-01 6.26180828e-01 4.35541123e-01 1.41035765e-01
-7.98240662e-01 -2.46404558e-01 2.85203010e-01 1.74595088e-01
3.07735711e-01 -2.07249179e-01 -5.79338670e-01 -1.22515726e+00
6.10326290e-01 -1.11486781e+00 4.06466156e-01 -3.71814400e-01
-6.08108521e-01 2.75077879e-01 3.93414497e-01 -8.11779439e-01
-5.37994027e-01 -8.05628061e-01 -6.34908617e-01 5.70666790e-01
-1.45652723e+00 -8.17627728e-01 -3.42236347e-02 3.53659242e-01
2.87534475e-01 3.58859688e-01 1.14177549e+00 1.05672643e-01
-4.57871646e-01 1.74426988e-01 -3.68196726e-01 -4.27785873e-01
-9.59748775e-02 -1.20023656e+00 -3.22597504e-01 8.38518381e-01
7.00505227e-02 5.26113808e-01 7.86921561e-01 -2.89723307e-01
-1.65200675e+00 -5.62155366e-01 1.23841166e+00 -1.11878201e-01
6.03918731e-01 -4.12084967e-01 -6.50505722e-01 6.73437297e-01
-1.54744878e-01 -7.68493041e-02 5.01223803e-01 5.00067890e-01
-1.55075595e-01 -2.39185765e-01 -1.11704671e+00 5.88571489e-01
9.45849538e-01 -1.87825575e-01 -6.98032022e-01 5.43394566e-01
1.30550981e-01 -3.87777090e-01 -7.14660466e-01 5.43894768e-01
1.94764182e-01 -9.99430776e-01 7.53502548e-01 -1.06368375e+00
4.22722787e-01 -4.85813707e-01 -1.09934933e-01 -9.41294611e-01
1.18299775e-01 -4.78580981e-01 -2.44201720e-01 7.72663057e-01
5.83895087e-01 -5.10686278e-01 7.77619839e-01 1.09261715e+00
-9.25993696e-02 -9.93416190e-01 -1.16511714e+00 -7.55305767e-01
-7.41278604e-02 -6.46273434e-01 7.03614414e-01 8.73698056e-01
6.34301662e-01 4.40415516e-02 -2.22943397e-03 2.33017758e-01
3.51119250e-01 9.10238504e-01 6.18025839e-01 -1.48330939e+00
-6.48925602e-01 -5.36381245e-01 -5.01841068e-01 -3.75122100e-01
3.57813179e-01 -1.06602681e+00 -8.44202042e-02 -1.42443311e+00
-1.18469760e-01 -8.89356196e-01 1.43560216e-01 7.18045294e-01
4.76767093e-01 -8.50841030e-03 2.88069308e-01 -3.83298367e-01
-6.73869431e-01 2.54055206e-02 1.13799548e+00 -2.46067151e-01
-1.84799463e-01 3.22397262e-01 -3.85688156e-01 8.22670639e-01
6.44786358e-01 -5.49992919e-01 -4.95440990e-01 -1.21054195e-01
1.03017437e+00 3.87703478e-01 3.54722738e-01 -8.47395957e-01
2.70254642e-01 -8.47689271e-01 -2.67822891e-01 -2.27610394e-01
1.47567317e-01 -1.09263611e+00 4.38507289e-01 6.63857877e-01
-3.49298358e-01 -3.20858061e-01 2.59042412e-01 6.66614398e-02
-2.79539734e-01 -1.01532567e+00 4.65784729e-01 -1.54531628e-01
-4.43181008e-01 -1.62185222e-01 -5.34495473e-01 -5.61657846e-02
1.39362788e+00 -1.04632959e-01 7.60634169e-02 -1.70594212e-02
-9.53345895e-01 2.16306731e-01 4.52668160e-01 -1.19787648e-01
4.15683180e-01 -9.12259996e-01 -1.55818000e-01 1.76136598e-01
8.94582272e-02 5.47535121e-01 -2.11380664e-02 8.97466421e-01
-7.33109772e-01 7.26215541e-01 -1.36483401e-01 -2.04315558e-01
-1.22100711e+00 9.14697647e-01 2.95703888e-01 -6.45001352e-01
-3.19991082e-01 4.46957976e-01 -2.95975953e-01 -3.17061812e-01
9.72243100e-02 -8.05425286e-01 -2.11435910e-02 -3.96313593e-02
3.42439115e-01 4.02157158e-01 4.04451191e-01 -1.27567887e-01
-5.24854481e-01 5.04145324e-01 1.70953691e-01 -7.42582008e-02
1.55445302e+00 2.97392517e-01 -6.69306576e-01 3.61379329e-03
8.06723416e-01 1.13897383e-01 -5.05874157e-01 1.95134923e-01
4.61421430e-01 -3.44822973e-01 -1.59795865e-01 -7.82157481e-01
-7.63136506e-01 7.39586234e-01 -2.11528152e-01 4.71401632e-01
1.61169732e+00 -2.99642440e-02 1.88814417e-01 1.01218939e+00
7.17538476e-01 -1.29984736e+00 -5.18571019e-01 4.77525294e-01
8.06476116e-01 -8.36064517e-01 2.73715228e-01 -1.00930941e+00
-1.20430119e-01 1.45536041e+00 3.05670768e-01 1.11400135e-01
1.21108346e-01 4.04625863e-01 -5.59710503e-01 -3.63746248e-02
-8.09044898e-01 -1.17989108e-01 1.64009213e-01 2.44366020e-01
-3.79416645e-02 -8.00372809e-02 -1.07860374e+00 7.28440225e-01
-2.16375440e-01 3.08488637e-01 5.45244157e-01 1.01818192e+00
-2.08866581e-01 -1.85574973e+00 -4.49022800e-01 2.16034055e-01
-1.99932411e-01 4.85600941e-02 -5.64389586e-01 7.78387129e-01
5.14694452e-01 8.61643076e-01 -2.83669561e-01 8.22399836e-03
2.12880984e-01 2.32211173e-01 1.11254811e+00 -9.01394367e-01
-4.02889699e-01 -3.96749526e-01 5.32010555e-01 -6.89847708e-01
-4.03092057e-01 -7.26872504e-01 -1.41451955e+00 -2.03458145e-01
-5.09081662e-01 3.46425861e-01 7.70457745e-01 1.20844197e+00
-6.88982010e-02 2.41875321e-01 4.02215034e-01 -5.33218384e-01
-5.82316458e-01 -2.98352778e-01 -1.66171446e-01 1.30363017e-01
-1.35372058e-01 -6.83345377e-01 -1.89856693e-01 -2.34421313e-01] | [8.608901023864746, 6.663442611694336] |
a7504d07-6c4f-4770-add7-75e99999f7b6 | image-deconvolution-with-deep-image-and | 1910.08386 | null | https://arxiv.org/abs/1910.08386v1 | https://arxiv.org/pdf/1910.08386v1.pdf | Image Deconvolution with Deep Image and Kernel Priors | Image deconvolution is the process of recovering convolutional degraded images, which is always a hard inverse problem because of its mathematically ill-posed property. On the success of the recently proposed deep image prior (DIP), we build an image deconvolution model with deep image and kernel priors (DIKP). DIP is a learning-free representation which uses neural net structures to express image prior information, and it showed great success in many energy-based models, e.g. denoising, super-resolution, inpainting. Instead, our DIKP model uses such priors in image deconvolution to model not only images but also kernels, combining the ideas of traditional learning-free deconvolution methods with neural nets. In this paper, we show that DIKP improve the performance of learning-free image deconvolution, and we experimentally demonstrate this on the standard benchmark of six standard test images in terms of PSNR and visual effects. | ['Zipei Wang', 'Zhunxuan Wang', 'Hakan Bilen', 'Qiqi Li'] | 2019-10-18 | null | null | null | null | ['image-deconvolution'] | ['computer-vision'] | [ 3.64469051e-01 -1.55457288e-01 2.75238961e-01 -1.01855762e-01
-4.34612900e-01 -1.28343076e-01 4.74817067e-01 -8.11753511e-01
-2.77427584e-01 8.95271122e-01 4.65557277e-01 -5.70384078e-02
-1.21572025e-01 -7.35429823e-01 -1.02680016e+00 -9.47955906e-01
1.95267662e-01 -4.25012857e-02 1.64759472e-01 -2.84593910e-01
2.63373703e-01 2.95151085e-01 -1.35824656e+00 4.92591351e-01
1.01543534e+00 9.51961696e-01 5.08417606e-01 4.38011438e-01
4.51713987e-02 1.17390823e+00 -1.40166476e-01 -3.16279501e-01
1.79632738e-01 -6.17029846e-01 -7.50527799e-01 3.34963463e-02
9.74272639e-02 -7.25875854e-01 -1.05645001e+00 1.31528890e+00
4.20971185e-01 1.97533965e-01 5.91521859e-01 -9.32415783e-01
-1.26469696e+00 3.07445139e-01 -5.38150609e-01 1.97925806e-01
4.86038849e-02 -1.44524416e-02 4.50021565e-01 -1.22141957e+00
2.82990009e-01 1.12941086e+00 1.03703845e+00 4.70833421e-01
-1.38862860e+00 -3.04799855e-01 -4.94941801e-01 5.50381064e-01
-1.16760457e+00 -5.31428039e-01 9.08386230e-01 -2.28996783e-01
5.41260183e-01 6.34220615e-02 2.19691068e-01 1.00604773e+00
3.38136405e-01 7.69181848e-01 1.43588865e+00 -5.43119669e-01
1.72670543e-01 -7.99235478e-02 -4.65980098e-02 5.10892928e-01
-4.15248517e-03 3.68820190e-01 -5.81618607e-01 3.88325863e-02
1.28957558e+00 1.71388790e-01 -9.27932799e-01 7.02205375e-02
-9.97968912e-01 6.73106909e-01 6.08329952e-01 1.19826660e-01
-5.61491549e-01 3.21848541e-01 -1.12154491e-01 1.22188568e-01
4.83254492e-01 5.10878228e-02 -1.35002673e-01 2.58480102e-01
-1.06142342e+00 2.84481943e-02 6.47457421e-01 5.84287167e-01
9.28468943e-01 2.89331973e-01 -2.92149305e-01 9.53366935e-01
2.37618700e-01 2.35067680e-01 5.85899770e-01 -1.46317005e+00
-4.34734821e-02 6.60385564e-02 4.19821262e-01 -8.11612010e-01
-3.51964049e-02 -4.01511371e-01 -1.29917252e+00 6.45094514e-01
2.01818764e-01 6.97139800e-02 -1.02593315e+00 1.60844731e+00
-1.39846548e-01 7.19109118e-01 3.17507625e-01 1.14298093e+00
8.35828841e-01 8.79073024e-01 -2.42582679e-01 -1.81979746e-01
1.15817249e+00 -1.12484014e+00 -9.54711676e-01 -2.95271724e-01
-1.64301917e-01 -8.02251935e-01 8.57678115e-01 5.93878746e-01
-1.32110465e+00 -7.28350639e-01 -1.05704188e+00 -2.77511090e-01
8.00494626e-02 1.67945519e-01 4.85523909e-01 4.60226029e-01
-1.38041258e+00 1.09861314e+00 -6.10301733e-01 1.25549063e-01
4.48960245e-01 5.31384796e-02 -3.42650086e-01 -4.92149562e-01
-1.22994387e+00 9.77780640e-01 3.14407319e-01 3.57131779e-01
-1.20925355e+00 -8.17765892e-01 -6.71654999e-01 2.41935432e-01
1.98364809e-01 -7.26599097e-01 1.03878045e+00 -1.03768003e+00
-1.80280614e+00 5.10308981e-01 2.50453837e-02 -5.08583605e-01
4.33952749e-01 -5.01096606e-01 -2.52324939e-01 3.94414932e-01
-2.36463040e-01 4.36143398e-01 1.41087508e+00 -1.56212401e+00
-8.86587128e-02 -2.30180006e-02 -5.50445588e-03 -1.52615935e-01
-7.64375106e-02 -1.01690926e-01 -4.66444731e-01 -9.13909853e-01
9.46751460e-02 -6.58691347e-01 -8.70179087e-02 2.08238766e-01
-1.78674147e-01 2.93810397e-01 7.02240407e-01 -1.31279945e+00
7.18355894e-01 -2.41931939e+00 3.32032889e-01 -2.70327210e-01
2.12690622e-01 3.38637769e-01 -2.59934485e-01 2.73452461e-01
-4.00652111e-01 -1.31654873e-01 -6.98182285e-01 -4.80498195e-01
-2.22884305e-02 5.42265296e-01 -4.53773916e-01 5.81887662e-01
2.53812701e-01 9.33514416e-01 -7.75311470e-01 -1.21735767e-01
2.14454189e-01 1.16562402e+00 -4.34269279e-01 4.49412882e-01
-4.30963049e-03 4.59825993e-01 -3.54100578e-02 4.32523847e-01
1.21230400e+00 -3.17926109e-01 -2.55940109e-01 -6.98956788e-01
-1.52215376e-01 -3.20407599e-01 -9.74334180e-01 1.82470274e+00
-5.49548328e-01 8.03339362e-01 4.32392448e-01 -1.11455178e+00
4.76063162e-01 4.00880933e-01 3.02854002e-01 -6.94746315e-01
-1.70626976e-02 2.24698469e-01 -4.19956535e-01 -6.01720452e-01
3.02769244e-01 -3.58658105e-01 7.03391373e-01 1.43184319e-01
1.46564126e-01 6.04147166e-02 -1.77822322e-01 6.50056973e-02
1.02917874e+00 5.02248228e-01 -2.39981383e-01 -2.18046486e-01
5.78621566e-01 -2.23990932e-01 4.64759231e-01 6.38288319e-01
1.70099646e-01 1.40953505e+00 1.26481220e-01 -1.92517430e-01
-1.08445060e+00 -1.25626266e+00 -2.26681262e-01 5.67727327e-01
2.64255911e-01 6.04989789e-02 -9.37834740e-01 -1.30125836e-01
-2.67053753e-01 7.55611777e-01 -5.13627946e-01 -3.35631788e-01
-4.58597720e-01 -1.09614003e+00 4.94022459e-01 3.75841737e-01
1.05740404e+00 -1.07889318e+00 -2.39765316e-01 2.73686469e-01
-5.10094762e-01 -1.33123648e+00 -4.77128446e-01 2.87882775e-01
-7.90658236e-01 -9.76920843e-01 -1.37789285e+00 -8.32903802e-01
7.37211108e-01 5.23562849e-01 1.02777040e+00 8.77133086e-02
-2.57766902e-01 4.07290637e-01 -2.89722651e-01 1.17982835e-01
-5.45337260e-01 -9.27913249e-01 -2.24307209e-01 3.46708119e-01
-3.19055676e-01 -8.94767761e-01 -8.45620632e-01 2.34598473e-01
-1.41041732e+00 2.25136951e-01 9.58302140e-01 1.23524535e+00
5.03798723e-01 4.66968209e-01 4.00187075e-01 -4.96104896e-01
5.67089379e-01 -3.66712958e-01 -5.14565647e-01 2.29252279e-01
-5.22160470e-01 2.50297427e-01 5.41331232e-01 -4.02201772e-01
-1.62173617e+00 -1.26730993e-01 -2.40362868e-01 -8.54474902e-01
2.45123301e-02 6.67303026e-01 -7.52779245e-02 -5.73570728e-01
6.25790894e-01 8.30698133e-01 2.89728612e-01 -8.67822289e-01
2.69070387e-01 3.85437697e-01 1.15377176e+00 -4.74014103e-01
8.60242009e-01 6.53071940e-01 4.62146848e-02 -7.51096070e-01
-8.47602487e-01 -1.55397639e-01 -2.96997070e-01 -9.40495804e-02
9.74345505e-01 -1.13110137e+00 -5.53465128e-01 8.48204374e-01
-1.37978792e+00 -4.49319512e-01 -1.19593352e-01 5.23771167e-01
-6.05167806e-01 8.19999754e-01 -9.38631356e-01 -4.55634654e-01
-1.72594786e-01 -1.09181345e+00 6.45026445e-01 3.69623393e-01
6.02424562e-01 -1.07612038e+00 -1.93885759e-01 3.06318015e-01
8.04657996e-01 2.01856747e-01 7.72590458e-01 2.37309560e-01
-6.99719191e-01 2.46874705e-01 -6.82196498e-01 1.18515849e+00
5.28844558e-02 -6.03225172e-01 -1.21330130e+00 -2.82351464e-01
6.28603101e-01 -8.72323811e-02 1.30956686e+00 5.61823726e-01
1.31556070e+00 -2.59557188e-01 8.96509141e-02 1.05919349e+00
2.04844189e+00 -1.77482054e-01 1.38130498e+00 1.28043011e-01
6.99925780e-01 3.23378444e-01 -1.13576531e-01 3.06034386e-01
-5.12884511e-03 5.37851036e-01 6.45840466e-01 -3.49611878e-01
-7.69140601e-01 1.27403527e-01 4.91873413e-01 6.13645434e-01
-4.75248516e-01 -8.12000781e-02 -4.95221287e-01 5.01978576e-01
-1.83347332e+00 -9.12462711e-01 -3.44204485e-01 2.01138568e+00
1.04109848e+00 -2.25745663e-01 -5.27725399e-01 -1.12406626e-01
6.96599185e-01 -1.26566544e-01 -4.64786649e-01 -1.85501371e-02
-3.73537868e-01 3.90045941e-01 6.09595716e-01 4.67348814e-01
-8.98359179e-01 4.72805351e-01 6.43546963e+00 8.39497149e-01
-8.84061337e-01 4.00775075e-01 6.06914997e-01 3.81711483e-01
-1.33963034e-01 -6.02686033e-02 -2.74122953e-01 6.94294453e-01
7.96249032e-01 2.33611204e-02 1.06806612e+00 4.60265011e-01
2.23266006e-01 -1.78417608e-01 -9.55463827e-01 1.28472495e+00
1.21162623e-01 -1.53750014e+00 4.05210517e-02 -2.63446905e-02
8.29939246e-01 2.86983568e-02 1.58821478e-01 1.20007060e-01
1.78266894e-02 -1.27512538e+00 6.71819985e-01 8.33774626e-01
6.74943864e-01 -6.60606503e-01 8.59500885e-01 4.14469659e-01
-6.04235709e-01 1.86429862e-02 -6.35680556e-01 -1.70801673e-02
2.83620834e-01 9.24844027e-01 4.39424105e-02 7.41454422e-01
8.94564390e-01 7.56389558e-01 -8.97653252e-02 1.32529819e+00
-4.24600273e-01 7.83974230e-01 1.05548389e-01 1.03893030e+00
1.33903161e-01 -5.33446670e-01 4.81009126e-01 1.09168422e+00
5.03784239e-01 2.99618870e-01 -2.49388859e-01 1.30748713e+00
-2.28365496e-01 -5.66515863e-01 -1.69100851e-01 2.83979416e-01
-2.16590121e-01 1.17715073e+00 -4.67391670e-01 -1.82183251e-01
-5.93663216e-01 1.51541579e+00 -1.19599923e-01 8.07218790e-01
-1.03213108e+00 -2.74995327e-01 5.61014831e-01 -1.14465440e-02
5.55565536e-01 -1.92563459e-01 -5.89333624e-02 -1.17557538e+00
-6.70126546e-03 -6.51260614e-01 -1.34846926e-01 -1.29569185e+00
-1.48292422e+00 5.04108012e-01 -8.25155675e-02 -1.19340265e+00
1.91546395e-01 -8.51022005e-01 -6.66654110e-01 1.07926261e+00
-2.25457263e+00 -1.09224069e+00 -7.15898216e-01 7.73827851e-01
2.84649372e-01 1.63253233e-01 5.69771767e-01 5.65750778e-01
-3.55553776e-01 1.11693308e-01 6.42180026e-01 1.36857808e-01
7.40588188e-01 -1.16147339e+00 -1.39376177e-02 8.87166321e-01
-2.10724816e-01 4.00918990e-01 7.77136922e-01 -3.46499681e-01
-1.57133973e+00 -1.08290815e+00 1.98114410e-01 3.61495428e-02
4.28027004e-01 1.29491597e-01 -1.33113992e+00 5.02986073e-01
5.89643121e-01 3.65233094e-01 3.29495400e-01 -7.50745475e-01
-2.00024620e-01 -4.82606664e-02 -1.21242249e+00 2.80788809e-01
8.26856673e-01 -5.64590931e-01 -6.38403475e-01 4.21334416e-01
7.52674699e-01 -3.28647822e-01 -7.36019731e-01 3.19565654e-01
2.56762683e-01 -1.26446748e+00 1.55536664e+00 3.03826574e-02
7.82856762e-01 -4.31620270e-01 -2.44167835e-01 -1.39745057e+00
-5.66538870e-01 -6.53414905e-01 -2.51742929e-01 1.15806890e+00
-1.32176250e-01 -5.90916753e-01 5.36146641e-01 3.80726457e-01
-5.75704098e-01 -4.09536958e-01 -8.29178512e-01 -9.59126294e-01
2.80166995e-02 -4.19041961e-02 2.49851927e-01 7.44091988e-01
-6.23044014e-01 -1.25479952e-01 -8.04067850e-01 3.70320290e-01
1.09310043e+00 -2.30247498e-01 1.26398385e-01 -1.02067316e+00
-6.58801854e-01 -2.75229871e-01 -1.69439781e-02 -1.04953837e+00
1.93908185e-01 -6.87541008e-01 3.22268605e-01 -1.76816511e+00
3.88020158e-01 -1.60782866e-03 -4.57395285e-01 3.97275656e-01
-9.73924920e-02 6.05325162e-01 -8.32933858e-02 4.59511489e-01
-1.38623625e-01 7.57596672e-01 1.47424078e+00 -2.18574867e-01
9.13467631e-03 -2.56440103e-01 -5.85928082e-01 7.81714201e-01
5.41725695e-01 -4.35030609e-01 -2.99408585e-01 -7.00388610e-01
1.81454450e-01 1.89932466e-01 9.64444458e-01 -1.12427485e+00
2.54254818e-01 1.72397066e-02 6.49754524e-01 -2.90056109e-01
4.89671797e-01 -8.29667330e-01 3.91868412e-01 4.05967385e-01
4.83746044e-02 -5.51724255e-01 2.71356612e-01 6.80049837e-01
-5.65838575e-01 -3.65860254e-01 1.05792308e+00 -4.06552732e-01
-1.01170170e+00 8.11566114e-02 -1.21846899e-01 -2.39091769e-01
5.52424550e-01 -2.99379081e-01 -3.49940151e-01 -3.59471202e-01
-7.16260374e-01 -2.50381559e-01 4.43016171e-01 -6.40954301e-02
9.97419596e-01 -1.29130065e+00 -9.38922644e-01 -1.67918317e-02
-3.76954257e-01 -1.53495260e-02 5.03842652e-01 1.10037422e+00
-7.06259310e-01 -1.28930852e-01 -4.14444178e-01 -4.02602553e-01
-7.34472811e-01 6.51890337e-01 4.97760952e-01 -3.03639448e-03
-1.00350964e+00 8.13063025e-01 5.44485807e-01 2.51103938e-02
1.18309677e-01 -1.28280237e-01 -4.60998230e-02 -4.67256635e-01
7.83199728e-01 3.93016636e-01 -9.02000219e-02 -4.33504641e-01
-8.79693255e-02 5.41231871e-01 1.65009916e-01 1.63508113e-02
1.70513713e+00 -1.87289953e-01 -6.25384867e-01 -1.70376211e-01
1.20325387e+00 -2.11657226e-01 -1.73864818e+00 -2.85079122e-01
-2.44682774e-01 -5.41782618e-01 6.57161891e-01 -9.03741181e-01
-1.33879519e+00 9.51522291e-01 7.10471332e-01 -5.35368398e-02
1.53499126e+00 -1.89353079e-01 1.01947892e+00 1.19589657e-01
2.10650831e-01 -7.55793750e-01 3.65274429e-01 3.92644674e-01
1.25351870e+00 -1.15244019e+00 8.39328989e-02 -3.92512679e-01
-2.35921219e-01 1.26868868e+00 2.73509055e-01 -3.15143883e-01
7.41519690e-01 2.29046166e-01 -2.75956810e-01 1.43772438e-01
-2.43025526e-01 -4.37110960e-02 2.08420068e-01 5.99986792e-01
1.21758983e-01 -4.19497341e-01 -1.47702307e-01 6.84393883e-01
5.05673707e-01 4.26811099e-01 7.70372510e-01 7.58595943e-01
-4.48268384e-01 -9.04566646e-01 -7.37266898e-01 -6.66653439e-02
-5.31132996e-01 -3.81947577e-01 1.62375882e-01 3.53211343e-01
2.24147841e-01 8.27029765e-01 -1.67990223e-01 -1.27036631e-01
4.75555435e-02 -1.27252281e-01 5.64491153e-01 -1.59479231e-01
-1.80281773e-01 7.56784305e-02 -3.43522102e-01 -5.81072152e-01
-6.05668664e-01 -1.25573769e-01 -1.25504684e+00 -3.17201197e-01
-1.57891184e-01 -6.17329478e-02 7.71521688e-01 9.31673527e-01
3.18840832e-01 7.96916008e-01 4.18058604e-01 -1.17715001e+00
-6.38665199e-01 -8.41173887e-01 -7.16721594e-01 4.29688692e-01
6.38058424e-01 -6.56643450e-01 -6.54112577e-01 4.60520893e-01] | [11.481878280639648, -2.430739164352417] |
71bdce57-ee23-4e4a-8376-ea22838b2104 | observability-blocking-for-functional-privacy | 2304.07928 | null | https://arxiv.org/abs/2304.07928v2 | https://arxiv.org/pdf/2304.07928v2.pdf | Observability Blocking for Functional Privacy of Linear Dynamic Networks | This paper addresses the problem of determining the minimum set of state variables in a network that need to be blocked from direct measurements in order to protect functional privacy with respect to {\emph{any}} output matrices. The goal is to prevent adversarial observers or eavesdroppers from inferring a linear functional of states, either vector-wise or entry-wise. We prove that both problems are NP-hard. However, by assuming a reasonable constant bound on the geometric multiplicities of the system's eigenvalues, we present an exact algorithm with polynomial time complexity for the vector-wise functional privacy protection problem. Based on this algorithm, we then provide a greedy algorithm for the entry-wise privacy protection problem. Our approach is based on relating these problems to functional observability and leveraging a PBH-like criterion for functional observability. Finally, we provide an example to demonstrate the effectiveness of our proposed approach. | ['Yuanqing Xia', 'Ranbo Cheng', 'Yuan Zhang'] | 2023-04-17 | null | null | null | null | ['blocking'] | ['natural-language-processing'] | [ 5.71147859e-01 4.64320719e-01 8.67848191e-03 -1.30202189e-01
-5.80047190e-01 -1.26120329e+00 1.41623691e-01 -3.62848900e-02
-3.74917567e-01 6.69905901e-01 -2.14226142e-01 -8.86496007e-01
-5.97029686e-01 -6.65483415e-01 -7.51899898e-01 -1.08036816e+00
-3.29074562e-01 -1.45034924e-01 -5.68716265e-02 -5.64973988e-02
2.74469674e-01 8.25480938e-01 -8.70502114e-01 -3.83871257e-01
4.13183659e-01 9.02316213e-01 -6.53697193e-01 8.92290175e-01
6.49711192e-01 4.59841520e-01 -5.36367059e-01 -1.76700652e-01
9.96931911e-01 -8.04153919e-01 -8.81158769e-01 1.83744490e-01
2.04564795e-01 -3.19126695e-01 -7.27268577e-01 1.61226285e+00
2.03462824e-01 -1.80695683e-01 3.19270581e-01 -1.75920320e+00
-2.47889310e-01 6.12213671e-01 -1.93330660e-01 4.80797552e-02
3.56673360e-01 1.91712111e-01 1.01660037e+00 -6.55766800e-02
3.11351329e-01 8.42123866e-01 2.94014543e-01 5.25631785e-01
-1.59259117e+00 -7.60146916e-01 -3.22328322e-02 -1.79513559e-01
-1.61458051e+00 -7.39431620e-01 5.39987862e-01 -2.96710312e-01
5.75368404e-01 1.04131043e+00 2.73745447e-01 6.26006603e-01
1.24905936e-01 8.37358907e-02 1.18066871e+00 -3.07437509e-01
3.95599812e-01 4.83656049e-01 3.56296957e-01 6.87496006e-01
8.62087607e-01 4.74982172e-01 -7.80285671e-02 -6.99558258e-01
6.79984272e-01 -1.06323056e-01 -8.19503784e-01 -7.09735513e-01
-8.91770184e-01 6.80596888e-01 9.70072821e-02 1.54780582e-01
2.13550404e-02 2.90204883e-01 5.93209779e-03 9.92309332e-01
-6.23859726e-02 6.42366409e-01 -3.09319049e-01 4.48824346e-01
-3.25864792e-01 -1.05911884e-02 1.46954679e+00 1.04527354e+00
6.30350590e-01 2.06016116e-02 -3.37719694e-02 -2.69466013e-01
3.25062662e-01 7.81571746e-01 -5.67654133e-01 -7.47682333e-01
4.60664660e-01 4.57376659e-01 5.25736570e-01 -1.14214623e+00
-2.17876956e-01 -2.35744789e-01 -7.88638234e-01 4.13196355e-01
6.36701167e-01 -6.34015799e-01 -4.39009249e-01 2.07083917e+00
3.16095382e-01 1.04794195e-02 2.53116727e-01 7.92280436e-01
-1.02901570e-01 6.46317840e-01 -5.77482283e-01 -6.32584631e-01
1.00116348e+00 -3.69780660e-01 -7.18366861e-01 5.86218201e-02
4.98204738e-01 -1.79774478e-01 3.76128405e-01 1.00981444e-01
-9.19032872e-01 2.23454267e-01 -1.37731838e+00 3.97488892e-01
-1.77442580e-02 -1.39279515e-01 2.02758580e-01 1.57894671e+00
-1.14275098e+00 3.46632242e-01 -7.94001043e-01 -1.46639094e-01
-6.02264963e-02 8.97942603e-01 -7.42168009e-01 3.42914999e-01
-9.96741116e-01 6.37579441e-01 3.96954149e-01 3.03528130e-01
-1.00525665e+00 -3.84272754e-01 -8.34498525e-01 2.75863975e-01
6.92423105e-01 -3.95097524e-01 8.01908255e-01 -6.44844949e-01
-1.33204639e+00 5.06324172e-01 -8.98717195e-02 -6.57105386e-01
5.55378437e-01 3.16171229e-01 -4.28596735e-01 1.92031428e-01
-3.30789685e-01 -6.09742939e-01 7.53634274e-01 -1.38878477e+00
-5.13021886e-01 -5.63415766e-01 6.83697045e-01 -1.67317986e-01
-4.83860761e-01 -3.57306935e-02 5.13551012e-02 4.02731523e-02
2.32614294e-01 -1.22534347e+00 -6.22711599e-01 1.55324206e-01
-1.10327029e+00 5.48789799e-01 7.88793385e-01 -3.61746252e-01
1.31018019e+00 -2.26936483e+00 8.49568024e-02 9.95441914e-01
5.02919316e-01 2.31012598e-01 1.11881368e-01 8.79832447e-01
-1.62564546e-01 4.05530095e-01 -2.85732806e-01 -2.42138192e-01
1.67183861e-01 8.72056782e-02 -4.10648674e-01 1.27322602e+00
-1.57597125e-01 3.44829053e-01 -4.91786867e-01 1.43877417e-01
2.02462241e-01 -3.45875919e-02 -3.88995498e-01 1.94707751e-01
1.92182213e-01 6.61608636e-01 -5.84823132e-01 1.41172931e-01
9.79584634e-01 -1.62034351e-02 6.24291241e-01 5.64884208e-02
-1.97683915e-01 -2.21541841e-02 -1.59505308e+00 7.87282526e-01
-2.03721568e-01 3.35111052e-01 6.56507969e-01 -8.51269901e-01
4.77829307e-01 6.79577887e-01 3.73567134e-01 -1.23514310e-01
6.76828802e-01 9.64229330e-02 8.11132044e-02 -1.32911354e-01
1.11094318e-01 -4.49550450e-02 -4.86018717e-01 8.21416497e-01
-2.14959025e-01 4.60914522e-01 -2.53749877e-01 2.07575634e-01
1.27913928e+00 -7.56694436e-01 5.26672602e-01 -6.86655521e-01
9.18491840e-01 -3.92130017e-01 7.03960836e-01 1.00579464e+00
-1.31123737e-01 1.01544715e-01 1.18912113e+00 8.93627387e-03
-9.30712819e-01 -8.20293188e-01 1.37485519e-01 3.73071373e-01
3.90381843e-01 -3.33767503e-01 -8.01266074e-01 -5.64912379e-01
1.15277454e-01 5.51482141e-01 -8.76235723e-01 -3.91926736e-01
-2.82314494e-02 -3.28998536e-01 6.61167264e-01 -1.90279379e-01
3.62591863e-01 -1.46217704e-01 -3.00800383e-01 -1.46976843e-01
3.57870191e-01 -8.89176071e-01 -5.85433483e-01 3.68214518e-01
-5.18227756e-01 -1.21347904e+00 7.97051266e-02 -4.30505097e-01
1.29990220e+00 3.67738098e-01 2.38460228e-01 1.56690240e-01
1.12124026e-01 4.41314131e-01 7.40450844e-02 -1.96759269e-01
-6.30849242e-01 -7.72252604e-02 2.19303459e-01 5.13799191e-01
-2.93754876e-01 -5.60093939e-01 -4.25226957e-01 4.43735868e-01
-1.07756805e+00 -4.33440417e-01 2.57372230e-01 4.25883383e-01
4.72320884e-01 4.60676104e-01 1.14776351e-01 -1.24122453e+00
7.03485548e-01 -4.12835240e-01 -1.28402507e+00 4.09706533e-01
-5.61773241e-01 3.21302056e-01 9.97242033e-01 -1.73468754e-01
-5.80333531e-01 5.70092380e-01 2.19973207e-01 -2.59561568e-01
1.63412482e-01 1.13594793e-01 -7.92192459e-01 -9.26787257e-01
5.26238501e-01 4.32798207e-01 7.50622973e-02 -9.75925922e-02
3.24009240e-01 4.52319860e-01 3.73129636e-01 -2.08000094e-01
1.42157209e+00 6.06771946e-01 7.05850661e-01 -7.18335211e-01
-6.76611722e-01 -3.56914967e-01 -3.17355275e-01 5.47238886e-02
4.17613417e-01 -5.53526163e-01 -1.51126730e+00 3.00634295e-01
-7.16844916e-01 1.95223078e-01 -1.23723432e-01 1.38860479e-01
-4.04434651e-01 6.05188727e-01 -3.18086296e-01 -1.40453041e+00
-1.55495599e-01 -1.11784792e+00 5.10388196e-01 -7.29305223e-02
2.33857304e-01 -1.03374159e+00 6.83576316e-02 -3.61689068e-02
3.87828320e-01 6.89508617e-01 5.12946784e-01 -9.55057740e-01
-8.21062863e-01 -8.10451984e-01 6.22764938e-02 3.46014887e-01
1.06479749e-01 -3.44205976e-01 -8.11010003e-01 -8.50348175e-01
5.00065565e-01 1.22414291e-01 2.53081203e-01 -4.30594832e-02
9.72840130e-01 -1.17731893e+00 -3.07106197e-01 7.24902928e-01
1.68138981e+00 2.36377567e-01 4.36688632e-01 -5.32527901e-02
6.12063587e-01 2.85933107e-01 3.07721108e-01 6.26666427e-01
-1.71892699e-02 2.94727087e-01 6.68172717e-01 5.98196685e-02
8.86980951e-01 -1.01386912e-01 1.85585231e-01 3.86822164e-01
3.48336428e-01 -5.12634039e-01 -2.64908075e-01 3.65734428e-01
-1.69176400e+00 -8.48547041e-01 -4.83350217e-01 2.78622293e+00
5.71494043e-01 4.23331112e-02 -3.14050145e-03 3.92255485e-01
6.43848658e-01 2.09528543e-02 -4.70779598e-01 -5.98005474e-01
-6.41423091e-02 -8.54298845e-02 1.44667351e+00 8.93122852e-01
-1.08931613e+00 3.70337307e-01 6.67299795e+00 2.87936091e-01
-8.20171356e-01 -2.37746108e-02 3.33452791e-01 1.18586540e-01
-4.80873853e-01 5.69007695e-01 -6.59528434e-01 2.60439903e-01
1.11632299e+00 -6.58639550e-01 6.94854498e-01 5.32643437e-01
4.88356501e-02 3.02911755e-02 -1.24702537e+00 6.19507730e-01
-1.10557191e-01 -9.45595622e-01 -4.26410347e-01 5.17666280e-01
6.62502110e-01 -7.52335727e-01 1.80477530e-01 -2.34490484e-01
4.48515058e-01 -8.23901892e-01 5.04610419e-01 2.75909036e-01
7.84834802e-01 -9.73807454e-01 3.89355093e-01 4.60022718e-01
-1.10267591e+00 -2.20560610e-01 -3.20609957e-01 -1.71209931e-01
1.67603612e-01 3.34521383e-01 -5.57370961e-01 7.46850908e-01
8.34282488e-02 1.00021139e-02 -2.54479907e-02 9.33404028e-01
-3.09713095e-01 5.06927729e-01 -6.40079141e-01 -5.52223772e-02
6.43688962e-02 -3.24701697e-01 9.08564925e-01 7.69259751e-01
2.60552227e-01 4.27510560e-01 1.44472674e-01 7.48168766e-01
-1.20581836e-01 -3.17539603e-01 -8.16581249e-01 -1.36704162e-01
8.05498779e-01 1.10188282e+00 -3.78505290e-01 1.08588487e-01
-1.86772406e-01 1.05589807e+00 -2.41026804e-02 3.81010175e-01
-6.62848771e-01 -7.18832433e-01 1.05367684e+00 -1.93438008e-01
1.59446284e-01 -3.23331982e-01 -3.66358221e-01 -9.19396996e-01
1.63437232e-01 -8.58497381e-01 3.75554860e-01 2.74295006e-02
-9.90999997e-01 4.73474592e-01 -2.21306339e-01 -1.23293304e+00
-2.57984530e-02 -3.15800220e-01 -4.71154720e-01 1.02201450e+00
-8.64454746e-01 -6.84891284e-01 1.33344978e-01 9.93486226e-01
-6.32800281e-01 1.92302480e-01 1.00985789e+00 -4.82703298e-02
-7.98929989e-01 8.12350094e-01 3.61207306e-01 -1.02825435e-02
8.38706419e-02 -1.23540151e+00 3.45708609e-01 1.58153820e+00
-1.28262639e-01 9.32612777e-01 1.08001220e+00 -4.06021148e-01
-2.12490535e+00 -9.96529937e-01 9.10392404e-01 -4.40013140e-01
8.40787828e-01 -7.49712646e-01 -3.69805872e-01 1.10656631e+00
2.84760147e-02 1.06298747e-02 7.27889419e-01 -2.77356863e-01
-3.56736898e-01 -8.38766322e-02 -1.62343812e+00 5.15862405e-01
1.00639522e+00 -7.12265491e-01 9.01224930e-03 3.99706095e-01
7.69247234e-01 -4.20213670e-01 -8.43261361e-01 -8.83015469e-02
3.40995461e-01 -7.34408915e-01 7.46239722e-01 -7.29266942e-01
-4.72915858e-01 -6.55169070e-01 -7.52245337e-02 -1.00854266e+00
-4.13899392e-01 -1.40069115e+00 -1.11171782e-01 9.71246481e-01
5.53959787e-01 -1.14192033e+00 7.18142807e-01 8.63219559e-01
4.29865718e-01 -1.99356541e-01 -1.12811708e+00 -9.91209388e-01
-3.34743053e-01 -6.96601421e-02 5.65649092e-01 9.16989625e-01
2.30700001e-01 1.33582473e-01 -9.85493600e-01 1.14063227e+00
7.53950059e-01 1.10358829e-02 9.06158864e-01 -9.91900623e-01
-4.85309064e-01 9.37078074e-02 -5.73383689e-01 -7.55398631e-01
6.76011816e-02 -6.80653811e-01 3.31054698e-03 -9.64432240e-01
1.01647429e-01 -3.54418188e-01 -4.37526792e-01 4.91620779e-01
1.00719221e-01 -6.52260259e-02 2.42732778e-01 -2.02253714e-01
-4.63263214e-01 1.67536754e-02 9.28663850e-01 1.63228169e-01
-1.70194611e-01 4.99551862e-01 -1.06323445e+00 1.21092141e-01
8.23399842e-01 -7.67784119e-01 -6.46494746e-01 2.10044496e-02
1.07901223e-01 5.91304064e-01 4.23431009e-01 -8.68959188e-01
3.65541220e-01 -3.05066973e-01 -3.30644727e-01 -1.29298300e-01
3.50464791e-01 -1.56304264e+00 7.63789296e-01 9.21474636e-01
-4.18029398e-01 -2.01331098e-02 -5.61889559e-02 8.85528564e-01
2.38985762e-01 -1.87474206e-01 7.37584174e-01 4.58706558e-01
-1.51892945e-01 5.75852156e-01 -6.09370708e-01 -1.64486438e-01
1.38989735e+00 -1.58140913e-01 -4.81377900e-01 -7.28284180e-01
-8.83829117e-01 3.28565985e-01 5.23063481e-01 -8.30200985e-02
1.32106885e-01 -9.30475652e-01 -3.40452164e-01 4.26255196e-01
1.32069141e-01 -6.74599528e-01 2.47708693e-01 9.51113641e-01
-4.41054642e-01 6.31794512e-01 -1.30817905e-01 -5.97448647e-02
-1.61493611e+00 8.22484076e-01 5.07684588e-01 -9.84104723e-02
-2.17387691e-01 8.06209922e-01 5.47626950e-02 -1.36571288e-01
3.25632989e-01 -1.47492543e-01 3.80637199e-01 -6.89348161e-01
4.77613270e-01 3.53230506e-01 -1.51145712e-01 -5.04486144e-01
-3.46319050e-01 1.39149144e-01 1.61291048e-01 -3.83695930e-01
9.05061126e-01 -5.40749609e-01 -3.78312141e-01 -1.65815819e-02
1.40508831e+00 3.23292702e-01 -1.16263235e+00 -2.24847749e-01
-1.13666266e-01 -7.73473263e-01 -1.71447128e-01 -3.56330842e-01
-1.26105464e+00 4.25720125e-01 5.37698746e-01 7.32986629e-01
1.31479442e+00 -3.47772658e-01 2.88889319e-01 6.72082782e-01
7.46333003e-01 -5.19840598e-01 -8.17865372e-01 1.85121536e-01
4.44323123e-01 -5.77787101e-01 -1.15739614e-01 -8.13669801e-01
-6.88965768e-02 7.94467926e-01 9.67631564e-02 -2.65811741e-01
7.28569806e-01 4.69335318e-01 -1.73898205e-01 1.36178717e-01
-6.58743680e-01 3.84807363e-02 -1.66702885e-02 5.44394314e-01
-2.69473404e-01 3.15077960e-01 -5.54553509e-01 3.78794551e-01
-1.67442754e-01 -4.80894208e-01 8.79708350e-01 1.12756133e+00
-3.93695444e-01 -1.36497128e+00 -4.78323638e-01 3.55923772e-01
-6.20957017e-01 2.31236264e-01 -7.38694012e-01 4.96277243e-01
-3.57207119e-01 1.23686850e+00 -4.23194230e-01 -5.29132605e-01
3.95556957e-01 -3.96846652e-01 3.02640587e-01 -3.69518191e-01
-5.73287845e-01 -3.92862082e-01 3.07946861e-01 -6.83299541e-01
3.75339687e-02 -4.32372183e-01 -7.68615961e-01 -7.45156467e-01
-5.37396669e-01 2.88657278e-01 4.37502325e-01 8.16244483e-01
2.48061389e-01 2.82624274e-01 1.29956686e+00 -6.64057434e-02
-1.09640765e+00 -4.04451162e-01 -1.01893580e+00 1.21860439e-02
7.43411183e-01 -1.46880820e-02 -8.78776491e-01 -3.06563944e-01] | [5.913848400115967, 6.930883884429932] |
3dd9ddc5-ce82-4b37-a0fe-f877c5820ab3 | generic-dependency-modeling-for-multi-party | 2302.10680 | null | https://arxiv.org/abs/2302.10680v1 | https://arxiv.org/pdf/2302.10680v1.pdf | Generic Dependency Modeling for Multi-Party Conversation | To model the dependencies between utterances in multi-party conversations, we propose a simple and generic framework based on the dependency parsing results of utterances. Particularly, we present an approach to encoding the dependencies in the form of relative dependency encoding (ReDE) and illustrate how to implement it in Transformers by modifying the computation of self-attention. Experimental results on four multi-party conversation benchmarks show that this framework successfully boosts the general performance of two Transformer-based language models and leads to comparable or even superior performance compared to the state-of-the-art methods. The codes are available at https://github.com/shenwzh3/ReDE. | ['Ke Yang', 'Xiaojun Quan', 'Weizhou Shen'] | 2023-02-21 | null | null | null | null | ['dependency-parsing'] | ['natural-language-processing'] | [-1.31182507e-01 3.67412090e-01 -5.20899072e-02 -8.76913488e-01
-1.01280737e+00 -5.50207973e-01 7.54523814e-01 -1.15352087e-01
-3.60107832e-02 5.86098373e-01 8.13323498e-01 -5.02848744e-01
3.03305417e-01 -5.06369710e-01 -5.21159708e-01 -4.76154864e-01
-6.46925867e-02 7.42906094e-01 1.68601990e-01 -5.94687819e-01
-6.96488395e-02 3.53489220e-02 -1.19436574e+00 6.84147775e-01
5.05760074e-01 7.12984741e-01 7.84563795e-02 8.24886203e-01
-3.78897667e-01 1.10880196e+00 -5.50242662e-01 -1.00670838e+00
-2.10622832e-01 -4.48521733e-01 -1.36980915e+00 -2.11741358e-01
-4.86820191e-02 -1.58346683e-01 -3.79736722e-01 7.61802077e-01
5.06506145e-01 5.93997650e-02 2.54157633e-01 -1.26161230e+00
-8.39868784e-01 1.42494249e+00 6.15590289e-02 5.59261262e-01
6.88553333e-01 -3.53825629e-01 1.39729083e+00 -9.18173194e-01
2.63310462e-01 1.73823822e+00 6.08463168e-01 5.74555516e-01
-1.08525252e+00 -3.84971172e-01 4.90123123e-01 2.50417173e-01
-1.06984627e+00 -8.37140918e-01 7.33470738e-01 -1.05696157e-01
1.58925617e+00 5.52437425e-01 2.86981523e-01 1.29591715e+00
1.59818590e-01 9.81612921e-01 9.89624441e-01 -4.89365250e-01
-3.18507522e-01 5.54169714e-02 6.60034418e-01 6.86850309e-01
-5.15312254e-01 -1.79232568e-01 -8.04677725e-01 -2.57527024e-01
2.97469676e-01 -3.58010501e-01 -2.76215404e-01 1.96167946e-01
-1.03004205e+00 8.05983841e-01 2.51643449e-01 5.15775561e-01
-1.18831985e-01 7.69094825e-02 5.94698846e-01 5.64943612e-01
7.26161659e-01 -2.02656940e-01 -7.27492869e-01 -3.16893905e-01
-1.38670981e-01 1.01416111e-01 1.16933846e+00 1.20026076e+00
4.44962502e-01 -4.66904759e-01 -2.98042506e-01 9.52322304e-01
3.45890284e-01 8.59491378e-02 2.00904518e-01 -7.86971688e-01
7.96786785e-01 4.88989681e-01 -1.74312875e-01 -3.61486852e-01
-4.74346191e-01 -5.21193296e-02 -6.42219305e-01 -5.64268768e-01
2.32692063e-01 -6.79723360e-03 -5.75555801e-01 1.86672628e+00
3.69835287e-01 9.40533578e-02 3.28135252e-01 4.26113963e-01
1.26059151e+00 7.07385659e-01 2.29645371e-01 -2.31065720e-01
1.58746994e+00 -1.39179277e+00 -1.22476089e+00 -3.82315785e-01
7.61769414e-01 -9.32073355e-01 1.02539754e+00 -1.48842111e-01
-1.27526796e+00 -5.15418530e-01 -5.46726763e-01 -4.59064633e-01
-4.52655643e-01 -2.83758461e-01 9.50071216e-01 7.33246028e-01
-1.09067583e+00 4.79743332e-01 -8.29097688e-01 -2.38380075e-01
6.64325897e-03 3.57608259e-01 -3.72267276e-01 1.38250872e-01
-1.67075467e+00 1.07504582e+00 1.52930245e-01 1.90582171e-01
-6.17439985e-01 -5.81656992e-01 -1.13973355e+00 1.27256572e-01
2.61474233e-02 -6.70248866e-01 1.94122601e+00 -5.99222124e-01
-1.97603393e+00 8.22444379e-01 -5.28006136e-01 -1.97123021e-01
3.76634672e-02 -4.12199229e-01 -3.20791721e-01 -2.39857614e-01
-1.51105165e-01 4.43338364e-01 3.19996148e-01 -9.06416535e-01
-5.87412059e-01 -1.23334773e-01 6.91352785e-01 2.84567147e-01
-1.79968834e-01 8.31192672e-01 -8.55260968e-01 -3.87977988e-01
-1.22626118e-01 -1.13366282e+00 -1.46641573e-02 -8.13895285e-01
-4.98095870e-01 -7.17433870e-01 6.97552919e-01 -6.98861837e-01
1.43671834e+00 -2.01327896e+00 4.23477650e-01 -4.05533671e-01
-1.20111287e-01 1.57133475e-01 -1.61265686e-01 9.27449822e-01
-3.56198132e-01 2.76653141e-01 -3.99883866e-01 -1.07252562e+00
1.87845334e-01 5.97903728e-01 -3.08004171e-01 1.93451285e-01
1.21038049e-01 1.09843612e+00 -8.12021375e-01 -4.24053758e-01
2.09224388e-01 8.43994558e-01 -5.18303096e-01 8.51754963e-01
3.22814827e-04 6.95214510e-01 -4.10067022e-01 5.98492384e-01
6.09736621e-01 -2.38190249e-01 6.84915245e-01 -2.84818172e-01
-1.85759693e-01 1.41019261e+00 -6.62473440e-01 1.81742251e+00
-9.42511141e-01 5.30938447e-01 1.69744492e-01 -6.84652865e-01
7.02471018e-01 7.46116579e-01 -3.89929712e-02 -5.15257001e-01
3.46077234e-01 -1.10082060e-01 -4.05265354e-02 -5.53075731e-01
4.02873129e-01 -1.77686989e-01 -5.42579353e-01 4.79483277e-01
3.45773339e-01 -2.22912893e-01 2.33035266e-01 3.61768305e-01
9.56800699e-01 -4.81446534e-02 2.82805413e-01 -3.17121804e-01
9.01996017e-01 -5.80912590e-01 5.62643170e-01 4.30605710e-01
-1.17781937e-01 2.09864765e-01 6.94378257e-01 -4.59405720e-01
-6.12590015e-01 -7.07787454e-01 -1.75809100e-01 1.80525732e+00
-1.71240211e-01 -8.53732407e-01 -7.93443322e-01 -9.05161619e-01
-3.14775199e-01 8.08886111e-01 -7.35906839e-01 2.33982310e-01
-9.53749955e-01 -8.67739618e-01 6.58310473e-01 7.43279576e-01
4.68763918e-01 -1.19276178e+00 -1.01227336e-01 2.78636843e-01
-8.79147589e-01 -1.53269529e+00 -5.93038619e-01 4.93700415e-01
-4.60900068e-01 -7.45385468e-01 -2.36101747e-01 -9.36650395e-01
1.82262376e-01 -7.87349045e-02 1.57715476e+00 2.97807485e-01
2.17102647e-01 2.39065066e-01 -7.44297326e-01 -2.57213235e-01
-6.68620467e-01 2.10256651e-01 -1.44387990e-01 -1.45599529e-01
3.76902103e-01 -7.07067549e-01 -1.42861098e-01 1.91265777e-01
-4.67588037e-01 2.95869522e-02 2.39133567e-01 8.14435184e-01
1.67978421e-01 -3.58560145e-01 4.82460290e-01 -1.17878854e+00
8.83617580e-01 -3.72292131e-01 -2.32901171e-01 2.77922958e-01
-2.31456324e-01 2.26473227e-01 5.41461825e-01 -4.05924320e-02
-1.34628367e+00 -2.63856292e-01 -8.68808508e-01 1.39148787e-01
-1.66833084e-02 3.31634402e-01 -3.60403836e-01 1.85067818e-01
-5.61247692e-02 -3.02281547e-02 -3.87200594e-01 -6.79368377e-01
5.48087358e-01 9.37209427e-01 3.87750566e-01 -7.93908358e-01
3.59439224e-01 1.71775371e-01 -4.75398153e-01 -4.98839319e-01
-1.01586568e+00 -4.57425117e-01 -7.86719203e-01 3.94052193e-02
8.33039820e-01 -7.62640178e-01 -8.00256014e-01 3.91658485e-01
-1.54871035e+00 -5.66197395e-01 1.71029702e-01 2.55476356e-01
-5.53769648e-01 4.87950236e-01 -1.26900327e+00 -8.09263468e-01
-5.41073799e-01 -1.18787146e+00 9.90186036e-01 3.56627591e-02
-1.80272281e-01 -1.20920444e+00 2.83649474e-01 6.10662162e-01
4.67584163e-01 -2.89876461e-01 1.03162706e+00 -8.11869264e-01
-2.73565054e-01 2.57000655e-01 1.57783031e-01 4.51211780e-01
6.34406582e-02 5.93948998e-02 -1.23556566e+00 -7.53945336e-02
3.97974206e-03 -3.59906882e-01 6.83690071e-01 -2.18351539e-02
8.17832708e-01 -2.61672348e-01 -3.18009049e-01 5.01474082e-01
8.75867665e-01 1.14680141e-01 6.05626881e-01 8.89862236e-03
5.21204472e-01 8.85212958e-01 4.08854634e-01 3.71039271e-01
1.11639869e+00 8.99086416e-01 3.45436454e-01 3.38908024e-02
-2.18372211e-01 -1.94521829e-01 4.90538448e-01 1.63733208e+00
-5.65303192e-02 -4.62596625e-01 -8.49907756e-01 6.67800546e-01
-2.12416625e+00 -7.95161784e-01 -3.46330911e-01 1.74211800e+00
9.50202465e-01 1.81603849e-01 -3.00308950e-02 -6.84315190e-02
7.80923665e-01 3.92383993e-01 5.43321148e-02 -1.06653631e+00
-8.53685215e-02 2.68080473e-01 1.09943278e-01 9.56677616e-01
-1.07629395e+00 1.15726435e+00 7.36601019e+00 2.07127288e-01
-5.09523928e-01 6.25209153e-01 5.22757649e-01 3.10424179e-01
-2.90714025e-01 7.62183964e-02 -1.05989158e+00 1.56847745e-01
1.67509031e+00 -1.71428025e-01 4.40701336e-01 5.95826387e-01
-1.40766755e-01 3.93333256e-01 -1.41368365e+00 6.85717165e-01
-5.51016517e-02 -1.21189761e+00 -1.26305044e-01 -4.71346319e-01
1.76917061e-01 3.02587599e-01 -1.73776358e-01 6.05440378e-01
6.08847678e-01 -7.91164219e-01 5.81799805e-01 9.65499356e-02
4.03306782e-01 -7.65713811e-01 9.60400462e-01 3.25000644e-01
-1.57251453e+00 -3.24907824e-02 -2.17719808e-01 -3.80460650e-01
5.15305579e-01 2.69818932e-01 -6.60673261e-01 8.08980942e-01
9.91633713e-01 5.60953200e-01 -2.34602541e-01 2.22632438e-01
-6.87901199e-01 6.24896586e-01 -1.40589163e-01 -1.76564604e-02
2.03770429e-01 4.62504290e-02 3.44415843e-01 1.91812360e+00
-1.54383942e-01 3.11240435e-01 -1.26526386e-01 4.50056136e-01
-2.14937791e-01 5.58821820e-02 -3.72543424e-01 1.45861626e-01
6.88384354e-01 1.26027167e+00 -3.01579803e-01 -5.45339108e-01
-9.03024733e-01 8.93500030e-01 8.20723951e-01 5.70546463e-02
-9.87064123e-01 -1.24281041e-01 9.86998379e-01 -3.11927021e-01
4.90405828e-01 -1.56715155e-01 1.63285226e-01 -1.06268287e+00
2.32297387e-02 -9.92931485e-01 6.26646101e-01 -5.55066645e-01
-1.44256485e+00 1.22160745e+00 2.86564618e-01 -7.04924703e-01
-4.16599005e-01 -5.24810791e-01 -7.47111380e-01 9.06502247e-01
-1.79400945e+00 -1.29802549e+00 -7.90567920e-02 7.30905473e-01
7.52984583e-01 2.56633759e-01 1.51514912e+00 4.12875444e-01
-7.59827673e-01 7.89659262e-01 -4.27335471e-01 3.92561078e-01
5.30727625e-01 -1.24463391e+00 1.10901582e+00 6.90653980e-01
1.62588805e-01 6.99963093e-01 5.74769497e-01 -7.69009292e-02
-1.42360711e+00 -7.04583406e-01 1.70997632e+00 -7.42916703e-01
8.56499851e-01 -7.66640902e-01 -9.95619476e-01 1.31675017e+00
9.57582235e-01 -8.45390782e-02 9.92150962e-01 5.23144186e-01
-6.18903220e-01 1.69764474e-01 -8.52765322e-01 3.92366767e-01
1.22215843e+00 -8.96412313e-01 -1.03493381e+00 4.47055012e-01
1.08586979e+00 -6.95989907e-01 -1.22617352e+00 4.53360349e-01
4.12903488e-01 -1.25568557e+00 8.61060202e-01 -6.06809616e-01
1.83749348e-01 1.24664940e-02 -4.41029519e-01 -1.31478798e+00
-5.72650552e-01 -1.01252985e+00 -1.58718079e-01 1.66234422e+00
6.93542719e-01 -9.03216779e-01 1.30597353e-01 7.77061641e-01
-4.52723145e-01 -6.66406572e-01 -1.18252003e+00 -4.93796915e-01
1.56939045e-01 -5.11010587e-01 8.69150221e-01 9.34759200e-01
5.10527790e-01 7.83392608e-01 -3.17393839e-01 3.58912945e-01
3.34648222e-01 2.51512736e-01 6.75181627e-01 -7.30454147e-01
-4.96657163e-01 -2.28181973e-01 -2.70162020e-02 -1.53187370e+00
7.81174242e-01 -9.31792200e-01 -8.63659382e-02 -1.61715639e+00
1.59246802e-01 -4.51653749e-01 -3.86303931e-01 6.58591449e-01
-3.08478802e-01 1.65414065e-01 2.07434624e-01 -2.53184661e-02
-5.48183560e-01 5.79293549e-01 9.39349234e-01 -3.11182495e-02
-1.31167192e-02 1.48537651e-01 -9.13656771e-01 6.40781701e-01
9.27691102e-01 -6.95603907e-01 -1.46256268e-01 -6.55847788e-01
1.65696338e-01 3.61667156e-01 -6.00146838e-02 -4.03492182e-01
2.97394454e-01 1.63733959e-01 -3.96525711e-01 -6.33991301e-01
5.19909143e-01 -4.94424909e-01 2.95925979e-02 1.39854357e-01
-5.25749683e-01 3.89115334e-01 2.84771591e-01 9.29518789e-02
-5.01615465e-01 -1.69941887e-01 4.56795812e-01 -1.42716005e-01
-5.02068639e-01 -5.22230901e-02 -3.07785660e-01 6.03012182e-02
4.75723922e-01 6.67981029e-01 -5.96982062e-01 -4.51720804e-01
-7.20308721e-01 2.19139680e-01 -2.00942248e-01 7.95195341e-01
2.84278333e-01 -1.10822487e+00 -9.43846524e-01 1.78828537e-01
1.49393618e-01 -3.31206918e-01 1.43473431e-01 7.43199229e-01
3.44838127e-02 7.39679933e-01 1.49700478e-01 -3.84965062e-01
-1.57389617e+00 5.15852630e-01 3.99467289e-01 -6.71201289e-01
-5.80864906e-01 1.31010532e+00 2.48001203e-01 -8.06846797e-01
2.56789386e-01 -6.10689521e-01 -3.33841711e-01 -1.86177328e-01
6.37217820e-01 -9.16509852e-02 2.87417442e-01 -9.87169504e-01
-7.81765223e-01 3.30182105e-01 -3.09358507e-01 4.88594100e-02
1.22382689e+00 -3.70259345e-01 -4.78233337e-01 5.28015196e-01
1.27678514e+00 1.99497074e-01 -9.72396076e-01 -4.46533263e-01
-1.52470982e-02 -2.50014633e-01 -1.14223041e-01 -6.70633495e-01
-8.98359060e-01 6.87504351e-01 3.29730250e-02 6.51579201e-01
1.03729284e+00 5.76224625e-01 7.15297043e-01 5.50280094e-01
2.74266809e-01 -6.77406251e-01 -1.99134782e-01 9.97542560e-01
1.08731246e+00 -9.55021322e-01 -3.22284341e-01 -8.44368219e-01
-8.96085501e-01 9.18934703e-01 2.48781517e-01 1.15399428e-01
9.45965648e-01 8.77492070e-01 4.69572186e-01 -5.22232577e-02
-1.35060489e+00 -3.50841254e-01 -3.51330154e-02 4.97258604e-01
1.16256940e+00 2.70318180e-01 -3.73079628e-01 7.82216787e-01
-4.04484540e-01 -4.56876636e-01 4.34882492e-01 8.89180720e-01
9.09129754e-02 -1.79558837e+00 -5.04132453e-03 -2.09423557e-01
-8.10825765e-01 -6.32886350e-01 -3.98258805e-01 8.12478840e-01
-2.82623678e-01 1.40201569e+00 6.63046762e-02 -4.50708151e-01
5.76793492e-01 5.12826264e-01 4.08245802e-01 -6.41205609e-01
-1.20677125e+00 2.03936156e-02 9.23982024e-01 -6.42935753e-01
-7.99599230e-01 -4.25444394e-01 -1.26259434e+00 -5.27036250e-01
-2.89864421e-01 3.84045273e-01 4.94580567e-01 8.81708503e-01
3.82782876e-01 7.07601309e-01 7.63086200e-01 -8.09020758e-01
-2.92710185e-01 -1.32301080e+00 -1.81984246e-01 4.14481997e-01
3.24624985e-01 -4.72419232e-01 -4.93332565e-01 6.81494270e-03] | [12.488269805908203, 7.757236957550049] |
b687f8ae-005f-4755-b947-af6729c1094b | rng-kbqa-generation-augmented-iterative | 2109.08678 | null | https://arxiv.org/abs/2109.08678v2 | https://arxiv.org/pdf/2109.08678v2.pdf | RnG-KBQA: Generation Augmented Iterative Ranking for Knowledge Base Question Answering | Existing KBQA approaches, despite achieving strong performance on i.i.d. test data, often struggle in generalizing to questions involving unseen KB schema items. Prior ranking-based approaches have shown some success in generalization, but suffer from the coverage issue. We present RnG-KBQA, a Rank-and-Generate approach for KBQA, which remedies the coverage issue with a generation model while preserving a strong generalization capability. Our approach first uses a contrastive ranker to rank a set of candidate logical forms obtained by searching over the knowledge graph. It then introduces a tailored generation model conditioned on the question and the top-ranked candidates to compose the final logical form. We achieve new state-of-the-art results on GrailQA and WebQSP datasets. In particular, our method surpasses the prior state-of-the-art by a large margin on the GrailQA leaderboard. In addition, RnG-KBQA outperforms all prior approaches on the popular WebQSP benchmark, even including the ones that use the oracle entity linking. The experimental results demonstrate the effectiveness of the interplay between ranking and generation, which leads to the superior performance of our proposed approach across all settings with especially strong improvements in zero-shot generalization. | ['Caiming Xiong', 'Yingbo Zhou', 'Kazuma Hashimoto', 'Semih Yavuz', 'Xi Ye'] | 2021-09-17 | null | https://aclanthology.org/2022.acl-long.417 | https://aclanthology.org/2022.acl-long.417.pdf | acl-2022-5 | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-1.61560193e-01 1.65286317e-01 -2.91039258e-01 -4.31031048e-01
-1.58388317e+00 -8.63329589e-01 3.29475284e-01 4.59482163e-01
-3.58131915e-01 1.09574306e+00 1.46184281e-01 -1.25123784e-01
-7.09939778e-01 -1.16784012e+00 -1.02046442e+00 -3.88278723e-01
1.24171913e-01 1.18906748e+00 8.02987397e-01 -8.52407515e-01
-8.95346701e-02 -5.94201423e-02 -1.53316176e+00 5.89607477e-01
1.53585815e+00 1.04388976e+00 -1.59358755e-01 1.92564636e-01
-3.06995600e-01 8.43631327e-01 -4.64853317e-01 -1.13480747e+00
1.17434867e-01 -1.44598156e-01 -1.09160805e+00 -4.33033586e-01
1.04940975e+00 -1.02350451e-01 -4.24685210e-01 1.03632665e+00
5.58468342e-01 1.95197791e-01 3.56098741e-01 -1.13656867e+00
-9.45228636e-01 9.41015661e-01 -1.97392032e-01 2.53288448e-01
7.35325098e-01 -2.03705519e-01 1.82896352e+00 -1.13061261e+00
8.29943120e-01 1.27773964e+00 5.32753289e-01 6.75212801e-01
-1.11846173e+00 -6.10354781e-01 -6.15066253e-02 5.90818942e-01
-1.53014898e+00 -1.23340644e-01 3.21532011e-01 -1.15004130e-01
9.90091681e-01 3.35598052e-01 2.40742460e-01 1.02535880e+00
-3.13793480e-01 1.13579226e+00 1.06291735e+00 -2.52749801e-01
2.92777896e-01 8.72090086e-02 6.73913062e-01 8.25998068e-01
4.85290468e-01 -3.80013496e-01 -8.40304554e-01 -5.04560649e-01
2.46918917e-01 -6.15228951e-01 -4.63464975e-01 -7.05824673e-01
-9.86671448e-01 8.41635823e-01 3.81335437e-01 -1.51377171e-01
-2.91576505e-01 -4.38655727e-03 2.63619512e-01 5.07373273e-01
2.02231035e-01 8.72811675e-01 -6.44060552e-01 2.18466595e-02
-7.90650725e-01 6.74121618e-01 1.00819862e+00 1.24412823e+00
7.72148609e-01 -4.66803819e-01 -5.90908170e-01 8.57630551e-01
8.08372647e-02 6.05649710e-01 1.98275074e-01 -6.77706599e-01
9.21861231e-01 8.07467878e-01 1.38995692e-01 -4.58322823e-01
-2.86582440e-01 -6.44916236e-01 -3.29900175e-01 -5.43959081e-01
3.18615735e-01 1.00973636e-01 -9.73765433e-01 1.86967587e+00
5.36123812e-01 6.77164420e-02 4.99828070e-01 7.71532714e-01
1.14897156e+00 6.83069468e-01 2.22317576e-01 -8.84119049e-02
1.41178727e+00 -1.09002650e+00 -7.22096026e-01 -1.88941300e-01
6.68714762e-01 -1.16632596e-01 1.31882668e+00 6.22428775e-01
-1.09700799e+00 -2.61495888e-01 -1.09932339e+00 -2.65485466e-01
-4.19408768e-01 -2.22965881e-01 6.62674487e-01 7.77715027e-01
-1.02177763e+00 4.07032162e-01 -4.24003541e-01 -3.98530215e-01
3.06031942e-01 2.16892347e-01 -4.58841383e-01 -4.88038301e-01
-1.75914645e+00 8.57989907e-01 6.65670097e-01 -2.78987795e-01
-1.08999240e+00 -1.18608570e+00 -6.74010277e-01 1.74365327e-01
1.05081701e+00 -1.13261354e+00 1.28643942e+00 -2.62243658e-01
-1.26467371e+00 4.79977310e-01 4.03528698e-02 -6.46038115e-01
3.61226946e-01 -5.30437768e-01 -5.85545182e-01 1.72829539e-01
1.86818421e-01 5.05551100e-01 2.11941823e-01 -1.14326358e+00
-9.51372623e-01 -3.02200735e-01 5.82566679e-01 4.71443176e-01
-3.78341317e-01 -5.03313780e-01 -8.40652645e-01 -3.84538293e-01
2.51810793e-02 -7.23450840e-01 8.28735456e-02 -4.75388676e-01
-2.53319919e-01 -8.15876961e-01 2.73804605e-01 -4.86525059e-01
1.56718254e+00 -1.83711445e+00 4.54105437e-01 1.39749110e-01
-3.18331295e-03 2.87670672e-01 -2.24744633e-01 7.59354472e-01
3.37502062e-01 1.83140114e-01 -1.70704618e-01 -2.82490235e-02
3.31772029e-01 3.12229007e-01 -6.70280755e-01 -2.24982183e-02
2.47432396e-01 1.08631241e+00 -1.25639641e+00 -4.99776155e-01
-4.88609523e-01 -4.27335538e-02 -6.66512549e-01 1.31486848e-01
-7.33704448e-01 -2.08399460e-01 -5.82836688e-01 7.50865877e-01
4.63297009e-01 -4.57488656e-01 2.54682004e-01 -2.35743657e-01
5.22762358e-01 7.28163719e-01 -1.21726251e+00 1.98380864e+00
-1.65809765e-01 -1.17933519e-01 -3.46383989e-01 -6.04544818e-01
7.32974291e-01 2.31173635e-01 1.63157105e-01 -9.01207864e-01
-5.39392471e-01 5.34043968e-01 -1.52091995e-01 -4.51257914e-01
7.64190912e-01 -1.78352937e-01 -3.39621305e-01 1.46005198e-01
4.89397556e-01 -1.96819663e-01 6.81063473e-01 8.77539158e-01
1.21012330e+00 1.49716914e-01 2.51965374e-01 -4.04426068e-01
4.72384572e-01 1.53005704e-01 6.00664914e-01 1.09094501e+00
2.24950209e-01 2.30828390e-01 5.16916573e-01 -1.59706205e-01
-5.24518847e-01 -1.42168427e+00 -1.08449705e-01 1.35142624e+00
4.58036780e-01 -7.02596307e-01 -4.80024636e-01 -1.29003358e+00
2.48618141e-01 1.02984357e+00 -6.38476610e-01 -2.34848976e-01
-4.43256646e-01 -7.06939816e-01 8.00949752e-01 5.39214671e-01
4.33936834e-01 -8.93621802e-01 -5.33599332e-02 3.19156587e-01
-4.35218841e-01 -1.17638946e+00 3.50719169e-02 -3.54261626e-03
-7.35748112e-01 -1.08080363e+00 -2.75192648e-01 -5.29428303e-01
2.58019418e-01 3.86691391e-02 1.55303383e+00 -1.85630322e-01
1.70187026e-01 4.66694683e-01 -6.65689766e-01 -4.00492758e-01
-2.00590074e-01 5.56162536e-01 -7.14134499e-02 -2.46934056e-01
3.30859035e-01 -3.50430369e-01 -4.36602116e-01 2.89424956e-01
-1.09231436e+00 -2.38655820e-01 7.13627994e-01 8.56126249e-01
7.05983937e-01 -5.39942645e-02 1.07219708e+00 -1.23226988e+00
7.74338126e-01 -7.27341592e-01 -5.08091927e-01 9.56700683e-01
-8.70326459e-01 3.40911597e-01 5.36423564e-01 -7.04620481e-02
-1.22850406e+00 -3.80103081e-01 -7.08633568e-03 3.12548541e-02
2.44216144e-01 9.13134634e-01 -4.45823133e-01 1.20749593e-01
1.01834095e+00 1.30055755e-01 -6.85792685e-01 -4.19081777e-01
7.46835649e-01 2.90159643e-01 6.15719080e-01 -1.07322216e+00
8.85964811e-01 3.62154782e-01 -1.43365443e-01 -4.05412644e-01
-1.40949845e+00 -7.16717899e-01 -3.19186568e-01 1.90177679e-01
5.22138536e-01 -1.00599098e+00 -3.55550230e-01 8.63079727e-02
-7.99464762e-01 -1.68993801e-01 -4.62233365e-01 1.17814302e-01
-3.81991386e-01 4.88837510e-01 -6.48605287e-01 -5.18538833e-01
-4.98407513e-01 -6.92265391e-01 1.03260255e+00 2.90043563e-01
2.62158453e-01 -7.45023727e-01 3.49239707e-01 7.44861662e-01
1.83317900e-01 -6.45240694e-02 1.31678402e+00 -1.10505784e+00
-8.45380604e-01 -1.20076649e-01 -1.60031065e-01 1.85489431e-01
-2.08416373e-01 -4.75384802e-01 -7.18336284e-01 -3.09568584e-01
-5.71691275e-01 -9.17374909e-01 1.11933982e+00 -2.97456354e-01
8.46452594e-01 -3.14232767e-01 -2.34002382e-01 3.60746920e-01
1.57120275e+00 -1.21308327e-01 6.72479928e-01 4.54582155e-01
5.41888058e-01 3.96636695e-01 1.10609591e+00 1.56909928e-01
7.35051036e-01 7.46723115e-01 5.54513454e-01 2.65761048e-01
-9.84081849e-02 -5.12224674e-01 2.69585043e-01 8.29292178e-01
1.12959214e-01 -7.15722203e-01 -9.88508999e-01 9.22042251e-01
-2.09317684e+00 -7.15991914e-01 -9.63462517e-02 2.24816942e+00
1.23336017e+00 9.70992520e-02 -1.28109045e-02 -1.44001976e-01
1.80847645e-01 -9.34044942e-02 -4.85758990e-01 6.23874366e-03
-5.22699237e-01 6.16621315e-01 3.37460399e-01 4.19593930e-01
-1.04345179e+00 1.17780495e+00 5.78680992e+00 1.11946225e+00
-4.28584486e-01 1.21645555e-01 6.66248798e-02 -2.44626440e-02
-6.93589747e-01 9.12555959e-03 -1.23536015e+00 6.90271705e-02
8.60403538e-01 -4.27342296e-01 4.77252245e-01 7.49353528e-01
-6.74186289e-01 1.21663377e-01 -1.17476523e+00 5.10325313e-01
2.68204480e-01 -1.29073560e+00 6.24083698e-01 -1.92104205e-01
9.91653264e-01 4.38762344e-02 3.54597904e-02 8.83503079e-01
5.72247565e-01 -9.30840671e-01 5.98736703e-01 4.23044264e-01
5.48852444e-01 -7.42075741e-01 8.49910080e-01 3.32825452e-01
-9.44656134e-01 -2.66141862e-01 -6.36722505e-01 4.49102938e-01
1.05407842e-01 4.60641831e-01 -1.04511392e+00 1.39058650e+00
7.20781147e-01 2.77796179e-01 -8.24499965e-01 1.06238174e+00
-6.03757143e-01 7.54555225e-01 -2.07746491e-01 -1.47037372e-01
2.85979569e-01 1.02710322e-01 4.99142885e-01 1.07428384e+00
3.39157164e-01 3.48893791e-01 1.69792950e-01 5.83401561e-01
-4.81482655e-01 4.88744706e-01 -2.09425479e-01 -1.10782925e-02
7.34674573e-01 1.21808672e+00 3.29884738e-02 -5.79031289e-01
-4.32622343e-01 7.30854928e-01 8.58888030e-01 3.61214072e-01
-9.12595212e-01 -3.56987745e-01 2.79585004e-01 9.34977308e-02
5.43746114e-01 1.13789365e-01 3.04471910e-01 -1.19025373e+00
4.33513701e-01 -1.22098374e+00 1.22561848e+00 -6.98159993e-01
-1.59004569e+00 5.66907406e-01 3.29022646e-01 -8.17339242e-01
-1.77543223e-01 -4.98214841e-01 -1.06147863e-02 5.02087533e-01
-1.71854937e+00 -1.27023733e+00 -4.13330436e-01 6.27154648e-01
4.00449067e-01 -6.50306698e-03 8.08485210e-01 4.36414659e-01
-3.23095888e-01 7.97485411e-01 3.90466861e-02 -1.31284073e-01
9.69327092e-01 -1.73684049e+00 2.97551662e-01 7.54079282e-01
4.29358929e-01 8.54182363e-01 5.58952153e-01 -8.46772552e-01
-1.65712357e+00 -1.12643147e+00 7.57408619e-01 -9.44475889e-01
8.45693886e-01 -3.17799509e-01 -1.25612915e+00 7.39133060e-01
1.75190821e-01 1.90698430e-02 6.80322289e-01 4.22191739e-01
-8.56850803e-01 -4.40752029e-01 -9.23013151e-01 4.09972101e-01
1.27165389e+00 -4.15538132e-01 -1.14798748e+00 3.75552595e-01
9.06110942e-01 -5.51820874e-01 -9.89684224e-01 8.29306185e-01
3.78070682e-01 -7.41153240e-01 9.53875005e-01 -1.12162101e+00
3.00513357e-01 -4.56193507e-01 -2.66200572e-01 -1.29198050e+00
-3.73925447e-01 -4.77128774e-01 -5.14019787e-01 1.28796625e+00
8.85291338e-01 -4.62841332e-01 8.44943881e-01 4.02054548e-01
-2.26670980e-01 -8.93633306e-01 -1.14425266e+00 -9.63309884e-01
1.03664078e-01 -1.74631089e-01 6.79008901e-01 8.85701358e-01
1.01661436e-01 6.46957755e-01 -2.43469730e-01 5.03067553e-01
5.98656893e-01 1.57594502e-01 8.64244640e-01 -1.19899309e+00
-4.87445384e-01 -6.28079996e-02 -2.18503937e-01 -1.14189494e+00
3.86286937e-02 -1.05573773e+00 -1.01626769e-01 -1.76150787e+00
4.11240876e-01 -5.11910617e-01 -6.37097061e-01 5.03559053e-01
-6.47495747e-01 2.43415609e-02 -1.89931486e-02 2.10709702e-02
-1.30524611e+00 5.46513915e-01 1.12377167e+00 -1.39462754e-01
-4.31633443e-02 -3.34426701e-01 -8.08284998e-01 1.65928051e-01
4.86176491e-01 -4.08723950e-01 -8.31804335e-01 -4.93653834e-01
9.34048533e-01 2.07507629e-02 1.73301786e-01 -9.39417362e-01
5.01726210e-01 -2.12946758e-02 -1.73661008e-01 -5.94932675e-01
3.54032248e-01 -3.85102570e-01 -2.70612240e-02 6.60824776e-02
-4.41006869e-01 5.05745001e-02 2.75612921e-01 9.06100869e-01
-2.54451126e-01 -2.11000741e-01 3.52585495e-01 2.29171105e-02
-9.73565102e-01 4.35482413e-01 2.93761164e-01 6.65337324e-01
7.13057280e-01 1.91650957e-01 -1.04138708e+00 -3.44549417e-01
-4.87082481e-01 6.20653450e-01 1.95015162e-01 4.82563764e-01
4.42695141e-01 -1.33375967e+00 -9.67693150e-01 -2.56145269e-01
6.93943322e-01 1.92692176e-01 3.77413273e-01 8.27115238e-01
-2.22498924e-01 6.46228254e-01 -1.60706453e-02 -3.33424151e-01
-9.60625052e-01 6.03971958e-01 1.75830528e-01 -7.66236424e-01
-2.96885699e-01 1.06599164e+00 2.98238844e-01 -4.28629369e-01
2.05483824e-01 -1.96478397e-01 -1.55783623e-01 7.62575939e-02
4.04223502e-01 3.94672155e-01 5.01958966e-01 -4.81145382e-02
-4.06700015e-01 2.01370642e-01 -3.87186229e-01 -8.37465152e-02
1.21959853e+00 1.67893693e-01 -1.12591855e-01 2.28797853e-01
5.81294775e-01 2.99595386e-01 -6.68674350e-01 -5.69968402e-01
4.30079639e-01 -2.32132807e-01 -3.24477106e-01 -1.38268447e+00
-7.07452655e-01 5.62872887e-01 2.64574885e-01 2.40125880e-01
9.82671857e-01 3.08900386e-01 8.85286510e-01 8.39520156e-01
6.83470130e-01 -1.00986850e+00 2.76832491e-01 5.08561075e-01
8.86771262e-01 -1.04451847e+00 3.48148011e-02 -7.19114065e-01
-7.35018849e-01 8.21628988e-01 9.10072207e-01 1.48496121e-01
2.89734285e-02 -2.95630097e-01 -6.23996072e-02 -5.37362933e-01
-1.14481425e+00 -3.17704886e-01 7.36742079e-01 4.02764201e-01
1.96775213e-01 1.66070685e-02 -4.57469195e-01 9.95272696e-01
-2.25692213e-01 -1.40160441e-01 3.27474356e-01 8.23432684e-01
-5.39649487e-01 -1.20940232e+00 6.99267611e-02 3.98988783e-01
-4.15964693e-01 -3.14051747e-01 -5.79422414e-01 1.10235000e+00
-6.03224784e-02 7.45984077e-01 -3.92500162e-01 -2.44110703e-01
7.38593936e-01 3.30246329e-01 6.12426281e-01 -7.42804766e-01
-5.48141897e-01 -4.96359199e-01 6.15536690e-01 -7.47333467e-01
-1.16209269e-01 -4.59264427e-01 -1.13508642e+00 1.14118673e-01
-5.71416497e-01 6.55139625e-01 2.40689099e-01 7.18085945e-01
5.00756741e-01 1.84021324e-01 1.08420916e-01 4.05317485e-01
-1.08191180e+00 -8.74891937e-01 -5.12803853e-01 7.15017021e-01
9.29675400e-02 -6.87896311e-01 -1.21799305e-01 -4.20714021e-01] | [10.499441146850586, 7.951272487640381] |
e7a8b529-2125-4cd5-b7b9-ada70b07723b | iterative-zero-shot-llm-prompting-for | 2307.01128 | null | https://arxiv.org/abs/2307.01128v1 | https://arxiv.org/pdf/2307.01128v1.pdf | Iterative Zero-Shot LLM Prompting for Knowledge Graph Construction | In the current digitalization era, capturing and effectively representing knowledge is crucial in most real-world scenarios. In this context, knowledge graphs represent a potent tool for retrieving and organizing a vast amount of information in a properly interconnected and interpretable structure. However, their generation is still challenging and often requires considerable human effort and domain expertise, hampering the scalability and flexibility across different application fields. This paper proposes an innovative knowledge graph generation approach that leverages the potential of the latest generative large language models, such as GPT-3.5, that can address all the main critical issues in knowledge graph building. The approach is conveyed in a pipeline that comprises novel iterative zero-shot and external knowledge-agnostic strategies in the main stages of the generation process. Our unique manifold approach may encompass significant benefits to the scientific community. In particular, the main contribution can be summarized by: (i) an innovative strategy for iteratively prompting large language models to extract relevant components of the final graph; (ii) a zero-shot strategy for each prompt, meaning that there is no need for providing examples for "guiding" the prompt result; (iii) a scalable solution, as the adoption of LLMs avoids the need for any external resources or human expertise. To assess the effectiveness of our proposed model, we performed experiments on a dataset that covered a specific domain. We claim that our proposal is a suitable solution for scalable and versatile knowledge graph construction and may be applied to different and novel contexts. | ['Sandro Gabriele Tiddia', 'Livio Pompianu', 'Alessandro Sebastian Podda', 'Leonardo Piano', 'Alessandro Giuliani', 'Salvatore Carta'] | 2023-07-03 | null | null | null | null | ['graph-construction', 'graph-generation', 'knowledge-graphs'] | ['graphs', 'graphs', 'knowledge-base'] | [ 5.70500828e-02 4.83001202e-01 -1.81091636e-01 2.66431272e-01
-5.06975770e-01 -8.84714007e-01 8.98823559e-01 5.85638702e-01
-1.54126598e-03 6.02869570e-01 1.54076681e-01 -5.02684534e-01
-6.01017475e-01 -1.14566076e+00 -5.12917399e-01 -3.27353835e-01
6.35590628e-02 5.56389093e-01 2.10261926e-01 -3.91320556e-01
3.67598623e-01 6.01963997e-01 -1.54938257e+00 2.94191204e-02
1.18205678e+00 7.26422012e-01 4.19536501e-01 2.21080735e-01
-7.23481715e-01 8.35455060e-01 -5.21128297e-01 -5.65622509e-01
2.87756678e-02 -4.18657839e-01 -9.81456041e-01 2.11050108e-01
2.27301363e-02 1.78574696e-01 4.70934436e-02 8.07017207e-01
3.47699285e-01 6.33049533e-02 4.87715602e-01 -1.01075113e+00
-6.34192228e-01 7.81736314e-01 -1.59201562e-01 -9.61509049e-02
6.06706679e-01 1.33412912e-01 8.94478738e-01 -7.81528890e-01
9.85697687e-01 9.99303341e-01 3.25650036e-01 1.58801898e-01
-1.06681645e+00 -2.51745641e-01 3.74396533e-01 3.45109314e-01
-1.41656673e+00 -1.93182841e-01 9.27748084e-01 -6.82779551e-01
8.22477937e-01 2.57266998e-01 1.02635646e+00 1.05340612e+00
-1.33503556e-01 4.72494990e-01 8.91589582e-01 -5.77748418e-01
3.15687478e-01 4.51116174e-01 2.17430517e-01 6.31800532e-01
6.30081713e-01 -4.56480116e-01 -6.90554500e-01 -2.14921921e-01
7.04068124e-01 -1.05387770e-01 -2.15183496e-01 -5.04633188e-01
-1.09609675e+00 5.60245454e-01 1.68055117e-01 7.29714811e-01
-6.67126954e-01 -2.37609986e-02 1.37125760e-01 2.07863718e-01
4.11881417e-01 7.10401237e-01 -9.04759243e-02 -2.74718076e-01
-9.74898517e-01 1.25154138e-01 1.09304678e+00 1.08611715e+00
9.82353568e-01 -1.27945051e-01 -2.74659634e-01 3.96429867e-01
2.80321687e-01 2.21646756e-01 2.00269207e-01 -5.62755287e-01
3.88973802e-01 1.21595287e+00 2.40730032e-01 -1.29144800e+00
-2.82545656e-01 -6.99039876e-01 -6.34318292e-01 -3.15795034e-01
2.17794165e-01 -5.09695262e-02 -6.59730315e-01 1.47478676e+00
4.29515153e-01 1.76705480e-01 -1.83371510e-02 4.58822459e-01
1.03635478e+00 4.43404227e-01 3.01710099e-01 -1.37375921e-01
1.50181830e+00 -6.75382912e-01 -7.40195632e-01 -2.89836854e-01
4.66864198e-01 -7.49496818e-01 1.12004495e+00 2.21174881e-01
-9.32595313e-01 -3.87881041e-01 -8.97537470e-01 3.22994702e-02
-7.09617913e-01 5.98210990e-02 8.42659056e-01 4.71094579e-01
-1.22331774e+00 2.98125446e-01 -4.62081045e-01 -7.25915670e-01
2.95535356e-01 5.26273288e-02 -1.97351366e-01 -2.16106221e-01
-1.09236825e+00 8.36952567e-01 6.95032597e-01 1.45404682e-01
-5.91502845e-01 -6.78196013e-01 -6.57915592e-01 2.92567372e-01
9.15555239e-01 -1.07649708e+00 8.07271481e-01 -7.23855555e-01
-1.33361673e+00 5.35319805e-01 -1.25402495e-01 -1.65292397e-01
3.81228745e-01 -2.27319047e-01 -2.67131567e-01 1.15347929e-01
8.64849538e-02 2.30991781e-01 7.26422191e-01 -1.37534702e+00
-2.71960795e-01 -1.92954108e-01 2.80555755e-01 1.68677524e-01
-6.47697866e-01 -7.78920278e-02 -7.98365772e-01 -5.00515759e-01
-2.46940479e-02 -6.74332738e-01 -2.90531695e-01 -5.26008487e-01
-4.05473411e-01 -3.06176424e-01 6.04320824e-01 -5.24407923e-01
1.69893587e+00 -1.84150410e+00 3.36482286e-01 4.73812670e-01
4.04134393e-01 4.95778829e-01 -2.42123231e-02 1.18990827e+00
3.43924642e-01 3.48179400e-01 -1.92434322e-02 1.41342515e-02
1.23632275e-01 -6.96053952e-02 -4.24858600e-01 -2.15323031e-01
2.53972352e-01 1.24641895e+00 -1.17450047e+00 -5.13650894e-01
2.71739900e-01 5.69691479e-01 -1.89251408e-01 1.20314896e-01
-5.33580482e-01 3.21740836e-01 -6.91713393e-01 5.60376883e-01
3.10217440e-01 -6.79419816e-01 5.50329447e-01 -1.34888589e-01
-1.69291437e-01 -2.46018246e-02 -1.37613690e+00 1.62242377e+00
-5.21813273e-01 2.68384874e-01 -2.02218756e-01 -8.18759024e-01
1.18803358e+00 2.51358896e-01 2.66916901e-01 -5.70542932e-01
-2.81658839e-03 1.52961478e-01 -2.96868742e-01 -5.56427121e-01
6.97053015e-01 2.00845115e-02 -1.28254652e-01 5.35307467e-01
2.15961084e-01 -1.69912323e-01 5.63673556e-01 6.86497092e-01
1.12700951e+00 2.51900762e-01 4.84716982e-01 -1.16872057e-01
5.77182770e-01 1.08726196e-01 1.16114505e-01 7.27250397e-01
3.05838794e-01 5.55580575e-03 6.66628480e-01 -3.39587063e-01
-6.82219446e-01 -7.21246362e-01 3.93188864e-01 6.26278400e-01
2.81193051e-02 -9.10692394e-01 -5.64174891e-01 -5.69761455e-01
-1.26563475e-01 7.51288652e-01 -3.78241152e-01 -6.41766936e-02
-3.10421973e-01 -4.39507872e-01 3.39538276e-01 2.61845529e-01
2.56970465e-01 -1.13962507e+00 -5.36302149e-01 2.90272206e-01
-2.90648490e-01 -1.20125341e+00 2.31897086e-02 -2.41570964e-01
-8.53058577e-01 -1.19709635e+00 -4.93599057e-01 -5.55621564e-01
8.04490030e-01 3.85773510e-01 1.30246508e+00 1.55883372e-01
-1.68279648e-01 8.23742151e-01 -6.42958105e-01 -3.40645969e-01
-4.93707359e-01 3.27823281e-01 -4.43208724e-01 5.96082322e-02
2.58538216e-01 -6.49487019e-01 -4.66840178e-01 1.73540376e-02
-1.06655693e+00 2.63613909e-01 7.24279642e-01 3.52049708e-01
5.29262125e-01 1.69062857e-02 9.21007097e-01 -1.02462375e+00
1.13091516e+00 -7.43420899e-01 -5.86327672e-01 5.24027944e-01
-8.15474093e-01 1.90350473e-01 6.87492669e-01 -1.73341006e-01
-1.08079803e+00 -1.77139938e-01 1.61387905e-01 -8.02545324e-02
-6.34621009e-02 1.05530095e+00 -1.67755783e-01 4.83604223e-02
6.43849194e-01 3.33811015e-01 -1.74662814e-01 -5.84881663e-01
8.78450036e-01 3.75282407e-01 2.33568460e-01 -6.92956388e-01
9.36413348e-01 3.52945715e-01 1.19189277e-01 -8.10880840e-01
-5.24356127e-01 -4.51573789e-01 -5.34044087e-01 -4.16948438e-01
5.78874528e-01 -6.41013443e-01 -5.88794768e-01 1.11641333e-01
-1.01643586e+00 -9.80371609e-02 -4.69539106e-01 1.89688072e-01
-2.24704251e-01 6.93855166e-01 -1.96929395e-01 -7.24944234e-01
-5.91925442e-01 -8.66697371e-01 7.77480543e-01 4.42149788e-01
-3.35753113e-01 -1.29116714e+00 1.85526963e-02 4.65202510e-01
5.97191155e-01 2.89108723e-01 1.06985593e+00 -7.26396799e-01
-1.04380167e+00 -2.65050083e-01 -1.95074975e-01 6.10755198e-02
2.57180393e-01 1.44066587e-01 -7.40891278e-01 1.01542706e-02
-4.01920915e-01 -2.14969724e-01 4.62492734e-01 -1.80297568e-01
8.80814135e-01 -2.77168512e-01 -3.06725383e-01 8.64400938e-02
1.49455833e+00 9.60496292e-02 5.37876189e-01 4.67060924e-01
7.75921106e-01 6.78416669e-01 4.50182676e-01 4.82370257e-01
6.72434866e-01 7.24368095e-01 3.09379816e-01 -6.60534203e-02
-2.65489638e-01 -5.15032709e-01 1.57803744e-01 1.00035846e+00
-2.28422508e-01 -2.50199020e-01 -1.14673841e+00 7.87828326e-01
-2.07511711e+00 -9.67794955e-01 -1.49808660e-01 2.17993498e+00
6.43155932e-01 7.39627285e-03 -9.76411849e-02 2.10774001e-02
4.17873651e-01 1.26701713e-01 -1.70387268e-01 -2.89350122e-01
-1.00237012e-01 4.48005289e-01 -7.22206905e-02 3.96859497e-01
-4.43164498e-01 1.15445292e+00 5.71241474e+00 8.39336812e-01
-9.35785115e-01 -1.97684184e-01 2.76187789e-02 2.62558043e-01
-7.09534109e-01 4.70696390e-01 -6.70740902e-01 2.74177402e-01
8.37386429e-01 -6.78561330e-01 3.24219555e-01 7.44540930e-01
2.67274678e-01 -9.52897966e-02 -8.25179398e-01 7.55984187e-01
5.75268082e-02 -1.68679905e+00 5.48219621e-01 6.92550465e-02
5.63649952e-01 -3.25117946e-01 -3.04475695e-01 1.42881215e-01
2.75626808e-01 -7.73613036e-01 5.68807364e-01 8.81796002e-01
4.35291976e-01 -7.15739787e-01 5.91914713e-01 4.22975719e-01
-1.23809612e+00 9.02540460e-02 -3.80742177e-02 -3.04666813e-02
2.47773618e-01 7.96202779e-01 -1.08446336e+00 1.28075290e+00
2.45702147e-01 5.79491377e-01 -8.28529537e-01 9.56786096e-01
-5.07157624e-01 5.92287719e-01 -9.12331268e-02 -1.73064038e-01
1.85190290e-01 -3.30512047e-01 5.37594080e-01 1.35483885e+00
5.60372293e-01 -9.83995572e-02 1.55563816e-01 9.59769547e-01
6.26560301e-02 5.43933749e-01 -8.35961401e-01 -3.99883986e-01
6.24710858e-01 1.46351302e+00 -1.05647993e+00 -5.23504615e-01
-3.33882004e-01 4.67526853e-01 3.68310302e-01 4.36539203e-01
-4.66101676e-01 -4.47333157e-01 9.07497481e-02 2.84839809e-01
3.65964353e-01 -2.91190714e-01 -2.15960205e-01 -1.16032469e+00
1.37939259e-01 -8.46174002e-01 3.97906035e-01 -7.07985640e-01
-9.54559505e-01 5.66830635e-01 1.55131876e-01 -9.15345550e-01
-4.21510279e-01 -2.29228377e-01 -4.97399360e-01 7.69593120e-01
-1.46092546e+00 -1.35126162e+00 -6.61456764e-01 5.51859200e-01
1.57051712e-01 -9.19818599e-03 9.12621975e-01 2.76255578e-01
-5.16365170e-01 1.48720846e-01 -2.35382557e-01 -2.47142091e-01
4.64212507e-01 -1.15046823e+00 3.59579355e-01 1.09003246e+00
4.00316149e-01 1.02249682e+00 5.78108072e-01 -8.14724505e-01
-1.64376032e+00 -9.17597890e-01 1.16311288e+00 -4.83521670e-01
9.04105663e-01 -3.06691915e-01 -1.00010872e+00 5.21138251e-01
1.51123494e-01 -4.52372402e-01 8.05783987e-01 2.69118994e-01
-4.66869734e-02 -1.16648160e-01 -7.22542405e-01 6.10959470e-01
1.05811501e+00 -4.70135242e-01 -5.61905563e-01 2.80733138e-01
6.59753859e-01 -1.10498168e-01 -9.50317204e-01 8.17234889e-02
2.02838093e-01 -8.13502610e-01 7.05770373e-01 -4.56506252e-01
2.19925389e-01 -5.05449951e-01 2.51868099e-01 -1.08633208e+00
-3.65793675e-01 -1.03721607e+00 -5.17429173e-01 1.49293303e+00
4.55326736e-01 -6.76024377e-01 5.22990406e-01 5.40457845e-01
-9.59575921e-02 -8.00979257e-01 -5.80404580e-01 -6.09284520e-01
-4.14092511e-01 -6.45919859e-01 7.02648282e-01 9.54404294e-01
2.03088507e-01 5.25634170e-01 -2.95416862e-01 -3.63105461e-02
3.90213072e-01 2.87688583e-01 1.10024118e+00 -1.45546246e+00
-1.87092215e-01 -3.81849885e-01 -3.70381087e-01 -7.30495095e-01
-1.39982790e-01 -1.10513091e+00 -5.71439862e-01 -2.32137394e+00
1.95770845e-01 -3.89698565e-01 -2.57446706e-01 4.74772036e-01
-1.88435316e-01 -2.75235593e-01 2.98553765e-01 2.26562202e-01
-7.26057112e-01 4.15323257e-01 1.26407492e+00 1.39768228e-01
-4.21521097e-01 -1.83963105e-01 -1.27042067e+00 5.47687232e-01
6.81547523e-01 -3.28336358e-01 -8.75699759e-01 -1.85184747e-01
8.72139335e-01 -1.94834352e-01 4.92839813e-01 -7.69801319e-01
4.65275466e-01 -2.18143851e-01 -9.12588686e-02 -3.20303738e-01
9.72637013e-02 -7.23548532e-01 5.91463208e-01 2.71506786e-01
2.58890539e-01 -4.96835746e-02 3.15984696e-01 5.74902594e-01
-2.35562399e-01 -1.78893760e-01 1.85902059e-01 -2.53899097e-01
-9.69669163e-01 4.35606800e-02 -1.92614019e-01 3.38019207e-02
1.16936624e+00 -2.77009070e-01 -3.89268845e-01 -3.02138746e-01
-5.29701293e-01 1.54787138e-01 5.50777793e-01 5.46981275e-01
4.54331458e-01 -1.06226981e+00 -4.91795838e-01 6.27577752e-02
3.13390791e-01 -4.99622859e-02 2.29307920e-01 6.31617904e-01
-3.70755672e-01 5.86407006e-01 4.78528738e-02 -2.89009392e-01
-8.42019737e-01 5.96001327e-01 -1.25015438e-01 -5.66351652e-01
-6.51687264e-01 3.47339630e-01 -1.93968534e-01 -2.35003382e-02
3.14113591e-03 -2.63259590e-01 -4.48061556e-01 4.04673368e-01
3.66822332e-01 4.38462883e-01 3.22152786e-02 -4.27037895e-01
-2.56437182e-01 5.44161856e-01 -2.91005131e-02 -8.96842033e-02
1.42013848e+00 -1.53437704e-01 -3.27193737e-01 3.33740354e-01
3.81127089e-01 3.11132282e-01 -7.15306699e-01 -2.26726949e-01
3.01022142e-01 -2.68078238e-01 -1.82366043e-01 -8.85524988e-01
-6.78417504e-01 5.56973159e-01 4.86306287e-02 4.93337184e-01
1.06777239e+00 3.55846763e-01 4.79077399e-01 3.67537022e-01
5.56612313e-01 -9.42541897e-01 1.64060414e-01 2.99258500e-01
8.51734877e-01 -7.76183248e-01 1.17531545e-01 -6.22951925e-01
-5.24604440e-01 1.09764183e+00 3.29017907e-01 3.73972833e-01
5.07093132e-01 -7.16519058e-02 -3.51931043e-02 -6.62210464e-01
-8.34188044e-01 -3.82401437e-01 4.59328383e-01 5.54242790e-01
4.71320182e-01 -2.20699143e-02 -6.35308862e-01 6.85563803e-01
-5.56591786e-02 3.03657025e-01 3.89017910e-01 1.02860177e+00
-4.44030344e-01 -1.44746256e+00 -1.41992986e-01 3.19427848e-01
-1.24469616e-01 -3.63984741e-02 -7.08529174e-01 9.08339620e-01
8.76595378e-02 1.02779508e+00 -5.67923427e-01 -2.27652490e-01
5.49078405e-01 2.65037149e-01 2.86643326e-01 -8.23424816e-01
-6.67808473e-01 -5.04243746e-02 2.06662253e-01 -4.86115009e-01
-4.20996904e-01 -3.86839211e-01 -1.15007126e+00 -2.30917901e-01
-1.55526772e-01 3.33793253e-01 6.74381077e-01 8.63221347e-01
9.92703378e-01 5.37854254e-01 1.63270637e-01 -5.21674991e-01
-1.44033864e-01 -8.31547320e-01 -2.33857974e-01 3.73551100e-01
-2.67648548e-01 -6.68926120e-01 -1.53138917e-02 1.17792629e-01] | [9.326194763183594, 8.125299453735352] |
f17632d0-28dc-4d42-8da9-9e3929333278 | cross-domain-few-shot-graph-classification | 2201.08265 | null | https://arxiv.org/abs/2201.08265v1 | https://arxiv.org/pdf/2201.08265v1.pdf | Cross-Domain Few-Shot Graph Classification | We study the problem of few-shot graph classification across domains with nonequivalent feature spaces by introducing three new cross-domain benchmarks constructed from publicly available datasets. We also propose an attention-based graph encoder that uses three congruent views of graphs, one contextual and two topological views, to learn representations of task-specific information for fast adaptation, and task-agnostic information for knowledge transfer. We run exhaustive experiments to evaluate the performance of contrastive and meta-learning strategies. We show that when coupled with metric-based meta-learning frameworks, the proposed encoder achieves the best average meta-test classification accuracy across all benchmarks. The source code and data will be released here: https://github.com/kavehhassani/metagrl | ['Kaveh Hassani'] | 2022-01-20 | null | null | null | null | ['cross-domain-few-shot'] | ['computer-vision'] | [ 7.24643469e-02 1.09308742e-01 -6.45214498e-01 -2.64100641e-01
-9.23098326e-01 -4.88270730e-01 7.42037654e-01 2.54141778e-01
-1.12366505e-01 7.56162882e-01 2.50453621e-01 1.67728271e-02
-2.32693478e-01 -9.43952203e-01 -7.25241542e-01 -3.69902939e-01
-2.56111473e-01 5.58821142e-01 2.37803221e-01 -3.03045481e-01
2.91745305e-01 4.76394780e-03 -1.40607238e+00 4.38893586e-01
8.12247574e-01 1.01925969e+00 5.06118350e-02 8.20535302e-01
-1.25835270e-01 8.35531771e-01 -4.44185704e-01 -9.67379749e-01
1.94010168e-01 -2.71317273e-01 -1.13327885e+00 -1.44147009e-01
6.88043177e-01 -6.21353984e-02 -8.19219112e-01 1.11846256e+00
4.41918105e-01 2.36871138e-01 8.00377071e-01 -1.44217932e+00
-1.32769191e+00 4.39690381e-01 -4.28674370e-01 4.25831228e-01
2.73264974e-01 2.07507700e-01 1.37518489e+00 -8.11710417e-01
9.71673250e-01 1.01062477e+00 6.11541152e-01 6.16474032e-01
-1.30161285e+00 -3.76771003e-01 1.11588918e-01 7.71047294e-01
-1.27598989e+00 -4.02847737e-01 1.05413175e+00 -5.63347578e-01
1.22429836e+00 -2.65960991e-02 4.68942046e-01 1.62905192e+00
3.57722491e-01 5.51504970e-01 8.25415194e-01 -3.79703999e-01
1.76323205e-01 -6.43892512e-02 3.68471771e-01 1.05268741e+00
5.45207500e-01 -1.86509173e-02 -4.03532505e-01 -3.45773011e-01
5.39787054e-01 -5.09167202e-02 -2.68131793e-01 -8.92563462e-01
-1.17173386e+00 1.05962324e+00 6.03290558e-01 3.49839926e-01
4.29102685e-03 2.18226746e-01 6.83915317e-01 5.87920487e-01
8.30806077e-01 5.61189413e-01 -4.24928278e-01 -3.73794511e-02
-2.51396239e-01 3.08916941e-02 7.90301323e-01 1.35447335e+00
8.19146633e-01 -4.36429195e-02 -2.55462736e-01 1.16067410e+00
-1.85786009e-01 1.57868236e-01 6.02153778e-01 -8.36288035e-01
7.48396456e-01 7.98986673e-01 -3.97355974e-01 -1.06727242e+00
-3.50236416e-01 -4.81711328e-01 -8.31650972e-01 -8.81405324e-02
1.30240858e-01 -2.74357088e-02 -9.19465959e-01 1.87607515e+00
9.08149928e-02 2.98334420e-01 -4.33283783e-02 6.46497071e-01
1.13436830e+00 3.15443367e-01 -1.69200562e-02 2.68860996e-01
1.10940719e+00 -1.36819124e+00 -4.48220819e-01 -3.53163391e-01
9.77786899e-01 -2.31137693e-01 1.17447639e+00 -3.13641168e-02
-9.07005608e-01 -5.41313112e-01 -1.32130122e+00 -3.62925172e-01
-8.33881259e-01 -2.30073825e-01 6.03664398e-01 2.01404542e-01
-9.73887146e-01 6.71290755e-01 -6.39829516e-01 -6.09644592e-01
7.12612569e-01 -9.27892327e-02 -6.59257531e-01 -3.99623990e-01
-1.28272986e+00 9.20440495e-01 3.81720364e-01 -5.16031325e-01
-9.74370837e-01 -7.16337740e-01 -1.08596623e+00 -3.00563425e-02
3.33443612e-01 -9.04552937e-01 9.32714403e-01 -7.76715815e-01
-1.21940422e+00 1.14986420e+00 2.14002982e-01 -3.45218778e-01
3.84707838e-01 1.17671646e-01 -4.54765022e-01 7.13640228e-02
2.13011861e-01 2.83912808e-01 5.51041365e-01 -8.29965055e-01
-2.53319681e-01 -5.09775281e-01 2.63720691e-01 9.59977508e-02
-4.63993013e-01 -3.65323752e-01 -4.71631944e-01 -6.87903404e-01
-4.01700318e-01 -7.92557716e-01 -2.21555158e-02 -1.25918135e-01
-3.14714402e-01 -1.89894870e-01 5.69495678e-01 -6.18204594e-01
1.10566735e+00 -2.03135729e+00 5.36790192e-01 -2.64568448e-01
5.56978881e-01 8.03494975e-02 -6.57205701e-01 6.85758948e-01
-1.89113244e-01 1.15132809e-01 -4.92494732e-01 -1.67977706e-01
-6.06273152e-02 4.57898378e-02 1.24770015e-01 5.31593621e-01
2.73661673e-01 1.30663121e+00 -1.06320059e+00 -3.99908811e-01
1.14856891e-01 3.01019102e-01 -3.69284719e-01 2.48936340e-01
-1.99140936e-01 1.07439265e-01 -3.49596679e-01 6.75442457e-01
5.86562335e-01 -7.80249476e-01 3.98715019e-01 -2.33294979e-01
6.46440089e-01 1.47872135e-01 -8.04908633e-01 2.25628805e+00
-5.35338581e-01 6.31141067e-01 -1.80817768e-01 -1.37224495e+00
7.73964226e-01 4.43162657e-02 3.12819541e-01 -1.05593240e+00
1.17421210e-01 -1.05874538e-01 -3.16620022e-01 -2.63495862e-01
1.36540487e-01 2.31750384e-01 -2.89139241e-01 3.67954403e-01
1.04860902e+00 1.64727211e-01 3.15227509e-01 3.49569499e-01
1.62330389e+00 2.48443126e-03 6.50744438e-01 -2.85780698e-01
3.03657651e-01 2.91405730e-02 4.23588336e-01 6.59085572e-01
-3.29863727e-01 5.17861605e-01 5.88154137e-01 -6.43440068e-01
-1.12913871e+00 -1.22406864e+00 -2.16211323e-02 1.45144260e+00
2.19473392e-01 -6.86753750e-01 -4.14904922e-01 -1.17639530e+00
2.10682601e-01 7.02915370e-01 -9.97063458e-01 -5.00331223e-01
-2.53861099e-01 -3.98500234e-01 3.21339101e-01 5.39893508e-01
3.90600592e-01 -7.09028542e-01 -1.25132427e-01 -2.27173865e-02
-1.76102016e-02 -1.10129237e+00 -4.28414375e-01 1.22901767e-01
-7.83128977e-01 -1.50080478e+00 -8.09005082e-01 -8.56547236e-01
4.27560121e-01 3.91333640e-01 1.60451543e+00 1.66421115e-01
-5.61717749e-01 6.88332260e-01 -5.72977006e-01 -6.26329407e-02
-2.16259792e-01 4.12257552e-01 -4.35446769e-01 -2.14267924e-01
2.00829923e-01 -7.38227069e-01 -4.89857644e-01 1.14749372e-01
-4.26981121e-01 1.96807548e-01 2.77255267e-01 1.03865016e+00
4.03703392e-01 -5.32650650e-01 6.88500464e-01 -1.21826899e+00
6.69434845e-01 -8.08550537e-01 -5.57206869e-01 4.99491483e-01
-4.85047460e-01 2.21007049e-01 6.48985863e-01 -9.70345512e-02
-7.02622414e-01 -2.95797437e-01 2.28643313e-01 -6.20096385e-01
-1.15956038e-01 5.15004516e-01 -1.36935517e-01 -2.59899527e-01
9.13828552e-01 6.85251057e-02 -8.75530913e-02 -3.62165898e-01
7.15588868e-01 4.63428825e-01 2.56052226e-01 -6.18035614e-01
5.83236277e-01 2.56980211e-01 -1.04996301e-01 -6.93697929e-01
-8.93250763e-01 -4.27163631e-01 -8.23733628e-01 -8.47602785e-02
7.52934575e-01 -9.35784280e-01 -5.06619141e-02 2.14199856e-01
-1.04291105e+00 -5.70662677e-01 -2.90335774e-01 1.89548790e-01
-9.15092885e-01 3.07024896e-01 -6.02283835e-01 -1.05085962e-01
-3.82486045e-01 -9.32400048e-01 9.55945313e-01 -2.44219616e-01
1.69543605e-02 -1.52206111e+00 6.52701914e-01 3.03303391e-01
3.24137688e-01 5.09536445e-01 1.06767392e+00 -1.05215371e+00
-4.25069213e-01 -1.62088707e-01 -3.40354145e-01 2.31703103e-01
1.76353157e-01 -3.02878946e-01 -9.65508997e-01 -5.14122903e-01
-5.80832183e-01 -7.31123686e-01 1.10785460e+00 9.33456346e-02
1.29188561e+00 -3.90655071e-01 -4.31041092e-01 8.23041797e-01
1.65730417e+00 -2.71959394e-01 4.44508582e-01 2.39116400e-01
9.10161555e-01 3.32437426e-01 3.61096054e-01 3.39865208e-01
6.92028344e-01 7.76178002e-01 3.35358948e-01 3.04936498e-01
-5.78515172e-01 -2.35672876e-01 2.74137072e-02 9.70542133e-01
-5.50849326e-02 -3.92843336e-01 -1.17638266e+00 8.12496185e-01
-2.02643633e+00 -1.09696841e+00 3.45577568e-01 1.97043979e+00
5.69544554e-01 2.15749443e-02 1.92724764e-01 -2.84892589e-01
8.70866418e-01 6.14349365e-01 -6.42029166e-01 -4.18882966e-01
2.83134892e-03 3.65315557e-01 5.30465662e-01 4.01319236e-01
-1.23228765e+00 8.69127333e-01 6.23075676e+00 7.70432711e-01
-8.48039269e-01 5.15268445e-01 3.72259617e-01 -9.65795889e-02
-3.87441844e-01 -1.57859847e-01 -2.40117297e-01 3.86129677e-01
1.14184105e+00 -6.40750587e-01 5.31030834e-01 9.95079219e-01
-6.94125235e-01 5.43980896e-01 -1.27034843e+00 9.26092386e-01
3.02022815e-01 -1.69813597e+00 1.34364694e-01 2.43645683e-02
8.41719389e-01 6.26140356e-01 -3.89962010e-02 6.21358454e-01
5.83927810e-01 -8.08729231e-01 3.20386291e-01 4.73597050e-01
9.47212338e-01 -6.29388690e-01 5.29185474e-01 6.82300851e-02
-1.37810898e+00 -2.40387574e-01 -5.54897130e-01 -4.71001007e-02
-2.79644012e-01 4.31724250e-01 -5.99773884e-01 1.00147712e+00
5.01009941e-01 1.30141401e+00 -9.77380633e-01 9.04407620e-01
-1.02389067e-01 3.04228574e-01 3.47369820e-01 7.32143670e-02
-3.73998061e-02 1.18085518e-01 5.98351538e-01 1.25361693e+00
2.10850835e-01 -1.83373630e-01 2.01908156e-01 8.18421066e-01
-5.01537502e-01 8.93922448e-02 -1.29961491e+00 -1.62406653e-01
5.57093501e-01 1.21969628e+00 -3.57919991e-01 -4.96199250e-01
-7.43455470e-01 1.06583297e+00 1.20371163e+00 3.78178030e-01
-9.55922961e-01 -6.91555142e-01 6.99109018e-01 -1.49339205e-03
3.13752353e-01 -2.85820514e-01 -1.73266724e-01 -1.58572173e+00
-3.42208408e-02 -6.99131131e-01 9.59061444e-01 -5.33663154e-01
-1.84265053e+00 5.11955678e-01 6.69922456e-02 -1.08344030e+00
-1.84293851e-01 -8.22384834e-01 -7.23272979e-01 5.19710958e-01
-1.46468318e+00 -1.34635091e+00 -6.37218833e-01 5.94868422e-01
5.19913077e-01 -5.18901229e-01 9.97905254e-01 2.65856773e-01
-5.56340158e-01 8.82601321e-01 2.23146141e-01 3.68557245e-01
7.70628810e-01 -1.33080816e+00 9.21505153e-01 4.84940946e-01
4.68952805e-01 1.20661333e-01 2.07025677e-01 -6.16376340e-01
-1.56707501e+00 -1.34512556e+00 4.67889309e-01 -7.48220682e-01
8.37808251e-01 -6.02627218e-01 -1.06090021e+00 9.71008956e-01
3.76692265e-01 4.84442055e-01 7.98848450e-01 4.48216468e-01
-8.73502672e-01 -4.57126163e-02 -9.56338346e-01 3.43192816e-01
1.67008579e+00 -6.73564494e-01 -6.49140656e-01 5.29991448e-01
8.88746798e-01 -2.05813214e-01 -1.12800193e+00 4.34307009e-01
3.43485504e-01 -7.99185216e-01 8.35421920e-01 -1.05373871e+00
5.15619993e-01 3.28600228e-01 -3.15481514e-01 -1.65539706e+00
-8.45993638e-01 -5.01516722e-02 -3.50645125e-01 9.58857536e-01
5.21589160e-01 -8.79338801e-01 6.15226567e-01 7.52084702e-02
-2.18422860e-01 -7.81435251e-01 -1.01772428e+00 -1.03373659e+00
1.90286964e-01 -1.75503567e-01 3.81227165e-01 1.32260513e+00
3.01158011e-01 5.91026068e-01 -2.74353057e-01 -6.01183809e-02
8.72688890e-01 3.97592485e-01 8.26021254e-01 -1.28317940e+00
-2.96566397e-01 -4.97727960e-01 -9.04941976e-01 -2.36887828e-01
5.89109957e-01 -1.63670886e+00 -4.94912356e-01 -1.83748996e+00
5.02540588e-01 5.32343276e-02 -6.32792771e-01 5.75173736e-01
-1.03637882e-01 1.02919810e-01 -6.30287547e-03 -6.79422989e-02
-1.12284827e+00 7.28493392e-01 1.19695401e+00 -4.55618739e-01
3.05380791e-01 -4.47547913e-01 -6.21824086e-01 2.55241483e-01
7.96149731e-01 -4.06505615e-01 -6.07164025e-01 -6.21476233e-01
2.05776896e-02 -4.22356613e-02 5.49934387e-01 -1.08136559e+00
1.69373885e-01 -1.27995253e-01 3.28903168e-01 1.01544298e-01
2.63991982e-01 -4.22893763e-01 -8.30410644e-02 3.43338072e-01
-4.38204557e-01 2.30402902e-01 1.92419544e-01 9.00481522e-01
-8.18935782e-02 1.30627736e-01 8.45676303e-01 -2.31081814e-01
-9.99190509e-01 7.53607631e-01 3.31339210e-01 7.21069336e-01
1.16292119e+00 1.43738249e-02 -9.32914555e-01 -2.99120605e-01
-8.75929415e-01 1.83914080e-01 6.88627481e-01 7.15485156e-01
5.73423266e-01 -1.64446008e+00 -8.16751361e-01 1.12348437e-01
8.83165121e-01 -5.98269522e-01 4.35560763e-01 5.70252419e-01
-3.12391758e-01 2.71318644e-01 -6.54982686e-01 -3.58059555e-01
-8.42779577e-01 8.46259892e-01 3.63813043e-01 -3.55905831e-01
-7.21976817e-01 7.61262178e-01 8.10634568e-02 -6.29578054e-01
1.26430159e-03 6.94488361e-02 2.64438484e-02 3.70525718e-02
4.08367008e-01 5.45487344e-01 2.51280963e-01 -3.09820235e-01
-5.23495138e-01 5.00436664e-01 -7.46565461e-02 2.30904013e-01
1.32075584e+00 9.27236602e-02 2.10957333e-01 5.85695148e-01
1.64558566e+00 -4.33273613e-01 -1.00269341e+00 -5.21944702e-01
1.24654144e-01 -5.58808446e-01 -5.12018688e-02 -7.27961302e-01
-1.10352242e+00 9.69399869e-01 4.50706571e-01 1.71340436e-01
8.66663337e-01 2.54190534e-01 4.38607961e-01 3.52326274e-01
6.15106225e-01 -9.68725324e-01 4.45771724e-01 5.69102883e-01
1.05365777e+00 -1.50670099e+00 -3.01397946e-02 -1.13007776e-01
-6.90373659e-01 8.82411838e-01 8.37953985e-01 -4.13928002e-01
8.07977557e-01 -6.36622086e-02 -4.02488649e-01 -4.05230075e-01
-1.35274899e+00 -3.87656808e-01 5.57078063e-01 9.67971981e-01
5.35472333e-01 1.78035766e-01 -8.69851708e-02 6.66595697e-01
7.84085244e-02 -2.21608460e-01 2.99283862e-01 8.31876993e-01
-1.90429434e-01 -9.86630738e-01 5.16162038e-01 6.92719519e-01
-4.00997922e-02 -2.83972509e-02 -5.52670896e-01 9.33348417e-01
-2.78701544e-01 5.67029595e-01 1.43049687e-01 -6.53268397e-01
3.96069109e-01 2.86553055e-01 8.02578866e-01 -9.13959622e-01
-7.03700259e-02 -6.61543250e-01 3.30657274e-01 -8.46498847e-01
-2.72647440e-02 -1.62619814e-01 -6.79579437e-01 -4.30839866e-01
1.72817633e-02 1.80002209e-03 2.33158082e-01 5.14418066e-01
8.79871845e-01 6.32033288e-01 5.36340296e-01 -6.34349167e-01
-4.88070965e-01 -9.43998694e-01 -4.10057008e-01 5.46878219e-01
2.42694631e-01 -9.96258080e-01 -2.91868418e-01 -3.57900918e-01] | [9.963494300842285, 3.0825862884521484] |
61cdf053-016c-4c2b-a596-b5a2a759a060 | pixel-level-equalized-matching-for-video | 2209.03139 | null | https://arxiv.org/abs/2209.03139v2 | https://arxiv.org/pdf/2209.03139v2.pdf | Pixel-Level Equalized Matching for Video Object Segmentation | Feature similarity matching, which transfers the information of the reference frame to the query frame, is a key component in semi-supervised video object segmentation. If surjective matching is adopted, background distractors can easily occur and degrade the performance. Bijective matching mechanisms try to prevent this by restricting the amount of information being transferred to the query frame, but have two limitations: 1) surjective matching cannot be fully leveraged as it is transformed to bijective matching at test time; and 2) test-time manual tuning is required for searching the optimal hyper-parameters. To overcome these limitations while ensuring reliable information transfer, we introduce an equalized matching mechanism. To prevent the reference frame information from being overly referenced, the potential contribution to the query frame is equalized by simply applying a softmax operation along with the query. On public benchmark datasets, our proposed approach achieves a comparable performance to state-of-the-art methods. | ['Sangyoun Lee', 'Chaewon Park', 'Minhyeok Lee', 'Seunghoon Lee', 'MyeongAh Cho', 'Woo Jin Kim', 'Suhwan Cho'] | 2022-09-04 | null | null | null | null | ['semi-supervised-video-object-segmentation'] | ['computer-vision'] | [ 0.4289758 0.07187465 -0.5807999 -0.39339277 -0.8473154 -0.50261086
0.528147 -0.06335836 -0.60399777 0.58947873 -0.08850454 -0.01882804
0.21327367 -0.6287555 -0.9672189 -0.65958136 0.07787567 0.12231856
0.81202894 0.2519071 0.33928034 0.24879882 -1.6237096 0.21430178
0.82571524 1.2713555 0.36371246 0.41677994 -0.32669815 0.86797774
-0.746272 -0.44752094 0.52508587 -0.40340164 -0.9609956 0.1724678
0.549124 -0.49676967 -0.610086 1.3444686 0.4314826 0.33507833
0.35271326 -1.360993 -0.47040674 0.41222587 -0.70470417 0.4344357
0.44676945 0.13337268 0.9468193 -0.85140073 0.7076755 1.0960307
0.17900528 0.47901288 -1.1377106 -0.9062852 0.33794594 0.43188012
-1.4542997 -0.6613715 0.82064915 -0.33678362 0.7140228 0.34869137
0.5657386 0.8369449 -0.1622955 1.0529124 0.70034105 -0.25932473
0.15675244 0.205911 -0.01625686 0.52153224 -0.20719449 0.20545931
-0.5778213 0.11601374 0.6510573 -0.22965719 -0.3829621 -0.44146383
-1.0743526 0.5294578 0.45749637 -0.00788717 -0.17667894 -0.06982219
0.42277 0.28121498 0.07026756 0.23012361 -0.165832 -0.35438675
-1.2034012 -0.03896128 0.4349625 1.3331486 0.76945275 -0.17530088
-0.6283318 0.8065623 0.27685344 0.26705596 0.49593636 -0.9002118
0.7037527 0.55615675 0.12036573 -0.7093345 -0.1277858 -0.18325259
-0.23372935 0.04566815 0.7667853 -0.04230575 -0.8763872 1.7854741
0.5060792 0.6059356 -0.11558297 1.2879606 1.052948 0.50568736
0.20499124 -0.20793295 1.2980015 -0.93673253 -0.5907355 -0.3827065
0.35754362 -1.0825068 1.0274947 0.11221735 -1.1655133 -0.6815404
-1.2504289 -0.13786292 0.01594931 -0.12312707 0.501221 0.5259006
-0.57874835 0.45523342 -0.87410885 -0.08213308 0.5363858 0.58146155
-0.44914874 -0.14199755 -1.2810471 0.66542774 0.44552112 0.01842962
-0.7250494 -0.5822822 -0.8882298 0.0847905 0.9045515 -0.42142805
1.1181039 -1.2993333 -1.6307256 0.9888513 -0.27925956 -0.356988
0.6456931 -0.19129479 -0.23302211 0.09234904 0.05734434 0.8997831
0.9992346 -0.99995464 -0.83894956 -0.17548586 0.3545408 0.46303087
-0.14318562 0.14308566 -1.3887346 -0.73750764 0.18168798 -1.0454963
0.07005267 0.03222927 -0.44856617 -0.09594391 0.7804089 -0.34420642
1.274564 -2.257509 -0.01408594 0.1538264 0.04208129 0.35017154
-0.23539448 -0.2959518 0.11361897 -0.11379807 -0.0428318 -0.14597173
-0.12715666 0.10862562 -0.04927181 0.58081335 0.2773505 0.83328366
-0.80189764 -0.7960299 0.30057424 0.1579456 -0.55818444 0.4026013
-0.06408601 0.18022667 -0.3860551 0.6050826 0.88989717 -0.1721429
-0.10268427 -0.39478257 0.17944033 0.4660122 -1.3665788 1.5847126
-0.11649497 0.6353284 -0.07867657 -1.1545122 0.76573026 0.19892552
0.6720333 -1.034878 0.3193203 -0.21826646 0.05123295 -0.52812165
0.42029616 0.25102368 0.15551353 0.2505456 0.08745122 0.15005381
0.3038389 -0.01554726 0.80030876 0.30577025 0.14276938 -0.2708995
0.71943957 -0.1771508 1.0221686 0.7661297 -0.5456261 0.6504521
0.5647392 -0.1061253 -0.718446 -0.9950575 -0.1303297 1.1264338
0.7892757 -0.09932826 -0.91214967 -0.9918479 -0.07469713 0.49117267
-0.5105905 -0.4210568 -0.6479506 -0.5901608 0.53288823 0.57613504
0.5428137 -0.90849423 -0.7115997 0.23471127 -0.3997346 -1.4048501
-1.0466683 0.08271737 -0.6791309 -1.1971818 -0.77258754 -0.6504695
0.7409132 0.5271673 1.1228251 0.04223874 -0.15433767 -0.04262261
-0.15436026 -0.14997469 -0.2638549 -0.07466115 -0.34113833 0.15273796
0.6948647 0.01736525 -0.7747964 0.73673785 -0.82679355 0.11872866
0.134243 0.7914948 0.81547576 -0.075541 0.50441176 -0.64986044
0.10009932 -0.2803862 -0.9073864 0.30872276 -0.39713776 0.1072095
0.3843986 -0.83616644 -1.0120844 0.12752199 0.04894057 -0.5369911
-0.08154388 0.05186403 -0.5371855 -0.18930668 0.33164367 0.19693255
-0.11171769 -0.13051052 0.2728717 0.6773138 0.6547211 -0.5429648
0.64509463 0.39900854 -0.18500896 -0.6098601 -0.6262744 -0.61259055
-0.52814084 -0.32178488 0.82070076 -0.88679034 -0.80781674 0.354733
-0.9526447 -0.11044458 -0.17288584 0.68095326 -0.508329 0.45391014
-0.46162087 -0.47422957 -0.06328114 -1.6503277 0.7868638 0.6030159
-0.28167772 -0.4846362 -0.43895984 0.5141144 0.31353414 -0.19717264
0.55041224 -0.78179276 -0.88728243 -0.2301323 -0.46515322 0.18513909
0.17503813 0.08503354 -0.9807612 -0.19590046 -0.07167525 -0.3012371
0.6472988 0.44465345 0.94733185 -0.00976627 -0.22407144 0.75475156
0.9458964 0.46433938 0.6535456 0.44255736 0.53475356 0.5653021
1.0028127 0.27331007 0.44287097 0.98828965 0.34308258 -0.13605952
-0.2404189 -0.19293542 0.2818073 0.22786184 0.21099216 -0.27201727
-0.42464456 0.27672797 -1.7841583 -0.9410436 0.15688989 2.863046
0.92907697 0.54965925 0.11199658 0.01856233 0.83409715 0.05976548
-0.7627648 0.04044941 -0.13062558 -0.15284614 0.63829124 0.5557205
-1.281958 1.1499829 6.0856895 0.81422806 -1.5129262 -0.06219586
0.5028043 -0.33667892 -0.06438056 -0.0502187 -0.82922006 0.63859844
0.47066736 -0.06936619 0.5422764 0.65031314 0.18560678 -0.31920233
-1.275131 1.1511782 -0.02568053 -1.0091754 -0.22958031 -0.16264771
0.35085607 -0.02531764 -0.10664919 0.36094528 -0.13274397 -0.67958385
0.9641609 0.3037304 0.64897114 -0.73719966 0.60779136 0.04583512
-1.2716094 0.21261092 -0.42457896 0.28122017 0.08114327 0.27849594
-0.7134602 0.34413582 0.7102627 0.35286236 -0.532245 1.2640487
-0.29191974 0.60799575 -0.4440918 0.31397545 0.13650638 -0.146774
0.67059857 1.250695 0.07072274 -0.25786343 0.3302053 0.84431505
-0.24558012 0.16743247 -0.24848703 0.10921801 0.71590644 0.9412097
-0.9187226 -0.46631068 -0.7205551 1.0696695 0.13441691 0.49676305
-1.0991985 -0.26343244 0.61423683 -0.04538507 0.27456412 0.16534805
-0.1916776 -1.1182829 0.21030581 -0.89451736 0.5568138 -0.5216803
-0.91073126 0.49990517 -0.02111398 -1.4497342 -0.3425297 -0.09735139
-0.43259794 0.72717124 -1.3942665 -0.8977212 -0.25480574 0.66793567
0.6141888 -0.01331222 0.44811237 0.61745733 -0.7002977 1.1879952
-0.44619396 0.15521872 1.0128961 -1.0498954 0.21833934 1.0863913
0.11083258 0.41771 0.67059296 -0.4612954 -1.2474595 -0.78661424
0.60772574 -0.14529136 0.3533879 -0.3738459 -1.2086288 0.3241845
-0.13285206 0.38893294 0.62140304 -0.12749141 -0.33976072 -0.16937587
-1.0945003 0.7233192 1.0032029 -0.6186762 -0.46589956 -0.09719464
0.76588464 -0.6276662 -0.81548005 0.5099058 0.7237158 -0.9094495
1.0014175 -0.33201417 0.12483187 -0.5177384 -0.1284806 -0.7693527
-0.07171219 -0.81109065 -0.23499897 1.4239916 0.3836921 -0.30583438
0.9801059 1.0624396 0.1405107 -0.66051364 -1.1394808 -0.78216714
-0.22826995 -0.5696541 0.7229992 0.86354196 0.00881505 0.2145016
-0.3782105 0.17336439 0.45813796 0.01778658 0.83647436 -0.7758165
-0.3646689 -0.6142419 -0.47492698 -1.4337914 0.21007292 -0.66707206
0.14416972 -0.8412181 0.2893791 -0.6259534 -0.46736374 0.2875693
-0.4680471 0.4106204 0.37520853 0.23368797 -0.5118569 0.37321618
1.3897305 -0.15348971 -0.48429355 0.26170996 -0.49703556 0.71561974
0.4019663 -0.50502515 -0.54223496 -0.34813395 -0.20458908 0.393361
0.13340391 -0.752512 0.4199202 -0.30728364 0.28996953 -0.43939376
0.3651334 -0.8951532 0.04236047 0.12431274 -0.39657387 -0.2710493
0.1891107 0.33932912 -0.47366187 -0.47140488 1.0013247 0.2924334
-0.6995244 0.35594538 -0.19805332 0.12087084 1.1871082 -0.6780334
-0.12255961 -0.32414728 -0.50262785 0.38377962 0.5491806 0.76060355
0.43688813 -1.1072261 -0.34631556 0.3417604 0.21518047 -0.08001532
0.22986841 0.8510658 -0.18093324 0.28187206 0.07448916 -0.6721416
-1.4317191 0.6783964 0.39828265 0.02247186 -0.61849153 0.79181653
0.49741018 0.04642141 0.68694025 -0.14877002 -0.1136979 0.01375331
0.64274853 0.32006896 0.04521172 -0.86974967 -0.46472967 0.5632197
-0.24493097 -0.05050215 0.6707169 -0.38827774 0.2910542 0.08285249
1.1780727 -0.06361838 -1.5963604 -0.5212832 0.01609071 -0.76509917
0.28661218 -0.52812475 -1.1841785 0.65910333 0.79390085 -0.09112249
1.1958711 -0.04040989 0.87870795 0.22809109 0.16357912 -1.2478154
-0.18433 0.2285289 0.52756363 -1.5442811 -0.11399098 -0.6447438
-0.69629914 0.9958038 0.85242784 0.04511443 0.5334675 0.45580643
0.11154221 0.0402322 -0.54064304 -0.31733024 0.6804748 0.38749793
0.38060594 -0.28682595 -0.196758 0.49179804 0.03894265 0.11623506
0.23612276 0.8737221 -0.30643272 -0.9383592 -0.45544282 0.37736785
-0.8539107 0.01894021 -0.07639612 0.7497595 0.06450441 1.0123245
0.2776422 -0.18500781 0.44156966 -0.06968857 0.4682358 -0.29319957
-0.33770314 0.42064345 -0.04560208 -0.80010074 -0.4957437 -0.45069832
-1.42907 -0.07167037 -0.74447745 -0.02640692 0.26080137 1.1228759
0.25070104 0.48513827 0.4849848 -0.8277094 -0.31976202 -0.6991322
-0.1728225 0.6706347 0.12371915 -0.85669243 -0.14291428 0.10262951] | [9.180779457092285, -0.11028552055358887] |
bdd2397e-f8f5-4a9d-9376-c2ee1c6ecbbc | continuously-index-domain-adaptation | null | null | https://icml.cc/virtual/2020/poster/5986 | http://wanghao.in/paper/ICML20_CIDA.pdf | Continuously Index Domain Adaptation | Existing domain adaptation focuses on transferring knowledge between domains with categorical indices (e.g., between datasets A and B). However, many tasks involve continuously indexed domains. For example, in medical applications, one often needs to transfer disease analysis and prediction across patients of different ages, where age acts as a continuous domain index. Such tasks are challenging for prior domain adaptation methods since they ignore the underlying relation among domains. In this paper, we propose the first method for continuously indexed domain adaptation. Our approach combines traditional adversarial adaptation with a novel discriminator that models the encoding-conditioned domain index distribution. Our theoretical analysis demonstrates the value of leveraging the domain index to generate invariant features across a continuous range of domains. Our empirical results show that our approach outperforms the state-of-the-art domain adaption methods on both synthetic and real-world medical datasets. | ['Hao Wang', 'Dina Katabi', 'Hao He'] | 2020-07-01 | null | null | null | icml-2020-7 | ['continuously-indexed-domain-adaptation'] | ['methodology'] | [ 6.76463425e-01 7.48980232e-03 -4.38361198e-01 -4.46925193e-01
-8.09865892e-01 -6.87347770e-01 3.89945060e-01 2.20011026e-01
-3.27262431e-01 1.21299338e+00 1.49736419e-01 -1.82799697e-01
-1.26878634e-01 -9.19099748e-01 -8.48575592e-01 -5.93791127e-01
-2.53831856e-02 8.89527082e-01 1.76017776e-01 -2.67085999e-01
-1.29467234e-01 2.05860078e-01 -8.68740916e-01 2.94179499e-01
1.17602098e+00 8.32757115e-01 -2.55096406e-01 4.47838306e-01
-2.97292834e-03 3.82149100e-01 -8.26988161e-01 -4.75656211e-01
3.18572432e-01 -7.37158358e-01 -7.51134217e-01 -2.31609583e-01
2.39005625e-01 -4.47589576e-01 -3.74975443e-01 9.72486079e-01
6.33920968e-01 1.10626221e-02 1.10564208e+00 -1.25888276e+00
-9.61165607e-01 3.27091604e-01 -3.97360146e-01 8.09459984e-02
4.12447155e-01 -5.07857390e-02 5.85928679e-01 -2.83604741e-01
7.45005846e-01 9.73153055e-01 7.62167633e-01 8.17962229e-01
-1.53406358e+00 -9.18205857e-01 6.17875904e-02 -4.21387404e-02
-1.07658470e+00 -6.75992444e-02 8.98293257e-01 -6.43674076e-01
4.17370856e-01 -7.48049766e-02 2.64297456e-01 1.54141283e+00
1.61109611e-01 5.99006295e-01 1.20170808e+00 -1.85159460e-01
5.47806025e-01 1.53994441e-01 -2.68870950e-01 2.76686519e-01
1.86920449e-01 2.38265038e-01 -3.01912904e-01 -5.55282593e-01
8.68056297e-01 2.02024296e-01 -8.29572454e-02 -6.77238524e-01
-1.42664933e+00 8.57570708e-01 2.23872706e-01 1.53226890e-02
-3.77963603e-01 -2.23038375e-01 6.35336637e-01 6.40583873e-01
5.87722063e-01 3.85756105e-01 -6.68039322e-01 3.38507444e-02
-7.01124847e-01 4.37001258e-01 8.75703037e-01 1.11975849e+00
4.56984788e-01 -2.79607356e-01 -1.78415328e-01 9.58081245e-01
-1.97047889e-01 5.51696718e-01 6.54461384e-01 -8.52667391e-01
4.73462880e-01 4.06811953e-01 2.55690932e-01 -6.16787136e-01
-2.16796398e-01 -1.58711612e-01 -1.15498006e+00 -2.41678245e-02
6.81550622e-01 -2.77173787e-01 -1.00922465e+00 2.13329816e+00
4.47699934e-01 4.57131088e-01 3.23274404e-01 5.37470281e-01
5.21632254e-01 1.94764599e-01 3.09399933e-01 3.54242027e-02
1.28551745e+00 -5.07677794e-01 -6.50990605e-01 -1.60785496e-01
4.21354204e-01 -4.23720479e-01 9.98315930e-01 2.31986478e-01
-8.47811103e-01 -3.89968753e-01 -9.40670848e-01 5.76652847e-02
-4.34671879e-01 -2.39960000e-01 4.69044626e-01 5.07730484e-01
-5.12395918e-01 5.28223455e-01 -8.79264235e-01 -5.64312458e-01
6.39152348e-01 3.91492933e-01 -5.23248136e-01 -2.11168259e-01
-1.70018053e+00 6.57502353e-01 5.24045825e-01 -8.56526494e-01
-6.83946490e-01 -1.11094308e+00 -7.95247257e-01 -1.37801692e-01
1.30786374e-01 -1.12568498e+00 1.08387804e+00 -1.15713811e+00
-1.44155657e+00 1.09551597e+00 6.39750808e-02 -6.35265410e-01
7.01226532e-01 -1.34966120e-01 -6.04708850e-01 1.56327099e-01
1.20366916e-01 4.14342821e-01 8.35124791e-01 -1.05549896e+00
-5.03821731e-01 -4.25317943e-01 1.48265921e-02 1.59687951e-01
-5.62621713e-01 -1.96592584e-01 -8.27742964e-02 -9.97602880e-01
-2.83428490e-01 -8.76740217e-01 -1.56093225e-01 4.28511918e-01
-2.55806655e-01 1.10487826e-01 5.67155302e-01 -7.72080958e-01
1.05969799e+00 -2.33339572e+00 2.00280547e-01 1.29695430e-01
1.52547732e-01 1.76959246e-01 -2.74388455e-02 3.52972537e-01
-1.66987255e-01 -7.24833533e-02 -7.89465129e-01 -4.66842428e-02
-5.40372450e-03 4.85637307e-01 -3.96098733e-01 3.72379929e-01
1.94526255e-01 7.16981649e-01 -1.12226045e+00 -6.20733619e-01
-1.89698055e-01 4.06684846e-01 -8.25656831e-01 3.39810103e-01
-1.38605729e-01 8.97474706e-01 -6.21005595e-01 4.70966488e-01
7.20575154e-01 -2.83979803e-01 2.53693372e-01 1.84351787e-01
6.45116687e-01 1.43870145e-01 -7.87495315e-01 2.11259270e+00
-3.01467299e-01 9.68085155e-02 -1.22095987e-01 -1.34182775e+00
9.50999081e-01 3.79542232e-01 7.80642867e-01 -4.64511007e-01
-1.94782928e-01 2.96480983e-01 1.21946335e-02 -2.09199920e-01
3.02224513e-02 -6.81367993e-01 -5.68556905e-01 3.41020823e-01
-8.58224090e-03 -7.44948536e-02 -1.63294002e-01 -1.75721217e-02
1.31808352e+00 4.54822630e-02 6.66099608e-01 -4.31282818e-02
5.52915275e-01 9.90825221e-02 8.23134422e-01 6.43666267e-01
-4.67706859e-01 7.68758893e-01 5.87966561e-01 -3.40148151e-01
-1.13776410e+00 -1.46877074e+00 -3.49063098e-01 9.87973273e-01
2.45746560e-02 3.30072820e-01 -6.65788889e-01 -1.09841204e+00
4.75444108e-01 5.11978388e-01 -9.24969196e-01 -6.48711264e-01
-4.93464887e-01 -5.82924426e-01 8.44321907e-01 7.57073045e-01
3.47152591e-01 -8.45194161e-01 -2.67125428e-01 2.97220379e-01
-2.26902068e-01 -9.95501518e-01 -5.98836422e-01 6.40460551e-02
-1.05852282e+00 -9.92166936e-01 -1.11960053e+00 -9.38125193e-01
7.08056211e-01 -4.74351197e-01 1.40980375e+00 -4.45706964e-01
-5.40070375e-03 3.38863581e-01 -2.74188936e-01 -5.09315014e-01
-8.49274278e-01 3.32459927e-01 2.05098465e-01 -3.58107798e-02
5.30976892e-01 -8.52654934e-01 -8.02359521e-01 3.68406892e-01
-1.14068198e+00 -2.32253149e-01 4.84927267e-01 1.31927049e+00
7.11541116e-01 -2.13277191e-01 1.03454471e+00 -1.55597043e+00
7.41794705e-01 -8.68869007e-01 -2.68846840e-01 3.38365883e-01
-4.95328099e-01 1.65339351e-01 9.52020109e-01 -8.82981837e-01
-9.70612466e-01 5.16872779e-02 2.97128167e-02 -4.79973942e-01
-4.93779570e-01 3.73488605e-01 -3.19419593e-01 3.16235662e-01
9.01554644e-01 3.29829931e-01 2.83000350e-01 -4.50194150e-01
2.56411493e-01 7.04777241e-01 8.40562642e-01 -8.23899627e-01
8.11787307e-01 6.01638734e-01 -9.79037508e-02 -2.71813095e-01
-6.03112817e-01 -3.01084638e-01 -7.48194754e-01 5.25143862e-01
7.14066684e-01 -9.46653605e-01 -2.13581324e-01 5.26594520e-01
-8.06562245e-01 -5.22529125e-01 -5.85547388e-01 4.41651046e-01
-9.87264931e-01 3.13224822e-01 -4.53309625e-01 -1.05448388e-01
-3.00346971e-01 -8.19785237e-01 9.94828463e-01 -1.50164291e-01
-3.67880613e-01 -1.46211183e+00 4.42159235e-01 3.99596617e-02
4.18120086e-01 6.59353375e-01 1.11302066e+00 -1.03478479e+00
1.42426789e-02 -1.83845669e-01 -1.18634617e-02 3.49111885e-01
6.81745410e-01 -6.20791256e-01 -6.59490943e-01 -4.38341409e-01
-2.06013441e-01 -4.78104979e-01 6.61604345e-01 1.79213077e-01
1.47513795e+00 -3.42127919e-01 -4.74424958e-01 8.86889219e-01
1.22302616e+00 3.44784379e-01 5.54232240e-01 2.31087431e-01
3.61886144e-01 3.65746707e-01 7.45253205e-01 5.60724676e-01
2.65876204e-01 6.57525957e-01 1.53915130e-03 -3.35716158e-01
1.62649562e-03 -3.98029089e-01 9.92801785e-02 3.62054348e-01
1.99031964e-01 -1.06599838e-01 -9.81735170e-01 7.83089936e-01
-1.63606012e+00 -7.99963355e-01 8.46088409e-01 2.33676267e+00
1.47118318e+00 3.75075415e-02 2.90242106e-01 -1.09313421e-01
7.13030279e-01 -2.51513332e-01 -1.22067773e+00 -3.52456540e-01
6.17860854e-02 5.13437152e-01 4.94194955e-01 1.43186510e-01
-1.32471883e+00 5.32020211e-01 6.50799465e+00 6.02929115e-01
-9.35132563e-01 9.37853977e-02 5.24759412e-01 8.21837559e-02
-2.37191901e-01 -4.05683845e-01 -3.50833178e-01 7.20241904e-01
9.38716590e-01 -6.34689748e-01 3.26161414e-01 8.55502784e-01
-3.99571747e-01 4.07095373e-01 -1.45546639e+00 6.50099158e-01
-8.46658424e-02 -8.82548213e-01 7.46817961e-02 1.59466445e-01
9.26876307e-01 -3.24389398e-01 4.22467947e-01 3.62915158e-01
6.72880292e-01 -9.84708071e-01 1.03617251e-01 4.71290559e-01
1.43446660e+00 -7.66946077e-01 5.82917631e-01 3.00233573e-01
-7.99247801e-01 -2.07895190e-02 -2.08106250e-01 2.30220690e-01
2.59976462e-02 4.97091144e-01 -9.30850565e-01 6.21328771e-01
4.09438789e-01 8.87139916e-01 -1.22504339e-01 7.19971895e-01
5.89277446e-02 6.31548047e-01 -2.16827378e-01 5.52548528e-01
-2.97139704e-01 1.21199880e-02 4.25997496e-01 1.11087775e+00
4.18475598e-01 1.62535638e-01 1.44938633e-01 7.29834080e-01
-3.08148623e-01 -1.74921572e-01 -1.01799417e+00 -3.31334770e-02
6.64375126e-01 4.92870539e-01 -2.35636726e-01 -4.78164583e-01
-5.25622725e-01 1.28393638e+00 1.64519474e-01 3.36360157e-01
-9.53534305e-01 -5.77378690e-01 1.15939486e+00 1.87425360e-01
2.96236038e-01 9.85899121e-02 -4.10763294e-01 -1.16359150e+00
-5.48348352e-02 -1.11760950e+00 8.19340825e-01 -2.23924175e-01
-1.97569811e+00 4.05906230e-01 1.50912598e-01 -1.59820676e+00
-4.36423004e-01 -4.36703622e-01 -2.80993164e-01 1.02462554e+00
-1.57965887e+00 -1.14163029e+00 -1.02588177e-01 1.08896172e+00
1.57017142e-01 -3.61278236e-01 1.24322379e+00 3.57735664e-01
-1.32166654e-01 1.19479358e+00 6.83491647e-01 4.47768688e-01
1.33216500e+00 -1.39678872e+00 4.81537223e-01 3.40472281e-01
-4.23678070e-01 6.78272069e-01 4.72260207e-01 -6.98420823e-01
-8.84603679e-01 -1.35377467e+00 5.95602691e-01 -4.72761393e-01
5.32737970e-01 -3.36566061e-01 -1.26957357e+00 7.36808896e-01
-6.46252632e-02 2.22268030e-01 1.10580206e+00 1.86493620e-02
-6.07453585e-01 -2.54683971e-01 -1.71184599e+00 3.66218537e-01
1.06268394e+00 -5.48092842e-01 -7.75332212e-01 3.78397137e-01
7.34007478e-01 -5.58793962e-01 -1.48849416e+00 4.82022703e-01
5.99778771e-01 -4.90807235e-01 1.39267182e+00 -1.05001640e+00
6.75043523e-01 4.52354699e-02 -7.83418268e-02 -1.63267422e+00
-2.31122613e-01 -3.17819476e-01 -1.99391350e-01 1.06633723e+00
2.23883644e-01 -1.07740569e+00 7.66112208e-01 6.52080953e-01
3.40577602e-01 -4.23906177e-01 -9.88843560e-01 -1.02260077e+00
7.84488678e-01 1.00353993e-01 9.93631244e-01 1.30843592e+00
-5.00166640e-02 8.55296291e-03 -2.81661183e-01 1.27043486e-01
6.08862936e-01 1.04335345e-01 7.18640089e-01 -1.28896272e+00
-6.03107631e-01 -1.23646684e-01 -5.22139728e-01 -7.67391026e-01
3.47886056e-01 -8.28255594e-01 -1.57953560e-01 -1.12040830e+00
2.39526644e-01 -7.92086720e-01 -7.78282881e-01 4.94351119e-01
-4.10607725e-01 7.02641308e-02 -1.29626647e-01 1.76851228e-01
-2.72362530e-01 3.73923004e-01 1.41428792e+00 -3.34864289e-01
-1.49784684e-01 1.34893253e-01 -9.26854789e-01 4.98129129e-01
1.06042707e+00 -6.63088381e-01 -6.83981657e-01 -2.34543681e-01
-2.13550240e-01 2.14806274e-01 2.68428206e-01 -9.39246833e-01
-1.85754858e-02 -5.29934406e-01 4.47428554e-01 -1.08556300e-01
1.94334060e-01 -8.71302664e-01 1.50793567e-01 4.88913387e-01
-4.85621065e-01 -1.77150562e-01 3.36994976e-01 8.23319077e-01
-3.21348667e-01 2.99285591e-01 9.60011184e-01 -9.55150351e-02
-4.90449816e-01 4.13200110e-01 1.05636828e-01 6.67608202e-01
1.12639940e+00 -4.06740680e-02 -2.37075239e-01 -2.71164387e-01
-9.50664043e-01 1.83174685e-01 7.94586718e-01 3.19939017e-01
5.38659573e-01 -1.44742012e+00 -8.90279114e-01 4.21491742e-01
3.67961764e-01 1.79172799e-01 1.17368624e-01 4.69941288e-01
-2.65662134e-01 8.46781954e-02 -5.57353973e-01 -5.13498664e-01
-1.10986423e+00 8.68285894e-01 3.11474085e-01 -6.50083482e-01
-5.03479779e-01 6.14669144e-01 6.34487808e-01 -7.35744476e-01
1.34080544e-01 -2.08492517e-01 4.83344495e-02 -1.62694380e-01
3.66208017e-01 1.20846689e-01 -1.27856836e-01 -2.32498512e-01
-4.24642980e-01 4.80689973e-01 -2.21047387e-01 -9.72699523e-02
1.19888628e+00 1.40607759e-01 2.78102666e-01 3.93000782e-01
1.09984720e+00 -2.96190351e-01 -1.33987546e+00 -6.61382079e-01
-2.44478896e-01 -4.76095140e-01 -6.88131452e-01 -1.18111444e+00
-8.48008573e-01 9.24832225e-01 6.76985383e-01 -1.46246180e-01
1.58862031e+00 2.26962790e-02 9.49904978e-01 1.67897180e-01
4.31005031e-01 -9.28374469e-01 3.97856943e-02 3.66818070e-01
7.50252008e-01 -1.27271950e+00 -1.24682225e-02 -2.84660459e-01
-8.93573344e-01 5.87725282e-01 4.87562835e-01 -1.62431568e-01
6.71190083e-01 2.37971976e-01 1.20034978e-01 3.81616920e-01
-5.47415316e-01 6.22831173e-02 2.05885321e-01 1.20729721e+00
2.89187133e-01 3.86852294e-01 -3.11888397e-01 8.61692429e-01
-1.65845200e-01 4.42155153e-01 1.67514250e-01 1.03605545e+00
1.32910252e-01 -1.69383442e+00 -4.33874041e-01 5.42553127e-01
-7.10349441e-01 -7.86992488e-04 -3.20865601e-01 7.61507571e-01
1.51127279e-01 3.89437854e-01 9.38734114e-02 -1.98087186e-01
4.75588292e-01 3.64299268e-01 4.27457184e-01 -6.16173685e-01
-3.44614744e-01 -4.14147586e-01 -3.20155680e-01 -2.79309154e-01
-3.37532818e-01 -9.02090013e-01 -1.18816996e+00 -2.17014775e-01
3.77426356e-01 -1.20541900e-02 2.34985918e-01 5.66914439e-01
5.84571242e-01 5.48160374e-01 5.17544448e-01 -1.79753393e-01
-9.30639446e-01 -7.90759563e-01 -7.10651577e-01 1.12402892e+00
6.32898211e-01 -8.45884323e-01 1.44975595e-02 4.73677814e-01] | [10.35343074798584, 3.112499713897705] |
e8c2d446-70ff-4bfa-b1a1-8075325b6020 | extracting-pico-elements-from-rct-abstracts | 1901.08351 | null | http://arxiv.org/abs/1901.08351v1 | http://arxiv.org/pdf/1901.08351v1.pdf | Extracting PICO elements from RCT abstracts using 1-2gram analysis and multitask classification | The core of evidence-based medicine is to read and analyze numerous papers in
the medical literature on a specific clinical problem and summarize the
authoritative answers to that problem. Currently, to formulate a clear and
focused clinical problem, the popular PICO framework is usually adopted, in
which each clinical problem is considered to consist of four parts:
patient/problem (P), intervention (I), comparison (C) and outcome (O). In this
study, we compared several classification models that are commonly used in
traditional machine learning. Next, we developed a multitask classification
model based on a soft-margin SVM with a specialized feature engineering method
that combines 1-2gram analysis with TF-IDF analysis. Finally, we trained and
tested several generic models on an open-source data set from BioNLP 2018. The
results show that the proposed multitask SVM classification model based on
1-2gram TF-IDF features exhibits the best performance among the tested models. | ['Wu Jinfa', 'Xia Yuan', 'Shi Qinwen', 'Li Ke', 'Li Shilei', 'Liao xiaoli'] | 2019-01-24 | null | null | null | null | ['pico'] | ['natural-language-processing'] | [ 2.53252298e-01 -2.66075313e-01 -7.25478947e-01 -3.45003068e-01
-7.33608961e-01 9.48846992e-03 4.58268136e-01 7.91121840e-01
-2.73960471e-01 1.00295281e+00 9.58747864e-02 -4.10260201e-01
-6.88659489e-01 -3.06145966e-01 -3.13429207e-01 -8.13376188e-01
-1.54996822e-02 4.91788596e-01 8.52569938e-02 1.53702021e-01
5.33147037e-01 3.70854177e-02 -1.52835405e+00 5.78419864e-01
1.23167455e+00 1.02272642e+00 3.09740603e-01 1.39970109e-01
-1.36578456e-01 5.31574547e-01 -4.61851329e-01 -1.49789110e-01
-2.60990560e-01 -4.76626486e-01 -9.47950900e-01 -3.49199057e-01
-6.80465475e-02 3.00799727e-01 2.68464267e-01 8.68563116e-01
5.79427481e-01 -3.67045030e-02 1.09492755e+00 -1.35836124e+00
-5.12677729e-01 3.88678730e-01 -2.49068782e-01 4.12251055e-01
3.88706177e-01 -2.51504958e-01 6.72365308e-01 -7.21039414e-01
4.74984258e-01 8.98220479e-01 5.09222090e-01 2.50677377e-01
-9.08355534e-01 -6.43025339e-01 -1.83486149e-01 5.86151719e-01
-9.71965194e-01 -5.02387695e-02 4.35965747e-01 -8.50622475e-01
7.48159647e-01 3.52939665e-01 5.91736972e-01 1.09323859e+00
9.92206872e-01 5.74349165e-01 1.43759680e+00 -3.96889091e-01
2.82279849e-01 2.60318667e-01 5.51895320e-01 7.13859618e-01
3.36892873e-01 -2.30375499e-01 -4.48828638e-01 -5.87516189e-01
2.09005773e-01 1.58372000e-01 -2.21271798e-01 -2.07264312e-02
-1.21975636e+00 8.08583677e-01 1.15308277e-01 5.05376995e-01
-4.01972771e-01 -3.37164819e-01 7.76538849e-01 5.02615012e-02
4.95319963e-01 2.94691980e-01 -8.95899594e-01 -2.35592257e-02
-9.45414186e-01 2.02860832e-01 8.35016251e-01 4.42247182e-01
2.34427050e-01 -3.96863520e-01 -4.27239716e-01 1.03028035e+00
2.98986375e-01 2.68569261e-01 9.62714195e-01 -5.26674688e-01
2.26314917e-01 5.35178840e-01 -1.78630333e-02 -1.02910173e+00
-7.12914705e-01 -5.71123242e-01 -8.06683540e-01 -1.01842351e-01
1.59866676e-01 -7.93001279e-02 -7.95349121e-01 1.25249660e+00
4.45579708e-01 2.80324042e-01 3.69221494e-02 8.18767250e-01
1.49053526e+00 2.79082119e-01 4.62948650e-01 -5.12688458e-01
1.80877280e+00 -1.23324907e+00 -1.32869971e+00 2.30598748e-02
8.53655100e-01 -7.79250085e-01 8.32107186e-01 7.54169106e-01
-5.84759414e-01 -3.43801081e-01 -9.57023203e-01 -3.11418455e-02
-6.80785656e-01 3.67779195e-01 5.54016590e-01 3.89858693e-01
-3.26146305e-01 6.90117240e-01 -4.37186033e-01 -3.92737210e-01
3.59162778e-01 1.27252564e-01 -6.04609787e-01 7.73605183e-02
-1.47127247e+00 1.42994666e+00 5.77594757e-01 -9.10652354e-02
-5.63963473e-01 -8.23226392e-01 -4.95564431e-01 -1.65885925e-01
3.04622173e-01 -8.76733840e-01 9.29069459e-01 -4.78364289e-01
-1.18993199e+00 9.06653225e-01 -1.65051371e-01 -5.46195917e-02
2.38666818e-01 -2.01375410e-01 -5.56104481e-01 1.96594089e-01
2.32498422e-01 -2.33658440e-02 5.80299735e-01 -8.87127280e-01
-6.85209811e-01 -5.11130929e-01 -4.17662948e-01 8.27860981e-02
-3.73164117e-01 3.50853086e-01 -4.64030216e-03 -6.89947605e-01
-6.91476837e-02 -7.78548360e-01 -5.76572791e-02 -2.25075006e-01
-4.13953573e-01 -4.89051282e-01 5.86060822e-01 -7.92567194e-01
1.49323177e+00 -1.82184184e+00 2.83506930e-01 2.12893654e-02
3.04790676e-01 4.48887557e-01 3.55135918e-01 4.50303674e-01
-4.17638481e-01 1.42395005e-01 -1.13568828e-01 3.58624160e-02
-4.73153353e-01 4.79985550e-02 1.55724198e-01 5.63297331e-01
-2.66308133e-02 8.02919805e-01 -8.50846112e-01 -1.04297078e+00
1.48994997e-01 1.98178902e-01 -1.63414672e-01 1.19778514e-01
-1.45002892e-02 5.20294547e-01 -8.13577235e-01 8.75503361e-01
4.24554080e-01 -6.29586875e-01 6.98000118e-02 -4.54387844e-01
2.43236573e-04 2.72694621e-02 -7.08732724e-01 1.65515351e+00
-2.61117041e-01 3.77338082e-02 -6.56585321e-02 -1.51304793e+00
7.26242602e-01 6.07648134e-01 9.45904732e-01 -2.02565596e-01
3.72795284e-01 4.80612576e-01 2.18325123e-01 -1.15118456e+00
-2.84018964e-01 -2.56309807e-01 6.90468848e-02 7.31608719e-02
1.56126946e-01 2.64698267e-01 -2.55365539e-02 -2.49799043e-01
1.00326943e+00 -5.90832122e-02 8.93599391e-01 -6.47523880e-01
7.22299337e-01 1.20427266e-01 6.17339849e-01 3.67565006e-01
-5.09111643e-01 3.08298796e-01 4.36798394e-01 -4.03806239e-01
-4.49098796e-01 -7.09434688e-01 -8.66909564e-01 8.69554877e-01
-6.44706041e-02 -4.32227284e-01 -5.96622705e-01 -8.06929767e-01
2.31015041e-01 4.88563001e-01 -8.91923726e-01 -6.01500347e-02
-1.80770829e-01 -1.11771297e+00 4.10367072e-01 9.86377150e-02
4.17427599e-01 -1.05282164e+00 -6.33915842e-01 1.82968915e-01
-3.06288332e-01 -8.20687413e-01 -3.28111589e-01 5.14053524e-01
-9.16970968e-01 -1.31617510e+00 -6.49576604e-01 -9.13375556e-01
3.04139286e-01 -3.81482020e-02 5.03211796e-01 -2.28547417e-02
-5.64605415e-01 -1.59360558e-01 -6.48589909e-01 -7.31568933e-01
-1.26988024e-01 2.65745908e-01 -2.23440640e-02 -7.29341283e-02
4.17902559e-01 -2.65847325e-01 -4.85599697e-01 1.74844310e-01
-7.45896399e-01 1.82495579e-01 8.78701508e-01 1.10194731e+00
5.79812706e-01 -3.55951458e-01 1.00653601e+00 -1.12741566e+00
7.70637810e-01 -9.09343839e-01 9.88072604e-02 5.95777631e-01
-1.03034449e+00 4.49872538e-02 6.38285637e-01 -5.23298085e-01
-8.17977667e-01 -2.07981944e-01 -1.86315119e-01 -2.11245075e-01
-1.15669288e-01 1.21314085e+00 8.77029076e-02 -1.99280903e-02
7.77493834e-01 2.42188454e-01 2.66545057e-01 -5.45409083e-01
-6.82467446e-02 1.05657232e+00 1.03449687e-01 -5.79537034e-01
1.11213461e-01 -1.21699804e-02 3.05247515e-01 -5.91569066e-01
-1.09422898e+00 -4.73141730e-01 -3.28473657e-01 -1.57741040e-01
1.08432508e+00 -6.44659460e-01 -8.47784817e-01 3.15244526e-01
-8.54165018e-01 1.34810343e-01 3.11503351e-01 9.56099987e-01
-4.86188322e-01 3.15474272e-01 -4.73026633e-01 -4.47644204e-01
-5.69203079e-01 -1.41074884e+00 9.81195509e-01 1.94204688e-01
-3.00681800e-01 -1.11971927e+00 9.64249671e-02 5.85364461e-01
3.94280136e-01 4.17127728e-01 1.36941302e+00 -1.13308215e+00
3.32204670e-01 -3.26841772e-02 -8.30649734e-02 1.83711484e-01
5.29700518e-01 -2.04255786e-02 -6.99506402e-01 -5.77949453e-03
3.01745474e-01 -2.73586899e-01 7.30781853e-01 5.33629596e-01
1.57736993e+00 -2.44941786e-01 -8.06746304e-01 4.94565040e-01
1.23008478e+00 4.17372525e-01 4.01805192e-01 4.63737994e-01
3.07119310e-01 5.47818422e-01 1.00096965e+00 3.53917897e-01
2.28500172e-01 6.05697393e-01 2.07378328e-01 -7.63592198e-02
1.63063824e-01 8.99479389e-02 -1.18252091e-01 8.60290468e-01
-2.43356228e-01 1.01548441e-01 -1.05691075e+00 2.81043857e-01
-2.02784824e+00 -6.35880113e-01 -3.75820696e-01 1.86675370e+00
1.12585247e+00 3.58006246e-02 1.08340420e-01 2.12727904e-01
7.20073104e-01 -2.75088936e-01 -3.14692110e-01 -4.65714514e-01
1.08629815e-01 3.59366506e-01 2.76556641e-01 1.14917636e-01
-1.30208731e+00 5.36903322e-01 6.91072750e+00 1.15932393e+00
-1.31954420e+00 3.71601671e-01 5.76034069e-01 4.87889908e-02
9.05137435e-02 -2.54351825e-01 -6.34937942e-01 8.03957224e-01
1.04710436e+00 -4.83220994e-01 -1.56440824e-01 9.55578387e-01
4.91371721e-01 -2.85173833e-01 -1.07230330e+00 9.79145944e-01
1.01775356e-01 -1.29089296e+00 -3.37984376e-02 -4.56183217e-03
3.53807330e-01 -2.86408633e-01 3.47586796e-02 1.20053537e-01
-2.80479610e-01 -1.28037322e+00 1.85785010e-01 7.03649223e-01
7.70515382e-01 -1.60510987e-02 8.90947580e-01 5.47180891e-01
-7.62887299e-01 -1.71082765e-01 -2.84090471e-02 3.87529172e-02
-4.27733399e-02 9.38261688e-01 -6.59985483e-01 1.22420573e+00
7.33455062e-01 1.01298201e+00 -3.35045010e-01 1.30527616e+00
-3.37570608e-02 6.94875479e-01 1.31058723e-01 -2.53587335e-01
-7.74499327e-02 -2.24675983e-01 2.08528861e-01 1.27097893e+00
3.67668450e-01 1.41337588e-01 2.97851950e-01 3.61163110e-01
3.86713713e-01 8.37119937e-01 -3.31998020e-01 1.02191903e-01
3.44382226e-01 1.40176165e+00 -6.49689198e-01 -4.85383600e-01
-4.06630427e-01 5.58305860e-01 -2.71419412e-03 -3.02744731e-02
-1.00281096e+00 -4.13383514e-01 3.52432400e-01 -2.76105180e-02
-7.10594803e-02 4.09799069e-01 -4.75425541e-01 -1.19108522e+00
-1.11792803e-01 -1.02446580e+00 6.14127457e-01 -5.50799310e-01
-1.49054408e+00 4.47898865e-01 2.41148025e-01 -1.34318089e+00
1.77354112e-01 -6.74902976e-01 -3.34407121e-01 1.05405402e+00
-1.41270614e+00 -1.08836758e+00 -1.99690089e-01 3.88280958e-01
4.02234614e-01 -2.68251806e-01 1.10572958e+00 6.07090354e-01
-8.30637872e-01 3.84292603e-01 2.41618872e-01 -1.41120315e-01
1.02965009e+00 -9.91563261e-01 -7.79149473e-01 -4.95679006e-02
-6.02303624e-01 8.12615097e-01 7.29176223e-01 -8.21425796e-01
-1.14806342e+00 -9.63349402e-01 1.16974485e+00 -2.00834632e-01
5.46858013e-01 2.21305609e-01 -8.68887007e-01 5.09996772e-01
2.20985506e-02 1.69058383e-01 1.20274270e+00 1.01707064e-01
7.09220301e-03 -1.59199134e-01 -1.49795759e+00 6.78732991e-02
6.98534191e-01 -7.15092346e-02 -1.02852249e+00 6.41252995e-01
3.27897757e-01 -3.45607996e-01 -1.43723142e+00 6.75874531e-01
7.50078201e-01 -4.19950306e-01 8.94102931e-01 -1.03643095e+00
8.25123191e-01 -1.99889675e-01 2.69098356e-02 -1.35721707e+00
-3.30917358e-01 -2.01213900e-02 -1.10164419e-01 6.80655003e-01
4.06922460e-01 -7.89705992e-01 4.36590940e-01 2.24733889e-01
-3.49742472e-01 -1.40237105e+00 -1.05420303e+00 -6.65033638e-01
2.86242247e-01 4.00887206e-02 2.08882794e-01 1.25303555e+00
5.87010443e-01 3.19292128e-01 -2.63757050e-01 -1.94040075e-01
4.27504629e-01 1.78245589e-01 1.24519080e-01 -1.65501010e+00
-1.68115988e-01 -3.09110701e-01 -2.00827897e-01 -2.46433347e-01
1.68478996e-01 -1.28868186e+00 -1.65154323e-01 -1.92276454e+00
4.80288655e-01 -4.00892884e-01 -6.42239273e-01 6.46463096e-01
-5.22752404e-01 -3.34892750e-01 -2.84918338e-01 2.35789001e-01
-2.21192166e-01 3.09990287e-01 1.31517422e+00 -1.06369466e-01
3.39889410e-03 -3.58121656e-02 -7.95408845e-01 6.42334521e-01
5.99178374e-01 -7.23411739e-01 -3.08070391e-01 2.01325417e-02
-8.25416818e-02 3.27477396e-01 -2.52606790e-03 -7.40344048e-01
1.30470231e-01 -5.81148863e-01 3.52144808e-01 -4.11135286e-01
1.91722199e-01 -6.69806838e-01 3.02307755e-01 9.71569002e-01
-4.66467112e-01 -1.21493393e-03 6.09219969e-05 3.36386442e-01
-3.06781590e-01 -3.33825678e-01 6.13486052e-01 2.36452417e-03
-2.59828568e-01 1.92112491e-01 -2.87709445e-01 -2.54858863e-02
1.36177981e+00 1.23858780e-01 -4.74096805e-01 3.08379114e-01
-1.06371045e+00 2.78062403e-01 -2.49578252e-01 3.35357666e-01
5.20202458e-01 -1.24508595e+00 -8.98830831e-01 -3.47422957e-01
3.52960497e-01 -5.65621018e-01 3.57729286e-01 1.33780134e+00
-4.20045882e-01 7.34509468e-01 -2.95916945e-01 -5.15341997e-01
-1.54547834e+00 7.42590010e-01 2.12624684e-01 -4.62592810e-01
-4.32895869e-01 3.98227930e-01 -9.31647643e-02 -2.96357542e-01
1.72445223e-01 -3.73349190e-01 -8.24676037e-01 3.86612564e-01
3.59447300e-01 4.14205402e-01 4.29938734e-01 -4.51734394e-01
-7.43530452e-01 6.10299230e-01 -2.30209082e-02 1.49938077e-01
1.37886441e+00 4.67636406e-01 -4.52335894e-01 7.14896679e-01
1.25513148e+00 -3.64247352e-01 -6.57664686e-02 -1.44010305e-01
1.46170050e-01 -2.24645272e-01 2.72097923e-02 -1.04554963e+00
-7.58257389e-01 6.16062224e-01 5.75254261e-01 5.85843399e-02
1.02286673e+00 -1.70284554e-01 5.94887316e-01 2.73784429e-01
2.43533239e-01 -1.17272937e+00 -7.34840855e-02 1.84409007e-01
9.10846591e-01 -1.22818887e+00 1.91072285e-01 -8.19794059e-01
-6.91569149e-01 1.28641665e+00 4.19310510e-01 5.81442565e-03
1.38216126e+00 9.58001986e-02 -3.13674659e-02 -2.71263689e-01
-8.34766388e-01 1.22213587e-01 5.09996593e-01 2.07085833e-01
6.85181141e-01 7.38637522e-02 -1.50542378e+00 1.21208930e+00
-3.06564998e-02 7.11692631e-01 2.25353718e-01 8.49232256e-01
-4.79848027e-01 -1.31207657e+00 -3.10779721e-01 1.14075494e+00
-8.41240048e-01 -7.42858052e-02 1.25710219e-01 6.12510741e-01
4.56409127e-01 1.01921022e+00 -5.45192659e-01 -5.49597979e-01
1.05711609e-01 3.77494007e-01 3.56291562e-01 -7.58728206e-01
-8.41229618e-01 8.54752809e-02 1.69619143e-01 -4.00924802e-01
-7.40487576e-01 -6.96565807e-01 -1.33266282e+00 1.46964565e-01
-2.31026858e-01 3.93405586e-01 9.21230793e-01 1.40529788e+00
2.66669571e-01 8.05519044e-01 3.53097826e-01 -7.21367449e-02
-6.96469784e-01 -1.15970361e+00 -6.03657961e-01 3.64478588e-01
1.49867475e-01 -1.15156305e+00 -3.91458422e-01 2.97405366e-02] | [8.370064735412598, 8.572443008422852] |
f09934c4-940a-45b3-a0fd-0399ce540392 | iterative-residual-image-deconvolution | 1804.06042 | null | http://arxiv.org/abs/1804.06042v2 | http://arxiv.org/pdf/1804.06042v2.pdf | Iterative Residual Image Deconvolution | Image deblurring, a.k.a. image deconvolution, recovers a clear image from
pixel superposition caused by blur degradation. Few deep convolutional neural
networks (CNN) succeed in addressing this task. In this paper, we first
demonstrate that the minimum-mean-square-error (MMSE) solution to image
deblurring can be interestingly unfolded into a series of residual components.
Based on this analysis, we propose a novel iterative residual deconvolution
(IRD) algorithm. Further, IRD motivates us to take one step forward to design
an explicable and effective CNN architecture for image deconvolution.
Specifically, a sequence of residual CNN units are deployed, whose intermediate
outputs are then concatenated and integrated, resulting in concatenated
residual convolutional network (CRCNet). The experimental results demonstrate
that proposed CRCNet not only achieves better quantitative metrics but also
recovers more visually plausible texture details compared with state-of-the-art
methods. | ['Qian Yin', 'Zijian Hu', 'Li Si-Yao', 'Junfeng Li', 'Furong Zhao', 'Dongwei Ren'] | 2018-04-17 | null | null | null | null | ['image-deconvolution'] | ['computer-vision'] | [ 5.42672992e-01 -7.02522993e-02 3.08571070e-01 -1.71160400e-01
-4.08595353e-01 -2.99798429e-01 3.67307812e-01 -7.20317066e-01
-1.65132564e-02 7.37113535e-01 4.88448769e-01 -2.56977439e-01
5.58549091e-02 -2.01461062e-01 -7.68042147e-01 -8.37789893e-01
5.61195388e-02 -7.41154909e-01 3.81240994e-02 5.34059666e-02
3.89167100e-01 3.18108320e-01 -1.22160161e+00 1.32301137e-01
1.15604436e+00 1.12010324e+00 3.91425878e-01 5.95469296e-01
4.09130782e-01 1.17772365e+00 -3.67846847e-01 -2.50325888e-01
1.24860048e-01 -5.42688727e-01 -6.61914051e-01 2.46882066e-01
5.22160769e-01 -1.02638113e+00 -8.16636741e-01 1.46360779e+00
4.67133880e-01 8.62252712e-02 6.17671549e-01 -7.74550736e-01
-1.46280694e+00 3.48481596e-01 -6.54045880e-01 1.52879745e-01
-1.14046507e-01 2.27916494e-01 5.70740700e-01 -1.14453340e+00
3.00638467e-01 1.23866189e+00 9.04864848e-01 4.49224085e-01
-1.17162108e+00 -5.58215201e-01 -8.42391476e-02 2.43568778e-01
-1.33398020e+00 -5.84999979e-01 8.13046396e-01 -1.43971309e-01
4.62360024e-01 2.12051675e-01 1.81757897e-01 9.16113615e-01
1.87797040e-01 8.98054600e-01 1.31445301e+00 -2.43938968e-01
-8.77345446e-03 -2.22858131e-01 1.75771371e-01 5.37633181e-01
5.55329204e-01 3.19834262e-01 -2.39350274e-01 2.94117570e-01
1.10368574e+00 1.42729968e-01 -1.04417133e+00 1.26645759e-01
-1.17421067e+00 3.07619184e-01 8.02198470e-01 9.93826017e-02
-4.71559972e-01 4.91212070e-01 2.13131309e-02 1.02576911e-01
5.76380432e-01 2.35034928e-01 -1.57662958e-01 3.36882710e-01
-1.15140975e+00 1.06855258e-01 3.14154804e-01 7.23130405e-01
7.93503940e-01 2.30665430e-01 -3.08844745e-01 1.00630236e+00
2.87408292e-01 4.87723380e-01 3.92141521e-01 -1.18342352e+00
2.41191849e-01 1.38157949e-01 3.27969909e-01 -1.09843290e+00
-1.58312038e-01 -5.64853311e-01 -1.57451177e+00 3.83454949e-01
3.85415591e-02 -2.69587129e-01 -9.14992511e-01 1.46698916e+00
-8.69105682e-02 6.26293242e-01 2.97782421e-01 1.47039759e+00
8.73634636e-01 5.12230456e-01 -3.32570493e-01 -1.46294519e-01
1.34508073e+00 -1.13712966e+00 -9.17975485e-01 -2.10053951e-01
-9.53537151e-02 -9.19728220e-01 5.05735695e-01 3.90137047e-01
-1.15076876e+00 -7.99845994e-01 -1.26381862e+00 -3.71685505e-01
1.54197916e-01 4.78077829e-01 4.76857871e-01 4.21257019e-01
-1.33981586e+00 8.68665814e-01 -6.02371514e-01 3.45366970e-02
6.17795646e-01 7.10230470e-02 -3.63061160e-01 -3.62483859e-01
-1.01580417e+00 8.02001715e-01 1.43224731e-01 8.88747692e-01
-1.00776887e+00 -5.08102596e-01 -8.00777376e-01 1.16730466e-01
1.12889975e-01 -8.25017750e-01 1.24997115e+00 -1.05677366e+00
-1.55978310e+00 4.41736758e-01 -3.42161000e-01 -4.04096812e-01
4.66955960e-01 -4.84625578e-01 -3.91118675e-01 2.71552891e-01
-2.41661936e-01 5.09299040e-01 1.26396954e+00 -1.72726822e+00
-3.98097783e-01 -1.09813660e-01 -2.21910402e-01 6.71116710e-02
-2.51607448e-01 1.30903631e-01 -6.03377819e-01 -9.47146714e-01
2.07507759e-01 -5.80667377e-01 -2.21074954e-01 4.52717245e-02
-7.05913723e-01 3.89512807e-01 6.55138850e-01 -1.12614346e+00
1.26259184e+00 -1.98545206e+00 1.90027699e-01 -2.34881759e-01
6.83313072e-01 5.25380969e-01 -2.63501942e-01 1.60532817e-02
-3.94715130e-01 6.74643442e-02 -5.05663097e-01 -6.60416543e-01
-1.70260787e-01 -1.39969662e-01 -5.47886789e-01 8.02459776e-01
3.77227217e-01 1.18663335e+00 -6.11166418e-01 4.55085607e-03
3.39125574e-01 8.21959555e-01 -2.71746635e-01 3.12534660e-01
8.69015530e-02 5.35745025e-01 -1.36735991e-01 6.89401269e-01
1.39321387e+00 -4.63191628e-01 -5.49868681e-02 -6.60740733e-01
-2.77433336e-01 -1.83519542e-01 -9.08699572e-01 1.31714940e+00
-1.45160928e-01 1.13590693e+00 2.54174352e-01 -7.95505822e-01
8.55619669e-01 1.53027758e-01 1.06301881e-01 -6.63438916e-01
3.04287165e-01 4.55362499e-01 -3.43179971e-01 -7.03265250e-01
7.74017394e-01 -4.22968827e-02 4.28143591e-01 1.75508097e-01
-1.53743029e-01 2.39090323e-01 -3.69668454e-01 -6.05048127e-02
8.67678702e-01 7.33404830e-02 1.10804528e-01 -1.36432782e-01
7.69782722e-01 -2.97514290e-01 4.17287678e-01 7.27588236e-01
-2.82322884e-01 1.34593666e+00 2.65566260e-01 -4.73702729e-01
-1.10630810e+00 -7.95870662e-01 2.97130831e-02 2.27459818e-01
6.95734262e-01 2.56041944e-01 -1.05575037e+00 -3.19666624e-01
-2.27692500e-01 4.05189186e-01 -5.59630275e-01 -1.12203777e-01
-6.68599725e-01 -8.94494772e-01 7.08394289e-01 3.86720687e-01
1.17518044e+00 -8.29579413e-01 -1.88893631e-01 7.15414658e-02
-5.08677542e-01 -1.23023260e+00 -9.00572956e-01 -1.82829008e-01
-9.01372671e-01 -1.12629616e+00 -1.35658157e+00 -9.04950023e-01
9.79579985e-01 1.10925186e+00 7.22638905e-01 2.12462634e-01
-1.62715435e-01 -6.70878142e-02 -2.65246272e-01 1.67276934e-01
-2.49162674e-01 -4.04339284e-01 -1.38149947e-01 3.57245266e-01
2.02512443e-02 -6.61025703e-01 -1.13253033e+00 2.23521546e-01
-1.07958400e+00 3.51918727e-01 1.11675107e+00 9.31872785e-01
3.69647384e-01 2.69983947e-01 3.65976930e-01 -4.65099096e-01
9.39017773e-01 -3.83809060e-01 -5.87728202e-01 2.70781517e-01
-7.40757585e-01 3.34008075e-02 6.46349072e-01 -4.51805145e-01
-1.44006062e+00 -1.34608403e-01 1.93626329e-01 -7.98475742e-01
-2.61547744e-01 6.11753345e-01 5.23547865e-02 -3.34770083e-01
6.04118407e-01 6.86781228e-01 1.90980315e-01 -7.66684413e-01
4.61470634e-01 9.19311702e-01 1.13498402e+00 -1.38113648e-01
8.73564124e-01 5.44906855e-01 -1.38257980e-01 -7.41987050e-01
-8.03969800e-01 -3.00844640e-01 -3.54045868e-01 -3.90250146e-01
9.34813142e-01 -1.23194110e+00 -8.95028591e-01 1.16055202e+00
-1.52368450e+00 -5.14951527e-01 4.37269926e-01 4.17614907e-01
-3.51922780e-01 9.00947034e-01 -7.77662635e-01 -7.17797697e-01
-4.60478365e-01 -1.10233653e+00 8.57199550e-01 5.91865599e-01
3.43147188e-01 -7.89222658e-01 -1.95354939e-01 3.97314847e-01
7.40768492e-01 5.33589311e-02 3.81250471e-01 1.17465228e-01
-9.02347028e-01 -6.01833835e-02 -1.11892211e+00 8.59400034e-01
2.16175094e-01 -3.82997453e-01 -1.16767180e+00 -3.29796970e-01
2.95602649e-01 3.72359133e-03 1.35561478e+00 6.52264476e-01
1.30541766e+00 -5.46644986e-01 3.78187224e-02 1.08256412e+00
1.63135517e+00 -2.82712072e-01 1.24852514e+00 3.33854526e-01
9.29290295e-01 3.70075762e-01 1.64172888e-01 1.37208119e-01
3.93424511e-01 4.74442244e-01 5.88363171e-01 -5.43733239e-01
-7.55398452e-01 7.44956313e-03 5.04894435e-01 6.87506557e-01
-1.76328003e-01 -2.51579374e-01 -4.18741167e-01 5.54806173e-01
-1.87626231e+00 -8.28390360e-01 -4.42628652e-01 1.94079459e+00
7.96122432e-01 -3.34327847e-01 -6.01077199e-01 -7.87289068e-02
9.93420541e-01 2.34663546e-01 -6.59183562e-01 1.38276175e-01
-5.13990164e-01 2.60937419e-02 8.55635643e-01 6.11033440e-01
-1.15028584e+00 8.81968796e-01 6.04429054e+00 7.36227155e-01
-1.23919284e+00 -1.10264175e-01 8.62812400e-01 4.63176489e-01
-2.02717617e-01 -8.45640898e-02 -2.57973164e-01 5.69616914e-01
5.45264661e-01 1.28564104e-01 8.17239046e-01 3.67393851e-01
5.60597956e-01 -1.78180728e-02 -5.93564868e-01 1.22317219e+00
2.27507576e-01 -1.41454792e+00 -4.50628586e-02 2.99889967e-02
1.07599747e+00 -1.35075655e-02 3.44124854e-01 -2.15454057e-01
2.13757679e-01 -1.30604231e+00 8.35313737e-01 8.19290459e-01
1.00902021e+00 -4.90268707e-01 9.62587118e-01 1.33118704e-01
-1.00153422e+00 -7.08438605e-02 -4.80711520e-01 2.79975887e-02
1.53005153e-01 8.21835577e-01 -3.15622061e-01 7.56219745e-01
7.65569568e-01 9.89250243e-01 -3.61575961e-01 1.29137075e+00
-3.51833701e-01 4.44692761e-01 4.38018709e-01 3.78549665e-01
4.65845875e-02 -3.06296706e-01 5.24905384e-01 1.21218169e+00
5.26714087e-01 3.54791671e-01 -4.90794778e-01 1.28320730e+00
-2.93404788e-01 -5.22961974e-01 -8.91700536e-02 1.65980831e-01
2.79142320e-01 1.28610826e+00 -3.81194741e-01 -3.08189303e-01
-3.62957448e-01 1.66364253e+00 1.19234346e-01 9.44117427e-01
-7.18287528e-01 -5.61021626e-01 1.00474703e+00 -3.15591753e-01
5.60543358e-01 -2.10786164e-01 -5.03121197e-01 -1.50319493e+00
1.47346139e-01 -8.10472667e-01 -2.88066268e-01 -1.23151064e+00
-1.42473674e+00 5.64495623e-01 -4.09476757e-01 -1.35149789e+00
2.96336055e-01 -7.05422819e-01 -6.49002016e-01 1.33032358e+00
-2.33729935e+00 -1.12794971e+00 -9.57510054e-01 4.57726568e-01
4.46912318e-01 1.49996594e-01 4.13155407e-01 3.14873546e-01
-7.97199309e-01 3.74708802e-01 3.58917147e-01 2.47679815e-01
8.21281731e-01 -1.03572774e+00 5.55545807e-01 1.26976585e+00
-6.34662211e-01 8.67724121e-01 8.02032173e-01 -6.39676929e-01
-1.37961590e+00 -1.29891896e+00 5.51034093e-01 7.78883621e-02
5.06944537e-01 2.04014421e-01 -1.12080348e+00 4.10750151e-01
4.04025346e-01 1.89838931e-01 3.18245851e-02 -6.28632545e-01
-4.81276155e-01 -8.33285749e-02 -9.69281137e-01 7.18068838e-01
8.03930759e-01 -6.07388914e-01 -4.09269303e-01 -1.71429530e-01
8.24373841e-01 -5.45097172e-01 -5.02277493e-01 3.83376539e-01
5.15675008e-01 -1.23890841e+00 1.17336631e+00 -1.86566129e-01
9.82348144e-01 -6.32412016e-01 -1.65119339e-02 -1.27718735e+00
-4.72763717e-01 -9.00907636e-01 -3.09814781e-01 9.83232856e-01
1.71362795e-02 -6.53916180e-01 4.63286340e-01 4.15650308e-01
-4.49275523e-01 -4.75488842e-01 -4.23735499e-01 -5.77076256e-01
-7.38805607e-02 -2.18390629e-01 4.01094824e-01 8.64604771e-01
-4.74092156e-01 -4.11279388e-02 -1.07515585e+00 6.07340515e-01
1.07333422e+00 -1.59211457e-02 5.47826707e-01 -8.93547475e-01
-2.59179473e-02 -5.72056472e-01 -1.04930453e-01 -1.73269880e+00
-1.22840613e-01 -4.21923667e-01 4.29805845e-01 -1.56303716e+00
4.22379911e-01 -3.10013145e-02 -2.55163491e-01 2.06931338e-01
-5.59382081e-01 5.95412672e-01 -9.64314118e-02 6.21373117e-01
-4.23182607e-01 6.29201829e-01 1.55148077e+00 -1.22653231e-01
-1.14954570e-02 -2.48016313e-01 -1.19796050e+00 4.80752409e-01
7.06913531e-01 -3.50887552e-02 -1.06711395e-01 -7.00487971e-01
-9.73776355e-02 3.44888531e-02 8.72203231e-01 -8.20330501e-01
3.88428122e-01 -2.39911899e-02 5.78886867e-01 -4.80441391e-01
5.42238122e-03 -6.22392058e-01 1.59768403e-01 3.25888395e-01
-3.54684979e-01 -5.33849418e-01 4.88351025e-02 7.53806651e-01
-4.26247299e-01 -1.25700340e-01 8.87138665e-01 -8.01189989e-02
-6.58445179e-01 1.55122936e-01 -3.36842120e-01 -3.47853661e-01
4.39451039e-01 -3.06439608e-01 -7.68937767e-01 -4.58941877e-01
-3.23275745e-01 -7.59550929e-02 4.60997075e-01 2.72630483e-01
9.55758810e-01 -1.24966836e+00 -9.17829335e-01 9.09636244e-02
-2.30181605e-01 6.79791858e-03 7.63990939e-01 1.05860901e+00
-8.55974197e-01 4.04160023e-01 -2.06613123e-01 -2.89918602e-01
-1.11848724e+00 4.41110224e-01 5.24803340e-01 2.02146888e-01
-7.85276711e-01 8.86143386e-01 3.90655041e-01 -2.00559087e-02
2.01057911e-01 -3.65778953e-01 -1.98816925e-01 -4.35467601e-01
9.43780601e-01 4.54936296e-01 -1.52174935e-01 -7.16699600e-01
8.09275508e-02 5.95053077e-01 -7.38691092e-02 3.27322148e-02
1.41947567e+00 -7.71754026e-01 -4.40339595e-01 -2.85067141e-01
1.36186135e+00 -1.46768227e-01 -1.77125120e+00 -3.93245488e-01
-3.13686579e-01 -6.06368124e-01 5.20046711e-01 -7.44198620e-01
-1.35854161e+00 8.27769876e-01 6.72829449e-01 4.83892597e-02
1.47357976e+00 -3.59591663e-01 1.08557284e+00 5.82634583e-02
-1.40440151e-01 -6.28440797e-01 2.57106692e-01 4.17792588e-01
9.85724509e-01 -1.32443964e+00 4.25287932e-02 -3.51250648e-01
-5.41131914e-01 1.43116474e+00 3.34555417e-01 -2.90171593e-01
4.55505043e-01 -3.59532870e-02 4.53966781e-02 -1.15617834e-01
-3.62280339e-01 -2.50013441e-01 4.75486666e-01 5.82274675e-01
2.35220939e-01 -1.73698604e-01 1.01844463e-02 8.00741255e-01
1.74911305e-01 3.22082609e-01 8.43758047e-01 2.95121461e-01
-6.24724686e-01 -4.80538249e-01 -6.55515254e-01 -4.96211387e-02
-4.51268286e-01 -4.42972511e-01 -2.36450478e-01 2.62132883e-01
7.54777268e-02 1.10326469e+00 -1.07721873e-01 -4.36348617e-01
1.06922928e-02 -5.46966851e-01 2.83071488e-01 -4.33856212e-02
-8.70472938e-02 2.19671234e-01 -2.55952030e-01 -4.69245315e-01
-4.31584716e-01 -2.48191714e-01 -8.45787823e-01 -4.67311740e-01
-4.86393243e-01 -4.27094221e-01 5.34182429e-01 7.63576150e-01
3.74251664e-01 7.12917149e-01 7.07499266e-01 -1.28844833e+00
-4.51128334e-01 -1.11295390e+00 -5.19294500e-01 1.75106987e-01
1.05125809e+00 -3.21890593e-01 -7.81270862e-01 4.18197960e-01] | [11.482413291931152, -2.5796139240264893] |
0305cfcf-7329-4af4-b2c6-89457d12bb77 | olenet-at-semeval-2019-task-9-bert-based | null | null | https://aclanthology.org/S19-2216 | https://aclanthology.org/S19-2216.pdf | OleNet at SemEval-2019 Task 9: BERT based Multi-Perspective Models for Suggestion Mining | This paper describes our system partici- pated in Task 9 of SemEval-2019: the task is focused on suggestion mining and it aims to classify given sentences into sug- gestion and non-suggestion classes in do- main specific and cross domain training setting respectively. We propose a multi- perspective architecture for learning rep- resentations by using different classical models including Convolutional Neural Networks (CNN), Gated Recurrent Units (GRU), Feed Forward Attention (FFA), etc. To leverage the semantics distributed in large amount of unsupervised data, we also have adopted the pre-trained Bidi- rectional Encoder Representations from Transformers (BERT) model as an en- coder to produce sentence and word rep- resentations. The proposed architecture is applied for both sub-tasks, and achieved f1-score of 0.7812 for subtask A, and 0.8579 for subtask B. We won the first and second place for the two tasks respec- tively in the final competition. | ['Yu Sun', 'Jiaxiang Liu', 'Shuohuan Wang'] | 2019-06-01 | null | null | null | semeval-2019-6 | ['suggestion-mining'] | ['natural-language-processing'] | [ 3.17056745e-01 6.22518182e-01 9.46706682e-02 -6.25374675e-01
-8.22911263e-01 -1.96430117e-01 6.12774432e-01 2.21375391e-01
-5.28634369e-01 7.37122476e-01 6.58961415e-01 -6.60012484e-01
1.35243028e-01 -7.70328104e-01 -8.84187162e-01 -2.50381291e-01
-1.04670592e-01 5.49164176e-01 -1.05129667e-02 -5.65153956e-01
4.83075261e-01 -2.49910921e-01 -1.40659881e+00 8.65912676e-01
7.76021063e-01 1.03359509e+00 4.86772358e-01 7.52171934e-01
1.16018867e-02 8.63050699e-01 -6.37659848e-01 -5.51588833e-01
-2.29292154e-01 -2.15423346e-01 -1.15547824e+00 1.00214176e-01
1.97559461e-01 7.12323282e-03 -3.40453535e-01 7.99334586e-01
4.54095304e-01 4.90895122e-01 9.94148076e-01 -5.86905003e-01
-1.10078549e+00 1.45977783e+00 -3.49265307e-01 3.83477926e-01
2.27763683e-01 -3.92575026e-01 1.23347425e+00 -1.38040185e+00
3.77779692e-01 1.24931026e+00 3.51364285e-01 6.98693275e-01
-9.40899014e-01 -5.12658060e-01 2.85035044e-01 4.77163732e-01
-1.12454975e+00 -4.62370872e-01 7.05657721e-01 -7.36072958e-02
1.51915538e+00 3.56167823e-01 -4.12047058e-02 1.54746783e+00
2.70111352e-01 1.18658888e+00 9.19961572e-01 -4.77153629e-01
2.64143595e-03 2.62964606e-01 3.28380466e-01 3.85675371e-01
-1.93993524e-01 -2.61649638e-01 -6.73343897e-01 1.35435939e-01
3.36309493e-01 -1.78705528e-01 -2.16515854e-01 3.91773820e-01
-1.00031245e+00 9.59148824e-01 4.47351694e-01 4.44943637e-01
-2.82340050e-01 -2.14660406e-01 7.96959519e-01 6.85350478e-01
7.30173409e-01 7.43984938e-01 -8.88392806e-01 -2.45348271e-02
-6.81145012e-01 2.67592788e-01 6.98896229e-01 1.08234799e+00
3.21418792e-01 8.45542327e-02 -6.34664595e-01 1.25581849e+00
2.87896037e-01 -7.89294764e-02 1.32055581e+00 -2.23474428e-01
1.07198131e+00 4.17423040e-01 -1.62943467e-01 -4.23526347e-01
-4.60779697e-01 -5.55561900e-01 -1.13323188e+00 -5.33218443e-01
-3.84088367e-01 -4.08232510e-01 -8.55110705e-01 1.58491933e+00
-3.42363536e-01 1.75791651e-01 4.30458814e-01 4.92232531e-01
1.49382520e+00 7.83865809e-01 1.13500185e-01 -2.37623781e-01
1.40887165e+00 -1.41654265e+00 -4.72057611e-01 -2.93360531e-01
7.34559834e-01 -6.06305480e-01 1.19001675e+00 4.35093373e-01
-1.07904327e+00 -1.00475621e+00 -1.20629025e+00 -2.73535728e-01
-3.80717337e-01 5.37173569e-01 2.80700564e-01 2.68320620e-01
-1.05370355e+00 5.37838161e-01 -4.45967406e-01 -4.55991000e-01
4.45280492e-01 4.20762569e-01 -2.88969368e-01 3.46454501e-01
-1.34005153e+00 8.07344854e-01 6.10303402e-01 -6.10765330e-02
-1.02684069e+00 -4.03077364e-01 -8.96691620e-01 3.94086659e-01
1.93910584e-01 -4.92590696e-01 1.50305378e+00 -7.63141096e-01
-1.76348889e+00 7.92218745e-01 9.31786299e-02 -1.00112116e+00
-5.14849983e-02 -6.10193491e-01 -5.87032974e-01 -3.08910251e-01
1.68017998e-01 6.18410885e-01 7.50254989e-01 -7.28135645e-01
-5.21417618e-01 -2.11831048e-01 6.07236922e-02 5.05886257e-01
-5.75206578e-01 3.12530696e-01 -2.87906732e-02 -8.64299595e-01
-3.42116803e-02 -4.45767313e-01 -1.69363558e-01 -1.13725936e+00
-9.52454329e-01 -1.01478446e+00 7.51885295e-01 -5.79326928e-01
1.20885909e+00 -1.99544203e+00 9.03826281e-02 -4.32421595e-01
-7.95211047e-02 1.91718519e-01 -2.56172150e-01 5.03802121e-01
-4.01170135e-01 8.19417089e-03 -1.35141639e-02 -6.19933605e-01
-1.40790055e-02 1.12638203e-02 -8.72252107e-01 -1.89315043e-02
4.72974479e-01 9.35308397e-01 -7.84347057e-01 -1.29843891e-01
-1.71528254e-02 -2.19525203e-01 -3.60367358e-01 7.85106182e-01
-4.33063030e-01 8.10526386e-02 -4.36944097e-01 1.17717728e-01
3.98289889e-01 -2.54210025e-01 7.41017759e-02 1.16937548e-01
1.10981479e-01 1.19478285e+00 -6.92837894e-01 1.83578038e+00
-7.97664464e-01 4.24701631e-01 -3.18504900e-01 -1.44599891e+00
1.38186741e+00 6.02107108e-01 -9.20296982e-02 -5.98162711e-01
1.89234704e-01 2.08511855e-02 9.74159762e-02 -5.57677686e-01
1.04825926e+00 -1.89912230e-01 -2.24503502e-01 5.69150627e-01
6.23756588e-01 -6.17951937e-02 2.96707340e-02 1.56219006e-01
1.10699761e+00 1.18500829e-01 3.37278426e-01 -3.63467813e-01
7.75013149e-01 -4.25809681e-01 1.97092786e-01 7.88688898e-01
3.90132874e-01 6.66410089e-01 4.73903239e-01 -4.36239272e-01
-7.06410825e-01 -8.31737638e-01 -1.39354005e-01 1.50821209e+00
-9.78498235e-02 -8.10462713e-01 -3.81818652e-01 -1.07154942e+00
-3.40942562e-01 1.13497877e+00 -7.31277943e-01 -2.92315394e-01
-5.41334331e-01 -6.98177516e-01 4.53334659e-01 6.71426713e-01
5.66148520e-01 -1.58638418e+00 -4.38867122e-01 3.67575109e-01
-1.57182232e-01 -9.37400043e-01 -2.18456253e-01 9.02384818e-01
-8.34012568e-01 -7.67827332e-01 -5.43272674e-01 -9.87575352e-01
3.65908980e-01 3.85912098e-02 1.38132370e+00 -2.41904333e-01
2.92461634e-01 -1.52783632e-01 -7.94153810e-01 -5.08004010e-01
-3.14491808e-01 4.78752464e-01 6.39511645e-02 -4.25643995e-02
6.41176701e-01 -5.42855918e-01 -2.56197333e-01 9.78899598e-02
-3.94860148e-01 2.07786635e-01 7.33328462e-01 1.08715916e+00
3.88805002e-01 -2.93600768e-01 1.13224554e+00 -1.36232209e+00
1.28532517e+00 -7.74067223e-01 3.02211624e-02 2.38250554e-01
-5.04402459e-01 7.97400847e-02 8.86118352e-01 -7.48477355e-02
-1.09905660e+00 -3.14208925e-01 -4.68851537e-01 -1.47207916e-01
-3.48324031e-01 7.02894807e-01 -6.15085773e-02 9.29308534e-01
9.06513512e-01 3.86314422e-01 -4.68115121e-01 -7.93800294e-01
4.70930278e-01 1.18352568e+00 5.63443244e-01 -4.96726096e-01
2.90377408e-01 -3.07801634e-01 -6.79089129e-01 -5.65226197e-01
-1.19950652e+00 -3.15961659e-01 -3.41892660e-01 2.80474544e-01
6.41771197e-01 -1.04240584e+00 -3.89904439e-01 5.58008216e-02
-1.21479285e+00 -1.63447738e-01 -4.67571676e-01 2.35761821e-01
-6.72842026e-01 1.94073603e-01 -4.57890540e-01 -7.04414964e-01
-9.67227280e-01 -9.18032825e-01 1.02388537e+00 2.44329661e-01
-2.00355813e-01 -7.70715237e-01 3.56038287e-02 3.17776263e-01
4.32995588e-01 -4.13556278e-01 1.02094209e+00 -1.52218497e+00
-2.90913712e-02 -3.22842672e-02 -1.16182245e-01 4.67154264e-01
2.65132710e-02 -5.49777865e-01 -1.24556184e+00 -1.63058028e-01
3.08414489e-01 -1.03079939e+00 1.33570516e+00 2.41232619e-01
1.88602316e+00 -3.72510761e-01 -4.24192622e-02 2.42472902e-01
8.91129315e-01 -2.11760774e-01 6.76134586e-01 2.50289112e-01
2.18527541e-01 6.40540957e-01 6.22588098e-01 3.61911297e-01
4.92244124e-01 7.80577898e-01 3.52244973e-01 3.56722146e-01
6.36219159e-02 -5.65063953e-01 7.26672590e-01 1.16835880e+00
8.23364183e-02 -5.62731802e-01 -4.97158617e-01 6.03819489e-01
-2.09404826e+00 -8.85159552e-01 1.36556220e-03 1.88066804e+00
1.00214481e+00 4.03262466e-01 3.85819674e-02 -1.09432511e-01
4.87544030e-01 3.10982198e-01 -3.23352456e-01 -1.06232429e+00
-1.98007345e-01 5.70263028e-01 8.96250755e-02 7.95849711e-02
-1.44541752e+00 1.09651566e+00 5.57851839e+00 1.13609350e+00
-1.01482058e+00 3.16934824e-01 1.10523009e+00 2.32581615e-01
-2.68756330e-01 -3.41732889e-01 -8.43114197e-01 3.42384964e-01
1.54481721e+00 -3.07705909e-01 -1.36794746e-01 1.04143941e+00
-1.64155457e-02 1.71622559e-01 -1.10167325e+00 6.32895708e-01
4.81221855e-01 -1.43241847e+00 1.36063606e-01 -4.77847338e-01
6.01641893e-01 4.38159049e-01 1.56753972e-01 1.12860334e+00
4.63043213e-01 -1.46548736e+00 3.88079047e-01 2.98462749e-01
8.74544561e-01 -6.43142402e-01 1.04926908e+00 5.84856629e-01
-8.81089747e-01 -1.58423468e-01 -8.79631281e-01 -3.08390528e-01
-7.00565726e-02 3.07859570e-01 -1.45003951e+00 8.08659971e-01
5.88466585e-01 1.12661338e+00 -5.15319228e-01 6.59738600e-01
-6.04932964e-01 8.92616451e-01 5.64402528e-02 -4.09444720e-01
1.96933091e-01 1.26268476e-01 2.27577060e-01 1.50076890e+00
2.03984946e-01 -2.20355570e-01 8.91440362e-02 7.85974741e-01
-6.15961313e-01 2.84441471e-01 -3.75007242e-01 3.03956792e-02
2.69735783e-01 1.38528144e+00 -2.22710893e-01 -5.44306815e-01
-4.26471591e-01 8.15363467e-01 7.60432482e-01 2.10643739e-01
-5.67983091e-01 -4.60079610e-01 8.02548900e-02 -1.71835527e-01
5.10566115e-01 1.03849806e-01 -5.56632459e-01 -1.33074725e+00
-1.37568384e-01 -7.79323459e-01 4.11166877e-01 -7.39628792e-01
-1.57343686e+00 1.07554722e+00 -7.71827176e-02 -1.19349980e+00
-5.93801737e-01 -6.23127937e-01 -1.08452654e+00 1.05120492e+00
-1.45188200e+00 -1.26225674e+00 2.80916631e-01 4.94845778e-01
1.34451079e+00 -8.03926706e-01 1.20035827e+00 -1.50674433e-01
-5.31315506e-01 8.13329756e-01 7.92507157e-02 6.38734698e-02
7.13159919e-01 -1.62009776e+00 6.90855324e-01 6.42257452e-01
5.37031770e-01 7.54289091e-01 5.09752452e-01 -4.72635776e-01
-9.86341298e-01 -1.44533646e+00 1.34033990e+00 -2.82049209e-01
6.28615856e-01 -6.48070395e-01 -7.92288959e-01 8.65310669e-01
6.87003911e-01 -3.33834678e-01 7.51317918e-01 6.82228148e-01
-2.34177291e-01 2.05351010e-01 -5.71915329e-01 4.26884949e-01
6.93146646e-01 -5.13064861e-01 -1.16052735e+00 7.61086702e-01
1.16488361e+00 -5.73421180e-01 -5.81426382e-01 4.06551749e-01
-1.15931109e-01 -7.24875093e-01 8.22803676e-01 -1.02740598e+00
1.11894691e+00 1.64281875e-01 -2.08736360e-01 -1.30798876e+00
-3.58368069e-01 -6.03762329e-01 -3.05455804e-01 1.19256401e+00
9.03072000e-01 -2.52997596e-03 6.85136259e-01 -3.02019902e-02
-7.53824055e-01 -1.30843830e+00 -9.45049405e-01 -5.35810649e-01
3.16687495e-01 -6.14379525e-01 5.06314814e-01 6.76179767e-01
3.98033202e-01 1.25729704e+00 -4.68870610e-01 -2.45971501e-01
-1.09785669e-01 2.51896381e-01 5.65096676e-01 -7.65446365e-01
-5.38095891e-01 -2.65882581e-01 1.13751963e-02 -1.61216152e+00
3.86929035e-01 -1.21073711e+00 2.57256180e-02 -1.32354391e+00
2.01436594e-01 -1.85695559e-01 -5.81109285e-01 4.63350356e-01
-3.04064065e-01 -4.44012247e-02 -1.63094610e-01 -6.84134215e-02
-8.33196998e-01 1.10278654e+00 8.84741485e-01 -2.97896087e-01
-6.50615990e-02 4.70143110e-01 -1.03071809e+00 5.06193161e-01
8.87577891e-01 -2.41336122e-01 -5.87666094e-01 -6.15064025e-01
7.63481632e-02 -8.31940249e-02 3.17369141e-02 -6.22519612e-01
1.85709149e-01 1.89973623e-01 2.65247315e-01 -8.81010950e-01
3.42178404e-01 -2.33703434e-01 -7.57030666e-01 1.20912731e-01
-9.04825926e-01 1.84786871e-01 5.93145750e-02 5.59048355e-01
-4.26919788e-01 -6.29233778e-01 2.95051277e-01 -1.98454574e-01
-6.11541271e-01 1.33595854e-01 -4.27648395e-01 4.13281620e-02
5.23603261e-01 2.30906665e-01 -2.98278302e-01 -3.89840424e-01
-7.34450817e-01 1.55452445e-01 -3.29350233e-01 8.67848933e-01
9.53781784e-01 -1.06698442e+00 -1.11343467e+00 3.07977170e-01
4.00570899e-01 3.86946537e-02 2.02954978e-01 4.93397802e-01
2.88274847e-02 6.67427123e-01 -1.22877639e-02 -2.40488425e-01
-9.46509421e-01 3.08410525e-01 1.69285044e-01 -6.89071655e-01
-6.62124276e-01 1.62146747e+00 1.76172659e-01 -8.10827196e-01
2.69594163e-01 -4.45815891e-01 -8.32812488e-01 -2.40724504e-01
4.43439126e-01 6.06775023e-02 3.78799945e-01 -3.46486568e-01
-1.28585607e-01 -5.61798774e-02 -8.11662674e-01 6.82635158e-02
1.65535426e+00 6.82823136e-02 -6.30660653e-02 5.49408197e-01
1.09077251e+00 -4.12730217e-01 -6.47371471e-01 -5.20255685e-01
2.06382871e-01 6.33340329e-02 -1.34717450e-01 -1.02243483e+00
-5.51385581e-01 1.19661081e+00 1.00916155e-01 3.75060737e-01
9.13354695e-01 2.95122594e-01 6.56264484e-01 6.52829111e-01
9.08009559e-02 -1.18390632e+00 2.97446698e-01 1.12541986e+00
1.49068177e+00 -1.18366945e+00 -1.63351163e-01 -2.05388479e-02
-9.54603791e-01 1.31138897e+00 9.98851061e-01 -4.90287811e-01
5.91412425e-01 -1.26614138e-01 -5.12597263e-01 -3.09842795e-01
-1.38947141e+00 1.18042499e-01 3.67827803e-01 4.96859372e-01
1.10514963e+00 3.38573813e-01 -4.65794474e-01 1.34672105e+00
-6.53177440e-01 -2.28739008e-01 6.40298843e-01 4.76564556e-01
-5.99305212e-01 -1.12140417e+00 2.56210536e-01 8.75352621e-01
-3.46665293e-01 -3.97878021e-01 -4.62402314e-01 2.51468301e-01
2.72228532e-02 1.25144291e+00 6.66832328e-02 -7.63877094e-01
2.94951946e-01 4.66863438e-02 5.23062460e-02 -1.34362531e+00
-9.60802078e-01 -5.72896264e-02 5.55405080e-01 -1.61255866e-01
-2.08880454e-01 -4.15587664e-01 -1.00156760e+00 1.66939288e-01
-5.63635051e-01 4.91917372e-01 4.18282121e-01 1.10094321e+00
5.34067869e-01 9.54686165e-01 7.74032652e-01 -3.67026240e-01
-9.49785054e-01 -1.83693218e+00 -5.57028234e-01 2.71801144e-01
1.77374966e-02 -3.72595042e-01 9.37064141e-02 -6.98374882e-02] | [10.946457862854004, 7.663675785064697] |
80c3a9d2-415b-46f9-887f-b895c6f6629d | embodied-navigation-at-the-art-gallery | 2204.09069 | null | https://arxiv.org/abs/2204.09069v1 | https://arxiv.org/pdf/2204.09069v1.pdf | Embodied Navigation at the Art Gallery | Embodied agents, trained to explore and navigate indoor photorealistic environments, have achieved impressive results on standard datasets and benchmarks. So far, experiments and evaluations have involved domestic and working scenes like offices, flats, and houses. In this paper, we build and release a new 3D space with unique characteristics: the one of a complete art museum. We name this environment ArtGallery3D (AG3D). Compared with existing 3D scenes, the collected space is ampler, richer in visual features, and provides very sparse occupancy information. This feature is challenging for occupancy-based agents which are usually trained in crowded domestic environments with plenty of occupancy information. Additionally, we annotate the coordinates of the main points of interest inside the museum, such as paintings, statues, and other items. Thanks to this manual process, we deliver a new benchmark for PointGoal navigation inside this new space. Trajectories in this dataset are far more complex and lengthy than existing ground-truth paths for navigation in Gibson and Matterport3D. We carry on extensive experimental evaluation using our new space for evaluation and prove that existing methods hardly adapt to this scenario. As such, we believe that the availability of this 3D model will foster future research and help improve existing solutions. | ['Rita Cucchiara', 'Lorenzo Baraldi', 'Marcella Cornia', 'Silvia Cascianelli', 'Federico Landi', 'Roberto Bigazzi'] | 2022-04-19 | null | null | null | null | ['pointgoal-navigation'] | ['robots'] | [-3.59464675e-01 -6.08621025e-03 4.53066766e-01 -1.53228238e-01
-1.40656590e-01 -7.56439626e-01 8.44998062e-01 -2.83221733e-02
-6.99046612e-01 7.56187499e-01 5.05124032e-01 -2.72300243e-01
1.62248854e-02 -9.74193811e-01 -6.86285436e-01 -6.37874365e-01
-5.02774179e-01 6.49386764e-01 4.52241361e-01 -7.53281593e-01
3.21889520e-01 6.36325836e-01 -1.80834460e+00 -1.93597138e-01
5.40675521e-01 5.65809667e-01 5.58623672e-01 5.75814545e-01
1.90196514e-01 5.66419661e-01 -5.09157121e-01 -1.12452328e-01
4.91577387e-01 -2.24899482e-02 -8.14309835e-01 5.94235621e-02
2.61188745e-01 -4.95768994e-01 -6.72444820e-01 6.00731194e-01
5.96651793e-01 7.40599871e-01 4.98723507e-01 -1.26090598e+00
-6.95878148e-01 2.20056981e-01 -3.41377974e-01 2.12103426e-02
9.71213639e-01 5.03268301e-01 8.54521930e-01 -7.36260772e-01
9.33040917e-01 1.36339605e+00 7.47951865e-01 4.68971431e-01
-8.24555635e-01 5.28905205e-02 6.18916690e-01 1.91442415e-01
-1.37307954e+00 -3.46641034e-01 4.27503854e-01 -2.78643012e-01
1.19443631e+00 4.93189573e-01 1.08229315e+00 1.63863027e+00
-2.06264794e-01 1.07918024e+00 9.62316513e-01 -1.76598385e-01
7.18693495e-01 -2.44203985e-01 -1.79877102e-01 8.09354424e-01
1.02632657e-01 1.63552880e-01 -3.02575201e-01 7.32715949e-02
9.87565041e-01 3.73740971e-01 -5.32465160e-01 -1.02296758e+00
-1.49231076e+00 5.65541506e-01 9.89119887e-01 2.97661215e-01
-2.24978641e-01 9.71392840e-02 2.97809020e-03 -8.47790986e-02
8.32393542e-02 3.65728170e-01 -1.19622327e-01 -4.12673533e-01
-1.82255954e-01 6.09723747e-01 9.41729248e-01 1.58728981e+00
6.56204581e-01 -3.88476461e-01 1.24185286e-01 5.30790985e-01
2.96606570e-01 5.09911180e-01 2.60189235e-01 -1.28097391e+00
5.72399139e-01 7.46438324e-01 6.53258801e-01 -1.04804528e+00
-9.38292980e-01 -1.25904188e-01 -8.53507757e-01 3.56782675e-01
5.99650562e-01 3.25451232e-03 -9.66186166e-01 1.49459314e+00
7.33721673e-01 -5.52335158e-02 8.54991935e-03 1.17210639e+00
1.03831613e+00 3.85665089e-01 -2.58057058e-01 5.33815503e-01
1.23534572e+00 -1.43292356e+00 -3.94327492e-01 -3.01139086e-01
7.28642702e-01 -1.24064639e-01 1.41280115e+00 -9.01525561e-03
-8.42771947e-01 -2.86040157e-01 -9.32000458e-01 -4.71247733e-01
-7.61785686e-01 -2.16169029e-01 1.01093996e+00 4.65796083e-01
-1.14776361e+00 3.05284262e-01 -1.06988800e+00 -8.63245785e-01
4.06011522e-01 7.85007328e-02 -6.78377450e-01 -2.28447422e-01
-6.58291578e-01 1.02092075e+00 6.73912913e-02 1.81188628e-01
-1.06297612e+00 -3.77103776e-01 -1.26046646e+00 -2.71527350e-01
3.39026332e-01 -7.43882477e-01 1.21380150e+00 3.33053946e-01
-1.24309349e+00 9.32973981e-01 9.91942175e-03 -1.63503751e-01
8.03749919e-01 -6.94893226e-02 3.66259478e-02 -3.66366297e-01
3.88723791e-01 7.29705811e-01 -1.00473734e-03 -1.50958157e+00
-8.85931909e-01 -6.91617250e-01 8.35617304e-01 5.46749175e-01
2.55730271e-01 -6.28156960e-01 -7.33083725e-01 -1.23759173e-01
2.88666010e-01 -1.14989293e+00 -6.00523293e-01 1.58989504e-01
-4.59322482e-01 -1.35475278e-01 4.92589831e-01 -3.21330339e-01
6.43898666e-01 -2.19180894e+00 3.20913255e-01 -9.12931561e-03
1.81449711e-01 -3.23097229e-01 1.35473460e-02 6.66808546e-01
6.81005776e-01 5.80146490e-03 -2.75993615e-01 -8.09105217e-01
6.65632069e-01 5.70822001e-01 4.32003103e-02 6.11103117e-01
-5.73677480e-01 1.02979398e+00 -1.36154842e+00 -1.68449178e-01
5.67361832e-01 5.44653416e-01 -6.94925308e-01 -1.15124539e-01
-1.19055197e-01 8.21288466e-01 -5.32695770e-01 8.34484518e-01
5.67143559e-01 2.32324749e-02 -1.98579624e-01 3.90625805e-01
-4.11270767e-01 2.53787488e-01 -1.18046141e+00 2.48524690e+00
-4.71088767e-01 5.98384917e-01 -1.08584836e-01 -2.10150644e-01
7.40647554e-01 -1.77199364e-01 2.98167080e-01 -8.37601840e-01
3.79425250e-02 1.03388369e-01 -5.55835843e-01 -5.35412908e-01
9.08855081e-01 4.28282648e-01 -3.81773919e-01 2.22467527e-01
-3.68838012e-01 -5.23966312e-01 2.98836112e-01 -3.93546978e-03
1.46312261e+00 4.41298306e-01 2.06737220e-01 -1.68981478e-01
1.08215287e-01 2.33166963e-01 2.06550688e-01 9.94242191e-01
-5.07597625e-01 7.50747681e-01 -7.16937110e-02 -8.07355285e-01
-1.11413825e+00 -1.16538215e+00 1.25365764e-01 9.76608634e-01
8.76405180e-01 -3.67797494e-01 -6.06234848e-01 -6.03121221e-01
-5.87648200e-03 6.18069768e-01 -8.08179319e-01 3.45170557e-01
-6.32680297e-01 -4.13569719e-01 2.69742250e-01 5.41269720e-01
6.31441057e-01 -1.14390099e+00 -1.10337293e+00 1.62445065e-02
-4.20327634e-01 -1.14805520e+00 -3.32214147e-01 1.13662504e-01
-4.57619339e-01 -1.22498322e+00 -7.55762935e-01 -9.73337293e-01
7.37322986e-01 6.40293658e-01 1.12278187e+00 -8.15667491e-03
-2.79528916e-01 7.10363030e-01 -5.80227554e-01 -3.61786842e-01
2.61025071e-01 7.67118558e-02 1.25007004e-01 -5.28604150e-01
4.04135197e-01 -7.53887236e-01 -9.23305929e-01 6.03846669e-01
-3.14862341e-01 1.33945122e-01 1.32988110e-01 4.29441452e-01
6.26414895e-01 5.11201695e-02 -2.61541456e-01 -1.74185976e-01
4.29309338e-01 -4.63955909e-01 -5.39721072e-01 -1.22534640e-01
1.81775674e-01 -2.30914518e-01 2.58904308e-01 -1.67180195e-01
-8.90926659e-01 -7.76457638e-02 -1.63763821e-01 5.50471758e-03
-4.99009520e-01 -4.86866944e-02 -2.32686669e-01 -1.14494361e-01
7.86443770e-01 -4.15127501e-02 -4.75694835e-01 -6.53500676e-01
4.83628482e-01 4.79151726e-01 6.42606080e-01 -5.90149283e-01
7.61849523e-01 1.05393410e+00 -2.87322029e-02 -8.68987918e-01
-6.21368885e-01 -6.19586945e-01 -8.33600342e-01 -1.77007958e-01
8.22180450e-01 -9.78956401e-01 -1.21882033e+00 3.99812013e-01
-9.40191150e-01 -8.47798526e-01 -6.09512389e-01 5.38284600e-01
-8.58969510e-01 1.73622683e-01 -4.64552194e-01 -9.42265391e-01
2.54896998e-01 -1.13252974e+00 1.46050572e+00 2.01347902e-01
-1.85500219e-01 -9.97045279e-01 3.77897292e-01 2.34488100e-01
1.77976653e-01 7.01176405e-01 2.98072457e-01 -8.20152760e-02
-8.89697194e-01 -3.46432514e-02 9.60851312e-02 -4.93502945e-01
1.97455168e-01 -5.06486237e-01 -7.82154560e-01 -2.92065293e-01
-1.15933537e-01 -1.73382849e-01 6.39378667e-01 2.23295659e-01
7.60286689e-01 -5.66659383e-02 -6.98416114e-01 6.81514740e-01
1.25916409e+00 2.96030641e-01 7.41451085e-01 1.03373909e+00
7.69529283e-01 5.56832671e-01 6.70346737e-01 5.95764816e-01
1.14143598e+00 8.52968216e-01 8.97306144e-01 -5.39612062e-02
5.14872894e-02 -5.20507216e-01 2.74848435e-02 3.85938525e-01
-5.50037920e-01 -5.23932159e-01 -9.16740119e-01 8.66500974e-01
-2.09700418e+00 -9.68941927e-01 -1.05612822e-01 1.87902105e+00
2.51716286e-01 -2.68981189e-01 9.31059942e-02 -1.19455375e-01
2.12391853e-01 2.25064486e-01 -6.38885975e-01 2.12951526e-01
-3.36557120e-01 -3.31408739e-01 4.75226730e-01 5.80281675e-01
-1.18767560e+00 9.53229845e-01 6.41829920e+00 2.09426686e-01
-1.88135028e-01 -5.36094718e-02 2.29190402e-02 -4.13675100e-01
-2.51281261e-01 -1.47971645e-01 -7.88649857e-01 2.32335418e-01
1.64045557e-01 3.99506003e-01 7.58896232e-01 1.01835060e+00
1.67900056e-01 -2.89791137e-01 -1.26917124e+00 1.09695685e+00
-3.55746932e-02 -1.05896091e+00 -3.64806265e-01 4.47116822e-01
7.74447918e-01 3.40210676e-01 -7.13616610e-02 5.53066254e-01
8.33247900e-01 -9.49811459e-01 9.57201719e-01 4.70563352e-01
1.58442616e-01 -5.64197242e-01 5.21473765e-01 6.09012127e-01
-1.20656955e+00 -2.15515435e-01 -5.77216804e-01 -4.66617405e-01
4.84271139e-01 -1.82826623e-01 -5.86133778e-01 3.82941395e-01
1.21967769e+00 7.26747751e-01 -4.12109196e-01 1.40231395e+00
-2.94236571e-01 -4.65392053e-01 -6.53358221e-01 -4.03561413e-01
7.28039682e-01 -3.27987015e-01 6.37231946e-01 7.04588890e-01
5.36449373e-01 1.70540899e-01 2.57512599e-01 6.19883001e-01
9.00912732e-02 -1.42333820e-01 -1.12987554e+00 5.08553147e-01
3.81047666e-01 9.68842149e-01 -7.54508555e-01 -1.42972469e-02
-1.73602536e-01 1.15508735e+00 5.62850416e-01 4.38019812e-01
-8.55362952e-01 -1.19826486e-02 9.97982562e-01 8.52359831e-02
4.46992844e-01 -7.82643557e-01 -4.70633656e-02 -1.09030211e+00
2.95506597e-01 -4.08882231e-01 1.00522172e-02 -9.81430650e-01
-9.33620870e-01 5.90050995e-01 -1.87136047e-02 -1.23852837e+00
3.52718122e-02 -5.63468158e-01 -1.86120868e-01 3.73352349e-01
-1.46628702e+00 -1.10966599e+00 -8.80422235e-01 7.17468262e-01
6.88789666e-01 9.44124758e-02 9.93827522e-01 1.47416934e-01
-2.97145545e-01 1.06011704e-01 3.58109951e-01 4.81517501e-02
6.63241595e-02 -1.57232642e+00 1.02407372e+00 6.16771996e-01
2.41213322e-01 5.63722551e-01 6.64019883e-01 -4.40185487e-01
-1.48743045e+00 -9.19020653e-01 4.56455380e-01 -1.22808075e+00
4.34615284e-01 -7.95096874e-01 -2.89005786e-01 1.03913701e+00
3.83159965e-02 -3.17980230e-01 3.42023134e-01 1.97917864e-01
-9.19913203e-02 4.75424796e-01 -1.20188844e+00 1.23492110e+00
2.21895409e+00 -2.16766730e-01 -6.25505030e-01 5.95630229e-01
9.03520226e-01 -9.67489123e-01 -5.48268080e-01 1.77020177e-01
4.98185009e-01 -1.34684324e+00 1.25515950e+00 -2.98086971e-01
1.97715331e-02 -4.93724108e-01 -6.17442906e-01 -1.52011037e+00
-3.42507243e-01 -5.15193403e-01 -8.61914009e-02 6.30039454e-01
1.95830792e-01 -4.79146212e-01 9.63850975e-01 8.04350436e-01
-5.75692475e-01 -5.01415431e-01 -9.71294761e-01 -8.92420232e-01
-4.26335633e-01 -4.46725577e-01 1.16550481e+00 6.19098127e-01
3.31975110e-02 1.25239179e-01 -3.75555277e-01 3.44184369e-01
7.55486190e-01 1.04498349e-01 1.33643019e+00 -1.10389650e+00
1.68765515e-01 -4.55624759e-01 -7.15320587e-01 -1.70492792e+00
-4.63313535e-02 -5.25353312e-01 2.90301681e-01 -2.17623234e+00
-1.84870332e-01 -8.34321976e-01 3.12073737e-01 4.19456691e-01
1.68828547e-01 1.79083809e-01 3.55193391e-02 1.33810207e-01
-9.16377664e-01 1.02822244e+00 1.63396752e+00 -3.65323365e-01
-5.10935128e-01 -9.63994488e-02 -4.64503050e-01 8.17292035e-01
6.36289358e-01 1.58561040e-02 -4.44316834e-01 -7.66929924e-01
1.02193234e-02 -4.14683729e-01 5.50696790e-01 -1.16926789e+00
4.05012697e-01 -1.62025273e-01 1.79747894e-01 -8.82597923e-01
1.03960717e+00 -1.11169970e+00 2.61531949e-01 2.43970394e-01
1.70971647e-01 3.56752723e-02 6.89384295e-03 7.88484693e-01
2.03635052e-01 -1.30158998e-02 5.18423505e-02 -7.61450469e-01
-1.26461160e+00 4.71615046e-01 -1.99791446e-01 4.78205681e-02
1.17430401e+00 -6.94550514e-01 -4.19181049e-01 -3.23002726e-01
-8.86038363e-01 5.85002005e-01 1.18884802e+00 6.90513790e-01
6.81459904e-01 -1.40008700e+00 -3.69873434e-01 1.55009016e-01
3.37519318e-01 6.41103029e-01 4.85507727e-01 5.46525657e-01
-8.63856375e-01 3.71573120e-01 -2.49336243e-01 -7.54121423e-01
-9.58682716e-01 7.41131067e-01 2.92643934e-01 1.22398145e-01
-1.24480164e+00 8.13718438e-01 4.16581899e-01 -1.05426931e+00
6.05836272e-01 -7.38695264e-01 -3.05430561e-01 -2.95229495e-01
5.74769437e-01 5.47378540e-01 -2.12751165e-01 -8.20595503e-01
-3.82758081e-01 5.07162869e-01 4.50310975e-01 -4.71886779e-05
1.66612768e+00 -6.10521734e-01 1.55873567e-01 4.69237745e-01
7.82482862e-01 -1.26799375e-01 -1.60262835e+00 -6.06333651e-02
-2.06433773e-01 -7.94976473e-01 -3.93668532e-01 -3.87164801e-01
-6.10821307e-01 5.51728845e-01 5.40654898e-01 2.58902311e-01
5.75465262e-01 1.16343193e-01 5.44872284e-01 8.60987067e-01
1.40699375e+00 -9.86978769e-01 8.40134472e-02 7.34345198e-01
1.14299655e+00 -1.33032334e+00 -3.05467099e-01 -9.19960365e-02
-4.86555755e-01 4.09274548e-01 7.38443017e-01 1.03177749e-01
4.28217083e-01 6.15475141e-02 -1.24237746e-01 -5.72832406e-01
-2.95747012e-01 -6.23589158e-01 -3.07334512e-01 1.22107732e+00
-3.89138669e-01 1.63810685e-01 4.30848092e-01 3.01141143e-01
-7.64324963e-01 -4.25825626e-01 5.67672014e-01 1.23875165e+00
-4.00707066e-01 -5.41617334e-01 -4.71600920e-01 -1.43363610e-01
4.57251400e-01 1.72287509e-01 -2.90442318e-01 1.17203486e+00
1.34344816e-01 1.08614349e+00 9.12019834e-02 -2.06777990e-01
9.07323718e-01 -4.43150163e-01 7.64103293e-01 -4.45193112e-01
-2.02220261e-01 -5.16309142e-01 1.62605464e-01 -9.45327520e-01
-5.54113567e-01 -5.99414289e-01 -1.27425337e+00 -5.96037567e-01
-2.16324092e-03 -3.99130844e-02 8.42168689e-01 7.50693381e-01
3.05865198e-01 3.80253911e-01 2.85928875e-01 -1.70531666e+00
9.58864316e-02 -7.23351181e-01 -7.42325723e-01 4.97730106e-01
3.78406733e-01 -1.11470747e+00 -3.30819309e-01 -3.69100779e-01] | [4.703334331512451, 0.5293849110603333] |
14b819a6-6c7c-46eb-8e52-8f3b37c38590 | rethinking-video-salient-object-ranking | 2203.17257 | null | https://arxiv.org/abs/2203.17257v1 | https://arxiv.org/pdf/2203.17257v1.pdf | Rethinking Video Salient Object Ranking | Salient Object Ranking (SOR) involves ranking the degree of saliency of multiple salient objects in an input image. Most recently, a method is proposed for ranking salient objects in an input video based on a predicted fixation map. It relies solely on the density of the fixations within the salient objects to infer their saliency ranks, which is incompatible with human perception of saliency ranking. In this work, we propose to explicitly learn the spatial and temporal relations between different salient objects to produce the saliency ranks. To this end, we propose an end-to-end method for video salient object ranking (VSOR), with two novel modules: an intra-frame adaptive relation (IAR) module to learn the spatial relation among the salient objects in the same frame locally and globally, and an inter-frame dynamic relation (IDR) module to model the temporal relation of saliency across different frames. In addition, to address the limited video types (just sports and movies) and scene diversity in the existing VSOR dataset, we propose a new dataset that covers different video types and diverse scenes on a large scale. Experimental results demonstrate that our method outperforms state-of-the-art methods in relevant fields. We will make the source code and our proposed dataset available. | ['Rynson W. H. Lau', 'Huankang Guan', 'Jiaying Lin'] | 2022-03-31 | null | null | null | null | ['saliency-ranking'] | ['computer-vision'] | [ 2.38460019e-01 -2.04954341e-01 -3.06121379e-01 -3.89200777e-01
-5.02918780e-01 -1.66157648e-01 2.87756413e-01 2.77761638e-01
-2.10795745e-01 4.79680598e-01 4.96137440e-01 3.22611004e-01
-9.98964533e-02 -3.72969300e-01 -7.58063197e-01 -4.42390561e-01
-1.98842585e-01 -1.48133770e-01 1.13051200e+00 -3.78376603e-01
6.96597993e-01 1.15600027e-01 -1.99105382e+00 5.81897855e-01
7.73644328e-01 1.03545153e+00 9.78825390e-01 4.56844300e-01
2.36240834e-01 1.33379376e+00 -2.83441365e-01 -8.77524912e-02
1.20942734e-01 -5.07086277e-01 -8.16687584e-01 9.98336375e-02
6.52049482e-01 -4.65396732e-01 -4.30999696e-01 1.10226274e+00
3.76781106e-01 3.99060845e-01 2.80204922e-01 -1.21254981e+00
-6.75041139e-01 6.00265145e-01 -7.93178618e-01 9.24237013e-01
3.67012918e-01 -8.90601650e-02 1.12702286e+00 -1.10026097e+00
6.25977993e-01 1.25441945e+00 8.68944637e-03 4.08884734e-01
-7.03349650e-01 -3.93714011e-01 6.61209762e-01 8.57132554e-01
-1.24872780e+00 -4.07703638e-01 1.05862725e+00 -5.31841695e-01
2.79246807e-01 3.83831799e-01 8.50162864e-01 6.23761237e-01
1.40651865e-02 1.16332865e+00 9.24591660e-01 -1.18635356e-01
1.89442471e-01 -1.66351482e-01 1.42680973e-01 5.52660644e-01
-4.45240662e-02 -8.64515230e-02 -1.01397741e+00 2.64022276e-02
8.37844014e-01 4.28450942e-01 -2.42283210e-01 -5.65234601e-01
-1.52166188e+00 6.34600341e-01 7.75205135e-01 1.76353022e-01
-4.62887138e-01 3.77578028e-02 3.33905548e-01 -1.82135299e-01
4.68555003e-01 2.68576592e-01 -3.89404297e-01 1.06218137e-01
-9.78818595e-01 3.97305191e-01 1.76223572e-02 9.49868619e-01
9.23527718e-01 -1.06151894e-01 -7.04022229e-01 7.52488136e-01
3.44796926e-01 2.83482283e-01 5.55389881e-01 -9.48125005e-01
3.59176278e-01 6.13849044e-01 3.86570692e-01 -1.42121208e+00
-1.76870644e-01 -2.26056486e-01 -3.85613590e-01 -1.96570933e-01
-3.33467126e-02 3.65122288e-01 -7.28587568e-01 1.75099218e+00
4.63138551e-01 7.04886675e-01 -2.86339283e-01 1.54210842e+00
1.07678032e+00 6.74188197e-01 2.47018933e-01 -3.47933352e-01
1.23047185e+00 -1.20250106e+00 -4.99684662e-01 -3.80045801e-01
-7.24369362e-02 -8.15093458e-01 1.23798525e+00 -5.99508658e-02
-1.32590008e+00 -7.84460783e-01 -7.71051168e-01 -3.14413875e-01
-6.41925707e-02 1.01696514e-01 5.03684580e-01 -2.50487000e-01
-1.06376040e+00 3.45515102e-01 -5.72988033e-01 -2.86285847e-01
4.45938289e-01 1.49334908e-01 1.87508687e-01 2.18206756e-02
-1.29678369e+00 6.93820775e-01 4.44614500e-01 -9.13736783e-03
-1.28220522e+00 -7.35338628e-01 -7.78639197e-01 2.36106142e-01
4.46611911e-01 -5.20094991e-01 1.15116012e+00 -1.45938146e+00
-1.16413462e+00 6.72366321e-01 -5.78807771e-01 -2.97760129e-01
9.09879431e-02 -3.43631506e-01 -2.80711591e-01 4.56719309e-01
4.32619601e-01 8.98506284e-01 9.42086816e-01 -1.30243981e+00
-1.09348500e+00 -1.50999174e-01 3.95778477e-01 5.12311041e-01
-2.87829518e-01 3.79282087e-01 -6.08123124e-01 -7.82289684e-01
6.67522326e-02 -7.05445647e-01 -1.82350948e-01 -1.62053183e-01
-1.58696279e-01 -2.79507577e-01 8.47973168e-01 -5.08354306e-01
1.40083981e+00 -2.19289517e+00 3.57729644e-01 -1.39757231e-01
3.77383709e-01 9.43061560e-02 -2.23326758e-01 -2.19265353e-02
2.45629693e-03 -3.35504949e-01 5.56845143e-02 6.87808022e-02
-3.47214520e-01 -2.62090832e-01 -5.06905675e-01 3.38172197e-01
2.62055010e-01 8.79804373e-01 -1.52299500e+00 -8.33012998e-01
3.07721674e-01 3.91425312e-01 -5.54663539e-01 3.77271652e-01
-1.10941976e-01 4.20463860e-01 -6.46920741e-01 7.13560343e-01
4.53875452e-01 -3.47104341e-01 -1.59454495e-01 -2.22803906e-01
-2.87329853e-01 3.37456763e-01 -9.98535335e-01 1.74117005e+00
1.36153534e-01 5.97759187e-01 -3.46490592e-01 -7.55102873e-01
8.32567394e-01 -9.89827421e-03 6.24508262e-01 -7.64644265e-01
-1.04993299e-01 1.64141789e-01 -3.76662314e-02 -4.32146698e-01
9.67802405e-01 2.64218599e-01 1.43207774e-01 4.32333052e-02
-4.01092693e-02 4.38298494e-01 5.56518555e-01 4.41387057e-01
7.91874230e-01 1.58731014e-01 2.43433520e-01 -5.80215573e-01
6.11685812e-01 -2.37827539e-03 9.14693892e-01 6.69544995e-01
-5.33055425e-01 1.01803660e+00 3.11181396e-01 -6.34851217e-01
-8.81618917e-01 -1.25100744e+00 1.63899645e-01 1.65239966e+00
1.14821255e+00 -4.26059425e-01 -5.62520087e-01 -5.03353953e-01
-3.04583281e-01 2.76211411e-01 -7.22586095e-01 -1.34970352e-01
-6.40240312e-01 -3.44052404e-01 -4.33212429e-01 4.47557837e-01
4.46483791e-01 -1.42067516e+00 -1.00503039e+00 3.45111638e-02
-6.36135757e-01 -9.16992724e-01 -9.03669178e-01 -3.38829577e-01
-7.71196544e-01 -1.09054899e+00 -8.77206087e-01 -1.03624308e+00
7.49345839e-01 1.00432169e+00 1.24563122e+00 1.49423420e-01
-9.29013044e-02 1.03933595e-01 -6.13804162e-01 -3.50366861e-01
3.59965295e-01 -7.50888512e-02 1.43193109e-02 2.67246664e-01
2.55631804e-01 -1.32019252e-01 -1.04221404e+00 6.08967602e-01
-9.06997561e-01 5.68593681e-01 4.88067329e-01 5.24589717e-01
8.85957062e-01 -1.54716834e-01 5.05037189e-01 -6.08636558e-01
1.16230465e-01 -6.50147974e-01 -3.75846446e-01 3.05344969e-01
-2.83133332e-02 -2.79280007e-01 2.43459940e-01 -4.87895131e-01
-9.16142642e-01 7.74803311e-02 4.65627521e-01 -7.12360740e-01
1.19588934e-01 4.17090356e-01 5.89767331e-03 2.66140342e-01
2.74834156e-01 3.50447208e-01 -5.07643282e-01 -1.06301598e-01
1.51260123e-01 2.78523386e-01 6.02028489e-01 -3.10339302e-01
6.52861595e-01 5.23182213e-01 -1.84885100e-01 -4.60905135e-01
-1.48106575e+00 -8.04513872e-01 -7.56556690e-01 -5.63519657e-01
8.72310340e-01 -1.36244357e+00 -3.53539556e-01 2.41421416e-01
-1.11796093e+00 -1.80725366e-01 -2.69696951e-01 5.87517321e-01
-5.57418346e-01 2.65368372e-01 -3.83601040e-01 -4.61789578e-01
-2.49457985e-01 -1.08912635e+00 1.26535475e+00 6.37078762e-01
-6.63901269e-02 -5.25908113e-01 -1.06625475e-01 2.29834571e-01
3.34359020e-01 1.24990538e-01 4.74421531e-01 -7.96877071e-02
-1.01195037e+00 3.85817438e-01 -4.98872906e-01 -1.70208767e-01
1.87284023e-01 1.04785245e-02 -6.61057353e-01 -3.55258286e-01
-1.71203315e-01 -1.20158151e-01 9.72232401e-01 8.08347404e-01
1.39738905e+00 -2.59622335e-01 -2.38546506e-01 2.22639039e-01
1.18913639e+00 1.11863814e-01 6.67806447e-01 4.62872863e-01
9.30005431e-01 7.40451932e-01 1.37667358e+00 5.64948380e-01
6.16768420e-01 8.88716698e-01 5.21404922e-01 -2.03159362e-01
-1.37384802e-01 -2.74812609e-01 5.45285404e-01 7.72691846e-01
-1.58308864e-01 -2.16014516e-02 -5.80935240e-01 7.90250003e-01
-2.32836032e+00 -1.39734030e+00 -1.53545931e-01 2.18398786e+00
7.18781710e-01 2.58972235e-02 3.49271029e-01 -2.66048133e-01
1.19790316e+00 4.28573489e-01 -6.67563498e-01 8.66435934e-03
-7.33564645e-02 -4.44919288e-01 8.96241888e-02 3.76695126e-01
-1.26228309e+00 1.03147233e+00 6.14394140e+00 7.87023783e-01
-1.09805119e+00 1.47734955e-01 8.39648604e-01 -4.17101502e-01
-3.03563684e-01 3.29033099e-02 -7.18640685e-01 7.58277297e-01
3.85111004e-01 -2.79509962e-01 2.93408424e-01 9.03289199e-01
4.13752407e-01 -3.59725535e-01 -8.92005682e-01 1.04114175e+00
3.46319467e-01 -1.31826174e+00 1.63236260e-01 -3.29464436e-01
8.66941869e-01 -1.65732279e-02 2.04694048e-01 8.85097831e-02
-7.73839653e-02 -5.84499180e-01 1.20083082e+00 6.91579998e-01
3.28279078e-01 -7.00091302e-01 4.83503938e-01 1.12448605e-02
-1.52044773e+00 -2.03726545e-01 -6.68535054e-01 -1.96505710e-01
6.99785948e-02 5.85362732e-01 -4.42046463e-01 2.17823207e-01
1.13107514e+00 1.34983718e+00 -9.26283479e-01 1.26760852e+00
-1.21820033e-01 2.81233221e-01 1.42924637e-01 -3.33225653e-02
8.78646225e-02 5.47834076e-02 6.56464398e-01 9.28387702e-01
2.26472855e-01 2.29455590e-01 3.55352730e-01 7.08788216e-01
1.70897573e-01 9.59826857e-02 -1.86420381e-01 3.50221604e-01
3.82773697e-01 1.36898780e+00 -8.27782273e-01 -4.38417286e-01
-3.30363929e-01 8.84217501e-01 3.92315179e-01 2.90773690e-01
-1.02082586e+00 -5.65401651e-02 5.46549797e-01 4.08216089e-01
5.11346102e-01 -1.70208011e-02 -4.00142968e-02 -1.37290967e+00
1.19728312e-01 -5.96582115e-01 5.08607268e-01 -1.12311673e+00
-1.09535611e+00 4.24422443e-01 1.27446190e-01 -1.63599801e+00
8.26951712e-02 -7.39610344e-02 -6.63938463e-01 6.65022969e-01
-1.85240936e+00 -9.38766599e-01 -4.93649989e-01 6.95489645e-01
9.58643734e-01 -1.78084269e-01 9.65519175e-02 1.71952918e-01
-4.29739028e-01 1.16354883e-01 -1.70901552e-01 -3.81262861e-02
6.76572561e-01 -1.10103440e+00 1.36407688e-01 1.13039148e+00
1.12521917e-01 5.62961698e-01 6.99494243e-01 -5.95430791e-01
-1.14378774e+00 -1.23266625e+00 9.97199535e-01 -4.01134193e-01
4.89035845e-01 -1.64570287e-01 -8.21456313e-01 3.24844837e-01
8.79438296e-02 2.24964410e-01 4.21044171e-01 5.39186550e-03
-9.16953012e-02 -2.21835494e-01 -7.19834447e-01 7.77326286e-01
1.16283166e+00 -4.14632738e-01 -6.01365089e-01 1.61515623e-01
9.26499844e-01 -4.45714355e-01 -4.63922173e-01 5.98857820e-01
4.94295835e-01 -1.14242339e+00 1.12179828e+00 -3.35693955e-01
9.85270441e-01 -9.92030621e-01 -2.31068358e-01 -9.22114193e-01
-7.60759175e-01 -3.56642693e-01 -3.45645785e-01 1.20612907e+00
-2.06750687e-02 1.41858026e-01 6.37087405e-01 4.17880446e-01
-2.87461847e-01 -8.27578843e-01 -7.61588931e-01 -2.32155249e-01
-7.27881551e-01 8.26740265e-02 5.21525800e-01 7.71131098e-01
-2.48245075e-02 3.59912843e-01 -5.81335127e-01 2.77475238e-01
4.50891852e-01 6.59676790e-01 4.67538685e-01 -1.25300527e+00
-1.12826280e-01 -3.75485331e-01 -6.44441128e-01 -1.07757711e+00
2.84727514e-02 -6.06219232e-01 3.02162945e-01 -1.40535688e+00
9.52847183e-01 -2.62149513e-01 -9.69861686e-01 2.85221487e-01
-8.69125724e-01 2.99832374e-01 4.48199958e-01 5.60175061e-01
-1.52848542e+00 6.72640920e-01 1.37881184e+00 -8.39668885e-02
-1.84065104e-01 -2.40850329e-01 -9.16872442e-01 7.52550781e-01
5.97102284e-01 -3.60434741e-01 -6.01881921e-01 -5.23227036e-01
2.35666055e-02 2.77967248e-02 5.66878259e-01 -1.07969940e+00
9.80279222e-02 -6.08592808e-01 5.68409801e-01 -8.59695971e-01
1.50477529e-01 -5.82038522e-01 -1.88039929e-01 2.86738962e-01
-6.31170034e-01 1.60802543e-01 -7.04265088e-02 6.09722078e-01
-5.01226246e-01 -2.00049244e-02 9.06825662e-01 1.17087457e-02
-1.37627780e+00 6.33743465e-01 2.76608355e-02 2.11330816e-01
1.17988122e+00 -6.19395562e-02 -3.53843093e-01 -2.10061297e-01
-3.81177872e-01 3.60519648e-01 5.77777386e-01 8.34819257e-01
1.01021242e+00 -1.53220069e+00 -7.48149812e-01 -4.26213853e-02
3.02997351e-01 -2.76444014e-02 5.50259650e-01 7.34242141e-01
-1.72069252e-01 1.85854375e-01 -4.55874145e-01 -7.23423660e-01
-1.40606618e+00 7.35858262e-01 2.09380891e-02 3.97145189e-02
-3.41216505e-01 8.35306227e-01 8.96925390e-01 2.20196038e-01
1.24383427e-01 -1.26045153e-01 -7.08474100e-01 3.38726188e-03
9.14779305e-01 2.83083737e-01 -4.65474993e-01 -1.07104933e+00
-4.21960324e-01 5.97244263e-01 -2.42645547e-01 2.35094681e-01
1.17743111e+00 -4.86951023e-01 -2.04814672e-01 7.09871709e-01
9.52936590e-01 -1.58389837e-01 -1.65094602e+00 -3.03115785e-01
-1.82223216e-01 -9.80506063e-01 3.05338055e-02 -4.01732534e-01
-1.18377244e+00 6.14849031e-01 7.11414874e-01 2.71767844e-03
1.31183600e+00 2.45367751e-01 7.91538894e-01 -9.33392420e-02
4.37313914e-01 -1.07207084e+00 4.12342638e-01 5.00263035e-01
8.43921721e-01 -1.44861174e+00 1.49848863e-01 -5.57590842e-01
-1.05559886e+00 6.89795911e-01 8.97988081e-01 -2.14157715e-01
5.20859659e-01 -4.72761542e-01 -1.09912336e-01 -2.22426876e-01
-7.47377336e-01 -4.78924096e-01 7.17406571e-01 3.84664744e-01
4.56786811e-01 -3.35789248e-02 -4.15363520e-01 4.42680836e-01
1.99876398e-01 5.20709641e-02 5.03747225e-01 7.76875198e-01
-7.98957109e-01 -5.84070504e-01 -1.90663204e-01 4.85645413e-01
-4.26811397e-01 -2.59855092e-01 -1.73229218e-01 1.55346096e-01
1.86289892e-01 9.99392748e-01 2.72886455e-01 -3.42280984e-01
1.09827258e-01 -5.91996968e-01 1.92545936e-01 -6.93210542e-01
-3.46713364e-01 4.83454317e-02 -4.36897457e-01 -7.70419002e-01
-9.84847009e-01 -8.56650710e-01 -1.27512157e+00 1.25828059e-03
-1.29959822e-01 9.20546502e-02 2.88944575e-03 7.19849586e-01
5.06655812e-01 5.41544139e-01 8.82302761e-01 -1.26756763e+00
1.18559927e-01 -7.44203150e-01 -7.40327537e-01 7.53704727e-01
3.96980822e-01 -9.72527742e-01 -9.14958213e-03 3.63525391e-01] | [9.762688636779785, -0.16618628799915314] |
981dded1-04bf-4ec3-bbff-fccc93981485 | stop-hop-early-classification-of-irregular | 2208.09795 | null | https://arxiv.org/abs/2208.09795v1 | https://arxiv.org/pdf/2208.09795v1.pdf | Stop&Hop: Early Classification of Irregular Time Series | Early classification algorithms help users react faster to their machine learning model's predictions. Early warning systems in hospitals, for example, let clinicians improve their patients' outcomes by accurately predicting infections. While early classification systems are advancing rapidly, a major gap remains: existing systems do not consider irregular time series, which have uneven and often-long gaps between their observations. Such series are notoriously pervasive in impactful domains like healthcare. We bridge this gap and study early classification of irregular time series, a new setting for early classifiers that opens doors to more real-world problems. Our solution, Stop&Hop, uses a continuous-time recurrent network to model ongoing irregular time series in real time, while an irregularity-aware halting policy, trained with reinforcement learning, predicts when to stop and classify the streaming series. By taking real-valued step sizes, the halting policy flexibly decides exactly when to stop ongoing series in real time. This way, Stop&Hop seamlessly integrates information contained in the timing of observations, a new and vital source for early classification in this setting, with the time series values to provide early classifications for irregular time series. Using four synthetic and three real-world datasets, we demonstrate that Stop&Hop consistently makes earlier and more-accurate predictions than state-of-the-art alternatives adapted to this new problem. Our code is publicly available at https://github.com/thartvigsen/StopAndHop. | ['Elke Rundensteiner', 'Xiangnan Kong', 'Jidapa Thadajarassiri', 'Walter Gerych', 'Thomas Hartvigsen'] | 2022-08-21 | null | null | null | null | ['classification', 'irregular-time-series'] | ['methodology', 'time-series'] | [ 2.28330940e-01 -1.29703656e-01 -4.94728625e-01 -2.75483608e-01
-7.54234672e-01 -3.65855485e-01 3.52777034e-01 6.89377904e-01
-3.43671024e-01 4.94853050e-01 1.93036363e-01 -8.13150465e-01
-3.65473807e-01 -6.04821086e-01 -3.00909519e-01 -5.62214851e-01
-7.35310137e-01 4.24150318e-01 1.00547686e-01 -1.80177093e-01
2.26133466e-01 2.80916959e-01 -1.23537672e+00 5.84513068e-01
6.56231046e-01 1.15815961e+00 -2.58125186e-01 1.21406341e+00
1.30831257e-01 1.12831998e+00 -4.52347577e-01 2.53911465e-01
4.86079186e-01 -3.42904299e-01 -5.19692421e-01 -3.12635630e-01
-4.75307047e-01 -2.94073671e-01 -3.00429702e-01 3.11404526e-01
4.54913288e-01 -2.59105042e-02 3.67395371e-01 -1.36440802e+00
-8.12690258e-02 6.71629012e-01 -2.18128905e-01 6.62653983e-01
3.32789212e-01 4.18780029e-01 8.05446506e-01 -3.61276746e-01
2.54324913e-01 6.58380628e-01 1.26902592e+00 5.25097907e-01
-1.00697958e+00 -5.55025339e-01 9.77438241e-02 1.68030724e-01
-9.10346091e-01 -3.98761153e-01 6.64602637e-01 -6.72229707e-01
8.87363374e-01 4.86938059e-01 7.91860521e-01 1.24263561e+00
5.85372567e-01 7.70256221e-01 8.05392861e-01 -1.18925266e-01
3.17045659e-01 -1.98274046e-01 2.30580479e-01 3.72384638e-01
-1.23649739e-01 5.21038592e-01 -2.22160622e-01 -5.56016266e-01
5.11324346e-01 1.21211720e+00 -3.35176677e-01 2.60016948e-01
-1.46550667e+00 6.75395966e-01 1.80817440e-01 1.78924829e-01
-8.00566375e-01 -6.85463399e-02 6.45099163e-01 6.34124577e-01
6.16434395e-01 6.92162871e-01 -6.68937683e-01 -6.41327620e-01
-8.26461494e-01 8.59455541e-02 1.00823510e+00 6.22365355e-01
1.99106097e-01 -1.67592496e-01 -3.10620725e-01 5.03173769e-01
-8.01843777e-02 3.23690325e-01 6.78121090e-01 -9.02152658e-01
2.40077987e-01 3.24895084e-01 2.74925053e-01 -8.78870130e-01
-7.98750460e-01 -6.24163270e-01 -1.03599262e+00 -1.38642207e-01
5.71692944e-01 -3.58110070e-01 -8.32890868e-01 1.42684615e+00
2.69418687e-01 7.94102311e-01 9.30379778e-02 5.34964263e-01
1.02339610e-01 7.66187489e-01 -9.01911780e-02 -6.84119821e-01
1.10929525e+00 -7.57859111e-01 -5.74914455e-01 -8.79293978e-02
9.36630785e-01 -5.77290773e-01 5.70925891e-01 6.23454094e-01
-7.51389503e-01 -2.04843491e-01 -6.67207658e-01 5.10612130e-01
-3.59895080e-01 -5.12744546e-01 4.59574938e-01 5.09453379e-02
-7.80342042e-01 1.05118740e+00 -1.35612357e+00 -1.41146302e-01
3.80516171e-01 1.97376132e-01 1.54425144e-01 1.14976995e-01
-1.33083057e+00 6.16607487e-01 4.44533750e-02 1.32200420e-01
-5.90806603e-01 -1.06762302e+00 -6.06570005e-01 -1.14823945e-01
3.12763989e-01 -6.46066189e-01 1.87377667e+00 -8.60918641e-01
-1.22663200e+00 4.85677600e-01 -4.93999682e-02 -1.04644799e+00
8.22230399e-01 -2.43381768e-01 -6.03876352e-01 8.13271403e-02
-8.16715583e-02 -8.48183781e-02 7.70334899e-01 -6.74391150e-01
-9.75981116e-01 3.20935645e-03 -4.26206172e-01 -1.41838565e-01
-5.95747828e-02 -1.53520003e-01 3.09466779e-01 -8.03833187e-01
-1.78945154e-01 -8.44829202e-01 -7.50970960e-01 2.72450689e-02
-1.68924585e-01 -3.36858273e-01 7.64923871e-01 -5.50396383e-01
1.71732759e+00 -2.20753312e+00 -5.07289946e-01 2.86885500e-01
5.42643547e-01 3.22872341e-01 1.25596389e-01 8.21148694e-01
-3.41281861e-01 2.11546812e-02 -3.00287902e-01 -2.48378590e-01
-3.35598111e-01 3.96306753e-01 -7.16822326e-01 3.61581355e-01
1.48218244e-01 7.36287117e-01 -1.39525962e+00 -1.11539990e-01
2.23914638e-01 3.51241767e-01 -4.20921594e-01 6.07387841e-01
-1.17105260e-01 5.41356444e-01 -3.82440358e-01 5.75045586e-01
-2.20426451e-02 -6.66307926e-01 -2.46095791e-01 4.30225134e-01
-1.37868777e-01 1.55087471e-01 -7.16285706e-01 1.09731734e+00
-4.42630887e-01 6.00066841e-01 -5.36395609e-01 -1.39882970e+00
8.52896273e-01 6.74031556e-01 1.12089896e+00 -8.02789271e-01
1.24250546e-01 4.82976258e-01 -5.37763089e-02 -8.11597943e-01
4.53877971e-02 -1.34790286e-01 -5.40456176e-02 5.36406934e-01
-6.91436887e-01 2.87039906e-01 7.98088759e-02 -4.33827005e-02
1.73030198e+00 -2.52255142e-01 6.64358258e-01 2.51447350e-01
5.18776588e-02 2.73432843e-02 7.78282464e-01 9.34524596e-01
-3.62993360e-01 6.80455327e-01 5.52358270e-01 -1.06266785e+00
-1.02633417e+00 -9.72334445e-01 -5.35960756e-02 1.02431226e+00
-2.92041421e-01 -3.00307512e-01 -2.46522069e-01 -5.86804986e-01
1.58852711e-01 4.85175610e-01 -9.19820607e-01 -3.84518623e-01
-8.31277609e-01 -6.88783586e-01 1.66042820e-01 5.55249691e-01
-1.65553346e-01 -1.17890382e+00 -1.14742768e+00 8.44780207e-01
-1.87555730e-01 -8.60361040e-01 -5.08140504e-01 5.94694734e-01
-1.00875008e+00 -1.27602267e+00 -8.15200448e-01 -2.67741621e-01
4.40044284e-01 8.87292176e-02 1.22794449e+00 2.91298747e-01
-4.76305604e-01 4.04042423e-01 -6.03083789e-01 -7.87926078e-01
-7.17617273e-01 3.51633616e-02 2.82562338e-02 1.58027157e-01
2.63880402e-01 -6.31794691e-01 -1.10003340e+00 1.12587012e-01
-8.79004419e-01 4.22195494e-02 3.36374074e-01 1.07812130e+00
5.31167626e-01 -3.40149969e-01 9.75054264e-01 -1.16912234e+00
5.16501427e-01 -1.15693176e+00 -3.21621299e-01 1.94567963e-02
-9.78148818e-01 5.38634993e-02 1.20721555e+00 -6.82383895e-01
-1.63682759e-01 -1.49651185e-01 -3.46760631e-01 -5.35478890e-01
-6.54227734e-02 5.22058249e-01 1.00207460e+00 6.60602093e-01
8.50524545e-01 3.50085706e-01 1.01278678e-01 -2.67859012e-01
-2.77655423e-01 9.11459506e-01 5.16129315e-01 -2.33088568e-01
4.07998919e-01 3.81382316e-01 -2.64655799e-01 -5.16688824e-01
-9.92144942e-01 -8.49578798e-01 -2.59474546e-01 -1.11335248e-01
3.80889028e-01 -7.18730748e-01 -8.43236089e-01 4.02264535e-01
-8.89927089e-01 -8.85007501e-01 -6.44973457e-01 4.00176436e-01
-6.67590737e-01 -7.09589571e-02 -8.33515584e-01 -1.07544172e+00
-5.35209119e-01 -5.98663032e-01 1.01493871e+00 9.90260392e-02
-6.66861415e-01 -1.23689806e+00 3.53421152e-01 -2.35885009e-01
5.35673916e-01 6.90693557e-01 6.40383542e-01 -1.11963916e+00
-1.98766768e-01 -3.75421822e-01 9.77292955e-02 9.60579813e-02
3.84432286e-01 8.50059912e-02 -8.43183458e-01 -2.92036027e-01
4.94280607e-02 1.97719589e-01 7.56692410e-01 3.61471832e-01
1.40143931e+00 -8.20055783e-01 -5.11994481e-01 8.19720805e-01
1.14993560e+00 7.33856022e-01 3.19181979e-01 2.59778619e-01
2.98795849e-01 2.25083113e-01 7.02041924e-01 1.00968921e+00
4.17790413e-01 1.31235212e-01 3.24447930e-01 -2.67807424e-01
4.35943604e-01 -9.29442942e-02 2.25320369e-01 9.29232597e-01
-1.05782747e-01 -7.04922229e-02 -1.35163450e+00 6.29672825e-01
-2.03099966e+00 -1.19780147e+00 1.39781103e-01 2.36731362e+00
1.13220322e+00 4.16644156e-01 3.75490755e-01 5.60920835e-01
3.38815808e-01 -2.00186223e-02 -7.32538223e-01 -3.73503119e-01
4.42244709e-01 1.17233135e-01 3.66274625e-01 1.84049234e-01
-1.18509376e+00 2.05728173e-01 6.02074528e+00 3.12176406e-01
-1.58930135e+00 -1.09750114e-01 1.06112337e+00 -6.94221035e-02
2.00642958e-01 -3.34341168e-01 -3.54495168e-01 8.98758471e-01
1.74078000e+00 -3.78084630e-01 4.82548326e-01 8.24256837e-01
7.38264024e-01 1.95447773e-01 -1.52046776e+00 1.04413140e+00
-5.41279376e-01 -1.58505619e+00 -5.60486317e-01 -2.24252403e-01
4.64574337e-01 4.26185995e-01 -3.05121429e-02 3.59493107e-01
3.63927871e-01 -9.27595258e-01 2.89338082e-01 8.82083654e-01
6.92657590e-01 -6.40228450e-01 7.17068315e-01 6.34560049e-01
-1.16259038e+00 -6.12440050e-01 2.85918683e-01 -2.72867352e-01
8.73761773e-02 9.34030414e-01 -1.15119684e+00 1.08179420e-01
6.26977205e-01 1.21224165e+00 -2.21469209e-01 1.37290180e+00
2.64876693e-01 1.01820302e+00 -5.31453848e-01 4.24538925e-02
1.27394855e-01 3.83730710e-01 5.60854316e-01 1.42421436e+00
5.29246569e-01 3.21415722e-01 5.08207262e-01 2.28644386e-01
3.90002489e-01 -2.10708559e-01 -8.50516319e-01 -2.00222775e-01
6.38182878e-01 8.72123480e-01 -7.16807127e-01 -6.65472507e-01
-2.52093762e-01 6.57340050e-01 -1.12740584e-01 4.89950746e-01
-7.37199664e-01 -5.15076458e-01 6.19169056e-01 3.80023599e-01
-1.39586359e-01 -8.56123939e-02 -3.50010157e-01 -1.11134195e+00
-4.34185341e-02 -9.85152960e-01 8.94692183e-01 -4.56331700e-01
-1.46270502e+00 6.85258329e-01 -3.87621582e-01 -1.81331003e+00
-7.95467615e-01 -3.67017925e-01 -7.77791083e-01 2.34156147e-01
-1.53869629e+00 -4.56635445e-01 -1.58070996e-01 5.87811947e-01
8.46466660e-01 3.43425035e-01 9.89594638e-01 1.68670624e-01
-5.30799031e-01 4.09756213e-01 4.04124498e-01 3.37978363e-01
7.03373730e-01 -1.31468844e+00 5.11597574e-01 5.58741271e-01
-1.94517970e-01 3.69193584e-01 8.91700387e-01 -4.13633913e-01
-1.18435621e+00 -1.20804691e+00 8.49084020e-01 -5.14242947e-01
8.16599786e-01 -1.28659204e-01 -1.26760280e+00 7.09571242e-01
-1.37406066e-01 3.76716793e-01 7.30642259e-01 -1.30096033e-01
-2.03198940e-01 -2.72736520e-01 -9.09239709e-01 4.03186083e-01
8.15362751e-01 -3.61146033e-01 -5.51922619e-01 5.94458818e-01
7.72277713e-01 -5.70877254e-01 -9.23615634e-01 4.56593394e-01
5.55369198e-01 -8.84265125e-01 7.93393373e-01 -5.96901178e-01
3.90212864e-01 2.99180239e-01 2.94714570e-01 -1.51516247e+00
-1.12356074e-01 -1.34504914e+00 -6.22569799e-01 3.78820688e-01
4.28132892e-01 -1.16683936e+00 6.10887229e-01 4.57777917e-01
-2.50612170e-01 -1.38062966e+00 -6.69911683e-01 -6.65774286e-01
-1.78026929e-01 -5.18188655e-01 5.25297821e-01 1.12073565e+00
3.26004297e-01 -1.05451047e-01 -4.06937957e-01 2.83320118e-02
3.96747112e-01 2.51702785e-01 3.18260342e-01 -1.42418349e+00
-3.74396324e-01 -5.97798765e-01 -1.33810267e-01 -7.61216998e-01
-5.82804203e-01 -6.40191913e-01 1.96592420e-01 -1.27219510e+00
-9.05547217e-02 -7.61425972e-01 -7.96309471e-01 8.33406568e-01
-2.91006029e-01 -2.13832464e-02 -6.19397163e-02 2.84938008e-01
-5.49358964e-01 1.75000355e-01 8.65446448e-01 1.19981378e-01
-5.18075347e-01 5.75391710e-01 -3.92191261e-01 6.72376454e-01
9.65064645e-01 -6.46662593e-01 -3.07395309e-01 1.94791742e-02
2.73689836e-01 7.15045750e-01 3.18576425e-01 -1.09854507e+00
4.87214208e-01 -5.59531510e-01 2.32597709e-01 -4.65964377e-01
-5.41730262e-02 -8.40542614e-01 -1.29050156e-02 9.84695375e-01
-5.80982745e-01 4.99410063e-01 1.10630766e-01 6.91449344e-01
-3.85437198e-02 4.17290181e-01 7.10300684e-01 -1.94571093e-02
-3.68587852e-01 7.66557515e-01 -5.34164071e-01 4.96463686e-01
1.17147112e+00 -1.82798177e-01 -2.22950220e-01 -4.88627344e-01
-8.99476528e-01 6.07420444e-01 -6.75360858e-02 5.32897532e-01
6.58328116e-01 -1.04592252e+00 -7.96259224e-01 4.93983865e-01
1.02924123e-01 1.34467036e-01 2.02879325e-01 1.23832238e+00
-3.78752172e-01 1.56123921e-01 1.77507967e-01 -9.09457922e-01
-9.28292692e-01 6.89085543e-01 4.66061890e-01 -4.07412738e-01
-9.33033943e-01 4.50920194e-01 -1.09224707e-01 -6.96568787e-02
4.09095585e-01 -7.96413660e-01 2.46802680e-02 1.33062586e-01
8.70875716e-01 3.01829547e-01 -3.29336263e-02 1.39428720e-01
-2.08272874e-01 1.98248088e-01 -1.44200563e-01 2.02433929e-01
1.73971879e+00 9.25900117e-02 3.37393969e-01 9.60367560e-01
1.24698639e+00 -4.24399108e-01 -1.50996935e+00 -4.32752013e-01
4.07305121e-01 -1.71984166e-01 -3.84255201e-01 -7.00167656e-01
-7.54821181e-01 8.09521914e-01 5.42820871e-01 8.46465051e-01
1.33051991e+00 -2.87743837e-01 1.26856017e+00 4.50334430e-01
2.30442882e-01 -8.61566365e-01 9.07615721e-02 5.32743871e-01
6.60888910e-01 -1.31820524e+00 -3.52961212e-01 1.92478746e-01
-5.32965660e-01 1.22259903e+00 9.17211697e-02 -2.76521742e-01
9.10365820e-01 4.76628959e-01 3.83474112e-01 1.82209373e-01
-1.43163681e+00 1.53847709e-01 -1.20147273e-01 4.02408481e-01
2.99221039e-01 1.96259081e-01 4.58040573e-02 5.48985124e-01
-1.65029436e-01 4.54878181e-01 4.85048473e-01 1.07893586e+00
-3.65581244e-01 -8.58252287e-01 -2.88896680e-01 7.94487536e-01
-7.12415934e-01 -3.93189415e-02 2.85512924e-01 4.35600668e-01
-1.03491947e-01 8.02819908e-01 3.16269994e-01 -4.12866503e-01
2.18493387e-01 8.64461958e-02 -3.16942811e-01 -5.84172785e-01
-6.66624665e-01 -3.95297110e-02 -1.47842810e-01 -8.09493721e-01
1.48179919e-01 -7.12714255e-01 -1.59669745e+00 -2.41124645e-01
1.32994324e-01 2.39263624e-01 3.15464169e-01 9.84219849e-01
6.12796545e-01 7.74448156e-01 1.29902923e+00 -7.93165624e-01
-9.35437262e-01 -7.06681073e-01 -8.08110163e-02 3.78748715e-01
1.32284951e+00 -2.16369525e-01 -6.15724444e-01 1.34962246e-01] | [7.284955024719238, 3.259550094604492] |
59acb3f8-9e38-435c-974f-cd51d825c1eb | parameter-estimation-in-ill-conditioned-low | 2208.04471 | null | https://arxiv.org/abs/2208.04471v1 | https://arxiv.org/pdf/2208.04471v1.pdf | Parameter Estimation in Ill-conditioned Low-inertia Power Systems | This paper examines model parameter estimation in dynamic power systems whose governing electro-mechanical equations are ill-conditioned or singular. This ill-conditioning is because of converter-interfaced power systems generators' zero or small inertia contribution. Consequently, the overall system inertia decreases, resulting in low-inertia power systems. We show that the standard state-space model based on least squares or subspace estimators fails to exist for these models. We overcome this challenge by considering a least-squares estimator directly on the coupled swing-equation model but not on its transformed first-order state-space form. We specifically focus on estimating inertia (mechanical and virtual) and damping constants, although our method is general enough for estimating other parameters. Our theoretical analysis highlights the role of network topology on the parameter estimates of an individual generator. For generators with greater connectivity, estimation of the associated parameters is more susceptible to variations in other generator states. Furthermore, we numerically show that estimating the parameters by ignoring their ill-conditioning aspects yields highly unreliable results. | ['Oliver Kosut', 'Lalitha Sankar', 'Rajasekhar Anguluri'] | 2022-08-09 | null | null | null | null | ['connectivity-estimation'] | ['graphs'] | [ 2.61446591e-02 1.55440360e-01 -2.64904916e-01 5.59034765e-01
1.17895072e-02 -9.96858954e-01 3.01171869e-01 -3.54297966e-01
2.40692094e-01 8.32979620e-01 -3.15432906e-01 -3.18484902e-01
-7.96353877e-01 -4.63861406e-01 -1.65189564e-01 -9.60635602e-01
-3.64525139e-01 4.12743390e-01 -1.61162585e-01 -4.38039184e-01
-1.53843313e-01 5.07639349e-01 -8.03745568e-01 -7.76448071e-01
8.11681390e-01 5.02036572e-01 -3.70933488e-02 7.56607592e-01
6.77657843e-01 4.45253879e-01 -8.02723825e-01 2.97648311e-01
1.25788406e-01 -7.20469356e-01 -4.12411213e-01 2.68325973e-02
-2.72697270e-01 -3.61256480e-01 -5.39999247e-01 9.69014943e-01
4.15728033e-01 1.26071885e-01 1.12305272e+00 -1.64683008e+00
1.20180376e-01 6.04169607e-01 -3.32801342e-01 4.41821106e-02
3.06959897e-01 1.68062765e-02 8.58253956e-01 -4.11542028e-01
3.09516490e-01 8.96879435e-01 9.37123716e-01 1.22428834e-01
-1.88250422e+00 -5.06562352e-01 -5.91458343e-02 -2.87765801e-01
-1.72180080e+00 -2.62808800e-01 1.10386205e+00 -5.43759823e-01
9.61819947e-01 5.57155669e-01 9.94993627e-01 7.42665708e-01
3.94191414e-01 1.81697041e-01 6.64731979e-01 -3.99568468e-01
2.67203718e-01 3.87136608e-01 1.41251653e-01 5.57851315e-01
6.94355011e-01 1.52468532e-01 -1.03074513e-01 -5.36417544e-01
1.05301416e+00 -3.98801863e-01 -3.11224699e-01 -7.56206334e-01
-1.15440464e+00 5.75720012e-01 -9.10238847e-02 6.06867671e-01
-2.42183954e-01 3.51885945e-01 -6.69404445e-03 3.50278020e-01
7.97102228e-02 6.73156500e-01 -3.39826524e-01 -1.97855413e-01
-1.00084984e+00 -2.43633725e-02 1.29777431e+00 1.02236092e+00
4.30952787e-01 8.20228636e-01 4.96135920e-01 5.61028898e-01
3.31547469e-01 7.63316333e-01 2.08887577e-01 -8.37543190e-01
2.59185463e-01 3.45334649e-01 2.41396070e-01 -1.07803512e+00
-8.13872695e-01 -8.68220747e-01 -1.06237876e+00 3.83740216e-02
4.27450061e-01 -7.98555136e-01 -7.82622993e-02 1.71840870e+00
9.34969634e-02 -3.99610400e-02 -3.93207334e-02 7.20958829e-01
9.43972841e-02 6.94327593e-01 -3.10975403e-01 -7.72405088e-01
1.05730736e+00 -4.68599439e-01 -9.73507583e-01 -2.01230392e-01
6.70742154e-01 -5.34141481e-01 5.80331445e-01 3.34907562e-01
-1.05661190e+00 -2.55429655e-01 -1.45391273e+00 6.19824767e-01
-4.11629044e-02 6.19598925e-01 1.40920669e-01 7.62365758e-01
-1.02283192e+00 8.92955303e-01 -9.30398345e-01 -3.16559374e-01
-3.77542198e-01 4.10773963e-01 -4.71807197e-02 1.15642619e+00
-1.23306811e+00 1.43005955e+00 2.91605536e-02 4.42899257e-01
-3.20666760e-01 -7.30949819e-01 -4.64024216e-01 2.19893456e-01
4.23681766e-01 -8.19322407e-01 8.74275565e-01 -4.62808549e-01
-2.01742172e+00 -2.88427293e-01 -1.59739591e-02 5.32911606e-02
5.25952220e-01 1.98994607e-01 -4.30623859e-01 1.30432978e-01
-2.39016503e-01 -4.16803360e-01 1.08928525e+00 -1.10727179e+00
1.88208610e-01 7.13931397e-02 -3.40036184e-01 2.64638752e-01
-5.09508908e-01 -5.80868304e-01 1.44768357e-01 -5.29305756e-01
4.50574368e-01 -1.29200494e+00 -3.54561418e-01 -1.33670002e-01
-7.75445342e-01 2.32828036e-01 7.78993070e-01 -5.97499847e-01
1.54710686e+00 -1.90770137e+00 5.91038823e-01 6.89783394e-01
9.58853588e-02 -2.94389781e-02 3.44667345e-01 1.21357119e+00
-1.21984944e-01 8.33426639e-02 -2.81579159e-02 2.18291074e-01
2.59865195e-01 -4.85771196e-03 -2.69056231e-01 8.60388696e-01
1.23194538e-01 4.07085091e-01 -5.94364822e-01 -3.07145804e-01
4.34099019e-01 6.46590292e-01 -4.61311489e-01 -2.40994751e-01
6.08630776e-01 3.12639236e-01 -8.51496279e-01 1.01049058e-01
3.24400395e-01 -2.00396582e-01 6.31566942e-01 -6.72919512e-01
-1.73978940e-01 -3.34707741e-03 -1.70667636e+00 1.14169264e+00
-9.04042482e-01 5.66222370e-01 4.68164295e-01 -1.14607692e+00
6.82159185e-01 5.26050031e-01 8.43243361e-01 2.21812017e-02
1.95011988e-01 9.74974260e-02 3.62229228e-01 -2.73918450e-01
2.71608710e-01 -9.68338102e-02 -1.63476691e-01 5.68963587e-01
2.52647698e-01 -6.87623501e-01 3.85049015e-01 4.66360688e-01
8.96779478e-01 -1.62864804e-01 4.00170326e-01 -9.23967004e-01
5.68782926e-01 -1.64180323e-01 4.44488049e-01 3.17470014e-01
8.51731673e-02 1.52519196e-01 8.40687394e-01 3.53488952e-01
-9.65373278e-01 -9.93506849e-01 -6.47767305e-01 2.92594843e-02
2.45920345e-01 -6.86875463e-01 -6.41016245e-01 2.04760879e-02
2.95273662e-01 7.83511519e-01 -2.43007943e-01 -5.91748953e-01
-2.64077932e-01 -9.25743341e-01 1.14260472e-01 2.74254858e-01
-2.26235256e-01 -3.33637334e-02 -5.57744563e-01 5.43444097e-01
4.84044105e-02 -7.89995551e-01 -3.40742975e-01 4.17642385e-01
-9.52271283e-01 -9.14020419e-01 -5.68994701e-01 -3.23608488e-01
9.31609333e-01 -3.49988520e-01 7.06288755e-01 7.62450024e-02
-2.75165468e-01 4.82673913e-01 2.44301900e-01 3.00667491e-02
-5.07278383e-01 2.67221659e-01 7.99170434e-01 -2.62857735e-01
-6.40373409e-01 -8.30284059e-01 -3.62311274e-01 7.48932958e-01
-2.80639529e-01 4.31005210e-02 3.67034823e-01 7.75948584e-01
1.28305554e-01 6.79721653e-01 7.95098245e-01 -4.94085401e-01
7.08399117e-01 -2.99477965e-01 -1.03562903e+00 1.33460656e-01
-8.91501606e-01 1.15256660e-01 9.97823775e-01 -6.69088483e-01
-8.15724134e-01 2.75449961e-01 2.97742039e-01 -4.02499028e-02
3.85464340e-01 3.21343929e-01 -1.24148533e-01 -3.03878188e-01
2.07190752e-01 1.53297270e-02 1.74229026e-01 -4.19534773e-01
1.55757666e-01 4.98331517e-01 2.33147606e-01 -3.07653189e-01
1.45493078e+00 1.37171254e-01 6.81843877e-01 -1.31369269e+00
8.28622803e-02 -1.19919688e-01 -6.45478547e-01 -3.18193793e-01
1.87555984e-01 -7.38252103e-01 -1.01721025e+00 3.60914081e-01
-8.18212509e-01 -2.68628925e-01 -3.59142512e-01 6.96968257e-01
-6.65604472e-01 3.76289785e-01 -6.31466031e-01 -1.29243803e+00
-8.23257715e-02 -1.17707193e+00 8.04801285e-01 -7.16594085e-02
-6.80957675e-01 -1.33817160e+00 2.24737227e-01 -3.37751120e-01
6.12167656e-01 1.35510340e-01 8.06493282e-01 9.77182388e-03
-1.15462787e-01 -4.99435902e-01 4.54982936e-01 2.47869920e-02
4.85384762e-02 4.31942344e-01 -5.25052667e-01 -7.24562943e-01
1.77957952e-01 4.61923569e-01 -3.18017304e-02 4.85289752e-01
2.80508816e-01 -4.74147618e-01 -6.92508042e-01 3.69933605e-01
1.75748539e+00 6.05310202e-02 6.23031259e-02 -1.12678923e-01
5.54730356e-01 4.45402443e-01 2.29304079e-02 5.82821071e-01
5.81586398e-02 6.89547837e-01 1.06282257e-01 -1.61901608e-01
4.64108109e-01 -5.03747389e-02 5.23731470e-01 1.21568477e+00
-1.12017624e-01 -1.83764413e-01 -6.80774093e-01 1.31197378e-01
-1.51620233e+00 -6.91535056e-01 -3.68067652e-01 2.09292388e+00
5.61497867e-01 7.47983456e-02 3.59710783e-01 5.01822770e-01
5.29855192e-01 -2.23953441e-01 -5.48884034e-01 -2.94689626e-01
-8.58793706e-02 -1.92191258e-01 8.39818418e-01 6.97635949e-01
-7.33563006e-01 -3.95875052e-02 7.63409472e+00 7.05891550e-01
-8.78710210e-01 -1.09329000e-01 8.92488658e-02 -1.49576247e-01
-1.06545337e-01 4.24027443e-01 -7.33921170e-01 6.66535318e-01
1.06953287e+00 -9.03245151e-01 5.86049855e-01 6.57124102e-01
4.82816100e-01 -5.03912568e-01 -8.82086515e-01 5.92626572e-01
-2.74916440e-01 -7.01023817e-01 -5.18595278e-01 2.49293804e-01
7.81460226e-01 -4.98780280e-01 -2.34040037e-01 1.90645602e-04
-1.79498687e-01 -4.03336406e-01 7.01327384e-01 6.51173651e-01
6.03363156e-01 -4.90618557e-01 5.08540511e-01 6.22576535e-01
-1.26167727e+00 -2.16292024e-01 -1.53521255e-01 -1.86711058e-01
6.05989516e-01 9.40629423e-01 -3.47343415e-01 4.71897215e-01
-4.42955047e-02 7.05853939e-01 -3.59309107e-01 7.66263545e-01
1.08265646e-01 7.62786567e-01 -1.09808564e+00 -3.15342188e-01
-3.74865085e-01 -8.26580048e-01 8.80739093e-01 6.19129896e-01
4.07461733e-01 -9.02396347e-03 -1.41940042e-02 1.13119376e+00
7.05974579e-01 -2.52673060e-01 -4.86140281e-01 -1.77589767e-02
4.70382512e-01 1.42735732e+00 -7.79751778e-01 -2.59527475e-01
-1.66292787e-01 5.32873869e-01 -4.84669805e-01 7.30907559e-01
-9.93249655e-01 -7.32527554e-01 5.15757143e-01 1.75716817e-01
-1.74097955e-01 -4.84128147e-01 -2.21044764e-01 -1.31481206e+00
1.80049408e-02 -4.93282646e-01 -1.63677514e-01 -6.08331561e-01
-9.07122850e-01 3.36249560e-01 3.88526887e-01 -1.64878452e+00
-1.10950541e+00 -3.41885954e-01 -7.30435252e-01 9.78093922e-01
-7.50065267e-01 -5.75300574e-01 3.83600533e-01 1.98227793e-01
-1.43587645e-02 -1.45737771e-02 8.41835976e-01 2.74011016e-01
-1.16485870e+00 2.43586794e-01 6.72776043e-01 -3.35805506e-01
1.44623652e-01 -1.30369043e+00 -1.66190505e-01 8.06094468e-01
-2.41112739e-01 7.09681153e-01 1.24083507e+00 -4.66861904e-01
-2.12402201e+00 -3.55952382e-01 5.55826128e-01 -3.57146800e-01
1.10806859e+00 -4.09863412e-01 -6.62666142e-01 4.03539956e-01
2.84876406e-01 -2.49067485e-01 1.54057771e-01 -6.99261483e-03
5.27698338e-01 -2.51085877e-01 -7.96525538e-01 5.79490244e-01
5.84732890e-01 -7.09525943e-01 -2.47812763e-01 4.22263652e-01
-2.55020767e-01 -3.30801159e-02 -1.35440910e+00 1.99664518e-01
6.23096347e-01 -3.17828774e-01 1.03547049e+00 -9.13426876e-02
-3.30658704e-01 -5.35791039e-01 3.56100976e-01 -1.75368226e+00
-4.68672872e-01 -1.29195464e+00 -4.56615835e-01 1.15983391e+00
6.04579926e-01 -9.14485097e-01 4.90157992e-01 4.79148000e-01
3.44130725e-01 -4.95712578e-01 -1.02499557e+00 -1.29167104e+00
2.11762771e-01 8.16688091e-02 1.42655492e-01 1.06421530e+00
1.05970562e+00 6.24204814e-01 -1.29562914e-01 5.03051996e-01
6.82310045e-01 4.06790338e-02 5.49560308e-01 -1.00484407e+00
-4.65628266e-01 -4.82828140e-01 -2.35524744e-01 -7.19459295e-01
1.21657066e-01 -4.12668914e-01 -8.78712311e-02 -1.31154370e+00
-9.12115350e-02 -4.13880408e-01 6.79358840e-02 3.20955366e-02
7.79312849e-02 2.46562108e-01 -1.02707550e-01 1.77994296e-01
-1.19261155e-02 4.78681773e-01 9.54368412e-01 5.98620400e-02
-2.58564949e-01 3.26683015e-01 4.44871932e-02 6.44555032e-01
8.89428377e-01 -2.38142461e-01 -7.40310609e-01 -1.29770990e-02
2.80766875e-01 6.12842739e-01 2.97137618e-01 -1.28882492e+00
4.07780826e-01 2.00589567e-01 1.86989605e-01 -4.42723393e-01
3.91784817e-01 -1.11107028e+00 1.06915057e+00 8.02492142e-01
3.55585255e-02 1.13729686e-01 2.63760000e-01 4.46476966e-01
-2.66464334e-02 -2.89804816e-01 5.42826533e-01 4.99630302e-01
2.81563550e-01 -2.61734188e-01 -1.18585587e+00 -1.86739162e-01
7.19814360e-01 -2.32156858e-01 -2.08815902e-01 -7.72451937e-01
-9.54487145e-01 2.79416233e-01 2.99513012e-01 -1.37965055e-02
-2.24036425e-01 -1.27398002e+00 -3.74022752e-01 3.47249776e-01
-5.11768222e-01 -5.77208400e-01 2.21697912e-01 1.21766853e+00
-4.07723337e-01 6.27779305e-01 -6.48351386e-03 -7.89111674e-01
-8.28181803e-01 7.05165565e-02 6.28217518e-01 -1.34983614e-01
-2.38030195e-01 8.70421678e-02 -2.31931597e-01 -1.06353253e-01
-2.91133642e-01 -4.49184030e-01 -2.85702273e-02 1.94398552e-01
-2.53132701e-01 6.96134806e-01 -2.06530932e-02 -7.25763023e-01
-3.46730888e-01 1.01179874e+00 6.37557507e-01 -1.44103527e-01
1.01697254e+00 -4.37814862e-01 5.82777038e-02 6.45527780e-01
9.79314089e-01 3.79220434e-02 -1.11871850e+00 3.46627682e-01
-4.82351512e-01 7.27611184e-02 4.19614822e-01 -3.97767991e-01
-1.23835886e+00 4.66817766e-01 3.76018584e-01 8.77422750e-01
8.48001480e-01 -3.89291167e-01 1.10907272e-01 5.73518813e-01
5.25868058e-01 -1.39579105e+00 -4.08807188e-01 2.59775907e-01
8.27291787e-01 -4.78701502e-01 5.82033098e-01 -7.37251699e-01
-9.92726311e-02 1.10815609e+00 3.46986681e-01 -3.00637782e-01
1.07605982e+00 7.63559699e-01 -1.67725772e-01 2.88129658e-01
-8.52020204e-01 3.47011060e-01 2.20605969e-01 4.19398040e-01
4.47292715e-01 1.91560499e-02 -5.12113631e-01 7.79709101e-01
-2.76757032e-01 -3.51224184e-01 9.34273183e-01 8.27314734e-01
-4.52587157e-02 -8.83436620e-01 -3.97559106e-01 4.41283435e-01
-1.21893600e-01 3.28699619e-01 -2.95105129e-01 1.08843005e+00
-4.50671405e-01 9.95804965e-01 1.87513173e-01 -1.71262428e-01
6.84330523e-01 -1.86427265e-01 4.64277834e-01 -1.08969495e-01
-4.37777728e-01 4.91615146e-01 1.71646759e-01 -4.32343096e-01
6.58470839e-02 -8.66089106e-01 -8.70256722e-01 -4.33032453e-01
-1.36069369e+00 4.33248013e-01 6.87702477e-01 4.83581156e-01
5.41064553e-02 7.73760855e-01 8.57661664e-01 -1.00645459e+00
-9.74024117e-01 -7.98621416e-01 -1.18725359e+00 -1.72173396e-01
4.19911772e-01 -9.37767088e-01 -1.00652874e+00 -1.68090373e-01] | [5.5233235359191895, 2.6724047660827637] |
ae585a73-9a94-49e2-b849-1c09c5d6cfc6 | multi-source-semantic-graph-based-multimodal | 2306.16650 | null | https://arxiv.org/abs/2306.16650v1 | https://arxiv.org/pdf/2306.16650v1.pdf | Multi-source Semantic Graph-based Multimodal Sarcasm Explanation Generation | Multimodal Sarcasm Explanation (MuSE) is a new yet challenging task, which aims to generate a natural language sentence for a multimodal social post (an image as well as its caption) to explain why it contains sarcasm. Although the existing pioneer study has achieved great success with the BART backbone, it overlooks the gap between the visual feature space and the decoder semantic space, the object-level metadata of the image, as well as the potential external knowledge. To solve these limitations, in this work, we propose a novel mulTi-source sEmantic grAph-based Multimodal sarcasm explanation scheme, named TEAM. In particular, TEAM extracts the object-level semantic meta-data instead of the traditional global visual features from the input image. Meanwhile, TEAM resorts to ConceptNet to obtain the external related knowledge concepts for the input text and the extracted object meta-data. Thereafter, TEAM introduces a multi-source semantic graph that comprehensively characterize the multi-source (i.e., caption, object meta-data, external knowledge) semantic relations to facilitate the sarcasm reasoning. Extensive experiments on a public released dataset MORE verify the superiority of our model over cutting-edge methods. | ['Liqiang Nie', 'Mengzhao Jia', 'Kun Ouyang', 'Xuemeng Song', 'Liqiang Jing'] | 2023-06-29 | null | null | null | null | ['explanation-generation'] | ['natural-language-processing'] | [ 1.72273785e-01 4.90535438e-01 -7.52818882e-02 -3.04274291e-01
-5.53536415e-01 -3.06876510e-01 4.75480974e-01 3.24202418e-01
-7.72335678e-02 3.86371404e-01 6.55838251e-01 6.19159788e-02
1.17855296e-01 -4.35437381e-01 -5.68442047e-01 -4.59193587e-01
9.52634156e-01 3.68856937e-01 6.68612644e-02 -4.39090222e-01
4.92926002e-01 -4.25038427e-01 -1.40526175e+00 7.90944695e-01
6.57681346e-01 7.03274012e-01 3.01690996e-01 4.09826607e-01
-5.42656481e-01 1.28275347e+00 -4.99107629e-01 -9.87291276e-01
-5.14292598e-01 -8.25801849e-01 -8.80394101e-01 4.08544093e-01
2.52183735e-01 -3.74049507e-02 -1.76650271e-01 1.27080715e+00
3.00738662e-01 1.42229861e-02 5.09375393e-01 -1.66421425e+00
-1.06532872e+00 7.88029730e-01 -5.54104686e-01 -1.40858740e-01
3.38484943e-01 1.00854412e-01 9.98533666e-01 -1.02303886e+00
7.88724959e-01 1.32120621e+00 4.60649520e-01 6.15144849e-01
-7.03601897e-01 -2.38172725e-01 7.56613025e-03 5.70147693e-01
-1.04233468e+00 -1.78558920e-02 1.22964787e+00 -3.48577321e-01
3.77466828e-01 1.95410207e-01 7.71077394e-01 1.11760843e+00
-2.28733599e-01 1.11552155e+00 8.97681594e-01 -3.74311358e-01
4.35586944e-02 4.41400319e-01 2.63836294e-01 8.03773463e-01
-3.70256044e-02 -6.30317569e-01 -9.19745743e-01 2.41282489e-03
5.26844740e-01 -1.98822450e-02 -1.62070915e-01 -6.64445519e-01
-1.39516652e+00 9.24237072e-01 6.26370609e-01 2.18422353e-01
-2.20622599e-01 3.00967097e-01 5.11956513e-01 -1.31042704e-01
1.11964971e-01 2.08697572e-01 -7.44469687e-02 -6.56138286e-02
-3.73387009e-01 -1.77280381e-02 4.45158213e-01 1.07251489e+00
6.93841815e-01 -5.65824993e-02 -2.50893444e-01 7.22927988e-01
5.15710592e-01 7.05883265e-01 4.40987140e-01 -8.36561084e-01
8.41496050e-01 1.22237289e+00 -1.02630801e-01 -1.74303055e+00
-4.39678252e-01 -3.52291167e-01 -8.13837469e-01 -3.18459988e-01
1.59840882e-01 1.62912875e-01 -5.36910057e-01 1.87934685e+00
3.61563861e-01 1.30944178e-01 6.96231306e-01 1.48738420e+00
1.71801007e+00 4.87782329e-01 2.63723284e-01 -1.26764402e-01
1.74838114e+00 -1.46949720e+00 -8.53369057e-01 -5.20875216e-01
5.55584669e-01 -8.79218519e-01 1.20856369e+00 -1.59659609e-01
-1.07715428e+00 -5.85088968e-01 -9.15979743e-01 -4.28146243e-01
-2.20152482e-01 4.36166584e-01 6.46323085e-01 1.21343680e-01
-5.18778682e-01 -4.78979424e-02 -1.62374765e-01 -4.76724803e-01
2.17854708e-01 -7.08970129e-02 -4.84485775e-01 -1.06370583e-01
-1.13955557e+00 7.26375818e-01 5.40120244e-01 1.12409718e-01
-6.32441163e-01 -4.38583940e-01 -9.81201410e-01 2.96288729e-02
6.95432782e-01 -1.23960340e+00 9.92983758e-01 -1.51834941e+00
-9.80615854e-01 9.19240832e-01 -1.04025126e-01 -1.23472914e-01
2.49936923e-01 -1.30645195e-02 -2.48572305e-01 5.35574257e-01
2.30018169e-01 7.87550628e-01 8.67632449e-01 -1.72582889e+00
-3.68945241e-01 -4.63087440e-01 3.37381750e-01 7.26102591e-01
-4.93060321e-01 -1.29767239e-01 -7.65009761e-01 -5.49667001e-01
3.82265419e-01 -7.78773844e-01 -3.24863009e-02 -4.00410712e-01
-6.49344862e-01 -6.63215369e-02 6.13556981e-01 -7.04315424e-01
9.59304810e-01 -2.32601476e+00 5.79085410e-01 -1.06896669e-01
5.64967096e-01 -2.78497189e-01 -2.47243986e-01 5.41197360e-01
3.35863195e-02 1.08417803e-02 -3.29221934e-01 -4.03599411e-01
1.95044354e-02 9.80775803e-02 -1.73256099e-01 -8.62250291e-03
-1.36936978e-01 1.27518296e+00 -1.13919425e+00 -1.04081392e+00
1.08988367e-01 4.10187662e-01 -2.85959393e-01 1.73938215e-01
-2.08571598e-01 3.68650377e-01 -6.81163609e-01 5.98976851e-01
5.32182813e-01 -6.93482637e-01 1.50282726e-01 -7.22362816e-01
3.89332145e-01 -3.41444761e-01 -7.07751095e-01 2.00368857e+00
-1.79292962e-01 3.86319131e-01 -2.28209421e-01 -8.28360796e-01
1.01251435e+00 2.42497578e-01 3.94715786e-01 -6.55909538e-01
3.80217046e-01 8.11220929e-02 -3.29765886e-01 -9.32509243e-01
6.79435968e-01 -3.67289245e-01 -2.69566834e-01 3.95235002e-01
-6.28528148e-02 -7.62815326e-02 -1.51931450e-01 9.42109704e-01
5.50234020e-01 1.37744233e-01 2.05958903e-01 6.38489351e-02
7.60601223e-01 4.60844874e-01 2.26713508e-01 4.61139262e-01
-4.46960926e-02 8.45958531e-01 7.12833762e-01 -3.82288456e-01
-1.00002313e+00 -9.15771365e-01 6.35939777e-01 8.00189435e-01
1.06916833e+00 -8.73767436e-01 -8.74928653e-01 -8.10043931e-01
-2.91665733e-01 6.59256816e-01 -7.09182501e-01 -3.34597081e-01
-2.96482772e-01 -5.39046705e-01 2.79878199e-01 5.26693761e-01
6.67101681e-01 -1.19401908e+00 -5.79164863e-01 -1.06588788e-01
-9.35992301e-01 -1.29563975e+00 -5.88504672e-01 -4.62070644e-01
-5.23833454e-01 -1.25955713e+00 -3.28928262e-01 -9.40767705e-01
9.77720618e-01 6.02794528e-01 1.23807895e+00 5.99261820e-01
-1.45665452e-01 7.47016072e-01 -6.27789438e-01 -1.32806584e-01
-3.11569005e-01 -3.83118540e-01 -4.73193228e-01 2.64420718e-01
5.74856818e-01 -9.66167673e-02 -6.90754116e-01 1.39595255e-01
-8.21417391e-01 1.09010971e+00 4.93772089e-01 8.80388021e-01
6.06879532e-01 -2.74899602e-01 5.55347204e-01 -6.15721643e-01
7.64256895e-01 -6.26298785e-01 2.98412323e-01 4.54944164e-01
-1.95649564e-01 -1.17069811e-01 3.13815176e-01 -3.51152092e-01
-1.18078923e+00 1.77605256e-01 3.12892646e-01 -3.97024363e-01
8.88865069e-02 7.11613774e-01 -2.56044000e-01 4.62748706e-01
2.55772531e-01 4.20956343e-01 2.14375630e-01 -2.83064336e-01
7.79521525e-01 6.63442492e-01 8.62841308e-01 -5.10105669e-01
6.11556172e-01 6.33662343e-01 1.73229799e-02 -4.59866703e-01
-1.33017409e+00 -5.38397133e-01 -4.82799798e-01 -5.64496756e-01
1.28949332e+00 -1.08237755e+00 -8.32911074e-01 5.28818034e-02
-1.49873543e+00 3.41985911e-01 -1.52463123e-01 2.60205179e-01
-9.30635273e-01 6.53398335e-01 -4.17328715e-01 -6.71365857e-01
-3.42116296e-01 -1.04857028e+00 1.20571721e+00 4.74727333e-01
-1.56210080e-01 -9.75756705e-01 -2.81830281e-01 1.29221666e+00
-1.18094586e-01 1.13762841e-01 1.10735774e+00 -4.39907819e-01
-5.52665174e-01 4.45125885e-02 -7.50992119e-01 9.35079679e-02
-2.76510686e-01 -4.42923814e-01 -5.50087631e-01 1.59397945e-01
1.04783773e-01 -5.98097980e-01 7.89560199e-01 6.76623955e-02
9.71022785e-01 -3.15543115e-01 -1.35259181e-01 1.42399803e-01
1.39740705e+00 -2.82216579e-01 5.36407709e-01 3.70810241e-01
1.06562948e+00 8.28077853e-01 8.81324232e-01 4.11746144e-01
9.71469760e-01 7.70071566e-01 8.12554181e-01 -4.46714580e-01
-4.60287780e-01 -7.01276243e-01 3.23254317e-01 1.01919174e+00
-1.97057184e-02 -1.88969806e-01 -6.88586116e-01 4.54599053e-01
-2.49553132e+00 -1.11696184e+00 -7.42512882e-01 1.66374743e+00
4.97332066e-01 -4.68696862e-01 2.93192379e-02 -2.53119946e-01
8.56201112e-01 -1.10976472e-01 -4.44482952e-01 -1.59122914e-01
-5.91233075e-01 -8.85777771e-01 -8.41771886e-02 4.28651363e-01
-7.59141684e-01 1.11889780e+00 4.91886759e+00 8.11988294e-01
-5.35466194e-01 3.58056933e-01 4.69620258e-01 1.66350409e-01
-7.49885738e-01 2.61807829e-01 -4.31959093e-01 3.68325382e-01
2.73029178e-01 -1.14593901e-01 1.63187593e-01 8.79971385e-01
1.24510728e-01 5.77452686e-03 -6.64850473e-01 1.43120754e+00
1.06739581e+00 -1.48032928e+00 4.21769649e-01 -3.79933625e-01
5.46084166e-01 -5.39068758e-01 -1.58119291e-01 1.79031506e-01
-1.69920534e-01 -8.71921539e-01 9.91916895e-01 7.12714136e-01
4.61602360e-01 -6.38689816e-01 1.13529861e+00 3.50802749e-01
-1.13533759e+00 -9.39285308e-02 -2.07392737e-01 4.98570837e-02
3.72576922e-01 2.11189896e-01 -5.65887570e-01 8.00255716e-01
6.24149859e-01 1.00584364e+00 -9.99126375e-01 5.34582138e-01
-4.34355557e-01 1.92091241e-01 2.34450087e-01 -2.13415176e-01
2.38654047e-01 -1.86487049e-01 4.99985218e-01 9.02336121e-01
2.15007156e-01 2.37945706e-01 1.69346735e-01 1.04077137e+00
8.10104236e-02 6.12881482e-01 -3.43804628e-01 -9.80618969e-02
9.90063604e-03 1.44513094e+00 -7.22067118e-01 -4.53947365e-01
-4.43125188e-01 1.20151293e+00 4.50658113e-01 4.72221375e-01
-1.01583457e+00 -3.71126123e-02 4.54910547e-02 -9.15786028e-02
-8.47725645e-02 5.57920616e-03 -4.12936419e-01 -1.47394812e+00
5.57754971e-02 -8.73138487e-01 5.74451268e-01 -1.62387848e+00
-1.26687860e+00 4.71203417e-01 -1.42403185e-01 -1.20617628e+00
1.52382851e-01 -4.14967954e-01 -4.78196442e-01 4.87418175e-01
-1.17113042e+00 -1.90955174e+00 -7.94997394e-01 8.06537509e-01
8.11500669e-01 -2.80346841e-01 4.14192855e-01 -1.07135527e-01
-3.08934480e-01 1.77696422e-01 -3.68270278e-01 8.60236306e-03
5.34751892e-01 -1.05550206e+00 -3.11790079e-01 5.63544691e-01
2.31419697e-01 4.25351769e-01 9.23315704e-01 -8.14951718e-01
-1.45099664e+00 -6.19100273e-01 1.03631163e+00 -6.67050660e-01
8.09902430e-01 -1.66828707e-01 -6.66600943e-01 4.64493245e-01
3.52283984e-01 -4.95784312e-01 6.71206415e-01 -8.41867849e-02
-3.83711994e-01 2.71076411e-01 -6.99789047e-01 8.54628205e-01
1.11564958e+00 -3.13457221e-01 -8.88327897e-01 4.89773273e-01
9.02216196e-01 -2.64540106e-01 -4.51050311e-01 3.55480075e-01
3.03998441e-01 -1.04681468e+00 1.03402829e+00 -8.25000703e-01
1.28704274e+00 -5.57544649e-01 -3.84336114e-01 -9.68625903e-01
-5.04160859e-02 -9.47674289e-02 -3.43473516e-02 1.35782802e+00
3.18462014e-01 1.14297964e-01 7.78510928e-01 2.82882303e-01
-2.89042652e-01 -5.60941339e-01 -7.48944938e-01 -2.63844579e-01
-4.43036079e-01 -4.33717132e-01 5.32689214e-01 9.24529731e-01
5.08667827e-01 9.44277823e-01 -8.07004929e-01 4.19461317e-02
6.88047945e-01 6.26144409e-01 9.88025546e-01 -8.62401187e-01
-1.37920231e-01 -2.83044279e-01 -4.92368311e-01 -9.59766150e-01
2.54780263e-01 -1.10602450e+00 6.80501461e-02 -1.98373473e+00
1.09716344e+00 -3.82621512e-02 -2.14494660e-01 5.72427869e-01
-3.43354881e-01 4.71584857e-01 5.60423136e-01 4.36550140e-01
-1.17512953e+00 8.36725116e-01 1.57704878e+00 -3.98148268e-01
1.55725535e-02 -6.26894355e-01 -1.03609228e+00 8.86748731e-01
6.60627902e-01 -3.69705260e-01 -7.14014411e-01 -5.41120768e-01
7.26102710e-01 3.36307704e-01 1.06173742e+00 -5.61385870e-01
1.73983067e-01 -4.43027169e-01 1.48213252e-01 -7.13900805e-01
5.07923245e-01 -8.25613201e-01 4.53343801e-02 2.96424985e-01
-5.14988840e-01 7.92925060e-02 -1.60799459e-01 8.40760946e-01
-5.13376296e-01 -4.37337518e-01 4.16175365e-01 -2.28508964e-01
-8.80198240e-01 -3.40030417e-02 1.38582453e-01 5.09457588e-02
1.01571226e+00 -2.91643262e-01 -8.91021967e-01 -6.66286469e-01
-8.83448303e-01 3.30518335e-01 5.82729161e-01 7.97204733e-01
9.35289979e-01 -1.55082095e+00 -7.49558628e-01 -2.04808071e-01
7.06287622e-01 -4.06618208e-01 9.29469466e-01 1.16220748e+00
-2.17058882e-01 2.00857054e-02 -2.21803144e-01 -6.45592093e-01
-1.66890895e+00 7.80395627e-01 7.09900185e-02 2.19172314e-01
-9.77785170e-01 5.57032287e-01 5.19985855e-01 -1.83390081e-01
-1.36651084e-01 4.99129683e-01 -6.00057185e-01 -7.91498497e-02
4.14528847e-01 -9.37123597e-02 -6.03009343e-01 -1.25238478e+00
-3.62787932e-01 7.89248049e-01 3.96350354e-01 -9.81388614e-02
9.25666749e-01 -7.55409956e-01 -3.77321929e-01 5.50522208e-01
8.85439932e-01 -4.44783755e-02 -7.20616281e-01 -1.91129431e-01
-2.18396008e-01 -4.37185466e-01 -2.73661286e-01 -7.46942878e-01
-1.10862923e+00 9.92252171e-01 1.87192857e-01 -4.17247303e-02
1.00906456e+00 5.97819626e-01 7.84712791e-01 1.77089214e-01
-5.17641231e-02 -1.22057903e+00 9.67540562e-01 2.02642336e-01
1.16672921e+00 -1.42559028e+00 3.46589908e-02 -7.78409302e-01
-1.48996782e+00 1.24656415e+00 9.72229838e-01 3.91517729e-01
1.75133735e-01 -4.35691148e-01 3.76259446e-01 -7.18486011e-01
-5.71897507e-01 -2.38234237e-01 3.32844347e-01 5.50421298e-01
7.98224956e-02 -3.68177369e-02 -4.35611546e-01 1.23210454e+00
-2.43170932e-01 -2.41423175e-01 7.46278822e-01 4.17832285e-01
-5.44085860e-01 -6.65176511e-01 -2.30370745e-01 -1.00158863e-01
1.19411379e-01 -2.34117031e-01 -7.67698288e-01 5.33124208e-01
3.07072908e-01 1.17106009e+00 -1.53640404e-01 -4.81059968e-01
2.20541403e-01 -9.51277390e-02 1.30433097e-01 -3.59688073e-01
-5.22258937e-01 -1.29098251e-01 1.61102578e-01 -3.67299229e-01
-8.13053906e-01 -3.69292587e-01 -1.79213142e+00 -8.75298493e-03
-8.06543678e-02 2.54362106e-01 9.36130524e-01 1.23723757e+00
2.69656628e-01 3.77308965e-01 3.02675068e-01 -2.84954101e-01
-6.95943981e-02 -6.87556565e-01 -2.88582385e-01 1.34332287e+00
-7.64112845e-02 -5.28741181e-01 -3.94793570e-01 3.97692531e-01] | [10.576823234558105, 1.2118016481399536] |
8b1443d0-c589-4ccb-9564-2f4b45b513b1 | a-neural-ode-interpretation-of-transformer | 2212.06011 | null | https://arxiv.org/abs/2212.06011v1 | https://arxiv.org/pdf/2212.06011v1.pdf | A Neural ODE Interpretation of Transformer Layers | Transformer layers, which use an alternating pattern of multi-head attention and multi-layer perceptron (MLP) layers, provide an effective tool for a variety of machine learning problems. As the transformer layers use residual connections to avoid the problem of vanishing gradients, they can be viewed as the numerical integration of a differential equation. In this extended abstract, we build upon this connection and propose a modification of the internal architecture of a transformer layer. The proposed model places the multi-head attention sublayer and the MLP sublayer parallel to each other. Our experiments show that this simple modification improves the performance of transformer networks in multiple tasks. Moreover, for the image classification task, we show that using neural ODE solvers with a sophisticated integration scheme further improves performance. | ['Biswadip Dey', 'Amit Chakraborty', 'Tongtao Zhang', 'Yaofeng Desmond Zhong'] | 2022-12-12 | null | null | null | null | ['numerical-integration'] | ['miscellaneous'] | [-1.22085243e-01 3.31241250e-01 2.02096373e-01 -2.35436112e-02
-1.51019841e-01 1.26443654e-02 4.49258655e-01 -1.06172472e-01
-4.81140435e-01 5.89899719e-01 -1.33521587e-01 -4.59453851e-01
1.70133427e-01 -6.51376963e-01 -8.81221771e-01 -8.08155000e-01
2.59757787e-01 1.02636375e-01 4.42984372e-01 -2.81638533e-01
1.39997900e-01 7.22588778e-01 -1.35348928e+00 4.06191200e-01
7.52033055e-01 1.38422179e+00 2.08080113e-02 5.34892738e-01
-7.05019683e-02 1.34488678e+00 -2.00742513e-01 -3.41167718e-01
1.73707649e-01 -3.25673819e-02 -9.35795188e-01 -2.42474541e-01
2.20291644e-01 -2.01358438e-01 -6.58769459e-02 8.38026285e-01
4.05492067e-01 1.74007297e-01 4.69226062e-01 -1.22209072e+00
-6.02138400e-01 4.25144166e-01 -2.61299372e-01 3.05593669e-01
-1.01057515e-01 -1.63991466e-01 9.53397930e-01 -1.14172101e+00
5.52335940e-02 1.19050503e+00 1.27876723e+00 2.91318089e-01
-1.27498221e+00 -4.05990332e-01 3.33412349e-01 2.15280250e-01
-1.16580963e+00 -3.17032009e-01 7.40747869e-01 -4.63802218e-01
1.30633450e+00 8.93315226e-02 5.68232834e-01 7.59178936e-01
6.19967401e-01 6.57523334e-01 9.69342351e-01 -3.41133267e-01
4.16025892e-02 5.42213798e-01 3.15623999e-01 7.95107782e-01
-1.22810155e-02 -3.58400822e-01 6.93204030e-02 -5.45300245e-02
9.66186523e-01 3.18995804e-01 -2.03517616e-01 -3.07186067e-01
-6.94403112e-01 1.05177236e+00 8.90629828e-01 5.16628027e-01
-5.53379774e-01 1.79090485e-01 1.30897462e-01 1.90487459e-01
5.69989741e-01 3.48039538e-01 -3.60938579e-01 4.42141294e-01
-7.58955479e-01 -2.24991352e-03 9.80682552e-01 5.28020918e-01
8.08615923e-01 2.72712499e-01 -9.63546708e-02 7.82385409e-01
3.78881335e-01 1.81443945e-01 6.65558994e-01 -9.28749740e-01
2.31268451e-01 7.81622827e-01 -4.50283214e-02 -8.82829666e-01
-6.88309908e-01 -7.75793493e-01 -1.15951657e+00 5.51028013e-01
3.21476400e-01 -3.23878944e-01 -6.41798437e-01 1.45512247e+00
1.83477849e-01 2.42987603e-01 -6.49675876e-02 7.51204193e-01
5.42940080e-01 7.29594767e-01 -2.18123640e-03 5.43691367e-02
1.22942233e+00 -1.17644429e+00 -4.43443954e-01 -1.94553688e-01
4.60664958e-01 -5.68189681e-01 8.91112268e-01 2.87539005e-01
-1.50364637e+00 -6.04531825e-01 -1.02424455e+00 -4.05422032e-01
-6.36939049e-01 2.63482690e-01 5.40922165e-01 2.21332878e-01
-1.49160743e+00 1.03921556e+00 -8.39252114e-01 -8.43340382e-02
2.69095302e-01 6.48837209e-01 -2.32261553e-01 4.70974892e-01
-1.10010171e+00 1.25152349e+00 1.10302679e-01 5.81297338e-01
-6.32299781e-01 -7.32527971e-01 -9.32716846e-01 4.90320325e-01
-4.56464738e-02 -7.74713278e-01 1.34255195e+00 -1.10265017e+00
-1.89406908e+00 6.45069957e-01 -1.72912851e-01 -6.78194642e-01
4.24614400e-01 -4.08118010e-01 1.69550285e-01 -2.42476128e-02
-3.12539667e-01 3.51636022e-01 1.05043864e+00 -7.64199257e-01
-5.07963955e-01 -1.66224629e-01 3.01630765e-01 -8.08286015e-03
-3.41977894e-01 -1.46145031e-01 -5.81301898e-02 -5.67651212e-01
-3.51385251e-02 -8.92050624e-01 -5.17408967e-01 -8.60172659e-02
-1.57345176e-01 -2.16008022e-01 7.64756620e-01 -5.77062011e-01
9.26567018e-01 -2.12859583e+00 4.96264488e-01 1.49759585e-02
4.23208237e-01 3.36385667e-01 2.18663394e-01 2.59224474e-01
-3.41499031e-01 -4.51123677e-02 -2.65587747e-01 -6.86838686e-01
-1.14822298e-01 1.60717174e-01 -3.27933848e-01 5.60059190e-01
3.64604056e-01 8.50844204e-01 -4.40430790e-01 -2.21367314e-01
2.78109372e-01 8.53584826e-01 -7.58463740e-01 1.23682745e-01
-6.32375777e-02 3.10619831e-01 -2.28795052e-01 7.79037997e-02
4.89102691e-01 -5.97729445e-01 -1.66681632e-01 -1.82042450e-01
-3.66399407e-01 5.01610160e-01 -1.07225263e+00 1.33514154e+00
-8.69658530e-01 6.05782509e-01 4.56900090e-01 -1.37461782e+00
6.04661644e-01 4.58570302e-01 2.17828289e-01 -5.76983213e-01
3.92933935e-01 1.22913145e-01 -7.87255466e-02 -3.45155269e-01
1.75265580e-01 -4.19718504e-01 1.78337574e-01 1.83287814e-01
2.11800739e-01 2.99434781e-01 3.33489478e-03 -2.11003721e-01
8.79081905e-01 -1.26388416e-01 9.67477709e-02 -4.61732268e-01
9.70225155e-01 -3.98234010e-01 6.30169690e-01 4.91756916e-01
2.29867592e-01 5.05468428e-01 8.38781714e-01 -6.19158208e-01
-9.26867843e-01 -7.30482280e-01 -1.54944286e-01 9.83007789e-01
-2.88888127e-01 -1.40043423e-01 -8.11961412e-01 -2.64912188e-01
3.86722311e-02 3.90691698e-01 -6.84494913e-01 -1.34315819e-01
-7.76841342e-01 -7.90885746e-01 4.35645044e-01 9.00486112e-01
7.44928122e-01 -1.10382247e+00 -8.30872774e-01 2.64153749e-01
6.66201860e-02 -1.10921574e+00 -4.96679880e-02 4.54489857e-01
-9.41164315e-01 -9.35393453e-01 -1.02284026e+00 -9.86682117e-01
4.30910021e-01 -2.26266578e-01 8.97682011e-01 1.29069120e-01
-1.04154134e-02 1.62465468e-01 1.14015870e-01 -1.21540777e-01
-2.98711061e-01 2.69174576e-01 -4.23032381e-02 3.22265923e-01
-2.46541560e-01 -9.54363704e-01 -2.78825223e-01 7.75708109e-02
-6.54410064e-01 2.59334803e-01 6.10725999e-01 8.03816438e-01
2.65172124e-01 -2.77781755e-01 3.25075805e-01 -7.24418342e-01
5.61169982e-01 -3.79162073e-01 -9.15563822e-01 -2.31103227e-02
-4.61818159e-01 4.69404250e-01 1.11486983e+00 -3.13649893e-01
-8.29033196e-01 7.95126776e-04 -5.26112616e-01 -7.22769201e-01
1.74792409e-01 3.35120350e-01 -6.56594411e-02 -4.68279809e-01
7.15488866e-02 -3.91374622e-03 -3.86512130e-02 -9.13899183e-01
5.12346029e-02 3.21683884e-01 2.77856529e-01 -2.49590084e-01
4.49126691e-01 2.21358642e-01 2.08946183e-01 -7.45173454e-01
-6.87698603e-01 -1.20925665e-01 -6.64539218e-01 1.06113538e-01
1.00184381e+00 -7.79806554e-01 -1.11627102e+00 8.26077461e-01
-1.48256922e+00 -5.86520076e-01 -2.60806650e-01 4.29806530e-01
-3.61742526e-01 -4.50488478e-02 -1.12905586e+00 -7.27525949e-01
-4.36091542e-01 -1.29012537e+00 7.46477246e-01 2.72372663e-01
1.73678279e-01 -1.40036452e+00 -1.21194450e-02 -1.57395303e-01
7.60992110e-01 6.89137802e-02 1.01077473e+00 -5.57349563e-01
-3.81921470e-01 1.75873967e-04 -1.41053110e-01 7.04903126e-01
-1.99871004e-01 -2.47738972e-01 -1.36796391e+00 -2.53181964e-01
4.31072325e-01 3.57880555e-02 1.06596768e+00 3.93999517e-01
1.09179914e+00 -4.99321431e-01 -3.03989589e-01 8.50524008e-01
1.40190625e+00 -6.05986454e-02 5.44985771e-01 1.77973464e-01
7.80916929e-01 3.79753470e-01 -3.00138652e-01 2.61625171e-01
3.54181468e-01 4.62350845e-01 5.78281224e-01 -4.01474565e-01
1.54746324e-01 9.86079425e-02 4.00327832e-01 8.25777173e-01
-2.25799993e-01 1.63376033e-01 -7.88019001e-01 4.39111322e-01
-1.87976193e+00 -8.04860771e-01 -1.09421536e-01 2.07835793e+00
4.99572158e-01 3.42343867e-01 1.52897671e-01 4.44124341e-01
5.04034519e-01 -3.39756645e-02 -4.25037622e-01 -8.20310891e-01
5.78851998e-02 4.59228098e-01 5.68031788e-01 8.59539092e-01
-1.02929258e+00 6.59336150e-01 6.81128740e+00 2.64235705e-01
-1.52276087e+00 2.00004965e-01 4.17525321e-01 1.35195732e-01
1.23339191e-01 -2.39599213e-01 -8.30020249e-01 2.44590387e-01
9.95435119e-01 9.26443934e-02 4.43769544e-01 8.26285303e-01
4.13234718e-02 5.69475703e-02 -1.19308174e+00 7.66281247e-01
-8.84409994e-02 -1.29518902e+00 -1.19032197e-01 -1.60409268e-02
2.51418769e-01 2.53478259e-01 1.94094270e-01 3.43391627e-01
1.63500100e-01 -1.12951338e+00 5.78437924e-01 4.66875732e-01
1.26879051e-01 -6.33407772e-01 9.01551127e-01 3.48864228e-01
-1.24930799e+00 -2.80968785e-01 -2.45209828e-01 -6.49759829e-01
1.33521914e-01 6.18702888e-01 -4.33518261e-01 3.36336404e-01
8.02934349e-01 7.29238153e-01 -5.55002570e-01 9.79306459e-01
-1.61024928e-01 3.06553423e-01 -3.93563688e-01 7.71670192e-02
5.82398891e-01 -2.51767755e-01 3.73946220e-01 1.25817788e+00
1.17837109e-01 -1.36432275e-01 -2.28744149e-01 9.80365872e-01
2.21743342e-03 -5.83769195e-02 -5.42395115e-01 3.61595303e-01
-2.78207600e-01 1.41140437e+00 -4.97346342e-01 -3.86101276e-01
-6.32256746e-01 9.56785500e-01 7.25230813e-01 5.50847352e-01
-9.56729114e-01 -5.24396062e-01 9.47073221e-01 2.75669307e-01
7.16054559e-01 -9.36735421e-02 -3.48096639e-01 -1.23927319e+00
1.54071882e-01 -5.53925395e-01 2.59545535e-01 -8.88649225e-01
-9.63373244e-01 8.03584933e-01 -3.00305784e-01 -7.76181579e-01
-1.46142438e-01 -1.07852650e+00 -8.30755174e-01 1.02418423e+00
-1.74652994e+00 -8.20566297e-01 1.78752991e-03 8.02545309e-01
1.58191472e-01 2.45500371e-01 7.83296764e-01 4.80179578e-01
-8.21659088e-01 4.96045232e-01 2.24764515e-02 3.08225125e-01
5.48006967e-02 -1.27118361e+00 1.42695248e-01 6.64325595e-01
-3.20840150e-01 6.76829636e-01 5.55286765e-01 -8.63691941e-02
-1.20904100e+00 -9.54355180e-01 1.11493790e+00 1.15579907e-02
7.35763073e-01 -4.89943743e-01 -1.16835380e+00 1.07946146e+00
4.47003752e-01 1.00816771e-01 1.83750764e-01 -2.39126775e-02
-8.10647234e-02 -2.73824841e-01 -1.05948806e+00 3.67431641e-01
5.04343629e-01 -4.50676829e-01 -6.62659585e-01 6.19477592e-02
6.40126348e-01 -3.77808034e-01 -9.58377361e-01 1.91867560e-01
3.82261485e-01 -1.21937835e+00 9.53092039e-01 -4.88171875e-01
3.62777889e-01 -2.70468257e-02 1.07119203e-01 -1.31521749e+00
-5.27194619e-01 -6.23082817e-01 -2.34679192e-01 9.86923099e-01
4.22856778e-01 -1.24708331e+00 4.85733986e-01 6.68154538e-01
-2.97486961e-01 -9.09756899e-01 -9.60781693e-01 -4.67050910e-01
4.54579949e-01 -1.52339950e-01 1.97135836e-01 6.41164124e-01
1.47543222e-01 6.82333291e-01 -1.74070045e-01 2.38134876e-01
4.65369672e-01 -9.93627831e-02 1.02466643e-01 -1.49656999e+00
-4.88002747e-01 -6.59438431e-01 -3.81689310e-01 -1.10171068e+00
3.23503494e-01 -9.12776709e-01 -1.57835543e-01 -1.48039198e+00
-1.42032653e-01 -2.98638135e-01 -4.71886337e-01 7.29137778e-01
1.24959968e-01 1.85927257e-01 2.45654941e-01 -1.04636751e-01
-1.92953825e-01 5.31980813e-01 8.90553653e-01 -7.69559108e-03
-1.29921377e-01 2.72602111e-01 -4.75414455e-01 1.04397678e+00
8.41777861e-01 -4.18214828e-01 -8.89477283e-02 -6.64139211e-01
2.67653763e-01 -1.01415709e-01 6.79303229e-01 -1.14126933e+00
4.22280431e-01 3.33155155e-01 4.16141629e-01 -2.39759862e-01
4.99073297e-01 -9.78508055e-01 -1.02262691e-01 8.33576858e-01
-1.82043821e-01 4.60696280e-01 5.38956821e-01 -8.22707042e-02
-1.80836916e-01 -2.65431792e-01 9.75080788e-01 -1.80944309e-01
-2.42746487e-01 6.48267344e-02 -6.91208363e-01 -1.62843063e-01
7.24142790e-01 1.59846351e-01 1.84588730e-01 -1.62222877e-01
-8.91503155e-01 2.22159028e-01 2.41184995e-01 1.30246073e-01
4.11162436e-01 -1.11563289e+00 -3.91578436e-01 4.44517523e-01
-7.24171340e-01 1.81442380e-01 -1.25485092e-01 1.32081521e+00
-4.55081671e-01 6.06253088e-01 -3.55699390e-01 -4.53318119e-01
-9.21770513e-01 5.77321172e-01 9.94175196e-01 -6.13975346e-01
-8.04765761e-01 7.50782549e-01 3.69105369e-01 -4.25910443e-01
2.73891538e-01 -8.42340708e-01 -3.10302228e-01 -4.90231402e-02
5.25877416e-01 5.14792025e-01 1.68650389e-01 -4.56782967e-01
-5.53359687e-01 7.35010862e-01 6.03844412e-02 -9.38800424e-02
1.50166738e+00 9.48702469e-02 -2.94552088e-01 7.04939961e-01
1.33272099e+00 -2.98345208e-01 -1.21257079e+00 -1.11075118e-01
-2.64501758e-02 3.86657119e-01 2.50342160e-01 -3.69918525e-01
-1.25213015e+00 1.33595252e+00 3.37566465e-01 4.92531270e-01
1.13219690e+00 -2.88515270e-01 6.80311203e-01 4.09812123e-01
-1.01466693e-01 -7.41503656e-01 -2.23759651e-01 8.39830577e-01
9.58407283e-01 -8.80003273e-01 -2.75838137e-01 -5.98589554e-02
-1.33484736e-01 1.17433560e+00 5.77289999e-01 -5.09623826e-01
8.92736256e-01 6.52101457e-01 -1.31594956e-01 9.09000337e-02
-8.17401230e-01 -1.80188101e-02 1.36553183e-01 9.00271386e-02
5.55325150e-01 -4.68250096e-01 -1.26892300e-02 6.87449396e-01
1.10767603e-01 1.16120279e-01 3.31265122e-01 7.41149187e-01
-3.64661008e-01 -7.87171543e-01 -3.13875109e-01 1.74326479e-01
-7.35051394e-01 -3.36443603e-01 -6.96553811e-02 6.97891235e-01
-7.31922463e-02 6.43859684e-01 2.71427006e-01 -9.81193036e-02
3.32051754e-01 3.58860224e-01 5.31324983e-01 -3.33767295e-01
-1.09747374e+00 -1.05259329e-01 -2.90136367e-01 -5.43040514e-01
-2.62249231e-01 -2.99650729e-01 -1.28101981e+00 -4.37753946e-01
-1.48031220e-01 2.64758140e-01 5.69896996e-01 1.21476281e+00
3.73902291e-01 8.74156535e-01 4.75123733e-01 -1.17167711e+00
-5.73714197e-01 -1.12280667e+00 -4.47530836e-01 -5.59640899e-02
8.15170288e-01 -6.56226635e-01 -4.73377556e-01 -7.22843558e-02] | [7.44025182723999, 3.4817287921905518] |
fe58dc0d-2beb-4bf8-a283-4cde673b9071 | tracking-emerges-by-looking-around-static | 2008.01295 | null | https://arxiv.org/abs/2008.01295v1 | https://arxiv.org/pdf/2008.01295v1.pdf | Tracking Emerges by Looking Around Static Scenes, with Neural 3D Mapping | We hypothesize that an agent that can look around in static scenes can learn rich visual representations applicable to 3D object tracking in complex dynamic scenes. We are motivated in this pursuit by the fact that the physical world itself is mostly static, and multiview correspondence labels are relatively cheap to collect in static scenes, e.g., by triangulation. We propose to leverage multiview data of \textit{static points} in arbitrary scenes (static or dynamic), to learn a neural 3D mapping module which produces features that are correspondable across time. The neural 3D mapper consumes RGB-D data as input, and produces a 3D voxel grid of deep features as output. We train the voxel features to be correspondable across viewpoints, using a contrastive loss, and correspondability across time emerges automatically. At test time, given an RGB-D video with approximate camera poses, and given the 3D box of an object to track, we track the target object by generating a map of each timestep and locating the object's features within each map. In contrast to models that represent video streams in 2D or 2.5D, our model's 3D scene representation is disentangled from projection artifacts, is stable under camera motion, and is robust to partial occlusions. We test the proposed architectures in challenging simulated and real data, and show that our unsupervised 3D object trackers outperform prior unsupervised 2D and 2.5D trackers, and approach the accuracy of supervised trackers. This work demonstrates that 3D object trackers can emerge without tracking labels, through multiview self-supervision on static data. | ['Paul Schydlo', 'Adam W. Harley', 'Shrinidhi K. Lakshmikanth', 'Katerina Fragkiadaki'] | 2020-08-04 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/5494_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123710596.pdf | eccv-2020-8 | ['3d-object-tracking'] | ['computer-vision'] | [-2.40777388e-01 3.32302190e-02 -2.75881052e-01 -1.44056335e-01
-6.18221045e-01 -9.86528397e-01 7.91135669e-01 -1.14693835e-01
-2.53515154e-01 2.80114532e-01 -2.17740580e-01 1.08234577e-01
7.66045302e-02 -4.03214514e-01 -1.27899647e+00 -5.25341392e-01
-3.25941950e-01 7.70782471e-01 2.51986802e-01 4.81225252e-01
3.41115482e-02 7.34894991e-01 -1.71921003e+00 -3.03388983e-01
1.46068543e-01 1.01034749e+00 2.62749970e-01 9.33845997e-01
2.21231028e-01 7.03882813e-01 -4.82592434e-01 1.23018727e-01
7.59914756e-01 -3.73423956e-02 -1.99950203e-01 5.65816522e-01
1.12751961e+00 -7.35254526e-01 -7.56310523e-01 1.02419209e+00
1.40760690e-01 5.80498353e-02 6.84493303e-01 -1.67123115e+00
-7.99727678e-01 -7.38966390e-02 -6.82838917e-01 1.14802599e-01
6.49097502e-01 6.61861718e-01 7.83512890e-01 -9.57603276e-01
9.34881449e-01 1.45787132e+00 6.78850651e-01 7.38037109e-01
-1.42998624e+00 -5.09958267e-01 5.01713157e-01 -3.05197001e-01
-1.20958626e+00 -3.88835460e-01 7.19402492e-01 -9.05277312e-01
8.90483975e-01 -7.45398253e-02 9.92705584e-01 1.35685575e+00
3.98315251e-01 7.91923344e-01 9.36710954e-01 8.96490458e-03
3.07158262e-01 -5.39484322e-02 -5.79294711e-02 8.80675256e-01
4.45833355e-01 4.83321577e-01 -7.85996795e-01 -7.76976272e-02
1.05350590e+00 4.77910191e-01 -6.71843663e-02 -1.35556257e+00
-1.68417215e+00 4.75230336e-01 4.21291202e-01 -2.60548562e-01
-1.13952674e-01 5.76711535e-01 4.22982685e-02 1.46366268e-01
3.88726205e-01 1.01429470e-01 -3.43483418e-01 5.25079295e-02
-5.73571622e-01 4.69344735e-01 5.03623068e-01 1.59596550e+00
7.81576335e-01 2.31635407e-01 1.04334705e-01 -1.27839476e-01
5.56640863e-01 1.26739728e+00 1.30539969e-01 -1.43765593e+00
3.15946817e-01 5.51675797e-01 5.81011593e-01 -8.76449943e-01
-4.09583688e-01 -1.98241770e-01 -5.12125671e-01 8.48803401e-01
3.99407089e-01 8.74831975e-02 -9.64606583e-01 2.00488067e+00
7.51546085e-01 3.79618049e-01 1.61444850e-03 1.21610117e+00
6.00246727e-01 3.60484064e-01 -2.35784903e-01 -6.92299902e-02
1.15829241e+00 -6.18941844e-01 -4.48371887e-01 -2.80131102e-01
4.08137619e-01 -4.18535620e-01 7.05810368e-01 6.10515289e-02
-1.15158939e+00 -7.44112849e-01 -9.61941242e-01 -8.12259093e-02
-2.29999766e-01 -1.27244800e-01 5.41739941e-01 3.07526350e-01
-1.16011178e+00 2.42036015e-01 -1.37406433e+00 -5.43139398e-01
4.37251359e-01 4.00663823e-01 -7.03832090e-01 2.16352373e-01
-4.83016878e-01 9.69767094e-01 1.57645985e-01 -1.52922586e-01
-1.86456370e+00 -7.82405794e-01 -1.02159536e+00 -3.44733357e-01
3.14929634e-01 -1.09150922e+00 1.20909059e+00 -5.35676897e-01
-1.34514081e+00 1.01827979e+00 -2.72038192e-01 -3.93725187e-01
6.83149934e-01 -3.98604602e-01 3.15806456e-03 2.01660916e-01
3.04998815e-01 8.67967725e-01 9.75583375e-01 -1.59483171e+00
-6.68452501e-01 -6.95944130e-01 2.91576624e-01 4.70783740e-01
1.77026793e-01 -4.22318876e-01 -4.51649129e-01 -1.95887193e-01
4.26253438e-01 -1.14316642e+00 -1.46113977e-01 6.39468431e-01
-3.75805855e-01 -7.20514879e-02 1.18260145e+00 -6.81812465e-02
7.45389536e-02 -2.14282441e+00 3.35027337e-01 -1.76978171e-01
6.08600914e-01 -3.61079454e-01 6.90540671e-02 -1.44680128e-01
2.53890932e-01 -1.95030883e-01 1.45320117e-01 -7.36418486e-01
2.55391419e-01 1.90906391e-01 -3.87947261e-01 1.18093050e+00
6.75748289e-03 1.01308858e+00 -1.13660645e+00 -4.35362309e-01
6.47166491e-01 6.11428201e-01 -5.32869458e-01 4.17029917e-01
-6.05949283e-01 7.18288302e-01 -6.03813350e-01 7.53188550e-01
5.25123298e-01 -6.51525319e-01 -2.27845252e-01 -9.89981592e-02
-2.88425684e-01 3.79778631e-02 -1.19955039e+00 1.99075043e+00
-9.06076208e-02 8.02390575e-01 4.22283933e-02 -6.33572042e-01
8.40077877e-01 2.75398374e-01 8.63453627e-01 -4.34094995e-01
-3.80488075e-02 -1.47874489e-01 -4.93052930e-01 -4.25921440e-01
4.59835321e-01 3.44650358e-01 -1.60132289e-01 5.79745531e-01
3.17344010e-01 -2.79167324e-01 -3.04265857e-01 3.19752574e-01
1.21918702e+00 7.26715922e-01 9.74204615e-02 6.65853918e-02
-4.40533087e-02 1.02532446e-01 5.76914668e-01 8.26032519e-01
-3.89431566e-01 4.41665083e-01 1.79010063e-01 -6.97646618e-01
-1.26697195e+00 -1.63089085e+00 -3.05350665e-02 5.49218178e-01
7.62246549e-01 -1.98318109e-01 -2.52356738e-01 -7.15385199e-01
3.53308886e-01 1.46477833e-01 -5.80627918e-01 6.10157251e-02
-5.73236942e-01 -1.27854710e-02 2.10665464e-01 2.70723403e-01
1.92011863e-01 -4.82006252e-01 -1.39618337e+00 -4.57256623e-02
1.45029739e-01 -1.30254388e+00 -4.87832665e-01 2.72843838e-01
-8.69869173e-01 -1.19029379e+00 -3.96741122e-01 -6.09144986e-01
8.36192012e-01 7.54253983e-01 9.98635530e-01 -3.77530515e-01
-1.97005585e-01 1.25504959e+00 4.26025828e-03 -1.33795112e-01
-7.73507133e-02 -3.61001104e-01 7.76121676e-01 -1.86732873e-01
3.22055101e-01 -6.02437317e-01 -5.90022743e-01 3.16873074e-01
-3.18725169e-01 8.82069841e-02 1.54808894e-01 4.88506585e-01
8.53408456e-01 -5.82287431e-01 -1.11252151e-01 -3.36492658e-01
-3.92504960e-01 -2.95177609e-01 -1.27210116e+00 1.86942723e-02
-1.17840320e-01 -1.03097282e-01 3.95790756e-01 -7.26219714e-01
-6.30294681e-01 6.32240176e-01 8.65277052e-01 -1.40171230e+00
-3.72758120e-01 -3.06669176e-01 -1.33219108e-01 -2.41808653e-01
6.68994188e-01 1.21326007e-01 1.58258021e-01 -3.29514712e-01
6.09681785e-01 -5.39861768e-02 5.91081679e-01 -5.48063338e-01
1.30139625e+00 9.22810018e-01 2.05246344e-01 -5.67676961e-01
-7.86240995e-01 -3.51429254e-01 -9.21590984e-01 -6.93334281e-01
1.12710392e+00 -1.35846174e+00 -1.39751685e+00 1.68886617e-01
-1.31179702e+00 -3.94723028e-01 -4.77219284e-01 8.61127734e-01
-9.76167262e-01 9.34773535e-02 -2.05276117e-01 -7.95056105e-01
2.43255705e-01 -1.25334942e+00 1.49827874e+00 1.40397906e-01
-1.43297613e-01 -8.82052660e-01 8.46092701e-02 -9.44117382e-02
-2.27228366e-02 7.32892871e-01 3.88262868e-01 -2.39381194e-01
-1.49834394e+00 2.05087624e-02 8.34412649e-02 -1.85325608e-01
1.33729950e-01 -8.73093531e-02 -1.13521075e+00 -5.50171196e-01
1.45433843e-01 -3.86355698e-01 4.25494909e-01 5.78682482e-01
8.68294418e-01 -2.98699260e-01 -7.17255533e-01 9.26566541e-01
1.27408886e+00 2.86803823e-02 -2.45155469e-01 2.66582429e-01
8.74189079e-01 3.53089720e-01 5.76364100e-01 4.10939693e-01
5.97254634e-01 8.07072759e-01 1.02752590e+00 1.14953652e-01
-1.99679598e-01 -4.74735111e-01 6.59151554e-01 4.98539150e-01
2.02260301e-01 -4.28018272e-02 -7.85139143e-01 4.67043132e-01
-1.87247479e+00 -9.11896467e-01 1.00764930e-01 2.29642558e+00
2.77368844e-01 1.85343012e-01 1.29921257e-01 -6.13125443e-01
6.57441735e-01 3.71236145e-01 -1.17706716e+00 3.55023682e-01
-1.97462514e-01 -4.59853202e-01 8.47937047e-01 5.32240927e-01
-1.18546724e+00 8.64302218e-01 6.02820396e+00 -1.94089636e-01
-9.62749779e-01 1.75045088e-01 1.20559782e-01 -5.77639759e-01
-2.35365301e-01 2.40316391e-01 -9.94233668e-01 2.94450969e-01
5.85648715e-01 -2.39218608e-01 3.69754672e-01 1.04841816e+00
2.75790840e-02 1.33434176e-01 -1.79197121e+00 1.10800755e+00
2.30883911e-01 -1.39230883e+00 -1.34399394e-02 3.48362833e-01
7.44347513e-01 4.47894901e-01 4.23840255e-01 3.80944237e-02
8.11745584e-01 -6.15815699e-01 1.10857546e+00 5.60795546e-01
5.27255118e-01 -1.32142901e-01 -1.30045533e-01 4.28304553e-01
-1.36372089e+00 2.92212367e-02 -3.50243330e-01 8.43503922e-02
3.97954494e-01 1.99338108e-01 -7.43236780e-01 1.75312713e-01
9.07054842e-01 1.26684403e+00 -4.39383179e-01 9.94886816e-01
-1.54983886e-02 -8.49669203e-02 -6.03682280e-01 2.37140134e-01
1.46307468e-01 -1.17250606e-01 1.01762688e+00 5.23950994e-01
3.82095009e-01 -2.36757025e-02 5.27642667e-01 1.02036655e+00
-6.85191900e-02 -7.05700636e-01 -1.27538610e+00 4.19929236e-01
6.05906725e-01 1.00347567e+00 -6.56897008e-01 -3.28657895e-01
-5.11736691e-01 8.45578551e-01 3.14241946e-01 6.21267557e-01
-1.01316869e+00 2.83420473e-01 9.37925160e-01 -3.45893539e-02
5.26324809e-01 -7.00143039e-01 2.31715590e-02 -1.37725151e+00
1.91662461e-01 -5.06008029e-01 1.23441443e-01 -1.02580285e+00
-1.10554326e+00 2.54763216e-01 1.42498687e-01 -1.93332136e+00
-2.88971037e-01 -6.49135292e-01 -1.71658128e-01 4.45311755e-01
-1.24756527e+00 -1.33439469e+00 -4.34741378e-01 7.97583878e-01
4.77466822e-01 -6.18068911e-02 5.29018998e-01 5.27882017e-03
-9.39193368e-03 1.83715060e-01 -4.98134084e-02 2.80269720e-02
5.42539835e-01 -1.27441144e+00 6.18802667e-01 7.34392345e-01
5.43293059e-01 6.33758485e-01 7.00983047e-01 -6.23046398e-01
-2.19886231e+00 -1.25173497e+00 3.85218710e-01 -1.36247683e+00
6.87280774e-01 -7.86015153e-01 -5.57887733e-01 1.33723891e+00
-6.13228567e-02 6.68542266e-01 1.90987349e-01 -1.34787574e-01
-6.95811629e-01 -9.33341384e-02 -9.83892262e-01 4.93929356e-01
1.55162323e+00 -7.11354136e-01 -6.24005556e-01 5.51001370e-01
1.06163335e+00 -9.00081575e-01 -9.09853578e-01 1.43413946e-01
7.34475911e-01 -7.91787744e-01 1.38400364e+00 -5.80697775e-01
-1.22515209e-01 -8.14633310e-01 -3.82267058e-01 -1.12787890e+00
-1.32982478e-01 -6.25757277e-01 -5.69520056e-01 6.87244475e-01
-6.88232202e-03 -4.43550557e-01 7.94806719e-01 5.25029719e-01
-1.84885696e-01 -1.35820001e-01 -1.24685299e+00 -1.02314138e+00
-2.66250074e-01 -3.45650822e-01 4.31164533e-01 7.29097784e-01
-4.26193893e-01 4.04646434e-02 -4.41028178e-01 8.48906279e-01
1.35207462e+00 5.30872881e-01 1.36627316e+00 -1.30862331e+00
-2.40837350e-01 -1.51231721e-01 -8.18056583e-01 -1.46452188e+00
4.11611021e-01 -8.71734738e-01 5.42882569e-02 -1.14813566e+00
1.48547083e-01 -4.49049234e-01 4.89549078e-02 2.77239978e-01
3.45220685e-01 9.54762772e-02 4.33017313e-01 4.51659292e-01
-1.12598813e+00 6.27403021e-01 1.26941049e+00 -2.86366612e-01
-1.03754550e-01 -2.20155746e-01 -3.62007134e-02 8.09551477e-01
3.34526300e-01 -4.98916805e-01 -2.95992851e-01 -7.77812660e-01
-5.53589799e-02 2.60304302e-01 9.77378011e-01 -9.45349097e-01
5.77677727e-01 -1.96263596e-01 8.07707250e-01 -9.33182299e-01
8.36041451e-01 -1.20284927e+00 4.05085504e-01 4.06272620e-01
-2.95968473e-01 3.08255643e-01 3.59295607e-02 9.05804813e-01
3.14810961e-01 5.08048713e-01 3.87036532e-01 -3.20333540e-01
-5.82576454e-01 8.09065402e-01 -1.80392936e-01 4.41411622e-02
1.25404704e+00 -3.91047269e-01 -4.38509613e-01 -1.69462517e-01
-7.21708536e-01 3.66259187e-01 1.01476383e+00 6.81434929e-01
7.14515567e-01 -1.60452378e+00 -1.99858591e-01 4.05763477e-01
2.38651365e-01 4.70930815e-01 8.46075267e-02 6.51296139e-01
-1.86650291e-01 3.20823222e-01 -2.70776629e-01 -1.65543795e+00
-1.02635944e+00 9.10540462e-01 4.87840950e-01 4.36657280e-01
-1.10385752e+00 4.42101896e-01 6.42529607e-01 -4.91893411e-01
5.97752512e-01 -5.63333631e-01 1.85334653e-01 -2.61976928e-01
4.21745569e-01 1.36059687e-01 -4.52767462e-01 -8.75514209e-01
-3.63350898e-01 1.06967068e+00 1.26278147e-01 -2.16684178e-01
1.17117167e+00 -2.72675663e-01 2.88939476e-01 8.47013354e-01
1.09944868e+00 -3.50787252e-01 -2.11069345e+00 -2.46109337e-01
-2.08490968e-01 -7.33780742e-01 -1.97240502e-01 -1.21221058e-01
-9.64762032e-01 6.56216264e-01 8.30434620e-01 8.67884606e-02
4.40440863e-01 3.84908795e-01 2.37380251e-01 6.28504038e-01
7.80130804e-01 -4.90111649e-01 3.94872963e-01 4.23931718e-01
5.03034949e-01 -1.36151791e+00 2.80447286e-02 2.09697396e-01
-2.10310742e-01 8.16316545e-01 8.75481129e-01 -4.11422640e-01
5.89808941e-01 3.69458675e-01 -1.07365750e-01 -4.44845825e-01
-8.86283398e-01 -7.33910054e-02 -2.66773216e-02 8.50615621e-01
-2.90907860e-01 -6.66018855e-03 8.25994670e-01 -2.97184348e-01
-1.41633302e-02 -3.05486739e-01 3.63065690e-01 8.99906218e-01
-4.36275989e-01 -5.42876661e-01 -4.35185164e-01 9.66741517e-02
1.02654859e-01 5.52937746e-01 -2.35147700e-01 1.01428795e+00
4.82981978e-03 5.58954000e-01 4.09515649e-01 -2.88714439e-01
2.96320677e-01 -1.52840719e-01 7.82586336e-01 -5.22722423e-01
-1.57165512e-01 7.69035593e-02 -5.18256247e-01 -9.91076946e-01
-8.41412842e-01 -1.05769181e+00 -1.17746651e+00 -1.96726009e-01
-7.14642256e-02 -2.60401398e-01 6.82614982e-01 7.43902743e-01
4.27943945e-01 2.76799917e-01 7.68235981e-01 -1.42825866e+00
-3.41554791e-01 -3.98032814e-01 -2.96495914e-01 4.54223722e-01
1.13917041e+00 -1.15156734e+00 -5.57884932e-01 3.92306596e-01] | [7.021667003631592, -2.343470573425293] |
d769eff8-a6b1-4cae-af7a-cbc472fc72ff | diffface-diffusion-based-face-swapping-with | 2212.13344 | null | https://arxiv.org/abs/2212.13344v1 | https://arxiv.org/pdf/2212.13344v1.pdf | DiffFace: Diffusion-based Face Swapping with Facial Guidance | In this paper, we propose a diffusion-based face swapping framework for the first time, called DiffFace, composed of training ID conditional DDPM, sampling with facial guidance, and a target-preserving blending. In specific, in the training process, the ID conditional DDPM is trained to generate face images with the desired identity. In the sampling process, we use the off-the-shelf facial expert models to make the model transfer source identity while preserving target attributes faithfully. During this process, to preserve the background of the target image and obtain the desired face swapping result, we additionally propose a target-preserving blending strategy. It helps our model to keep the attributes of the target face from noise while transferring the source facial identity. In addition, without any re-training, our model can flexibly apply additional facial guidance and adaptively control the ID-attributes trade-off to achieve the desired results. To the best of our knowledge, this is the first approach that applies the diffusion model in face swapping task. Compared with previous GAN-based approaches, by taking advantage of the diffusion model for the face swapping task, DiffFace achieves better benefits such as training stability, high fidelity, diversity of the samples, and controllability. Extensive experiments show that our DiffFace is comparable or superior to the state-of-the-art methods on several standard face swapping benchmarks. | ['Kwanghee Lee', 'Seungryong Kim', 'Kychul Lee', 'Jisu Nam', 'Junyoung Seo', 'Seokju Cho', 'Yunho Kim', 'Kihong Kim'] | 2022-12-27 | null | null | null | null | ['face-swapping'] | ['computer-vision'] | [ 2.35709175e-01 3.33390623e-01 -7.62785450e-02 -2.76662171e-01
-3.81161481e-01 -3.33108842e-01 5.05251586e-01 -6.60684884e-01
2.38532886e-01 6.35094762e-01 -1.41462818e-01 1.04143001e-01
1.08300015e-01 -9.02206898e-01 -7.74565041e-01 -9.74551678e-01
5.02245009e-01 4.19479996e-01 -6.66859522e-02 -2.02056512e-01
-3.09349783e-02 6.43660545e-01 -1.61277080e+00 4.14995477e-02
1.16547668e+00 1.27704358e+00 2.88818516e-02 4.41755578e-02
-1.77291445e-02 5.86821854e-01 -4.41714138e-01 -5.84555626e-01
5.61695695e-01 -8.09907973e-01 -3.73610049e-01 2.97625065e-01
6.87731147e-01 -4.65160459e-01 -1.96487188e-01 1.05547249e+00
6.33230925e-01 -7.14429840e-02 5.32726407e-01 -1.57159889e+00
-9.44667876e-01 1.67892560e-01 -8.10118496e-01 -5.71715236e-01
1.82340860e-01 3.37801754e-01 5.12390494e-01 -9.98984158e-01
7.06629336e-01 1.37717164e+00 5.64881802e-01 1.10424459e+00
-1.29750252e+00 -1.26464784e+00 2.02039957e-01 -8.73025432e-02
-1.28350103e+00 -6.50730193e-01 1.04825115e+00 -3.56501132e-01
4.28152010e-02 2.06797034e-01 7.77492106e-01 1.13853621e+00
-7.12368041e-02 6.39646828e-01 1.21068859e+00 -3.52928430e-01
1.41010523e-01 1.45347282e-01 -5.82852721e-01 9.01932061e-01
-9.41695422e-02 1.76832318e-01 -5.00077903e-01 -1.95154548e-01
1.00559437e+00 5.61222397e-02 -6.21487319e-01 -6.69626713e-01
-9.50820327e-01 5.20079195e-01 4.22355145e-01 5.54746687e-02
-4.06135112e-01 2.72496752e-02 -2.04576831e-02 3.57631922e-01
5.35485148e-01 -1.00302845e-02 -1.28436729e-01 2.56728530e-01
-1.01849961e+00 9.69188288e-02 6.34601057e-01 1.02896690e+00
9.04473126e-01 1.85459912e-01 -4.32239532e-01 9.24772739e-01
3.29593360e-01 6.39196277e-01 2.85646915e-01 -1.23833764e+00
2.66411662e-01 5.85443377e-01 8.26080069e-02 -8.71880114e-01
3.38732153e-01 -3.65313798e-01 -1.02841079e+00 5.23911655e-01
1.42530546e-01 -1.43094093e-01 -1.16349697e+00 2.08990359e+00
6.31070614e-01 4.82325673e-01 -1.34853542e-01 5.69713593e-01
5.16328096e-01 5.81639230e-01 -1.91761896e-01 -3.79895151e-01
1.14110136e+00 -9.99539554e-01 -8.07223201e-01 -1.36280833e-02
-5.37839644e-02 -7.74083436e-01 1.08348382e+00 1.90492988e-01
-1.09600806e+00 -6.80322230e-01 -9.71482754e-01 1.89076021e-01
1.19569354e-01 3.56854051e-01 3.17285776e-01 7.99008846e-01
-1.09045541e+00 6.08307362e-01 -7.06028938e-01 -2.09308729e-01
8.15245509e-01 5.09704053e-01 -5.71981609e-01 -1.83615535e-01
-9.03187275e-01 4.64528114e-01 -1.14866577e-01 1.66990414e-01
-1.13069534e+00 -1.04488027e+00 -6.53135180e-01 -1.28306132e-02
3.10513377e-01 -9.76552129e-01 8.99252057e-01 -1.54558921e+00
-2.20597315e+00 8.52084577e-01 -2.72966355e-01 5.31170592e-02
8.86358202e-01 5.03880493e-02 -1.75680310e-01 -1.63304769e-02
4.13632430e-02 8.40234518e-01 1.39714622e+00 -1.52754080e+00
-3.89784217e-01 -3.06279510e-01 -7.29665831e-02 -3.86409014e-02
-5.94861269e-01 -1.63444549e-01 -7.24488378e-01 -8.55990767e-01
-2.40709499e-01 -1.13731086e+00 2.05383986e-01 6.29324138e-01
-2.39078999e-01 1.19334221e-01 1.19518697e+00 -6.43939197e-01
9.08137977e-01 -2.52792954e+00 4.30792719e-01 3.07947934e-01
1.65077075e-01 4.26445127e-01 -3.54969472e-01 2.15388268e-01
-9.71431360e-02 -9.38138142e-02 -4.01117295e-01 -7.97189236e-01
-1.82016790e-01 2.39948094e-01 -3.19379658e-01 4.01716858e-01
3.25761676e-01 7.71510541e-01 -7.02166319e-01 -5.02957225e-01
-5.23238443e-02 8.62529397e-01 -6.94329143e-01 4.58874047e-01
-1.25698939e-01 8.60604644e-01 -2.98434377e-01 6.60701156e-01
1.15962636e+00 1.88355353e-02 2.04324365e-01 -3.17880005e-01
1.95836589e-01 -3.84628057e-01 -1.05985796e+00 1.56717503e+00
-5.34823596e-01 2.94931263e-01 4.59091038e-01 -6.94393635e-01
1.11020076e+00 3.38720948e-01 3.54577154e-01 -5.25003076e-01
1.69547960e-01 3.66682053e-01 -1.43598586e-01 -1.25228330e-01
-3.21725234e-02 -8.86907950e-02 4.69374895e-01 3.60147148e-01
7.38173500e-02 -2.80041248e-03 -1.84943572e-01 -1.12140551e-01
5.01550913e-01 2.72828817e-01 -1.83185756e-01 -2.52828807e-01
7.23533750e-01 -5.68821788e-01 9.29688573e-01 4.87612449e-02
6.84050173e-02 8.24073374e-01 5.48386753e-01 -1.32834464e-01
-8.08638871e-01 -9.67464268e-01 6.94211274e-02 6.88343942e-01
9.08875987e-02 -1.23728044e-01 -1.22046745e+00 -8.23343635e-01
6.89997151e-02 4.76905912e-01 -9.29277599e-01 -3.57646257e-01
-6.85908675e-01 -3.43971997e-01 3.66287589e-01 4.54315126e-01
9.69162166e-01 -7.76703835e-01 -3.90598215e-02 8.47779308e-03
-1.51536539e-01 -7.51412630e-01 -9.88438904e-01 -4.11280125e-01
-7.41111219e-01 -9.98317003e-01 -1.01563692e+00 -1.05960214e+00
1.16773200e+00 6.72448650e-02 6.03553236e-01 2.41629452e-01
8.04329757e-03 1.46166638e-01 -6.82940185e-02 -1.83113992e-01
-4.95440543e-01 -6.34465441e-02 -1.16952933e-01 6.80182397e-01
-3.32779795e-01 -8.68947923e-01 -8.07094395e-01 6.57874882e-01
-8.55137050e-01 1.70360580e-01 4.72937584e-01 1.08215380e+00
6.65847301e-01 4.91677085e-03 5.64556956e-01 -7.89738953e-01
3.21330369e-01 -3.14612299e-01 -7.12638080e-01 3.90360087e-01
-5.77522397e-01 -8.64292607e-02 8.46700728e-01 -8.04788411e-01
-1.31170952e+00 1.57299682e-01 -8.81277844e-02 -8.92738044e-01
1.64471954e-01 -1.31121337e-01 -7.79774666e-01 -5.25430202e-01
3.67712647e-01 2.85343528e-01 5.84124982e-01 -4.51787740e-01
3.06133896e-01 5.68449438e-01 4.63656306e-01 -6.77579820e-01
8.72187197e-01 7.30318844e-01 4.37550098e-02 -3.19440484e-01
-4.20084506e-01 3.26329678e-01 -4.49920624e-01 -2.19968721e-01
5.20890176e-01 -8.83832157e-01 -7.01019466e-01 1.03291774e+00
-9.23655570e-01 -4.53849882e-01 -3.97925109e-01 -2.67574061e-02
-5.77792108e-01 1.54572949e-01 -3.78926158e-01 -6.80572212e-01
-3.80898237e-01 -1.22466528e+00 1.13415897e+00 3.44361961e-01
1.73639506e-01 -6.77015781e-01 -1.25390157e-01 2.44873047e-01
7.06349194e-01 4.56817210e-01 8.44282687e-01 -5.36937676e-02
-6.71034276e-01 1.62048731e-02 -3.13660130e-02 5.27248323e-01
7.32470095e-01 1.78781867e-01 -9.35702384e-01 -5.20676017e-01
1.15114816e-01 -1.77471280e-01 7.14223564e-01 1.36764124e-01
1.22837734e+00 -3.87593776e-01 -5.88879347e-01 8.72387528e-01
1.11299884e+00 3.22050959e-01 8.81382227e-01 4.34159376e-02
6.33449554e-01 8.12570691e-01 6.75525367e-01 4.28384423e-01
2.54883915e-01 7.29581773e-01 4.68233466e-01 -2.40399852e-01
-4.34123695e-01 -6.01126611e-01 4.25700635e-01 2.34183863e-01
-6.83092400e-02 -1.46249503e-01 -3.96357656e-01 3.97955120e-01
-1.76209831e+00 -8.47901821e-01 4.73050863e-01 2.26638961e+00
9.93072391e-01 -3.92511606e-01 1.08419478e-01 -8.46240222e-02
9.48740363e-01 9.74284634e-02 -7.40828037e-01 2.24357352e-01
6.07028343e-02 2.82312721e-01 9.17238221e-02 4.69337702e-01
-7.93460071e-01 9.28878307e-01 5.55277061e+00 1.03385723e+00
-1.47865772e+00 2.03837484e-01 9.40112948e-01 -1.94240317e-01
-4.79202688e-01 2.34630574e-02 -6.91107154e-01 6.74876392e-01
1.97652087e-01 5.02623990e-02 7.83589184e-01 7.54218161e-01
8.39688443e-03 3.95899385e-01 -9.85586345e-01 9.80498850e-01
1.80345520e-01 -1.23336852e+00 7.94925913e-02 2.04527482e-01
9.13004398e-01 -6.38565123e-01 2.82851219e-01 -4.52675596e-02
2.16076389e-01 -9.72343326e-01 8.85298967e-01 6.82965934e-01
1.15775728e+00 -8.54856133e-01 3.77907217e-01 1.05112426e-01
-1.11464727e+00 -1.01301685e-01 -5.07049151e-02 4.77420539e-01
1.11851521e-01 6.65198743e-01 -2.76594549e-01 5.65505624e-01
5.94538331e-01 6.31474078e-01 -3.13375682e-01 6.20199442e-01
-3.77387285e-01 3.81534278e-01 -2.14999989e-01 4.42992687e-01
-3.23503137e-01 -4.48539644e-01 4.73199576e-01 4.27332699e-01
6.55090868e-01 -7.05142692e-02 -5.08187450e-02 1.08723021e+00
-4.11442399e-01 -4.02806737e-02 -4.56191629e-01 1.93976149e-01
6.52721405e-01 1.28091860e+00 -3.40982974e-01 -1.73598006e-01
-8.19892660e-02 1.31329036e+00 2.92183340e-01 3.15593481e-01
-1.05311227e+00 -2.19083294e-01 9.05392408e-01 3.60349089e-01
5.78238428e-01 8.92523378e-02 4.79465835e-02 -1.04287910e+00
2.76704133e-01 -1.16821527e+00 -2.11199317e-02 -6.12684131e-01
-1.35481346e+00 7.52599955e-01 -1.42624110e-01 -1.20773375e+00
-5.11807529e-03 -2.99258828e-01 -7.78719544e-01 9.62326407e-01
-1.50626612e+00 -1.67890501e+00 -6.05072379e-01 8.44133437e-01
1.41421795e-01 -3.85965288e-01 5.95742822e-01 3.37484956e-01
-6.66042864e-01 9.40378129e-01 -1.65898651e-02 -1.77858472e-02
1.03955448e+00 -7.83483088e-01 2.64655441e-01 5.80299914e-01
-3.11651230e-01 7.34867811e-01 3.96358907e-01 -7.16989040e-01
-1.30395961e+00 -1.22783124e+00 3.33076179e-01 3.49072069e-02
2.49561355e-01 -4.84383196e-01 -8.58508170e-01 5.89966297e-01
4.36990499e-01 2.59454578e-01 6.05219841e-01 -4.63661283e-01
-4.17544663e-01 -6.61880136e-01 -1.58809972e+00 6.30390525e-01
1.40237677e+00 -3.82503062e-01 -2.19876058e-02 1.75182417e-01
4.92549926e-01 -4.62041855e-01 -7.69638717e-01 4.66500312e-01
6.66750312e-01 -8.74999166e-01 8.53067994e-01 -1.17956303e-01
2.87339807e-01 -5.77046216e-01 1.35855138e-01 -1.47341990e+00
-3.60754430e-01 -9.81329799e-01 6.11519776e-02 1.91936028e+00
5.80956750e-02 -1.06945610e+00 8.19608867e-01 4.29628015e-01
6.35071099e-02 -7.85058320e-01 -1.03442812e+00 -8.38487089e-01
1.01640694e-01 3.31606925e-01 1.18675590e+00 8.82128417e-01
-6.00255191e-01 -1.51143089e-01 -6.59355879e-01 1.46285906e-01
6.29703164e-01 2.42963701e-01 9.57629085e-01 -1.10613847e+00
-3.98751557e-01 -3.52217317e-01 -1.33063063e-01 -9.23478961e-01
4.91857022e-01 -5.84168315e-01 -4.98391129e-02 -1.12923813e+00
6.26698462e-03 -7.34291732e-01 -2.74702646e-02 7.48078585e-01
-1.71597764e-01 2.28628933e-01 3.06275904e-01 2.65973181e-01
2.08715513e-01 9.23345864e-01 1.62480998e+00 -1.64753690e-01
-1.47839248e-01 -9.38483849e-02 -9.13312614e-01 3.89154613e-01
7.42130518e-01 -4.13199425e-01 -7.22963929e-01 -4.32563871e-01
-2.54806638e-01 -3.09508424e-02 3.60404521e-01 -8.97646606e-01
1.32701159e-01 -3.75960499e-01 2.77341753e-01 2.49546804e-02
4.93953526e-01 -1.03785658e+00 6.78012609e-01 4.85900342e-01
7.90514648e-02 -2.34605446e-01 1.21110462e-01 4.38960582e-01
-2.67802119e-01 1.26855955e-01 1.19787514e+00 2.06589520e-01
-1.71154380e-01 6.71241462e-01 2.94046942e-02 -1.94062814e-01
1.43821406e+00 -1.98713407e-01 -2.65521705e-01 -3.94155383e-01
-3.84667814e-01 1.58236250e-01 9.34632301e-01 3.93581122e-01
4.71219927e-01 -1.63023603e+00 -6.74076796e-01 8.02815855e-01
-6.69826791e-02 9.18124616e-02 3.40076029e-01 8.70673060e-01
-3.23512644e-01 -3.41803014e-01 -4.47004676e-01 -4.59475398e-01
-1.25841951e+00 5.60268104e-01 5.59050977e-01 -5.33168092e-02
-3.16822022e-01 7.63796806e-01 5.10588169e-01 -4.92086738e-01
-3.40359844e-02 1.73089400e-01 1.83335081e-01 -2.40101479e-02
5.54456294e-01 1.95171252e-01 -9.71856564e-02 -5.39341390e-01
-3.21041524e-01 9.50694263e-01 -8.26944131e-03 -1.00727156e-01
1.26282167e+00 1.08615421e-02 -3.82849157e-01 -1.99431613e-01
1.01015639e+00 4.55437332e-01 -1.79083371e+00 9.89122875e-03
-8.16565633e-01 -8.76835287e-01 -1.39372095e-01 -7.40463972e-01
-1.86485219e+00 7.58985341e-01 6.63870096e-01 -2.34197930e-01
1.55293262e+00 -2.35630959e-01 8.46787930e-01 -3.56086642e-01
4.36027616e-01 -8.05814862e-01 2.34476984e-01 -3.42027061e-02
1.12339211e+00 -8.57894003e-01 -3.48103672e-01 -8.35178971e-01
-6.44956708e-01 9.76530015e-01 9.05963778e-01 -1.25827506e-01
6.77183986e-01 3.41625154e-01 -3.64971645e-02 1.68171197e-01
-4.87716556e-01 2.71194577e-01 6.38451949e-02 7.73951232e-01
5.70723601e-03 -2.05696389e-01 8.79715979e-02 2.85587966e-01
-1.14194907e-01 2.87648559e-01 -2.91815307e-02 7.69623637e-01
-7.41188824e-02 -1.60067749e+00 -3.62189412e-01 8.63742232e-02
-1.57123879e-01 4.94046696e-02 -4.71990317e-01 6.82051778e-01
3.96250546e-01 6.51275694e-01 3.15532275e-02 -3.82106274e-01
3.19714278e-01 1.56668071e-02 6.81145668e-01 -3.11399043e-01
-4.87245798e-01 1.30800456e-02 -1.99685842e-01 -5.20166814e-01
-3.89662504e-01 -5.46729326e-01 -1.00956595e+00 -5.91231823e-01
-3.82533520e-01 -1.06320130e-02 3.44933987e-01 7.10737050e-01
9.46111560e-01 3.40311557e-01 1.02605903e+00 -8.30047429e-01
-3.90708625e-01 -7.25963175e-01 -5.00339389e-01 3.82003576e-01
3.02015007e-01 -8.12508106e-01 -2.70812839e-01 1.12795368e-01] | [12.72463321685791, -0.07322370260953903] |
06b73885-30b7-4837-b46e-6a3c7a5b4c3c | on-context-distribution-shift-in-task | 2304.00354 | null | https://arxiv.org/abs/2304.00354v2 | https://arxiv.org/pdf/2304.00354v2.pdf | On Context Distribution Shift in Task Representation Learning for Offline Meta RL | Offline Meta Reinforcement Learning (OMRL) aims to learn transferable knowledge from offline datasets to enhance the learning process for new target tasks. Context-based Reinforcement Learning (RL) adopts a context encoder to expediently adapt the agent to new tasks by inferring the task representation, and then adjusting the policy based on this inferred representation. In this work, we focus on context-based OMRL, specifically on the challenge of learning task representation for OMRL. We conduct experiments that demonstrate that the context encoder trained on offline datasets might encounter distribution shift between the contexts used for training and testing. To overcome this problem, we present a hard-sampling-based strategy to train a robust task context encoder. Our experimental findings on diverse continuous control tasks reveal that utilizing our approach yields more robust task representations and better testing performance in terms of accumulated returns compared to baseline methods. Our code is available at https://github.com/ZJLAB-AMMI/HS-OMRL. | ['Bin Liu', 'ZiHao Zhou', 'Chenyang Zhao'] | 2023-04-01 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [ 3.07738096e-01 -6.02553459e-03 -3.94774467e-01 -4.01647627e-01
-8.24987829e-01 -6.50767148e-01 6.45927548e-01 -6.80517331e-02
-4.54369903e-01 1.08523476e+00 -6.58972487e-02 -4.88290608e-01
-1.04412615e-01 -5.00527620e-01 -9.83645737e-01 -4.60250050e-01
-1.66111335e-01 3.58372003e-01 -1.37336284e-01 -2.01002717e-01
4.24807698e-01 1.13047786e-01 -1.51653779e+00 4.99096811e-01
9.16444421e-01 8.24367046e-01 7.68945992e-01 6.81792259e-01
2.80474816e-02 9.06668782e-01 -6.89218760e-01 -7.56016374e-02
3.51227880e-01 -4.31950271e-01 -8.85124266e-01 -8.19656402e-02
1.92958582e-02 -3.78493488e-01 -3.65626812e-01 8.05323899e-01
4.26719815e-01 5.66572726e-01 5.53467274e-01 -1.10607135e+00
-7.65195072e-01 6.82911396e-01 -2.00217545e-01 4.59371209e-01
2.48233840e-01 5.86630464e-01 8.47676218e-01 -7.86235332e-01
4.84404504e-01 1.04597902e+00 2.80880928e-01 8.42378020e-01
-1.08698893e+00 -6.88874125e-01 7.02179134e-01 4.62647945e-01
-9.03726757e-01 -3.32079738e-01 8.10451925e-01 -1.96137279e-01
1.02624691e+00 -8.37528557e-02 5.09907246e-01 1.68110883e+00
3.22472185e-01 1.05544770e+00 1.40809357e+00 -5.54370701e-01
6.48509800e-01 4.14553732e-02 -2.23921597e-01 4.10742670e-01
8.24379995e-02 6.79196119e-01 -6.09275103e-01 -6.18865378e-02
7.43815899e-01 4.50789034e-02 -4.87651452e-02 -2.45762706e-01
-1.16387796e+00 6.96407259e-01 2.89198041e-01 1.83166623e-01
-5.70368946e-01 5.06695688e-01 5.88362634e-01 5.45437157e-01
4.21043396e-01 8.14962208e-01 -8.37306261e-01 -4.25458968e-01
-3.70986044e-01 3.85706782e-01 5.46699166e-01 1.06269515e+00
7.00984895e-01 2.52757281e-01 -6.50134921e-01 8.10002208e-01
4.88426760e-02 4.43496704e-01 8.84012997e-01 -1.05129635e+00
7.64059484e-01 2.60300428e-01 1.21519439e-01 -2.40504801e-01
-8.66085365e-02 -4.62808162e-01 -9.42274854e-02 6.07542396e-02
1.08226284e-01 -5.45559943e-01 -9.72977161e-01 1.82454658e+00
2.77606457e-01 5.47122061e-01 1.98314458e-01 6.83752477e-01
6.59970194e-02 5.18780470e-01 4.06513333e-01 -1.52299494e-01
1.01326871e+00 -1.03769410e+00 -7.26304770e-01 -5.90499401e-01
5.73858619e-01 -2.51573265e-01 1.47049260e+00 4.68158424e-01
-1.02795064e+00 -6.37890995e-01 -1.02723575e+00 3.19534063e-01
-3.01880151e-01 1.07691072e-01 8.20544899e-01 3.76076847e-01
-8.50598276e-01 6.05763197e-01 -8.07795823e-01 5.19453101e-02
4.48588639e-01 2.00529814e-01 8.26564282e-02 -1.35989532e-01
-1.27930963e+00 7.51668811e-01 7.05216825e-01 7.57238865e-02
-1.54648387e+00 -5.94859183e-01 -8.65817130e-01 9.66222584e-02
7.86425591e-01 -4.92477179e-01 1.83878732e+00 -1.36884129e+00
-1.83679938e+00 2.69596487e-01 6.16386980e-02 -6.52197659e-01
3.84117365e-01 -2.87678361e-01 -2.42851108e-01 5.54512180e-02
-5.92052862e-02 5.25256693e-01 1.20317507e+00 -1.27290154e+00
-8.71220589e-01 -9.48066488e-02 1.30159616e-01 3.74377549e-01
-1.70737222e-01 -2.76415884e-01 -9.57579166e-02 -8.57095122e-01
-5.26753545e-01 -9.42307770e-01 -1.96089387e-01 -6.66204333e-01
4.90963273e-02 -5.23064792e-01 6.39963865e-01 -4.12527025e-01
1.13090742e+00 -2.07890487e+00 6.03313744e-02 -1.35932432e-03
-1.89930469e-01 2.93841749e-01 -4.89650548e-01 4.39995497e-01
-3.40955965e-02 -1.01463877e-01 -2.44718313e-01 -2.43738383e-01
2.51309853e-02 2.74774909e-01 -6.43093824e-01 -1.25940233e-01
4.20337141e-01 9.95185614e-01 -1.24622202e+00 -8.65384787e-02
-5.44037186e-02 1.19498491e-01 -6.45077705e-01 6.11922204e-01
-8.37668419e-01 5.63870192e-01 -6.37630820e-01 7.92036414e-01
1.44971967e-01 -1.94503859e-01 6.58006549e-01 2.84011424e-01
2.27275863e-01 3.82930994e-01 -7.91301429e-01 1.86692059e+00
-9.24770772e-01 4.51770425e-01 -2.67543018e-01 -1.12951970e+00
9.98294473e-01 2.48917937e-01 1.91482663e-01 -9.72719371e-01
1.25970189e-02 3.91500518e-02 2.27089170e-02 -6.41185164e-01
4.98083621e-01 2.40001991e-01 -9.69107449e-02 5.58010697e-01
2.42718801e-01 1.04636051e-01 1.02120914e-01 -1.31020829e-01
1.17188728e+00 6.18417740e-01 2.30482310e-01 -5.34676164e-02
2.35996172e-01 8.16326886e-02 6.28288865e-01 9.60732639e-01
-4.66000795e-01 -2.36010291e-02 2.78766900e-01 -3.82600754e-01
-7.89961100e-01 -9.86418307e-01 1.66214302e-01 1.64641345e+00
-1.79893449e-01 -1.43957317e-01 -5.36037326e-01 -1.17348266e+00
3.33752424e-01 1.12167823e+00 -7.76292205e-01 -6.01980567e-01
-7.54818380e-01 -4.41856652e-01 2.02868432e-01 7.05401421e-01
5.14041007e-01 -1.48266387e+00 -8.50633025e-01 4.18372810e-01
-9.21801329e-02 -7.79521585e-01 -5.56176841e-01 5.20526588e-01
-1.17839122e+00 -1.07058680e+00 -4.72567946e-01 -7.12409258e-01
6.72390103e-01 4.23348211e-02 1.00655723e+00 -1.45508692e-01
-1.97676644e-01 7.07821071e-01 -4.16527838e-01 -6.39306307e-01
-3.36406708e-01 1.97532237e-01 1.33391753e-01 -1.88785017e-01
3.56111586e-01 -3.19262981e-01 -6.92971110e-01 1.81285933e-01
-8.03036451e-01 -1.52661532e-01 7.34459519e-01 1.15611291e+00
5.51486492e-01 -2.35837460e-01 1.36040354e+00 -1.09677863e+00
1.02704597e+00 -8.42307806e-01 -7.90648937e-01 5.05886853e-01
-1.03751707e+00 2.76564837e-01 6.35516405e-01 -8.21615577e-01
-1.40156126e+00 -1.52466372e-01 2.90749758e-01 -6.17214322e-01
-1.10659532e-01 6.26649141e-01 3.16487104e-01 3.51844758e-01
8.05283666e-01 3.74225736e-01 -8.87036547e-02 -3.52439195e-01
3.37347478e-01 6.94766879e-01 3.26640129e-01 -1.26598430e+00
3.90647799e-01 -5.65584796e-03 -3.07108551e-01 -1.61726058e-01
-8.52779984e-01 3.68614390e-04 -2.65849531e-01 -2.24861413e-01
5.02504230e-01 -1.04880881e+00 -6.18047655e-01 1.31299093e-01
-5.87528586e-01 -1.47776306e+00 -5.03541112e-01 3.83268118e-01
-9.87186909e-01 -2.55628079e-01 -4.67992812e-01 -9.39512789e-01
-2.51099259e-01 -1.25386035e+00 8.63111854e-01 3.44334602e-01
1.38738394e-01 -1.16747320e+00 1.99321985e-01 2.27377027e-01
5.53876460e-01 -1.04385547e-01 8.97743583e-01 -7.40130246e-01
-5.81723213e-01 1.79614976e-01 1.57566443e-01 3.26436669e-01
1.83669254e-01 -4.13158983e-01 -1.12457454e+00 -6.16944432e-01
-4.78872769e-02 -8.29799771e-01 6.64830506e-01 2.73978591e-01
1.69058609e+00 -3.69193554e-01 -1.12430476e-01 4.48188275e-01
1.32525563e+00 5.50615132e-01 3.91438484e-01 7.01695383e-01
2.24895671e-01 3.28989357e-01 1.24284720e+00 6.79961562e-01
4.45637941e-01 4.95476007e-01 3.53986651e-01 5.48952401e-01
-7.28129596e-02 -5.29465854e-01 7.46526957e-01 4.39013153e-01
7.09984004e-02 1.77958235e-02 -8.38977456e-01 5.82187295e-01
-2.03266883e+00 -8.05015326e-01 9.19906139e-01 2.24262929e+00
1.33440828e+00 1.38138115e-01 8.45348462e-02 -5.05108178e-01
5.21439075e-01 -9.95652378e-02 -1.13518357e+00 -5.83023012e-01
5.21286428e-01 3.47512901e-01 2.95561403e-01 3.15380126e-01
-8.72125328e-01 1.09307432e+00 6.32756138e+00 6.38967216e-01
-1.16153061e+00 2.58517474e-01 6.47860646e-01 -3.41227829e-01
-1.44809201e-01 -8.87031183e-02 -8.01091075e-01 5.48732460e-01
1.18134558e+00 -2.83664286e-01 8.52239847e-01 1.05288506e+00
8.93570855e-02 -5.97537146e-04 -1.32284117e+00 7.06326425e-01
-1.73850507e-01 -1.26190829e+00 -1.16187930e-01 -1.44390613e-01
1.02150536e+00 1.27784073e-01 3.27849567e-01 1.22240841e+00
5.39748847e-01 -8.62361073e-01 6.34383321e-01 5.88756979e-01
8.53546202e-01 -7.79505849e-01 4.77767527e-01 3.56697708e-01
-8.74276817e-01 -6.71071529e-01 -4.04432505e-01 5.22663333e-02
-5.38020074e-01 5.26739769e-02 -1.28102481e+00 4.21810806e-01
5.15754998e-01 5.43580234e-01 -5.99126220e-01 6.42636597e-01
-4.35615033e-01 9.41102087e-01 2.82609373e-01 -8.58599767e-02
1.64200485e-01 7.21611306e-02 3.11716914e-01 1.08355975e+00
2.15182781e-01 -4.73226383e-02 4.93198782e-01 9.69902039e-01
-1.13929145e-01 -2.32311651e-01 -7.49942064e-01 -1.08606108e-01
8.42610776e-01 9.72304583e-01 -9.10370946e-02 -3.40028971e-01
-1.97431505e-01 8.34728062e-01 7.76214421e-01 7.51111507e-01
-7.57961690e-01 -2.77339250e-01 6.21805549e-01 -3.36924225e-01
2.96568274e-01 -3.19285870e-01 9.18034557e-03 -1.03412199e+00
-1.06479838e-01 -1.16446495e+00 5.81612468e-01 -6.56566918e-01
-1.26414919e+00 2.95249194e-01 3.57779622e-01 -1.12685502e+00
-7.65857518e-01 -5.68511426e-01 -6.95958018e-01 7.98666894e-01
-1.87027764e+00 -7.89264202e-01 -7.99866989e-02 5.84701300e-01
1.12433302e+00 -5.09303093e-01 5.73320866e-01 -1.76851138e-01
-6.73387468e-01 8.67524028e-01 6.97664618e-02 -1.71912953e-01
8.33076656e-01 -1.33331680e+00 2.61600643e-01 4.72958952e-01
-1.41085580e-01 6.23316765e-01 4.56476241e-01 -7.81177700e-01
-1.74678159e+00 -1.31604588e+00 2.53015757e-01 -4.46287572e-01
6.15254283e-01 -2.89428145e-01 -8.76194656e-01 1.06814277e+00
3.26098323e-01 -3.34563330e-02 6.34869576e-01 1.93098545e-01
-3.02729428e-01 -1.29181609e-01 -1.16511393e+00 7.37729311e-01
1.03831077e+00 -5.47114968e-01 -7.40759194e-01 4.09146160e-01
9.41575766e-01 -5.37221014e-01 -8.61102939e-01 2.52057582e-01
2.92412758e-01 -4.29809719e-01 7.65516281e-01 -9.04232979e-01
4.81975585e-01 -6.04509152e-02 -1.48879096e-01 -1.69637012e+00
-3.19049209e-01 -5.94429970e-01 -6.02112114e-01 8.69212687e-01
3.67838919e-01 -9.47358727e-01 5.63128471e-01 3.69596422e-01
-2.75677890e-01 -8.88031363e-01 -6.93670094e-01 -1.05357003e+00
1.65912583e-01 -2.83991873e-01 8.16153109e-01 1.00916505e+00
8.81031528e-02 2.79596914e-02 -2.85964105e-02 8.47845674e-02
5.46819985e-01 2.35751003e-01 6.84641957e-01 -6.25538349e-01
-7.40292013e-01 -1.50377408e-01 2.75717497e-01 -7.20340610e-01
5.18317521e-01 -1.02876639e+00 2.11178064e-01 -1.01972222e+00
2.20051296e-02 -7.15319753e-01 -8.17216218e-01 8.70604932e-01
-4.22740191e-01 -3.93141657e-01 2.57214367e-01 1.07784592e-01
-6.93228841e-01 7.59658337e-01 1.44374716e+00 -9.97493640e-02
-4.22916532e-01 1.14282399e-01 -8.37704122e-01 2.51739562e-01
1.36676610e+00 -4.21966523e-01 -8.65941703e-01 -5.56540191e-01
1.60945445e-01 1.02074362e-01 2.53235877e-01 -9.71206486e-01
-2.24712044e-02 -5.79710305e-01 4.98862863e-01 -1.17277183e-01
1.09250255e-01 -7.40079165e-01 -3.26358140e-01 5.12315512e-01
-8.77454519e-01 2.59572476e-01 5.08043230e-01 8.19010556e-01
4.43009958e-02 -5.06963432e-01 2.85300195e-01 -2.49106511e-01
-9.91212606e-01 2.24156618e-01 -2.80307204e-01 2.15497494e-01
1.11622143e+00 2.84244537e-01 -5.85054040e-01 -2.24046096e-01
-7.24640667e-01 5.36365032e-01 9.72783640e-02 6.52738631e-01
7.35605896e-01 -1.29555213e+00 -6.28584385e-01 2.04887852e-01
3.30522090e-01 -5.21935038e-02 1.39673173e-01 3.59965086e-01
5.07329293e-02 3.11812252e-01 -3.01839858e-01 -3.88516307e-01
-6.10602200e-01 6.48365378e-01 3.99340689e-01 -5.37677944e-01
-4.98035431e-01 5.76784074e-01 8.09845328e-03 -6.17068350e-01
3.51092368e-01 -4.41248655e-01 -7.07732141e-02 -2.84753710e-01
5.37552655e-01 1.41318038e-01 4.67415974e-02 3.89231026e-01
2.26417799e-02 -2.06341401e-01 -4.45286542e-01 -2.52602100e-01
1.14268613e+00 5.74401394e-03 5.25265276e-01 5.62532008e-01
7.65121698e-01 -5.29604733e-01 -2.03039861e+00 -3.31898302e-01
3.39559406e-01 -5.15101194e-01 -7.15586171e-03 -1.31584156e+00
-7.83273280e-01 6.21648371e-01 7.63306797e-01 -9.31683630e-02
1.10058260e+00 -2.90247321e-01 4.88315344e-01 8.30044091e-01
6.35306776e-01 -1.58940113e+00 6.32999122e-01 7.14296758e-01
1.09984553e+00 -1.26542270e+00 -1.40923619e-01 2.68191338e-01
-9.80430305e-01 1.01806009e+00 1.04243505e+00 -1.79352671e-01
4.12161708e-01 -5.70253842e-02 -9.10063088e-02 -1.76732969e-02
-1.37520099e+00 -1.61943302e-01 4.04842272e-02 8.00960839e-01
2.59312868e-01 2.21217141e-01 1.54942870e-01 5.93276799e-01
1.47425413e-01 3.25987846e-01 3.98861736e-01 1.39051223e+00
-4.38620210e-01 -1.34007740e+00 -2.47330442e-01 5.47291100e-01
-2.57088810e-01 -3.09630204e-02 -4.51754890e-02 5.90598226e-01
-2.30527431e-01 7.83231020e-01 1.14277214e-01 -3.04653436e-01
3.61673385e-01 4.74299699e-01 7.72103906e-01 -9.06705141e-01
-7.75925696e-01 -1.49079010e-01 3.73954065e-02 -6.79699063e-01
-6.46750107e-02 -6.93077028e-01 -1.30269492e+00 2.86182493e-01
-2.67722048e-02 7.27551803e-02 7.44898379e-01 7.94686317e-01
6.05286241e-01 8.92185628e-01 1.05748570e+00 -8.02315891e-01
-1.24963963e+00 -1.01186085e+00 -2.82579482e-01 2.84502983e-01
3.36181879e-01 -8.72612059e-01 -1.46823183e-01 3.33489887e-02] | [4.1221208572387695, 1.9004857540130615] |
8eb384da-f52e-4ebc-b40a-b4dbc3320425 | a-fast-and-robust-camera-imu-online | 2305.08247 | null | https://arxiv.org/abs/2305.08247v1 | https://arxiv.org/pdf/2305.08247v1.pdf | A Fast and Robust Camera-IMU Online Calibration Method For Localization System | Autonomous driving has spurred the development of sensor fusion techniques, which combine data from multiple sensors to improve system performance. In particular, localization system based on sensor fusion , such as Visual Simultaneous Localization and Mapping (VSLAM), is an important component in environment perception, and is the basis of decision-making and motion control for intelligent vehicles. The accuracy of extrinsic calibration parameters between camera and IMU has significant effect on the positioning precision when performing VSLAM system. Currently, existing methods are time-consuming using complex optimization methods and sensitive to noise and outliers due to off-calibration, which can negatively impact system performance. To address these problems, this paper presents a fast and robust camera-IMU online calibration method based space coordinate transformation constraints and SVD (singular Value Decomposition) tricks. First, constraint equations are constructed based on equality of rotation and transformation matrices between camera frames and IMU coordinates at different moments. Secondly, the external parameters of the camera-IMU are solved using quaternion transformation and SVD techniques. Finally, the proposed method is validated using ROS platform, where images from the camera and velocity, acceleration, and angular velocity data from the IMU are recorded in a ROS bag file. The results showed that the proposed method can achieve robust and reliable camera-IMU online calibration parameters results with less tune consuming and less uncertainty. | ['Jian Zhao', 'Bing Zhu', 'Pengxiang Meng', 'Xiaowen Tao'] | 2023-05-14 | null | null | null | null | ['simultaneous-localization-and-mapping'] | ['computer-vision'] | [-3.51116002e-01 -7.48355210e-01 -8.36604908e-02 -1.97537005e-01
-2.69236296e-01 -6.16415083e-01 4.94472474e-01 2.81680422e-03
-5.70581973e-01 4.75843310e-01 -2.77251571e-01 -5.49795777e-02
1.00115418e-01 -4.62187916e-01 -7.97032475e-01 -6.96493447e-01
4.99497473e-01 4.94178608e-02 3.56035858e-01 -3.55835706e-01
4.41636205e-01 6.53018773e-01 -1.49434185e+00 -7.99583972e-01
7.99262822e-01 9.79783118e-01 4.87866551e-01 6.17386818e-01
1.75272316e-01 4.24164206e-01 -4.96068329e-01 1.58334267e-03
3.24350506e-01 -4.02544159e-04 2.52273053e-01 8.47115964e-02
9.34505463e-02 -3.79588425e-01 -5.16088665e-01 1.42389953e+00
5.50460696e-01 1.18365087e-01 2.13006407e-01 -1.50744224e+00
-3.17062438e-01 -1.94505364e-01 -6.55997694e-01 -5.27316481e-02
3.87787431e-01 -2.92742252e-03 1.87582493e-01 -9.10129666e-01
6.14544809e-01 1.14143813e+00 7.59788573e-01 -1.63047969e-01
-8.14917684e-01 -8.00078988e-01 -3.42039883e-01 6.84819400e-01
-1.66137969e+00 -4.18396741e-01 7.21861362e-01 -5.98434329e-01
6.20346069e-01 1.70081481e-01 6.26563668e-01 5.90899646e-01
5.53071618e-01 1.62871614e-01 7.68092573e-01 -2.16697127e-01
2.85158828e-02 3.04577351e-01 1.85551196e-01 3.76776904e-01
9.32886660e-01 6.47603273e-02 -2.99223483e-01 1.71901226e-01
7.09783256e-01 4.78933334e-01 -2.06991971e-01 -6.32462621e-01
-1.54096699e+00 5.93552828e-01 3.90524358e-01 -1.87922075e-01
-2.02861845e-01 1.12160079e-01 2.03710258e-01 6.15624972e-02
-3.47279251e-01 -4.48988229e-02 -6.05577417e-02 -3.88084859e-01
-3.51126492e-01 3.36592048e-02 3.49126965e-01 1.44350088e+00
1.17365384e+00 2.53726274e-01 5.96635938e-01 4.42694694e-01
7.77813137e-01 1.38502395e+00 7.65765727e-01 -8.29955220e-01
6.70724928e-01 6.31086409e-01 4.49928224e-01 -1.64796257e+00
-5.59570730e-01 -1.39037175e-02 -6.20416164e-01 1.97652668e-01
-1.73454568e-01 -2.99472272e-01 -5.40436029e-01 1.17532909e+00
6.00473702e-01 4.24208343e-01 2.93332756e-01 1.14496565e+00
6.75507367e-01 6.29736304e-01 -4.87232298e-01 -2.78448761e-01
1.13147628e+00 -4.07730758e-01 -1.28954268e+00 -4.32154089e-01
5.42417049e-01 -1.18999159e+00 4.77655381e-01 1.82991222e-01
-3.77361715e-01 -9.58763957e-01 -1.86688435e+00 -4.20750454e-02
-4.15109009e-01 4.63284075e-01 1.70732528e-01 4.42772090e-01
-6.28122389e-01 7.48280436e-02 -9.33193207e-01 -4.04294580e-01
-4.44481194e-01 3.91435355e-01 -6.60329401e-01 4.31096442e-02
-1.24312484e+00 1.42167401e+00 5.03854156e-01 4.92504895e-01
-2.13735446e-01 -7.49192834e-02 -9.97472465e-01 -5.14662445e-01
3.10083151e-01 -2.06199840e-01 1.01330340e+00 -4.84784842e-01
-1.75349510e+00 1.98935583e-01 -3.06065828e-01 -3.73290181e-01
2.75000781e-01 -3.17187339e-01 -7.71069765e-01 -1.65060908e-01
1.18846245e-01 2.96714693e-01 8.04573596e-01 -1.09219837e+00
-8.09264660e-01 -4.38915342e-01 -4.83936161e-01 4.33988124e-01
2.54143864e-01 -2.38011479e-01 -5.77256739e-01 2.93700285e-02
8.35979760e-01 -1.31813610e+00 -4.54797596e-02 -4.78948534e-01
3.13983187e-02 3.06877732e-01 9.92749035e-01 -6.27894521e-01
1.01451492e+00 -2.27908516e+00 -4.67202701e-02 3.88493210e-01
-9.39851701e-02 2.54332513e-01 4.47428167e-01 1.44890413e-01
6.82286471e-02 -4.59177554e-01 5.50514936e-01 6.51502386e-02
-1.80186063e-01 4.65102255e-01 -1.97803602e-01 9.45102036e-01
-2.59818345e-01 4.83004987e-01 -6.87887013e-01 -3.41231853e-01
8.17301035e-01 5.75138748e-01 1.50436722e-02 1.47302598e-01
4.57136482e-01 4.55617398e-01 -3.53240937e-01 5.57751238e-01
9.37438846e-01 4.04137164e-01 -1.54507414e-01 -7.50298381e-01
-6.06209755e-01 -1.21103570e-01 -1.96444595e+00 1.38761544e+00
-2.25081146e-01 8.75439823e-01 1.61160007e-01 -6.84778154e-01
1.27979755e+00 1.80616513e-01 3.51579010e-01 -7.00539768e-01
7.48971105e-01 3.15180838e-01 -1.87912732e-01 -6.48733377e-01
8.06418419e-01 4.90263194e-01 2.15940662e-02 -2.88985103e-01
-2.66466141e-01 -4.75345403e-01 1.07236721e-01 -4.55693342e-02
4.96511787e-01 3.81420791e-01 6.23820782e-01 9.45289060e-02
9.54699695e-01 2.01805159e-01 1.00441849e+00 9.72403120e-03
-3.14895689e-01 3.75530452e-01 -4.85880822e-02 -9.09090042e-02
-1.15391815e+00 -6.61417842e-01 -1.91119805e-01 5.23844250e-02
9.63014185e-01 -2.85179198e-01 -3.78157049e-01 3.58014315e-01
3.13463181e-01 2.64261782e-01 -6.52363449e-02 -2.77294576e-01
-6.06589735e-01 -3.82368863e-01 -1.82618573e-02 3.96496177e-01
7.66216159e-01 5.26748151e-02 -6.67304993e-01 2.73230672e-01
-5.97975142e-02 -1.39994550e+00 -2.52086759e-01 -2.80722499e-01
-7.98080921e-01 -1.12156808e+00 -2.52725273e-01 -4.84287173e-01
8.17718506e-01 1.05857658e+00 1.72199368e-01 -2.04324618e-01
2.20553149e-02 4.30910587e-01 -3.42432201e-01 -7.38579094e-01
-4.80662510e-02 -4.88065600e-01 6.49591506e-01 2.61117578e-01
3.75696480e-01 -1.83387309e-01 -3.15687358e-01 8.18321943e-01
-5.15066087e-01 5.81516139e-02 4.89903867e-01 5.30098915e-01
8.32431793e-01 -1.59860104e-01 6.43102229e-02 -1.49673820e-01
3.01163644e-01 -3.17539930e-01 -1.47295761e+00 -2.44592391e-02
-6.57777905e-01 -1.87911034e-01 4.40713346e-01 -2.80686170e-01
-8.05516720e-01 5.62291563e-01 3.94480258e-01 -6.35705113e-01
1.96213052e-01 5.81433356e-01 -3.73476714e-01 -4.92635190e-01
6.15045965e-01 7.24599212e-02 5.58527708e-01 -1.63169995e-01
3.39202017e-01 1.02789712e+00 9.12250161e-01 7.00601488e-02
1.04638946e+00 4.05082881e-01 3.77920181e-01 -1.04224741e+00
-5.70037365e-02 -8.03599536e-01 -8.07934284e-01 -4.89248544e-01
9.18035328e-01 -1.53149152e+00 -8.25241625e-01 4.39009458e-01
-1.14707053e+00 6.37642860e-01 4.85047519e-01 1.21484947e+00
-2.60096043e-01 6.53037488e-01 8.64202827e-02 -7.38653004e-01
6.03370816e-02 -1.66615891e+00 7.12776482e-01 6.78086579e-01
1.50913537e-01 -5.40240824e-01 5.57022877e-02 2.22272262e-01
1.89821139e-01 4.64988440e-01 -1.95422798e-01 1.18094482e-01
-9.56462145e-01 -8.04389000e-01 2.11778954e-02 3.38832885e-01
4.70706224e-02 3.31067353e-01 -4.32549000e-01 -4.39039022e-01
3.01765710e-01 3.26828927e-01 1.29510239e-01 1.23814708e-02
1.65270075e-01 7.25850984e-02 -5.07556677e-01 9.48857069e-01
1.63690400e+00 5.70656061e-01 4.96998549e-01 7.77380228e-01
9.80729401e-01 9.62523669e-02 1.17406821e+00 2.60999680e-01
7.62609422e-01 7.81907618e-01 5.75718462e-01 2.64712185e-01
3.04950505e-01 -4.26691212e-02 5.73211551e-01 1.14599526e+00
-1.26812384e-01 4.33666527e-01 -7.57571399e-01 2.91231155e-01
-2.17811561e+00 -5.92509508e-01 -6.52847648e-01 2.39062333e+00
2.25018725e-01 3.06611527e-02 -3.54528010e-01 2.38003626e-01
9.56834018e-01 -1.21881612e-01 -5.15321195e-01 -1.83456898e-01
-8.91852379e-02 -5.89193940e-01 1.44774127e+00 6.99476540e-01
-8.30490351e-01 6.33652866e-01 5.18184566e+00 1.60158142e-01
-1.55542088e+00 2.07147226e-02 -4.29334641e-01 2.57821351e-01
1.65905401e-01 3.60998094e-01 -1.20200419e+00 6.57933533e-01
8.98807466e-01 -3.54855090e-01 4.12234962e-01 1.09465182e+00
3.54087800e-01 -5.34268618e-01 -6.78424835e-01 1.68438995e+00
1.65044755e-01 -1.18125474e+00 -5.95449030e-01 -3.36170122e-02
6.89916551e-01 3.34229708e-01 -2.05170110e-01 -1.50586054e-01
-7.23959357e-02 -2.67701715e-01 7.63806760e-01 5.05390227e-01
5.46303630e-01 -7.82326937e-01 1.05036724e+00 3.53501022e-01
-1.38338983e+00 -2.43591405e-02 -7.15261221e-01 -2.40122020e-01
2.34137326e-01 4.26664323e-01 -9.47466969e-01 7.94382930e-01
5.31819761e-01 9.32895124e-01 -6.28947914e-01 9.64547813e-01
-2.42885500e-01 6.72249794e-02 -6.37861133e-01 -2.02011257e-01
-1.50395587e-01 -7.49997020e-01 5.79510510e-01 6.71721876e-01
7.81263232e-01 5.02284318e-02 1.66151971e-01 2.71801203e-01
5.51738262e-01 5.85826114e-02 -7.30101347e-01 2.77081966e-01
7.84498990e-01 1.33944798e+00 -2.93932021e-01 -3.60119134e-01
-5.50251484e-01 3.82409781e-01 -2.39309803e-01 2.37575024e-01
-1.04356897e+00 -5.79335630e-01 7.79335141e-01 -1.15768619e-01
4.05180417e-02 -1.05379689e+00 -2.83349752e-01 -1.13815689e+00
1.43899515e-01 -5.13549030e-01 -2.12149590e-01 -1.00659192e+00
-2.37213582e-01 1.67116687e-01 -5.65414093e-02 -1.96582663e+00
-3.93213123e-01 -6.28804326e-01 -3.03598404e-01 8.81802022e-01
-1.35719311e+00 -8.20031047e-01 -7.85285771e-01 7.22991288e-01
5.55604063e-02 -3.18066686e-01 3.52983624e-01 4.53704327e-01
-7.66030669e-01 2.10261554e-01 5.61938941e-01 -1.36012748e-01
9.04607236e-01 -6.92587793e-01 -6.50790483e-02 1.28103161e+00
-2.43171528e-01 6.23971999e-01 9.98827457e-01 -7.26183891e-01
-2.16336370e+00 -9.05827940e-01 6.48214459e-01 -4.94174182e-01
6.15993738e-01 -5.44728898e-02 -4.22730744e-01 9.96437132e-01
-9.99547988e-02 1.37308389e-01 7.91598111e-02 -6.29846811e-01
2.69155443e-01 -7.38694847e-01 -9.02032912e-01 4.29908365e-01
2.24211156e-01 -4.24707413e-01 -5.12784719e-01 1.78644195e-01
6.21282995e-01 -1.13819051e+00 -8.22511792e-01 3.12369108e-01
5.13700128e-01 -6.76119506e-01 1.02112281e+00 4.23411101e-01
-5.64003408e-01 -1.27396393e+00 -4.07653540e-01 -1.08610630e+00
-1.73808768e-01 -5.85682929e-01 1.19785123e-01 1.13292193e+00
-6.52636886e-02 -1.00889671e+00 3.69353294e-01 4.52486396e-01
-1.22504622e-01 4.81867120e-02 -9.44412649e-01 -6.25726342e-01
-8.99704218e-01 -4.68456060e-01 6.85049891e-01 7.30641544e-01
-7.17855915e-02 3.99052620e-01 -5.27177215e-01 8.11848640e-01
5.97022772e-01 -4.04636353e-01 1.61152387e+00 -1.03694034e+00
2.74406195e-01 3.92723717e-02 -1.08747852e+00 -9.85522807e-01
-2.62356102e-01 -3.56738508e-01 1.36239618e-01 -1.31196165e+00
-5.47414362e-01 -7.23935589e-02 9.93639529e-02 -1.80638552e-01
-1.07302509e-01 1.47415958e-02 2.71808654e-01 6.59681857e-01
-3.06243032e-01 3.75599623e-01 9.19457614e-01 7.05097914e-02
-2.09792912e-01 6.98657259e-02 -3.95963229e-02 7.23107159e-01
5.78237236e-01 -1.71243116e-01 -4.44118023e-01 -6.29302621e-01
3.72028202e-01 2.34113872e-01 2.03229949e-01 -1.41136384e+00
8.02471101e-01 -1.46069616e-01 5.36095202e-01 -1.08390164e+00
4.06369299e-01 -1.29242921e+00 7.19716728e-01 4.52694178e-01
5.82150280e-01 6.71332598e-01 1.05356285e-03 4.36929882e-01
-5.50767601e-01 -1.20974518e-01 5.40593028e-01 2.69856304e-01
-1.14929461e+00 1.32684022e-01 -2.90754080e-01 -7.30178416e-01
1.43046248e+00 -6.90280020e-01 -2.66582519e-01 -2.61545181e-01
-2.09088117e-01 4.16705489e-01 5.82520902e-01 5.86955667e-01
5.73867261e-01 -1.79803836e+00 -2.45646194e-01 4.90700781e-01
2.43199781e-01 4.18527573e-01 1.70478344e-01 1.05166519e+00
-1.06273353e+00 5.65330446e-01 -3.39995086e-01 -1.15976024e+00
-1.31128013e+00 4.36536789e-01 2.20169827e-01 6.90028071e-01
-2.31874526e-01 1.26276195e-01 -4.92014110e-01 -4.52927768e-01
6.44230172e-02 -5.31004429e-01 -3.56914729e-01 7.68187791e-02
3.73956591e-01 7.06217468e-01 1.10348366e-01 -1.40313995e+00
-5.26478052e-01 1.24764550e+00 4.15704966e-01 -2.07249805e-01
6.56426013e-01 -8.77303243e-01 -8.86635557e-02 4.24802184e-01
1.32318056e+00 6.52252510e-02 -1.21202052e+00 -7.43954852e-02
-1.59518003e-01 -6.55091345e-01 2.38858715e-01 -1.03994846e-01
-7.30495334e-01 6.63528740e-01 1.02642238e+00 -7.21666068e-02
6.38907194e-01 -7.78765738e-01 6.66761279e-01 6.89008892e-01
5.41175544e-01 -1.44108486e+00 -4.70873922e-01 6.40484452e-01
6.97035968e-01 -1.56508386e+00 4.45336282e-01 -2.95946509e-01
-5.34622014e-01 1.21340644e+00 6.56445086e-01 -2.35793114e-01
6.44279063e-01 7.28794113e-02 4.17947263e-01 4.26342934e-01
-8.92130807e-02 -2.43583252e-03 6.27251938e-02 6.18745387e-01
-1.75006799e-02 1.24160610e-01 -3.00138861e-01 2.59798229e-01
-2.48129129e-01 -1.14837520e-01 7.34997749e-01 1.02302277e+00
-8.12205374e-01 -8.87883663e-01 -1.23782003e+00 -2.26596296e-01
1.42263338e-01 5.85078657e-01 1.37121722e-01 9.02794063e-01
3.64276260e-01 9.87956166e-01 1.30796535e-02 -8.16396058e-01
6.78875029e-01 -3.27431887e-01 2.28977740e-01 8.28561094e-03
1.44629076e-01 1.78550646e-01 -2.54297048e-01 -7.34998107e-01
-3.77199203e-01 -7.90881038e-01 -1.55513370e+00 -3.57189476e-01
-5.94896913e-01 1.81017979e-03 1.50016367e+00 8.33032072e-01
4.52082515e-01 2.24998415e-01 9.40973103e-01 -1.03700781e+00
-5.03978789e-01 -7.69195437e-01 -3.55134130e-01 2.19927669e-01
4.74283844e-01 -9.70077634e-01 -3.74831766e-01 -2.96480842e-02] | [7.514829158782959, -1.9747815132141113] |
6b959daa-c391-4e4e-98a2-a9e34556eaff | one-shot-detail-retouching-with-patch-space | 2210.01217 | null | https://arxiv.org/abs/2210.01217v3 | https://arxiv.org/pdf/2210.01217v3.pdf | One-shot Detail Retouching with Patch Space Neural Transformation Blending | Photo retouching is a difficult task for novice users as it requires expert knowledge and advanced tools. Photographers often spend a great deal of time generating high-quality retouched photos with intricate details. In this paper, we introduce a one-shot learning based technique to automatically retouch details of an input image based on just a single pair of before and after example images. Our approach provides accurate and generalizable detail edit transfer to new images. We achieve these by proposing a new representation for image to image maps. Specifically, we propose neural field based transformation blending in the patch space for defining patch to patch transformations for each frequency band. This parametrization of the map with anchor transformations and associated weights, and spatio-spectral localized patches, allows us to capture details well while staying generalizable. We evaluate our technique both on known ground truth filters and artist retouching edits. Our method accurately transfers complex detail retouching edits. | ['Cengiz Oztireli', 'Fazilet Gokbudak'] | 2022-10-03 | null | null | null | null | ['photo-retouching', 'one-shot-learning'] | ['computer-vision', 'methodology'] | [ 6.12753570e-01 1.15409821e-01 2.73091167e-01 -2.20443070e-01
-7.62823761e-01 -6.82081342e-01 7.21663952e-01 7.91955516e-02
-1.30187526e-01 6.42385900e-01 1.26270562e-01 3.33646595e-01
-8.42706934e-02 -7.47018933e-01 -1.07480228e+00 -3.40106726e-01
3.45499426e-01 -4.23500910e-02 5.08176625e-01 -2.03039721e-01
3.97309661e-01 6.31543875e-01 -1.59692919e+00 1.69691101e-01
8.49380851e-01 6.05412543e-01 4.30227995e-01 1.01335442e+00
-9.36444178e-02 6.17399752e-01 -5.03318489e-01 -7.13655710e-01
6.39552951e-01 -4.04735744e-01 -5.50195634e-01 3.37547809e-01
1.31291223e+00 -4.34226424e-01 -1.93441704e-01 1.04517937e+00
4.67848152e-01 5.40929317e-01 7.40557492e-01 -8.26323688e-01
-1.22097635e+00 2.55416095e-01 -8.21032345e-01 1.31062359e-01
3.34757060e-01 3.48759115e-01 7.88610995e-01 -8.32126081e-01
8.77053380e-01 1.00020051e+00 1.09567213e+00 4.18318748e-01
-1.83187413e+00 -5.66208541e-01 -1.25256538e-01 3.14540952e-01
-1.34942508e+00 -4.50863510e-01 1.00865448e+00 -7.60463178e-01
8.84428680e-01 3.45859051e-01 9.49921131e-01 8.28130484e-01
4.35266405e-01 1.41576141e-01 1.03379893e+00 -6.01527095e-01
2.76271552e-01 2.28883699e-01 -2.57238775e-01 5.23486853e-01
1.17071196e-01 2.99165666e-01 -5.42713523e-01 1.49025694e-01
1.26673329e+00 2.57602096e-01 -4.74289745e-01 -5.11082470e-01
-1.23501730e+00 5.11614561e-01 7.02488780e-01 2.83194989e-01
-3.03077519e-01 4.01663691e-01 -4.72100154e-02 2.11747110e-01
6.52249694e-01 7.89797544e-01 -1.22013442e-01 6.93776682e-02
-1.26916039e+00 1.66188776e-01 3.91375005e-01 9.28227782e-01
1.34923327e+00 -9.85875651e-02 -4.98699009e-01 9.58736360e-01
-4.21145648e-01 4.55747873e-01 1.10359676e-01 -1.14789820e+00
3.13949622e-02 2.79992104e-01 2.87245423e-01 -1.36502969e+00
-4.42464799e-02 -8.28828067e-02 -9.94174242e-01 6.53582573e-01
7.49157518e-02 -1.10910773e-01 -1.12137830e+00 1.32667458e+00
2.50317395e-01 4.33698624e-01 -3.89359683e-01 8.93735886e-01
4.87306774e-01 7.22287357e-01 -7.29678050e-02 4.06332836e-02
1.25488353e+00 -1.02551436e+00 -7.60740459e-01 -1.84214756e-01
-1.73041016e-01 -9.54958558e-01 1.44641364e+00 1.84212878e-01
-1.13415015e+00 -7.82327712e-01 -1.03090334e+00 -5.82582176e-01
-3.25313389e-01 -3.25253536e-03 3.29092741e-01 2.82855779e-01
-1.15309739e+00 8.79138470e-01 -6.16043031e-01 -5.68415046e-01
3.71185780e-01 4.50421050e-02 -4.78958398e-01 1.05285980e-01
-7.83289194e-01 9.62044239e-01 6.25293106e-02 -1.74900696e-01
-5.17707825e-01 -1.31735790e+00 -8.78436685e-01 1.70524463e-01
1.92154750e-01 -1.00068319e+00 1.24178636e+00 -1.26365745e+00
-1.72488332e+00 9.10508692e-01 3.32784802e-02 -3.21345896e-01
4.67054099e-01 -1.38696492e-01 -3.92835706e-01 2.05593586e-01
-8.25605616e-02 7.90274680e-01 1.52138507e+00 -1.44212008e+00
-5.01525044e-01 4.81885299e-02 7.38731325e-02 1.29183263e-01
-4.01851833e-01 -4.06333029e-01 -2.61927545e-01 -1.16522598e+00
-4.19159740e-01 -4.91539180e-01 -1.75092071e-01 6.51353538e-01
-2.63304204e-01 4.58799571e-01 7.48955131e-01 -8.44976902e-01
1.19537854e+00 -2.18763852e+00 1.82219028e-01 9.65114310e-02
4.25290257e-01 2.07208619e-01 -2.62625635e-01 5.55485547e-01
-9.84235108e-02 -1.15121901e-01 -5.42835236e-01 -1.99022129e-01
-2.22520009e-01 -3.39739434e-02 -4.19379056e-01 2.75830299e-01
3.01206410e-01 1.00184262e+00 -9.42532301e-01 -2.72936612e-01
4.68826443e-01 7.19514430e-01 -6.30606353e-01 3.94068211e-01
-1.70283109e-01 4.94171888e-01 2.02869341e-01 1.93915337e-01
7.90092349e-01 -1.54404029e-01 -3.22375834e-01 -7.69450366e-01
-2.10466623e-01 -4.61127043e-01 -1.02502143e+00 2.10395765e+00
-7.48151541e-01 1.04355025e+00 -2.55486786e-01 -4.53391016e-01
9.52951550e-01 -7.36143589e-02 2.05564395e-01 -5.17262101e-01
-4.11228463e-03 -7.10254461e-02 -5.99850655e-01 -6.27466798e-01
7.32793689e-01 -4.19434458e-01 2.16997430e-01 5.96196532e-01
2.54684180e-01 -6.53299570e-01 -1.42645314e-01 2.66811419e-02
9.38296258e-01 3.20915468e-02 5.90671837e-01 -1.81408644e-01
2.52729565e-01 -6.64342120e-02 1.31984979e-01 5.35760880e-01
2.60211110e-01 1.17351580e+00 -1.38087794e-01 -7.97244549e-01
-1.57003522e+00 -1.17761278e+00 4.35108058e-02 8.11706364e-01
3.54094714e-01 -2.58112192e-01 -9.24372792e-01 -2.01352984e-01
4.88019269e-03 8.13239336e-01 -8.54136348e-01 -1.79107592e-01
-5.08747399e-01 -3.39949876e-02 1.84567541e-01 3.54699194e-01
7.11987972e-01 -8.46942902e-01 -7.04481125e-01 1.07978083e-01
1.02143519e-01 -8.69203687e-01 -1.15419221e+00 -2.49682978e-01
-6.47526741e-01 -1.03602302e+00 -1.37388158e+00 -8.10837150e-01
8.89297962e-01 5.51103830e-01 1.00609481e+00 -1.07724831e-01
-8.45752239e-01 6.68763280e-01 -1.31162331e-01 -7.63434172e-02
-4.18624520e-01 -6.32314086e-02 -3.32460523e-01 3.97249818e-01
-1.92149550e-01 -1.06393099e+00 -7.90093720e-01 2.26698294e-01
-1.15068734e+00 5.00322044e-01 6.94570065e-01 7.17075944e-01
6.44768834e-01 5.06834723e-02 -1.09368928e-01 -8.28505158e-01
6.51060760e-01 1.36377409e-01 -6.11382663e-01 5.21190047e-01
-4.44899589e-01 2.15624154e-01 5.80832481e-01 -7.77959108e-01
-1.32030797e+00 1.66150346e-01 4.71269101e-01 -6.67467892e-01
-8.19195211e-02 -1.72918484e-01 2.29302928e-01 -6.17099047e-01
1.28796625e+00 -5.41595407e-02 -2.19678488e-02 -5.79369664e-01
8.99119973e-01 4.05771255e-01 9.08840299e-01 -3.17171872e-01
1.27669251e+00 6.32011354e-01 -2.36894712e-01 -8.73766601e-01
-5.43597162e-01 -1.61036909e-01 -8.68373990e-01 -4.55154806e-01
9.13262248e-01 -6.84411049e-01 -3.88267338e-01 3.36638868e-01
-1.24877977e+00 -5.60829639e-01 -8.01876962e-01 1.32889807e-01
-6.65737331e-01 4.89812851e-01 -3.63777041e-01 -3.65569532e-01
-4.00919259e-01 -7.06687808e-01 1.25913882e+00 3.34576398e-01
-3.43034863e-01 -8.91557276e-01 5.88518798e-01 -1.88193575e-01
6.29420280e-01 4.71875548e-01 7.12854505e-01 4.48742092e-01
-8.15969348e-01 -5.46836406e-02 -4.32060421e-01 3.76930416e-01
4.84619051e-01 1.87107950e-01 -1.11719203e+00 -2.38310263e-01
-3.02126110e-01 1.51804551e-01 7.98119187e-01 4.13351655e-01
1.14343822e+00 -5.43983817e-01 -2.07871974e-01 8.84715915e-01
1.56473148e+00 1.54247899e-02 8.10877442e-01 1.70463681e-01
9.20780599e-01 5.11274338e-01 1.36175722e-01 2.17562065e-01
1.08524188e-01 8.23222876e-01 -7.52943605e-02 -3.57842505e-01
-6.58461154e-01 -5.53818524e-01 5.03863990e-02 5.38209081e-01
-2.65759379e-01 -4.08831760e-02 -6.46140099e-01 5.80632865e-01
-1.74234593e+00 -1.12074673e+00 2.73019969e-01 2.23696518e+00
1.00439525e+00 -3.61157805e-01 -2.16775388e-01 -9.54606086e-02
9.95911241e-01 2.09626555e-01 -5.31932354e-01 -3.57926160e-01
1.38922753e-02 4.75615144e-01 6.77580237e-01 8.33861828e-01
-8.98426354e-01 1.00353241e+00 6.35884190e+00 9.33890820e-01
-1.18055475e+00 1.07908808e-01 3.77749771e-01 -1.16367564e-01
-5.88839054e-01 1.30245134e-01 -3.03102702e-01 3.78046185e-01
3.82087290e-01 -5.83463252e-01 9.78218973e-01 5.51308811e-01
1.11644700e-01 -2.03573510e-01 -1.06504834e+00 1.43145347e+00
4.57678676e-01 -1.83622968e+00 3.27083230e-01 -3.00160080e-01
1.17906427e+00 -6.30260766e-01 1.18839771e-01 -2.33191490e-01
2.23897636e-01 -9.15536344e-01 6.96434975e-01 1.10402262e+00
1.28127265e+00 -5.34294426e-01 8.61983933e-03 -6.68456852e-02
-1.14250124e+00 5.60381785e-02 -4.66789663e-01 7.73151889e-02
3.32960993e-01 6.06712580e-01 -7.49933362e-01 4.43201661e-01
7.03487098e-01 9.91738200e-01 -8.27816248e-01 1.28438056e+00
-2.11449251e-01 5.18081747e-02 -7.61332884e-02 2.89973140e-01
-3.20092469e-01 -1.48187354e-01 6.98702455e-01 1.13822711e+00
6.53882802e-01 1.64314017e-01 -1.90378115e-01 1.15822816e+00
-6.36894349e-03 -1.80153511e-02 -7.54871786e-01 1.62353009e-01
2.65548348e-01 1.39977133e+00 -6.37431085e-01 -3.65344316e-01
-2.40559895e-02 1.55475378e+00 2.54362196e-01 6.59099281e-01
-6.23564363e-01 -8.61460745e-01 5.16562164e-01 4.69137579e-01
3.09253544e-01 -1.36049896e-01 -2.94189602e-01 -1.23466456e+00
-1.99034512e-02 -4.59504485e-01 -1.21360973e-01 -1.33771682e+00
-1.51315498e+00 7.15298414e-01 6.34444654e-02 -1.47434831e+00
1.22172236e-01 -2.67147511e-01 -8.46781909e-01 8.17066610e-01
-1.29121625e+00 -1.31713057e+00 -9.05709445e-01 5.64161658e-01
6.77980363e-01 4.88335751e-02 6.63687587e-01 3.22194099e-01
-8.64322856e-02 5.82240760e-01 -1.11809768e-01 -1.68841451e-01
1.09344983e+00 -1.11710918e+00 7.11275816e-01 9.26111519e-01
3.60319406e-01 4.58914369e-01 9.21962142e-01 -6.09316170e-01
-8.86155963e-01 -1.11334729e+00 4.13920820e-01 -4.90761042e-01
4.17124629e-01 -4.16552246e-01 -1.04760420e+00 6.78777635e-01
4.93622124e-01 -1.86157927e-01 2.54356295e-01 -3.21352184e-01
-3.32619905e-01 -3.47302169e-01 -1.06139278e+00 7.77546048e-01
1.02146173e+00 -8.21411729e-01 -6.92257166e-01 2.25254074e-01
7.47153163e-01 -4.54054147e-01 -7.12254763e-01 -2.88644955e-02
6.25304878e-01 -1.10440361e+00 1.07429409e+00 -2.07682341e-01
4.41286623e-01 -5.56222975e-01 2.53785372e-01 -1.54470897e+00
-6.28561199e-01 -1.15194798e+00 1.80773005e-01 1.14637637e+00
1.69106096e-01 -4.66721743e-01 3.39204669e-01 6.46347046e-01
6.30354434e-02 -1.85957134e-01 -6.16905510e-01 -6.36271000e-01
-2.94041693e-01 8.09506141e-03 7.37404525e-01 1.11138070e+00
-2.40394324e-01 2.08719596e-01 -6.34232342e-01 7.85933435e-02
8.65859151e-01 -3.58950682e-02 1.09939158e+00 -9.90059018e-01
-5.31413913e-01 -4.08717960e-01 -6.13471508e-01 -6.58129513e-01
-1.81017652e-01 -6.48175478e-01 3.72768939e-02 -1.50335729e+00
2.39756972e-01 -9.89419371e-02 2.65283912e-01 3.83555889e-01
-1.90117508e-01 6.26769662e-01 1.41556486e-01 3.21273178e-01
-9.11232363e-03 5.03975749e-01 1.46064663e+00 -3.05362761e-01
-4.00364995e-01 -3.15213948e-01 -5.23890793e-01 5.65577149e-01
6.36354268e-01 -2.10374624e-01 -4.77967173e-01 -6.44016683e-01
2.23730698e-01 -4.15061057e-01 7.80705154e-01 -1.20006883e+00
3.07327956e-01 -1.20967999e-01 4.91603434e-01 -1.51289716e-01
3.69917423e-01 -8.41294050e-01 8.70233893e-01 -1.41684236e-02
-4.65593904e-01 -9.76870582e-02 3.14628452e-01 8.73941779e-01
-3.26755308e-02 -1.14054168e-02 1.05707669e+00 -5.77383861e-02
-8.46949875e-01 2.88769901e-01 5.24945483e-02 -1.76624358e-01
1.11316288e+00 -6.10222697e-01 -2.34974340e-01 -4.96777892e-01
-8.28653872e-01 -2.96213269e-01 9.65447903e-01 3.35521877e-01
6.11554384e-01 -1.43818438e+00 -3.87431532e-01 3.97738308e-01
1.02932848e-01 -1.94476515e-01 7.97364533e-01 5.61137855e-01
-9.46714520e-01 -1.60845265e-01 -6.51744843e-01 -4.27549869e-01
-1.12393045e+00 7.27050781e-01 4.53777730e-01 3.05838376e-01
-1.29915679e+00 8.63907576e-01 5.04360855e-01 -8.88919905e-02
6.76404014e-02 -4.92268592e-01 1.03059307e-01 -5.49220666e-02
7.47196317e-01 4.42841262e-01 -3.11691146e-02 -3.62138897e-01
2.03816056e-01 1.36303926e+00 9.80659425e-02 -3.10065925e-01
1.27270341e+00 -2.88342655e-01 6.57083616e-02 5.38370609e-01
1.17917609e+00 8.94531086e-02 -1.76625359e+00 -3.62815827e-01
-3.31980824e-01 -9.62584198e-01 1.35366507e-02 -7.80516744e-01
-9.49662030e-01 9.57765996e-01 7.06881523e-01 7.97218978e-02
1.16582274e+00 -1.60860375e-01 8.72908413e-01 2.68237203e-01
3.52985293e-01 -1.06484818e+00 2.47494802e-01 -2.28156909e-01
1.27998137e+00 -9.19785321e-01 1.52190417e-01 -3.71122628e-01
-5.68902016e-01 1.17677951e+00 3.45649451e-01 -5.88521600e-01
5.64781666e-01 3.72995399e-02 5.66291437e-02 -2.00891763e-01
-3.92035544e-01 -2.28897426e-02 6.07362747e-01 8.87309134e-01
-1.18343681e-01 -1.59665659e-01 1.17153049e-01 7.94575959e-02
-2.93975353e-01 -3.80413909e-03 6.48223400e-01 6.17953837e-01
-5.71610570e-01 -8.92274201e-01 -5.06214857e-01 1.85970232e-01
1.23705044e-01 -1.35965616e-01 -4.54869658e-01 6.11253977e-01
1.14425853e-01 4.94907290e-01 2.20785573e-01 -3.36052507e-01
5.12362003e-01 -2.81533957e-01 7.32562602e-01 -6.78205788e-01
-4.77694273e-01 -4.28080186e-02 -3.51286441e-01 -5.70921540e-01
-5.05354643e-01 -5.75883687e-02 -6.37989163e-01 -4.10481244e-01
-1.50883971e-02 -2.16751039e-01 4.02300090e-01 3.73299599e-01
6.32209480e-01 5.68243086e-01 4.76997405e-01 -1.44213104e+00
-1.05789760e-02 -9.03313518e-01 -6.99061334e-01 4.79722589e-01
6.08837485e-01 -6.77028418e-01 -4.09312516e-01 8.69070053e-01] | [11.443848609924316, -0.6696160435676575] |
0d14c6e1-59dd-4a42-b168-c7b8b0b8e035 | neural-cell-video-synthesis-via-optical-flow | 2212.03250 | null | https://arxiv.org/abs/2212.03250v1 | https://arxiv.org/pdf/2212.03250v1.pdf | Neural Cell Video Synthesis via Optical-Flow Diffusion | The biomedical imaging world is notorious for working with small amounts of data, frustrating state-of-the-art efforts in the computer vision and deep learning worlds. With large datasets, it is easier to make progress we have seen from the natural image distribution. It is the same with microscopy videos of neuron cells moving in a culture. This problem presents several challenges as it can be difficult to grow and maintain the culture for days, and it is expensive to acquire the materials and equipment. In this work, we explore how to alleviate this data scarcity problem by synthesizing the videos. We, therefore, take the recent work of the video diffusion model to synthesize videos of cells from our training dataset. We then analyze the model's strengths and consistent shortcomings to guide us on improving video generation to be as high-quality as possible. To improve on such a task, we propose modifying the denoising function and adding motion information (dense optical flow) so that the model has more context regarding how video frames transition over time and how each pixel changes over time. | ['Min Zou', 'Nathaniel Harris', 'Khoa Luu', 'Manuel Serna-Aguilera'] | 2022-12-06 | null | null | null | null | ['video-generation', 'culture'] | ['computer-vision', 'speech'] | [ 1.86078370e-01 -1.01035431e-01 2.34983936e-01 -3.90955210e-02
-3.23100209e-01 -5.47197282e-01 4.14345711e-01 -3.02813381e-01
-6.57171369e-01 1.02717364e+00 1.54818594e-01 -2.01355904e-01
7.69191459e-02 -5.92575133e-01 -7.47751236e-01 -1.04250598e+00
1.55688161e-02 1.32496044e-01 2.99039096e-01 4.35796380e-02
2.35288501e-01 4.99074131e-01 -1.41852725e+00 3.30615819e-01
4.73915398e-01 5.28868198e-01 4.94215637e-01 8.97641599e-01
-2.42000781e-02 1.26112032e+00 -5.78600168e-01 -2.30387405e-01
2.25075588e-01 -7.25447178e-01 -9.58652258e-01 3.09504986e-01
5.57968259e-01 -6.64480925e-01 -5.86359799e-01 9.19101655e-01
4.20108557e-01 1.00519307e-01 5.15452325e-01 -1.16565537e+00
-7.61051059e-01 1.14983767e-01 -4.40177143e-01 6.62408710e-01
2.00448498e-01 3.81404757e-01 4.63419884e-01 -5.51506341e-01
1.28027070e+00 1.00551474e+00 4.38913435e-01 9.54947293e-01
-1.09520292e+00 -2.73064613e-01 2.87688315e-01 3.35197240e-01
-1.04551530e+00 -6.26581490e-01 5.48828244e-01 -7.00362146e-01
8.01357508e-01 9.45223868e-02 8.90901089e-01 1.32589948e+00
3.43039781e-01 7.00841904e-01 9.01471078e-01 -4.31920022e-01
2.93157250e-01 -1.81214571e-01 -3.49857688e-01 7.04041719e-01
2.97400177e-01 -7.46451616e-02 -3.83599818e-01 2.18244970e-01
1.20874214e+00 6.25857785e-02 -6.18759453e-01 -2.54022688e-01
-1.47699213e+00 5.76458097e-01 8.10647383e-02 6.81803763e-01
-3.73102337e-01 1.82244092e-01 1.34718701e-01 3.52331221e-01
4.31163967e-01 3.86017919e-01 -3.16881478e-01 -4.37708318e-01
-1.14912391e+00 3.52795005e-01 8.76288533e-01 7.89872348e-01
5.93850315e-01 1.58030644e-01 1.49419263e-01 7.04014957e-01
4.43864502e-02 1.22808196e-01 5.29877782e-01 -1.63506222e+00
-5.82668046e-03 1.38315678e-01 1.30630717e-01 -1.05708301e+00
-2.93515950e-01 -1.13719665e-01 -7.67866373e-01 4.47014183e-01
1.08025301e+00 -2.69606262e-01 -9.40329075e-01 1.82294297e+00
1.37195349e-01 5.71458340e-01 -2.93757897e-02 9.26312149e-01
4.91088361e-01 8.02855611e-01 -1.86442986e-01 -5.63705206e-01
9.30680454e-01 -9.42252636e-01 -9.14527178e-01 -1.06748030e-01
8.16394567e-01 -6.23284698e-01 7.51060545e-01 5.23730636e-01
-1.35221708e+00 -3.41314077e-01 -9.44501102e-01 -1.78629994e-01
-2.64859647e-01 -4.02466029e-01 5.64709008e-01 4.58504796e-01
-1.26011872e+00 8.84162962e-01 -1.31728840e+00 -6.61348164e-01
5.54355860e-01 2.05965281e-01 -4.95824039e-01 -3.26336861e-01
-7.91045666e-01 1.02453339e+00 2.85471529e-02 -6.10721204e-03
-8.49292934e-01 -7.98013747e-01 -6.86191738e-01 -1.74137935e-01
1.46481857e-01 -9.42907333e-01 1.17809379e+00 -1.19968045e+00
-1.31492603e+00 7.84142554e-01 -2.98768878e-01 -3.26737851e-01
7.04976678e-01 1.59900427e-01 -6.35474101e-02 4.26218092e-01
-1.07415050e-01 9.00967777e-01 5.97809076e-01 -1.30854630e+00
-7.67487645e-01 -3.02740306e-01 3.09744794e-02 7.82519355e-02
-3.66726160e-01 -1.10000238e-01 -6.21160924e-01 -6.50090158e-01
-2.01282650e-01 -1.08395982e+00 -2.55913526e-01 4.62582707e-01
8.14872310e-02 3.10608685e-01 9.14111555e-01 -6.92728400e-01
1.05227971e+00 -2.02408504e+00 3.51759881e-01 -3.08386981e-01
4.16132003e-01 2.75339603e-01 -1.99121624e-01 1.88630730e-01
1.42423078e-01 4.38318178e-02 -2.33542070e-01 -3.11444938e-01
-4.87715274e-01 5.40737808e-01 6.30848184e-02 5.16635358e-01
1.81274861e-01 6.86857700e-01 -1.10206366e+00 -5.52187264e-01
7.76277184e-02 8.23858976e-01 -7.55647480e-01 -2.03138907e-02
-4.42483686e-02 8.38737249e-01 -2.40498170e-01 5.38036644e-01
4.84701961e-01 -3.98413032e-01 2.06225172e-01 -1.63978964e-01
-1.69996932e-01 -2.30493650e-01 -9.96472776e-01 1.71641409e+00
-1.62844166e-01 1.04118395e+00 2.99421400e-01 -1.26115942e+00
4.69309866e-01 3.55561942e-01 9.60095823e-01 -4.69704896e-01
1.45424098e-01 2.03215361e-01 3.11593175e-01 -9.17148173e-01
2.90075004e-01 -3.47687542e-01 5.88110209e-01 3.96862626e-01
1.19255885e-01 -1.26133502e-01 6.04648888e-01 2.31344923e-01
1.33539307e+00 1.89082459e-01 -1.21471718e-01 -2.20456883e-01
3.23404819e-01 2.71118302e-02 6.60779715e-01 4.88716841e-01
-5.07435560e-01 8.91025603e-01 4.51387376e-01 -4.22075063e-01
-1.42506874e+00 -8.00426722e-01 3.64896394e-02 4.84572858e-01
-1.36818916e-01 -1.31857634e-01 -1.00557935e+00 -3.02563190e-01
-3.00830513e-01 2.62314200e-01 -6.52629256e-01 -3.40617746e-02
-6.74232244e-01 -8.30768347e-01 3.90570670e-01 4.37481046e-01
2.49840483e-01 -9.73419249e-01 -7.19148397e-01 4.28258806e-01
-3.06323707e-01 -1.33188450e+00 -2.77475476e-01 -8.43775943e-02
-8.05010617e-01 -8.78127277e-01 -1.25681281e+00 -9.55362976e-01
8.21164429e-01 3.96458268e-01 9.97716129e-01 2.55823940e-01
-4.68055725e-01 3.76338094e-01 -3.83621812e-01 -3.52118164e-01
-5.01908183e-01 -2.72201955e-01 1.28627550e-02 -8.51168782e-02
2.08521545e-01 -6.76017165e-01 -7.16029823e-01 1.51297450e-01
-1.28681421e+00 1.69928223e-01 2.07036898e-01 8.57142448e-01
5.12584448e-01 1.87634692e-01 3.99088293e-01 -8.16217601e-01
3.54889125e-01 -4.69816506e-01 -2.73697793e-01 -1.41702155e-02
-3.75995815e-01 -1.65138066e-01 7.19075799e-01 -6.52759254e-01
-8.94854486e-01 2.47578081e-02 -1.55009225e-01 -5.65431654e-01
-4.06463863e-03 4.11102444e-01 1.97071716e-01 -1.61208779e-01
4.34050918e-01 2.15458021e-01 3.49954844e-01 -1.10258743e-01
1.70558259e-01 3.46118748e-01 5.31356215e-01 -4.01814193e-01
4.64663714e-01 1.09174693e+00 -9.79497507e-02 -1.10554659e+00
-5.79695582e-01 -1.96122333e-01 -7.45631278e-01 -4.82551157e-01
1.08425450e+00 -5.97950399e-01 -3.89331698e-01 6.84846282e-01
-1.13632476e+00 -6.08593524e-01 -2.56744295e-01 6.17193818e-01
-7.09408581e-01 5.30261099e-01 -9.18637156e-01 -4.62294102e-01
1.69435799e-01 -1.31041145e+00 7.14258492e-01 3.05623621e-01
-1.76605791e-01 -1.32201648e+00 2.58667052e-01 3.01844507e-01
4.59502310e-01 2.77690887e-01 6.56736851e-01 8.38078037e-02
-8.58963728e-01 2.38197714e-01 -8.35212097e-02 4.12607670e-01
2.44491190e-01 3.39787483e-01 -8.54140639e-01 -4.32315171e-01
2.79640496e-01 -1.09208554e-01 8.54984701e-01 7.61813164e-01
1.20311713e+00 -1.20372392e-01 -3.07685107e-01 8.06642592e-01
1.44320846e+00 4.22199816e-01 9.55984533e-01 5.46665967e-01
6.75184608e-01 7.74375200e-01 3.27463031e-01 1.87626734e-01
3.08821410e-01 4.49746013e-01 2.43019387e-01 -2.00496778e-01
-2.96113104e-01 1.36909187e-01 2.42495060e-01 1.06529176e+00
-4.00484234e-01 -3.82187247e-01 -7.86185205e-01 6.69598222e-01
-1.85795319e+00 -1.34776378e+00 -1.13145979e-02 1.87329364e+00
8.21217179e-01 -8.78490433e-02 3.46405841e-02 7.80022889e-02
4.48602259e-01 -1.77419949e-02 -5.19486129e-01 -1.31480277e-01
-2.27888271e-01 -1.95136204e-01 4.04430777e-01 5.09368300e-01
-8.55701506e-01 7.04180062e-01 6.74622965e+00 3.89838934e-01
-1.49437094e+00 6.35957122e-02 8.97132814e-01 -4.37606871e-01
-2.76229113e-01 -8.86044726e-02 -5.40241122e-01 6.65785372e-01
9.45260406e-01 -5.92146739e-02 5.53275645e-01 2.97445148e-01
6.09759152e-01 -2.40027949e-01 -1.31374347e+00 1.14554513e+00
2.38645464e-01 -1.63099086e+00 -1.66736320e-02 3.27705532e-01
8.07383478e-01 7.38776382e-03 -3.66440304e-02 -1.86981037e-01
1.59145430e-01 -9.74628508e-01 7.23728597e-01 7.13119566e-01
3.32946777e-01 -2.62865692e-01 5.10424793e-01 3.63032460e-01
-8.05844605e-01 -7.93084502e-02 -4.50069249e-01 -1.47700563e-01
5.55832148e-01 5.89610457e-01 -4.47928607e-01 3.07044059e-01
7.69072831e-01 9.20399487e-01 -3.95802021e-01 1.12979114e+00
3.32381845e-01 3.95422190e-01 -1.05350181e-01 2.03914344e-01
2.22140089e-01 -5.01976371e-01 4.11908150e-01 1.08682156e+00
7.38193274e-01 1.71595335e-01 -2.29632016e-02 7.00111985e-01
-6.49667829e-02 -2.95359433e-01 -9.61221874e-01 -1.72820434e-01
1.57954201e-01 1.22275519e+00 -1.01436913e+00 -3.92058611e-01
-8.25788140e-01 1.04307342e+00 2.86402911e-01 5.41176081e-01
-5.95851421e-01 8.28717276e-02 6.10635757e-01 1.69127673e-01
2.94152021e-01 -4.37381208e-01 -8.64902395e-05 -1.39187646e+00
-1.76253438e-03 -9.13797617e-01 1.10570550e-01 -1.04534721e+00
-1.19662011e+00 4.30900037e-01 -2.38603413e-01 -1.22563303e+00
-2.22874716e-01 -7.49176145e-01 -3.12535107e-01 4.34198171e-01
-1.48413765e+00 -8.53488684e-01 -2.34572887e-01 5.63342273e-01
5.95962465e-01 1.03844613e-01 4.96731192e-01 6.66577697e-01
-5.23680687e-01 1.35547504e-01 3.53266031e-01 7.70364925e-02
7.67971158e-01 -9.34876680e-01 4.25738655e-02 9.32652533e-01
1.59831911e-01 5.40994465e-01 8.29586685e-01 -4.66350049e-01
-1.45250738e+00 -7.65450120e-01 6.90375268e-01 -5.17358541e-01
7.88585544e-01 -8.11956450e-02 -1.02607405e+00 7.34534025e-01
2.51487404e-01 1.10592708e-01 4.74173546e-01 -5.52837789e-01
1.29210845e-01 1.05541989e-01 -9.10962820e-01 8.07902455e-01
1.06348479e+00 -3.47957611e-01 -2.71223336e-01 3.07809860e-01
3.81609708e-01 -2.30284080e-01 -7.85450161e-01 -2.58435216e-02
5.73357105e-01 -8.97646606e-01 8.89563024e-01 -5.07415414e-01
5.74608088e-01 -3.50240380e-01 -5.51855043e-02 -1.47220480e+00
-3.02068859e-01 -6.16948426e-01 -3.72474566e-02 1.11186862e+00
2.90859193e-01 -3.60897750e-01 1.03469586e+00 7.09405720e-01
-1.35994077e-01 -7.01023877e-01 -7.11335957e-01 -9.09411252e-01
4.29368645e-01 -1.49836943e-01 5.87095097e-02 1.08383775e+00
-1.15707275e-02 7.14902952e-02 -3.73682410e-01 -2.21614942e-01
4.25652474e-01 -2.76353329e-01 6.35925055e-01 -1.03484857e+00
-2.88497299e-01 -3.89588535e-01 -5.25248289e-01 -1.02227581e+00
7.15548173e-02 -5.75766742e-01 6.75521418e-02 -1.63461816e+00
3.45346332e-01 -1.78070620e-01 -1.37280285e-01 1.84566572e-01
-6.05894588e-02 3.44131529e-01 4.03103203e-01 4.19103324e-01
-3.75334710e-01 1.09996796e-01 1.91724086e+00 -2.29776040e-01
5.50749600e-02 -5.90283692e-01 -6.63885951e-01 6.71148539e-01
5.53740621e-01 -2.97949821e-01 -5.72231591e-01 -8.84591937e-01
6.64146692e-02 8.72232988e-02 2.81113356e-01 -1.10342467e+00
3.77341002e-01 -1.32520840e-01 4.62903947e-01 -1.37674049e-01
3.97357196e-01 -7.15557277e-01 3.14710468e-01 5.61791182e-01
-2.31536880e-01 1.35836482e-01 -3.64896506e-02 5.34699678e-01
-3.25256199e-01 -3.02053869e-01 1.03883171e+00 -4.92531240e-01
-6.15514457e-01 4.26370412e-01 -8.14700902e-01 1.23091668e-01
1.11239624e+00 -4.00274545e-01 -5.52573562e-01 -6.37345552e-01
-8.13668370e-01 6.46826029e-02 9.72565353e-01 2.33753845e-01
5.33897877e-01 -1.10207927e+00 -5.83033144e-01 7.98105597e-02
-2.62864321e-01 9.11131948e-02 4.35771704e-01 1.13590562e+00
-1.04781806e+00 2.28758335e-01 -6.35460854e-01 -5.76525509e-01
-1.11962509e+00 7.36673355e-01 2.72876084e-01 -7.69514814e-02
-7.61228681e-01 8.41607988e-01 2.54906565e-01 1.16016850e-01
1.24202013e-01 -3.83643627e-01 -2.63373911e-01 5.58244362e-02
7.98160255e-01 3.06452483e-01 -1.08522020e-01 -5.36294162e-01
-1.12744510e-01 6.29454255e-01 -2.40353376e-01 -2.41911277e-01
1.59315443e+00 -3.97357702e-01 -9.86786410e-02 3.43610674e-01
1.28208780e+00 -1.75286755e-01 -1.91309202e+00 2.34489024e-01
-1.95018589e-01 -4.66420114e-01 1.39995158e-01 -3.09635133e-01
-1.35689676e+00 9.54241455e-01 4.50535715e-01 4.72987056e-01
9.59656715e-01 -1.56064630e-01 1.05509067e+00 1.86975807e-01
2.14139968e-01 -1.12470114e+00 1.20606750e-01 4.68857288e-01
4.69497472e-01 -1.01084507e+00 -1.64726295e-03 -1.92635000e-01
-4.65265840e-01 1.23521030e+00 4.07901078e-01 -1.17039762e-01
7.11074412e-01 6.24357820e-01 3.05918038e-01 -1.85749665e-01
-8.84648204e-01 1.38669573e-02 -3.43367666e-01 7.76772320e-01
5.31751096e-01 -4.85417724e-01 -3.56131792e-01 -1.69829115e-01
1.74620569e-01 5.10239661e-01 1.01499796e+00 1.29571044e+00
-3.25015336e-01 -1.20603430e+00 -2.09287286e-01 4.22586650e-01
-7.10791886e-01 1.17515512e-01 -1.55688241e-01 6.94115758e-01
6.35019988e-02 8.91674578e-01 1.49360135e-01 -1.44562259e-01
-8.52126330e-02 -8.15028399e-02 8.60156476e-01 -4.49626625e-01
1.02402598e-01 2.11195484e-01 -3.65881734e-02 -5.59412003e-01
-1.01564419e+00 -8.36841047e-01 -1.24636328e+00 -4.81312186e-01
5.09177670e-02 -1.85769051e-01 4.98144716e-01 9.94364738e-01
2.57392913e-01 6.31030619e-01 2.73114294e-01 -1.23787117e+00
-2.08701175e-02 -5.92318356e-01 -6.13729477e-01 6.34857833e-01
5.20328104e-01 -6.36712193e-01 -4.40264702e-01 8.88340294e-01] | [10.785906791687012, -1.3900699615478516] |
41b9b7b3-4798-42a4-94d6-b2de3ebf7403 | adding-syntactic-annotations-to-flickr30k | null | null | https://aclanthology.org/L18-1716 | https://aclanthology.org/L18-1716.pdf | Adding Syntactic Annotations to Flickr30k Entities Corpus for Multimodal Ambiguous Prepositional-Phrase Attachment Resolution | null | ['Frederic Bechet', 'Alexis Nasr', 'Sebastien Delecraz', 'Benoit Favre'] | 2018-05-01 | adding-syntactic-annotations-to-flickr30k-1 | https://aclanthology.org/L18-1716 | https://aclanthology.org/L18-1716.pdf | lrec-2018-5 | ['prepositional-phrase-attachment'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.288529396057129, 3.716902494430542] |
0c748a81-74d0-43b0-9739-aee4ef6912e4 | pediatric-automatic-sleep-staging-a | 2108.10211 | null | https://arxiv.org/abs/2108.10211v3 | https://arxiv.org/pdf/2108.10211v3.pdf | Pediatric Automatic Sleep Staging: A comparative study of state-of-the-art deep learning methods | Background: Despite the tremendous progress recently made towards automatic sleep staging in adults, it is currently unknown if the most advanced algorithms generalize to the pediatric population, which displays distinctive characteristics in overnight polysomnography (PSG). Methods: To answer the question, in this work, we conduct a large-scale comparative study on the state-of-the-art deep learning methods for pediatric automatic sleep staging. Six different deep neural networks with diverging features are adopted to evaluate a sample of more than 1,200 children across a wide spectrum of obstructive sleep apnea (OSA) severity. Results: Our experimental results show that the individual performance of automated pediatric sleep stagers when evaluated on new subjects is equivalent to the expert-level one reported on adults. Combining the six stagers into ensemble models further boosts the staging accuracy, reaching an overall accuracy of 88.8%, a Cohen's kappa of 0.852, and a macro F1-score of 85.8%. At the same time, the ensemble models lead to reduced predictive uncertainty. The results also show that the studied algorithms and their ensembles are robust to concept drift when the training and test data were recorded seven months apart and after clinical intervention. Conclusion: However, we show that the improvements in the staging performance are not necessarily clinically significant although the ensemble models lead to more favorable clinical measures than the six standalone models. Significance: Detailed analyses further demonstrate "almost perfect" agreement between the automatic stagers to one another and their similar patterns on the staging errors, suggesting little room for improvement. | ['Mathias Baumert', 'Alfred Mertins', 'Huy Phan'] | 2021-08-23 | null | null | null | null | ['sleep-staging'] | ['medical'] | [ 2.87145879e-02 2.28305086e-01 -3.98628652e-01 -5.30524850e-01
-3.01868498e-01 -3.02854627e-01 -2.57436067e-01 3.66926134e-01
-4.90798235e-01 9.03008401e-01 1.61873624e-01 -1.53859228e-01
-4.25781250e-01 -2.86897212e-01 -5.97711317e-02 -8.53074968e-01
-2.08330780e-01 7.80992568e-01 8.94202739e-02 5.17503768e-02
-2.63433963e-01 1.31806582e-01 -1.87351215e+00 -5.18390574e-02
1.55442178e+00 1.11461151e+00 -1.52338609e-01 6.53818130e-01
1.72050312e-01 3.48903805e-01 -7.94521391e-01 -3.54509801e-01
7.44855925e-02 -1.87577978e-01 -5.63563406e-01 -2.76617974e-01
5.29792547e-01 -1.55770868e-01 1.52309015e-01 8.25024486e-01
9.75032747e-01 1.32641315e-01 3.47702354e-01 -9.37787771e-01
-2.54619837e-01 5.91716647e-01 -3.79390866e-02 5.87323129e-01
3.41002047e-01 1.17160492e-01 8.61995637e-01 -2.40229115e-01
-1.26836672e-01 2.72513330e-01 1.13434064e+00 9.67360675e-01
-1.27005589e+00 -7.85765409e-01 -1.63871422e-01 1.78127989e-01
-1.31149971e+00 -3.81365240e-01 2.04804093e-01 -4.76461381e-01
9.20262992e-01 3.53381664e-01 1.27653348e+00 8.92458200e-01
2.52153426e-01 3.98000568e-01 1.20151699e+00 -1.39325857e-01
3.94185692e-01 1.87476546e-01 2.50669301e-01 6.37380958e-01
8.30621839e-01 5.10146422e-03 -4.91135806e-01 1.92014024e-01
-3.86142954e-02 2.27881908e-01 -2.62858391e-01 -9.27755460e-02
-6.75315499e-01 4.66755778e-01 1.54162630e-01 6.99350536e-01
-3.97887498e-01 -2.30642319e-01 3.84113222e-01 1.17093824e-01
6.60191357e-01 6.56627297e-01 -6.61290705e-01 -4.65916306e-01
-1.68294418e+00 -1.88037664e-01 9.74739552e-01 4.86639351e-01
3.03489208e-01 8.36029798e-02 -9.39725339e-02 8.68062973e-01
1.18698217e-01 5.88020086e-01 9.03242350e-01 -8.10159028e-01
2.17254728e-01 8.91903460e-01 -8.44405070e-02 -4.40993965e-01
-1.14949369e+00 -9.77843940e-01 -8.94997001e-01 -5.29129095e-02
1.24354184e-01 -3.11774075e-01 -9.69042420e-01 1.73881066e+00
3.51709425e-02 1.72125041e-01 -1.93012193e-01 5.40911138e-01
1.02665293e+00 -1.26941381e-02 1.45477861e-01 -3.39360118e-01
1.21589172e+00 -9.81108844e-01 -5.60251653e-01 -2.56940901e-01
3.62363368e-01 -2.87550777e-01 6.12839818e-01 7.10764706e-01
-1.32080829e+00 -5.39412200e-01 -1.24184549e+00 3.18801552e-01
-2.81502873e-01 2.32961997e-01 5.95142245e-01 1.30058491e+00
-1.45465779e+00 1.06366527e+00 -1.22282004e+00 -5.71097612e-01
5.22772670e-01 1.05580080e+00 -1.43134207e-01 4.95727360e-01
-1.08426464e+00 1.01276934e+00 2.27815941e-01 1.21569492e-01
-6.14268601e-01 -1.04621971e+00 -4.64655608e-01 3.86864781e-01
-1.76103905e-01 -1.12939262e+00 1.25356984e+00 -7.73434341e-01
-1.45570755e+00 1.11455905e+00 -1.82692364e-01 -6.94713354e-01
2.98151135e-01 -2.91070789e-01 -6.40816867e-01 2.03478947e-01
1.10659972e-01 4.84369874e-01 6.43041730e-01 -5.29716969e-01
-8.34030151e-01 -6.45501554e-01 -2.49705657e-01 -3.32682190e-04
-3.10013980e-01 -1.74031794e-01 7.33876228e-02 -3.39760005e-01
1.16014145e-01 -1.05620706e+00 -3.12907994e-01 -4.08362567e-01
-5.86996563e-02 -2.47129217e-01 -1.32373050e-01 -5.63433707e-01
1.72667181e+00 -1.84313130e+00 1.46157995e-01 -1.60478324e-01
7.45873153e-01 5.67383945e-01 5.31318545e-01 1.63525358e-01
-3.27532500e-01 4.60133329e-02 -3.31267923e-01 -8.53242218e-01
-1.69238761e-01 4.00827289e-01 3.55055600e-01 4.97265339e-01
-3.99908334e-01 7.00798512e-01 -8.60910535e-01 -2.78402269e-01
3.88401806e-01 1.67850152e-01 -6.36960745e-01 2.58624822e-01
3.81060690e-01 5.79302311e-01 -1.61294952e-01 6.31772280e-01
4.01512980e-01 -3.41031373e-01 -1.12334797e-02 2.14002192e-01
-2.38126710e-01 5.24216592e-01 -7.26277053e-01 1.96774507e+00
-4.16674405e-01 5.33548772e-01 -1.33343056e-01 -8.34139884e-01
6.89170837e-01 4.60202992e-01 6.68220043e-01 -2.31409311e-01
2.93623865e-01 5.10285497e-01 5.80663383e-01 -3.98374021e-01
3.37277167e-02 -5.42297482e-01 2.30289027e-01 3.29265147e-01
3.61884207e-01 -9.25219059e-02 2.84647256e-01 -3.02635640e-01
1.17698634e+00 -4.16865677e-01 5.11267066e-01 -4.99449998e-01
6.61274195e-01 -3.01745713e-01 7.61127114e-01 8.48446608e-01
-6.02768064e-01 7.67046332e-01 1.36600286e-01 -5.06133139e-01
-5.10576367e-01 -1.25333631e+00 -4.83900726e-01 6.73291087e-01
-2.65516400e-01 -5.44315398e-01 -8.36462677e-01 -6.55104220e-01
-8.61102417e-02 6.73044026e-01 -8.17134619e-01 -3.08145970e-01
-2.10479647e-01 -1.09030759e+00 7.57713377e-01 7.10511684e-01
3.40508372e-01 -7.79504776e-01 -8.29140604e-01 2.07633242e-01
6.94880337e-02 -7.99856722e-01 -9.05534253e-02 3.38549703e-01
-1.31435168e+00 -1.08709335e+00 -8.59827161e-01 -3.92118007e-01
3.35991621e-01 -2.32500866e-01 1.29697371e+00 2.55157351e-01
-1.23659544e-01 3.97285849e-01 -2.43184194e-01 -5.45596063e-01
-3.42322856e-01 5.47048628e-01 6.70132875e-01 -2.77927905e-01
6.21365666e-01 -1.18014824e+00 -1.25126040e+00 1.50123656e-01
-3.82426947e-01 -3.36701453e-01 5.14049947e-01 6.04325235e-01
2.58776158e-01 -1.35090783e-01 4.95288312e-01 -5.60476899e-01
3.35086405e-01 -5.25390923e-01 -3.55449289e-01 1.30357161e-01
-1.66105425e+00 -6.65293187e-02 6.91743016e-01 -1.05626635e-01
-6.32634461e-01 -2.74884284e-01 -4.55157787e-01 -3.76519918e-01
-4.27707523e-01 2.18203291e-02 4.31095958e-01 1.98864147e-01
6.13258958e-01 1.49237573e-01 1.51640758e-01 -7.01798677e-01
-2.47259259e-01 5.84609985e-01 3.35588932e-01 -3.16491634e-01
5.29997528e-01 3.32621455e-01 1.58029526e-01 -6.03728116e-01
-1.19152892e+00 -7.11989403e-01 -8.68499637e-01 1.30235739e-02
1.22206342e+00 -8.77586484e-01 -6.65103376e-01 4.34804738e-01
-4.56079692e-01 -3.40573519e-01 -5.84790945e-01 6.20377839e-01
-3.47785026e-01 2.75154173e-01 -2.80678123e-01 -6.82153761e-01
-1.15461099e+00 -1.18357563e+00 8.23135972e-01 6.40523434e-01
-5.09209096e-01 -1.36959767e+00 5.97812057e-01 5.84702611e-01
6.06265366e-01 -2.57029012e-02 7.04488993e-01 -1.18832684e+00
4.22565490e-02 -1.41126946e-01 2.30567887e-01 5.69255531e-01
3.70387822e-01 -6.35471717e-02 -1.19538569e+00 -4.00401384e-01
1.65872037e-01 2.11392373e-01 7.46456861e-01 6.37047529e-01
1.02913904e+00 1.10082194e-01 -1.79228216e-01 7.34190166e-01
1.14128506e+00 4.46552843e-01 3.44254255e-01 2.84663975e-01
4.09769088e-01 4.33895946e-01 1.30845653e-02 2.00714394e-01
4.51914042e-01 4.03939039e-01 2.23644346e-01 1.14050306e-01
-1.92001551e-01 -1.07455023e-01 6.56483546e-02 1.11819017e+00
-4.77065861e-01 1.25315949e-01 -1.00125980e+00 5.03523767e-01
-1.48430216e+00 -8.33635926e-01 -1.36743090e-03 2.34574366e+00
5.49509466e-01 1.60664901e-01 4.35077041e-01 2.17832297e-01
5.49474597e-01 2.73317665e-01 -5.58202565e-01 -6.37519300e-01
1.97577149e-01 8.55542541e-01 4.05249864e-01 3.27026509e-02
-7.85427034e-01 2.92228878e-01 7.28280449e+00 3.60894710e-01
-1.25558245e+00 4.27817762e-01 4.64598656e-01 -5.52699208e-01
2.16327518e-01 -3.54013562e-01 -7.09404767e-01 9.09378409e-01
1.51856506e+00 -2.00512889e-03 4.24179941e-01 8.48096132e-01
2.88241625e-01 -8.93086269e-02 -1.11356759e+00 9.44020689e-01
1.46380842e-01 -1.08630538e+00 -7.57437706e-01 -1.81422359e-03
9.51504648e-01 5.39592862e-01 3.76733840e-02 4.02350813e-01
-2.74121910e-01 -1.05196369e+00 3.22037727e-01 5.14443815e-01
1.02340984e+00 -2.89566994e-01 1.06362391e+00 1.98005170e-01
-9.15624976e-01 -3.23195726e-01 3.48946825e-02 -2.02543840e-01
4.83502075e-03 5.56704581e-01 -9.81944978e-01 4.56197470e-01
9.93558764e-01 7.94871449e-01 -9.09592569e-01 1.31268263e+00
-3.71746689e-01 9.34985459e-01 -4.99433190e-01 -1.85709223e-01
-9.81037542e-02 -1.32298291e-01 4.80942160e-01 1.03794312e+00
3.87333512e-01 3.39780189e-02 -5.55944979e-01 7.98799157e-01
2.72539854e-01 4.86428477e-02 -2.60799706e-01 1.49856746e-01
2.03635201e-01 1.30626571e+00 -6.13467991e-01 -2.81902879e-01
-5.35488784e-01 5.90753019e-01 2.78324801e-02 -1.31157666e-01
-7.16564536e-01 -3.48603204e-02 5.95243037e-01 2.96894073e-01
1.81952178e-01 4.02622044e-01 -7.87261069e-01 -1.28586149e+00
-1.97262168e-01 -6.84561789e-01 5.90939641e-01 -5.98404467e-01
-1.09603918e+00 4.83044386e-01 7.67111480e-02 -1.17158413e+00
-1.30308136e-01 -5.37171483e-01 -1.12844908e+00 5.72087526e-01
-9.82562959e-01 -4.72120255e-01 -4.20539677e-01 1.76176548e-01
4.35274720e-01 -2.52689689e-01 7.80807137e-01 6.01061285e-01
-9.48638141e-01 8.35280120e-01 4.63773012e-02 -3.07868004e-01
4.42044020e-01 -1.78369796e+00 3.36293392e-02 7.99311459e-01
-1.53325170e-01 9.95019853e-01 6.69983029e-01 -4.55261409e-01
-5.81625938e-01 -7.17383742e-01 9.21229482e-01 -7.16895223e-01
6.37909055e-01 2.31596738e-01 -6.12096846e-01 4.48493868e-01
1.48521915e-01 -3.74394208e-01 1.19711173e+00 5.38787603e-01
2.48142153e-01 -6.36708796e-01 -1.27402329e+00 1.31779432e-01
1.07495487e+00 -1.75324827e-01 -9.65765655e-01 -8.00142065e-03
4.10970867e-01 -4.64019001e-01 -1.26906002e+00 7.79508293e-01
8.89908135e-01 -1.48533487e+00 6.31020188e-01 -4.88198012e-01
3.02921861e-01 -4.76381518e-02 1.22843064e-01 -1.30193865e+00
-9.73362774e-02 -6.23118043e-01 -1.92857012e-01 9.39188302e-01
5.30073822e-01 -9.24174726e-01 1.06052506e+00 9.06318069e-01
-5.97519279e-01 -1.14665949e+00 -1.16786289e+00 -4.60837066e-01
4.43866923e-02 -4.48493868e-01 7.40712583e-01 5.19774318e-01
-8.28144848e-02 3.91845673e-01 -2.28583869e-02 2.18598709e-01
3.76431406e-01 1.09656721e-01 3.17394942e-01 -1.69658768e+00
-3.27647924e-01 -5.50205469e-01 -6.56237662e-01 -4.47342217e-01
-6.84163868e-02 -8.11934471e-01 -3.42831522e-01 -1.51053548e+00
2.52116859e-01 -5.42515159e-01 -8.64584208e-01 3.48613203e-01
-2.88590074e-01 1.13079891e-01 -9.47929844e-02 6.01082742e-02
-6.45301938e-01 2.49213934e-01 9.05127227e-01 1.72451481e-01
-4.32187229e-01 9.28673089e-01 -8.13577592e-01 9.27799463e-01
8.84906411e-01 -5.45780301e-01 -5.81631124e-01 8.14159808e-04
1.99798867e-01 -6.98537603e-02 -9.24595818e-02 -1.50365520e+00
5.52000217e-02 3.76567453e-01 3.87487680e-01 -5.26316226e-01
3.30087483e-01 -7.31585026e-01 3.05500120e-01 9.01505113e-01
1.31636724e-01 1.84682727e-01 1.63052022e-01 1.64263308e-01
7.79120401e-02 -4.76340294e-01 7.59536088e-01 1.85350142e-02
-2.45304197e-01 4.14028108e-01 -3.56450528e-01 1.70467481e-01
5.42401373e-01 -3.53255868e-01 -2.79983848e-01 -3.20671022e-01
-1.21526158e+00 1.55184641e-01 4.81808037e-01 2.93180972e-01
8.23877826e-02 -6.93284452e-01 -1.51057631e-01 4.35708165e-01
1.68002807e-02 5.13554513e-02 4.01907712e-01 1.60440147e+00
-5.70619583e-01 6.95219100e-01 -1.19010419e-01 -8.43237579e-01
-1.22803020e+00 2.28041932e-01 7.28870332e-01 -3.99965346e-01
-4.07431871e-01 9.59492564e-01 -1.50767028e-01 -2.04074889e-01
3.52382988e-01 -8.37174416e-01 -2.91496038e-01 2.87019044e-01
3.45413536e-01 8.55598927e-01 3.24408829e-01 -3.46909016e-01
-4.51251894e-01 6.45635605e-01 1.05953276e-01 3.31174344e-01
1.39299619e+00 -1.14444517e-01 -1.76481679e-01 5.54933310e-01
6.47334635e-01 1.87571123e-01 -6.77522898e-01 4.21949178e-01
-6.17212057e-02 -7.75433891e-03 -2.66979545e-01 -1.00619292e+00
-1.14340580e+00 9.01862144e-01 1.21892166e+00 3.61366540e-01
1.47160590e+00 -3.07797473e-02 7.49567270e-01 4.99177128e-02
2.29004800e-01 -8.10802221e-01 -4.55877185e-01 -6.12806976e-02
2.83230931e-01 -1.19946373e+00 2.80198991e-01 1.89442664e-01
-4.06945497e-01 9.34616268e-01 6.16510808e-01 1.64687149e-02
5.93847752e-01 -2.63487071e-01 5.81617765e-02 -2.48281986e-01
-6.42390668e-01 -3.12763125e-01 6.14834130e-01 5.59327245e-01
2.60110736e-01 3.44697207e-01 -5.58092654e-01 9.95500982e-01
-7.30152845e-01 1.68785125e-01 3.34201813e-01 2.71844864e-01
-4.94820178e-01 -1.07657897e+00 3.55889052e-02 7.63807535e-01
-9.04526114e-01 -7.84231648e-02 -4.88864668e-02 7.44937003e-01
6.42495096e-01 1.15176702e+00 9.08713974e-03 -5.09392738e-01
2.73227245e-01 4.35788602e-01 4.77429837e-01 -8.69155526e-01
-8.54463220e-01 -3.08702022e-01 1.96615130e-01 -5.29011905e-01
-6.94509089e-01 -7.92887926e-01 -1.07788861e+00 -7.35102668e-02
-3.57663482e-01 5.01065016e-01 5.70431590e-01 1.01316738e+00
3.31974000e-01 5.29416561e-01 4.75565791e-01 -5.15317142e-01
-4.34570014e-01 -1.10616934e+00 -5.87981045e-01 6.91298693e-02
6.40726686e-01 -7.64644563e-01 -8.19558561e-01 -5.13805449e-01] | [13.506965637207031, 3.516704797744751] |
71f808e6-77fb-4a7d-b486-4d54d939e5cc | a-secure-federated-learning-framework-for | 2209.14547 | null | https://arxiv.org/abs/2209.14547v2 | https://arxiv.org/pdf/2209.14547v2.pdf | A Secure Federated Learning Framework for Residential Short Term Load Forecasting | Smart meter measurements, though critical for accurate demand forecasting, face several drawbacks including consumers' privacy, data breach issues, to name a few. Recent literature has explored Federated Learning (FL) as a promising privacy-preserving machine learning alternative which enables collaborative learning of a model without exposing private raw data for short term load forecasting. Despite its virtue, standard FL is still vulnerable to an intractable cyber threat known as Byzantine attack carried out by faulty and/or malicious clients. Therefore, to improve the robustness of federated short-term load forecasting against Byzantine threats, we develop a state-of-the-art differentially private secured FL-based framework that ensures the privacy of the individual smart meter's data while protect the security of FL models and architecture. Our proposed framework leverages the idea of gradient quantization through the Sign Stochastic Gradient Descent (SignSGD) algorithm, where the clients only transmit the `sign' of the gradient to the control centre after local model training. As we highlight through our experiments involving benchmark neural networks with a set of Byzantine attack models, our proposed approach mitigates such threats quite effectively and thus outperforms conventional Fed-SGD models. | ['Robin Doss', 'Abdun Naser Mahmood', 'Shama Naz Islam', 'Nasser Hosseinzadeh', 'Adnan Anwar', 'Muhammad Akbar Husnoo'] | 2022-09-29 | null | null | null | null | ['load-forecasting'] | ['miscellaneous'] | [ 9.91260037e-02 -2.61595491e-02 -1.18317641e-01 -5.71634948e-01
-1.00352013e+00 -1.21809804e+00 7.34423697e-01 1.72289491e-01
-8.42075348e-02 6.14245057e-01 2.34335303e-01 -6.80787683e-01
7.14084879e-02 -8.14413846e-01 -8.23508799e-01 -1.24417925e+00
-3.13913167e-01 -7.96058103e-02 -2.37347811e-01 1.25392139e-01
2.54758358e-01 5.71200907e-01 -9.73598003e-01 2.26181880e-01
6.49971187e-01 1.58674240e+00 -7.32669771e-01 2.96037614e-01
3.58841449e-01 9.16347325e-01 -5.98396122e-01 -8.47673774e-01
8.67782712e-01 -1.47206217e-01 -4.77979511e-01 -4.00877893e-01
2.75887787e-01 -9.09394205e-01 -4.92955804e-01 1.39474535e+00
6.78965092e-01 -3.54469985e-01 2.19096869e-01 -1.93727517e+00
-6.20215416e-01 1.02376509e+00 -2.89466232e-01 -1.18576571e-01
-4.67963591e-02 5.62832892e-01 8.75848591e-01 -2.67794460e-01
1.19645454e-01 7.05899954e-01 7.30382025e-01 3.96696746e-01
-1.19176948e+00 -1.16424096e+00 5.93345612e-02 2.96293676e-01
-1.37080634e+00 -6.34905338e-01 9.00526702e-01 -4.02356721e-02
7.23772407e-01 6.84345067e-01 7.82215595e-02 8.94029737e-01
4.90205854e-01 9.54050481e-01 1.17878354e+00 2.04826873e-02
9.49953079e-01 2.96403438e-01 2.28715725e-02 1.98897287e-01
3.82995993e-01 4.82617259e-01 -5.19954979e-01 -1.07263720e+00
-1.29583653e-03 1.44640148e-01 -4.80337083e-01 -5.06697536e-01
-6.78381145e-01 9.05656278e-01 3.90633821e-01 -2.02202782e-01
-4.08231765e-01 4.29808348e-01 9.02024090e-01 5.70995867e-01
3.47150177e-01 -2.73115277e-01 -1.06331718e+00 1.99815169e-01
-6.93748951e-01 9.22847986e-02 1.20616078e+00 5.37574470e-01
5.74130118e-01 3.41499716e-01 1.45171002e-01 -3.01241577e-01
5.36565781e-01 5.08281529e-01 6.42429173e-01 -5.89963377e-01
6.59562826e-01 3.11649382e-01 1.68876231e-01 -1.01172566e+00
-5.63357398e-03 -3.36613059e-01 -9.74133849e-01 3.91120046e-01
1.30033299e-01 -5.76787055e-01 -2.89817099e-02 1.58422542e+00
6.45532429e-01 4.41727847e-01 3.63546550e-01 6.50688827e-01
-4.98552574e-03 3.71640831e-01 2.28759289e-01 -3.61606963e-02
1.10140216e+00 -5.55575669e-01 -6.62374318e-01 1.25344202e-01
9.70461369e-01 -2.22193122e-01 3.52722943e-01 5.22192895e-01
-7.03156531e-01 1.65983096e-01 -1.21393096e+00 2.80735463e-01
-5.12326598e-01 -3.49093467e-01 6.36306465e-01 1.49729121e+00
-8.38265777e-01 5.84408104e-01 -8.83933902e-01 3.60366166e-01
1.01419032e+00 5.70246875e-01 -3.73365879e-01 9.86113399e-02
-1.35582113e+00 5.25149286e-01 2.71130085e-01 -9.64184999e-02
-1.01317537e+00 -9.57033873e-01 -7.77754843e-01 2.38044962e-01
-2.53154896e-02 -4.53681797e-01 1.20919418e+00 -7.09337234e-01
-1.43444514e+00 4.63508427e-01 6.24523997e-01 -1.31667972e+00
7.75532126e-01 4.80757281e-03 -7.84367800e-01 1.22254258e-02
-4.66243267e-01 -1.53217763e-01 1.07248878e+00 -1.10582578e+00
-8.68823349e-01 -8.98934126e-01 -1.31572619e-01 -3.56456667e-01
-5.78774214e-01 -3.35956202e-03 9.72935259e-01 -6.17691398e-01
-2.69286156e-01 -6.77817583e-01 -1.57146931e-01 1.63371712e-01
-3.97065043e-01 -9.60201249e-02 1.57739735e+00 -8.06266785e-01
1.11716735e+00 -2.38082862e+00 -9.58768547e-01 3.77180338e-01
1.84382629e-02 4.83068615e-01 2.84780651e-01 7.57821083e-01
-1.05913483e-01 7.08428994e-02 -2.76978761e-01 -4.73007619e-01
8.92092347e-01 9.26434398e-02 -8.40680838e-01 1.27108347e+00
-4.00680929e-01 1.26729286e+00 -6.29707336e-01 8.88406038e-02
3.00868690e-01 6.69633746e-01 -2.27954194e-01 1.98012460e-02
-2.82198519e-01 4.03947443e-01 -6.24601364e-01 5.19741774e-01
1.26787758e+00 -9.77305621e-02 1.92280382e-01 -2.18324229e-01
2.07256630e-01 3.09587181e-01 -1.30397975e+00 1.27996171e+00
-1.88555628e-01 3.19584124e-02 4.07612860e-01 -8.15807581e-01
9.63113964e-01 6.22998953e-01 5.63608766e-01 -5.64069986e-01
4.67357010e-01 1.85054734e-01 -5.24174511e-01 -1.39135465e-01
-2.66244225e-02 1.20759144e-01 -1.33059129e-01 1.14399505e+00
-4.58459646e-01 3.90736610e-01 -8.46622109e-01 7.52062500e-02
1.14353669e+00 -1.27016559e-01 1.44058004e-01 -1.91784516e-01
6.01998627e-01 -4.19975102e-01 6.32217169e-01 6.08898938e-01
-6.63428962e-01 -9.60790068e-02 9.46801156e-02 -5.99795818e-01
-5.03121436e-01 -7.59682298e-01 -5.60915843e-03 9.12213147e-01
-1.86048865e-01 -3.74121606e-01 -7.55636930e-01 -1.18154001e+00
5.86731076e-01 1.17388439e+00 -3.34607750e-01 -4.27338779e-01
-3.36681068e-01 -5.83763897e-01 1.14143109e+00 2.83157736e-01
9.07617748e-01 -6.99562073e-01 -7.17565656e-01 1.64318055e-01
3.56271029e-01 -8.37549925e-01 -6.29257083e-01 3.54378998e-01
-5.57327271e-01 -9.97754812e-01 -6.75125197e-02 -3.88937116e-01
4.48026538e-01 1.40596241e-01 5.77868998e-01 -2.62524873e-01
1.69632323e-02 4.11022544e-01 -1.11603431e-01 -4.68476117e-01
-5.84456146e-01 -1.84738725e-01 2.09654495e-02 6.74108684e-01
6.91649258e-01 -9.46931779e-01 -1.02208495e+00 1.04346566e-01
-1.10767293e+00 -6.67226732e-01 -3.60414609e-02 3.90879065e-01
2.69510984e-01 3.74886781e-01 8.83164108e-01 -8.99381578e-01
4.97061670e-01 -6.60171986e-01 -9.61740732e-01 1.84300944e-01
-1.10887325e+00 -2.64688462e-01 1.15798318e+00 -2.15350509e-01
-8.20852518e-01 2.21508756e-01 -7.50246793e-02 -2.89470375e-01
-2.24975064e-01 -1.96174551e-02 -6.50657058e-01 -6.75544798e-01
3.95160466e-01 5.40851712e-01 3.61509062e-02 -6.25639975e-01
7.59033442e-01 1.03721452e+00 5.53068817e-01 -1.01498835e-01
1.02484012e+00 7.11070538e-01 9.03639346e-02 -9.28860009e-02
-1.72388107e-01 -6.30550832e-02 -1.79570913e-01 2.57188588e-01
3.77480567e-01 -9.54885960e-01 -1.40336072e+00 1.01140368e+00
-7.47752011e-01 6.11142674e-03 -4.13026541e-01 1.75386950e-01
-3.92541498e-01 6.51099741e-01 -6.99709058e-01 -9.14196908e-01
-1.10788250e+00 -9.07170951e-01 6.54613912e-01 -5.28559573e-02
1.56677783e-01 -1.14517236e+00 -1.22699879e-01 3.79359424e-01
8.51846874e-01 6.34145319e-01 9.32597756e-01 -1.23656452e+00
-5.43188989e-01 -9.02496994e-01 1.63584784e-01 6.76775336e-01
3.07490647e-01 -7.00191379e-01 -1.37267637e+00 -8.12194467e-01
7.92622447e-01 -3.17456365e-01 1.89211965e-01 -1.48886710e-01
1.12263107e+00 -1.32019675e+00 1.87830292e-02 9.55961764e-01
1.59971988e+00 -4.58353050e-02 5.11216581e-01 2.65205055e-01
5.57021916e-01 2.06712425e-01 9.49655175e-02 1.09112740e+00
5.58869720e-01 1.42178694e-02 9.03821349e-01 3.02456319e-01
5.96003294e-01 -5.54462552e-01 5.05944431e-01 3.36686671e-01
9.63212311e-01 -1.80196404e-01 -2.01053098e-01 1.49331227e-01
-1.61357081e+00 -8.29699457e-01 1.31378055e-01 2.21462274e+00
6.94859862e-01 -2.12440714e-01 -7.36123919e-02 5.97615659e-01
3.99204433e-01 3.75587672e-01 -1.03721499e+00 -5.41806638e-01
-2.38773808e-01 3.27783339e-02 1.23089695e+00 1.54544041e-01
-1.09181786e+00 3.65970641e-01 4.75677872e+00 7.41210997e-01
-1.18877971e+00 3.99441332e-01 8.17449212e-01 8.38827416e-02
-2.94409245e-01 7.08166650e-03 -5.88324785e-01 7.33306527e-01
1.22398913e+00 -5.82712114e-01 5.52366495e-01 1.11830986e+00
8.54772478e-02 3.85394186e-01 -1.08999109e+00 8.97778034e-01
-6.58359230e-02 -1.35747290e+00 -2.45493606e-01 3.24515462e-01
6.71410978e-01 2.49972373e-01 3.09718549e-01 -7.66209587e-02
8.39186370e-01 -7.47634649e-01 7.12075770e-01 1.51016712e-01
4.23889488e-01 -1.23012137e+00 3.73023868e-01 6.73713088e-01
-9.39274549e-01 -5.30787706e-01 -2.24571541e-01 1.88595697e-01
1.42847106e-01 4.81342643e-01 -5.51599562e-01 6.89904332e-01
5.11083722e-01 1.91093221e-01 -3.75478119e-01 6.31781459e-01
-2.06673294e-01 1.00734758e+00 -4.95253474e-01 2.01765358e-01
4.12690610e-01 -9.80905592e-02 3.11160982e-01 8.65924478e-01
7.76380226e-02 -1.40537205e-03 4.80853170e-02 6.22802794e-01
-3.03357244e-01 2.42536161e-02 -6.09606564e-01 1.73851565e-01
6.69010460e-01 1.09805775e+00 -1.69548884e-01 1.80869410e-03
-2.28484184e-01 1.14446270e+00 -1.33154556e-01 1.88630253e-01
-5.03132761e-01 -2.71483064e-01 8.43741179e-01 -3.48412633e-01
7.12840259e-01 -3.29210842e-03 -4.72929955e-01 -9.98131752e-01
1.10039763e-01 -1.13686633e+00 7.92876244e-01 -1.44638851e-01
-1.73635662e+00 1.93432644e-01 -5.98330438e-01 -1.09577966e+00
-2.66272753e-01 1.54458489e-02 -6.55874610e-01 7.76988685e-01
-1.74785817e+00 -1.42320669e+00 4.23025578e-01 1.14448333e+00
-2.01397523e-01 -2.30540663e-01 1.17590177e+00 1.29329145e-01
-2.87058055e-01 1.14731228e+00 7.90228546e-01 2.08870709e-01
1.80187479e-01 -9.49321449e-01 9.08000290e-01 8.15963745e-01
3.68921570e-02 3.41638893e-01 5.19982338e-01 -5.89344859e-01
-2.08146930e+00 -1.44809687e+00 7.93705881e-01 -1.09107614e-01
8.10825229e-01 -5.63667059e-01 -8.21151018e-01 1.01004899e+00
2.34632269e-01 4.73860383e-01 9.17326927e-01 -9.50492382e-01
-7.99562991e-01 -5.66674829e-01 -2.15997815e+00 1.79588377e-01
3.58669490e-01 -8.83026779e-01 -9.40566137e-03 5.74147880e-01
8.60343158e-01 1.48206860e-01 -1.03325891e+00 -1.07572846e-01
4.48864818e-01 -1.04777431e+00 6.82356656e-01 -5.66501677e-01
-6.93137288e-01 -2.36737162e-01 -6.12808883e-01 -1.18716490e+00
-4.50275838e-02 -1.51011813e+00 -7.16408968e-01 1.61169195e+00
-2.06088088e-02 -1.55953717e+00 8.85586262e-01 1.12640083e+00
4.45619136e-01 -2.97570556e-01 -1.39102244e+00 -6.12468004e-01
3.27196985e-01 -4.86977071e-01 1.72769773e+00 1.09154975e+00
9.52579007e-02 -4.88439143e-01 -4.35605526e-01 7.27961242e-01
1.18707323e+00 2.50694575e-03 7.29041994e-01 -6.79919302e-01
-9.44638103e-02 4.86360341e-02 -5.37749887e-01 -7.37883866e-01
4.29870695e-01 -9.78322685e-01 -4.34706897e-01 -6.52878463e-01
-4.94877905e-01 -3.54977667e-01 -6.78033054e-01 5.96401930e-01
4.76960182e-01 2.01838195e-01 3.61540288e-01 7.94584677e-02
-2.75473475e-01 5.58377326e-01 4.55540895e-01 -2.32747674e-01
2.16475442e-01 4.61560279e-01 -9.35044229e-01 2.71663845e-01
1.08633947e+00 -6.83104396e-01 -4.35437202e-01 -2.60097474e-01
9.87510458e-02 1.26251876e-02 5.35572171e-01 -6.92691922e-01
5.87076247e-01 3.76402512e-02 1.71328992e-01 -6.10362887e-01
-1.28928006e-01 -1.48662233e+00 5.39009392e-01 7.91604519e-01
-2.93767989e-01 8.27965662e-02 -2.05828041e-01 7.01264679e-01
2.48674631e-01 6.88541383e-02 5.38722038e-01 -2.97273453e-02
-1.56423628e-01 5.68398356e-01 -8.03986415e-02 -3.76188964e-01
1.15112209e+00 1.01358965e-01 -5.19580126e-01 -4.17850852e-01
-3.07384104e-01 4.00239199e-01 5.04980624e-01 1.72298729e-01
2.70967931e-01 -1.19805682e+00 -7.00184762e-01 6.13393009e-01
-1.66904107e-01 -4.43617374e-01 2.50854433e-01 5.11902988e-01
-4.21589874e-02 4.89749253e-01 1.76781490e-01 -1.13844685e-01
-1.00878632e+00 9.76946592e-01 4.47386563e-01 -5.23808561e-02
-9.11575556e-01 3.92827511e-01 -2.19490290e-01 -5.12229085e-01
6.78098679e-01 -1.56817541e-01 4.48729783e-01 -1.27753347e-01
9.60185468e-01 7.32502341e-01 5.79574466e-01 -7.26252019e-01
-3.56463283e-01 -1.37651488e-01 -3.16502973e-02 1.65069133e-01
1.34302115e+00 -6.50749624e-01 -7.34764934e-02 -2.56330427e-02
1.66075599e+00 -2.57092863e-01 -1.49106741e+00 -3.93352032e-01
2.94162030e-03 -6.27760708e-01 3.98155153e-01 -1.04721010e+00
-1.40020406e+00 4.91652399e-01 1.01120603e+00 4.20495123e-01
1.16010654e+00 -6.38297260e-01 1.55276883e+00 2.00322762e-01
8.45832765e-01 -7.25386322e-01 -8.00003409e-01 -1.35994971e-01
2.62255281e-01 -8.44479442e-01 -1.86655253e-01 9.17949528e-02
-6.34786248e-01 7.64340162e-01 -1.92962453e-01 -1.31533682e-01
1.21957922e+00 6.17395401e-01 3.66980821e-01 1.15788609e-01
-7.80262351e-01 7.36674666e-01 -5.06018460e-01 8.17952454e-01
-5.60511827e-01 2.80821860e-01 1.41413743e-02 9.33054507e-01
-2.71313071e-01 2.31634244e-01 2.37106025e-01 1.34668624e+00
6.55797124e-02 -1.06743491e+00 -4.61963415e-01 2.13096470e-01
-1.03565156e+00 1.24203108e-01 -1.17237523e-01 1.05393425e-01
-1.18780755e-01 1.11865354e+00 -2.97327608e-01 -3.85370404e-01
3.15545406e-03 2.95291871e-01 -1.14521533e-01 3.23302060e-01
-1.31313097e+00 -3.16266000e-01 -3.48158479e-01 -6.78894937e-01
6.45130575e-02 -8.31476152e-01 -1.14005148e+00 -7.43299127e-01
-4.37552899e-01 1.25804529e-01 1.07976627e+00 8.30829859e-01
6.90784037e-01 -4.20177907e-01 1.73468173e+00 -4.52448010e-01
-1.78070879e+00 -4.04265404e-01 -1.19932950e+00 5.09225130e-01
6.08288169e-01 1.68205485e-01 -8.63919079e-01 -8.29318687e-02] | [5.816794395446777, 6.668407917022705] |
45ca7e0b-b05c-44e3-bc0d-5e7ab44de8c1 | unaligned-supervision-for-automatic-music | 2204.13668 | null | https://arxiv.org/abs/2204.13668v1 | https://arxiv.org/pdf/2204.13668v1.pdf | Unaligned Supervision For Automatic Music Transcription in The Wild | Multi-instrument Automatic Music Transcription (AMT), or the decoding of a musical recording into semantic musical content, is one of the holy grails of Music Information Retrieval. Current AMT approaches are restricted to piano and (some) guitar recordings, due to difficult data collection. In order to overcome data collection barriers, previous AMT approaches attempt to employ musical scores in the form of a digitized version of the same song or piece. The scores are typically aligned using audio features and strenuous human intervention to generate training labels. We introduce NoteEM, a method for simultaneously training a transcriber and aligning the scores to their corresponding performances, in a fully-automated process. Using this unaligned supervision scheme, complemented by pseudo-labels and pitch-shift augmentation, our method enables training on in-the-wild recordings with unprecedented accuracy and instrumental variety. Using only synthetic data and unaligned supervision, we report SOTA note-level accuracy of the MAPS dataset, and large favorable margins on cross-dataset evaluations. We also demonstrate robustness and ease of use; we report comparable results when training on a small, easily obtainable, self-collected dataset, and we propose alternative labeling to the MusicNet dataset, which we show to be more accurate. Our project page is available at https://benadar293.github.io | ['Amit H. Bermano', 'Ben Maman'] | 2022-04-28 | null | null | null | null | ['music-transcription', 'music-information-retrieval'] | ['music', 'music'] | [ 6.73657954e-01 -8.28615203e-02 -8.83254111e-02 -2.22394243e-01
-1.59992826e+00 -1.26037109e+00 3.24187130e-01 -2.71434337e-01
-2.25434467e-01 4.79516983e-01 3.64507616e-01 1.47292182e-01
-3.17548454e-01 -2.00246111e-01 -7.18109548e-01 -4.43986058e-01
8.98724943e-02 7.68888652e-01 -2.88503468e-01 -1.45148620e-01
1.72738731e-01 -7.81224621e-03 -1.59996712e+00 4.44419593e-01
4.75541025e-01 1.08353651e+00 4.44833897e-02 8.91741037e-01
2.46085405e-01 4.81442571e-01 -8.01329136e-01 -5.24468124e-01
6.25669956e-01 -6.85362101e-01 -8.65761757e-01 1.80012193e-02
8.01278114e-01 -1.66324023e-02 1.75524503e-01 7.91299462e-01
7.45777726e-01 6.58269003e-02 3.92708600e-01 -1.01436436e+00
-3.69858742e-01 1.10940135e+00 -2.52201557e-01 -2.71928400e-01
5.79525650e-01 -2.43872330e-02 1.62841427e+00 -7.64451027e-01
5.95536172e-01 6.43460870e-01 1.06202447e+00 4.27053660e-01
-1.63130569e+00 -9.70601678e-01 -5.07309437e-01 2.30820496e-02
-1.57993412e+00 -7.47008920e-01 8.78932297e-01 -5.00765920e-01
5.60359478e-01 6.33440614e-01 8.57168615e-01 1.24412167e+00
-5.42511523e-01 6.40838742e-01 1.08370304e+00 -6.01842403e-01
7.35533908e-02 -3.31931651e-01 -4.10126776e-01 2.73981988e-01
-4.02513832e-01 -2.00843010e-02 -1.03867793e+00 -1.51606739e-01
7.43708968e-01 -4.51880187e-01 -3.56626540e-01 -1.75178781e-01
-1.65057790e+00 4.48470622e-01 2.45535865e-01 2.96738893e-01
-2.79174089e-01 1.54631555e-01 5.92875004e-01 4.56173241e-01
1.46146521e-01 9.44177330e-01 -4.37902123e-01 -6.93078101e-01
-1.49167395e+00 3.79459590e-01 7.88791478e-01 8.73247445e-01
4.80003148e-01 7.01101497e-02 -1.07203700e-01 1.05930221e+00
-1.97428346e-01 4.39301759e-01 7.03568518e-01 -1.19738138e+00
5.39267659e-01 1.50158152e-01 4.15188409e-02 -6.91228986e-01
-2.04650804e-01 -6.75090194e-01 -4.88472700e-01 2.01902926e-01
6.66050315e-01 8.36406648e-02 -7.16983497e-01 1.93292451e+00
5.84482739e-04 3.12766999e-01 -1.33453235e-01 1.06155086e+00
4.78492409e-01 2.78019607e-01 -3.11158776e-01 -1.39879405e-01
1.11942863e+00 -8.66933405e-01 -4.71659273e-01 6.58501089e-02
5.43518186e-01 -1.11410701e+00 1.57982194e+00 9.49199438e-01
-1.10629463e+00 -7.38686860e-01 -1.10289931e+00 -1.28909603e-01
1.55496836e-01 5.62964797e-01 4.30270970e-01 4.19484466e-01
-8.55860174e-01 1.05369401e+00 -5.67756593e-01 -4.55692522e-02
3.25924158e-01 4.26595569e-01 -4.09314394e-01 3.92773867e-01
-8.45607758e-01 2.72073656e-01 4.47002590e-01 -4.39828774e-03
-8.34116995e-01 -7.74143457e-01 -5.50593197e-01 -1.59759745e-01
3.62500966e-01 -3.92762393e-01 1.66227210e+00 -1.26457489e+00
-1.59111476e+00 1.03131092e+00 2.48864040e-01 -3.92571837e-01
5.46115816e-01 -4.57027227e-01 -3.65324080e-01 1.54676855e-01
2.44027421e-01 7.22324967e-01 8.39800060e-01 -1.00837100e+00
-3.83212835e-01 -2.03474179e-01 -3.02666575e-01 2.76872605e-01
-3.47038090e-01 -1.69717930e-02 -4.18808371e-01 -1.04650962e+00
2.29559526e-01 -1.30928087e+00 2.22617969e-01 -4.27302957e-01
-6.91438854e-01 9.45216492e-02 2.21540764e-01 -9.65179980e-01
1.32528162e+00 -2.37863278e+00 3.88483435e-01 1.88869938e-01
-1.00307867e-01 2.55270898e-02 -1.97089583e-01 3.13568830e-01
-1.63815916e-01 -4.04395685e-02 -3.97278070e-01 -4.78103518e-01
3.16087216e-01 -2.75909193e-02 -6.11789882e-01 1.72663614e-01
-6.72851279e-02 8.42231989e-01 -7.69839883e-01 -3.71427298e-01
-1.01532429e-01 3.04574966e-01 -7.58843124e-01 1.49461925e-01
-2.58578807e-01 8.55398655e-01 -2.07004882e-02 7.37150371e-01
1.43856006e-02 1.02400547e-02 1.48237348e-01 -4.22417194e-01
-1.42376591e-02 8.92313123e-01 -1.38422668e+00 2.50377536e+00
-4.78865534e-01 6.15232170e-01 -2.17549466e-02 -5.32394826e-01
9.87191200e-01 6.33311212e-01 5.45413911e-01 -3.62099290e-01
1.15447901e-02 5.77242553e-01 7.44804665e-02 -2.03297719e-01
6.22678101e-01 -3.26868325e-01 -4.52494413e-01 5.52769423e-01
3.51818770e-01 -5.10698855e-01 1.97236881e-01 -1.90689221e-01
1.10461879e+00 5.29815257e-01 2.62833983e-01 1.04683056e-01
-2.74903458e-02 1.92856237e-01 5.18254697e-01 5.62562704e-01
1.33642927e-01 1.15309024e+00 9.15679783e-02 -1.45379648e-01
-1.30762184e+00 -1.16593611e+00 -1.43558815e-01 1.30971038e+00
-4.40656036e-01 -9.01555538e-01 -7.95475602e-01 -3.94091338e-01
-1.29081592e-01 5.26286483e-01 -3.87331784e-01 1.70924008e-01
-6.54044807e-01 -3.45172644e-01 1.12270033e+00 4.46117491e-01
1.53303057e-01 -1.26383483e+00 -3.39910001e-01 2.89834827e-01
-3.71671379e-01 -7.74054110e-01 -7.24986136e-01 3.59419465e-01
-6.22913659e-01 -8.83895755e-01 -5.89597285e-01 -6.98296666e-01
2.42250087e-03 -2.62550563e-01 1.19730186e+00 -2.48202726e-01
-2.38610297e-01 1.57519758e-01 -3.39863002e-01 -2.56682307e-01
-4.09026444e-01 4.73352879e-01 2.63859868e-01 4.33651544e-02
4.60926589e-04 -1.10517526e+00 -4.81985092e-01 3.21074188e-01
-6.79131389e-01 2.40862668e-01 5.42995691e-01 7.94056296e-01
9.22362268e-01 -3.59246910e-01 7.03639269e-01 -8.10527325e-01
4.43238080e-01 5.46764918e-02 -2.88292110e-01 -1.47219509e-01
-4.25374031e-01 -1.71431974e-01 6.46554291e-01 -6.17584705e-01
-5.53137720e-01 3.00500333e-01 -2.62464166e-01 -7.61608124e-01
-1.96245104e-01 3.71840894e-01 -1.33430704e-01 2.54682630e-01
9.55322027e-01 5.00898845e-02 -2.02110201e-01 -8.41288447e-01
4.76681113e-01 9.32980776e-01 1.18072164e+00 -8.05079162e-01
8.18072557e-01 1.94032803e-01 -1.66190177e-01 -3.64112556e-01
-1.17902517e+00 -4.29105043e-01 -8.51369321e-01 -2.16170147e-01
5.93458354e-01 -9.72579598e-01 -6.31639302e-01 1.36046916e-01
-7.57734001e-01 -4.63908017e-01 -7.18052089e-01 5.11239171e-01
-1.02014732e+00 -9.72419605e-02 -5.77872813e-01 -6.31226242e-01
-4.09117371e-01 -7.14109540e-01 1.38551915e+00 -1.50899529e-01
-9.34031665e-01 -5.03765941e-01 4.04420048e-01 7.05161095e-01
8.74503404e-02 3.47111881e-01 5.40116668e-01 -7.57675052e-01
-3.90121639e-01 -2.97343850e-01 2.46205509e-01 4.97799933e-01
2.14069784e-01 -2.06327572e-01 -1.43773925e+00 -1.93047315e-01
-2.05348983e-01 -8.39820504e-01 6.11378193e-01 -3.10410764e-02
1.06449282e+00 -3.31974655e-01 2.61697114e-01 8.44936788e-01
1.02124774e+00 -1.74015537e-01 4.24014866e-01 3.62014830e-01
8.48784149e-01 4.48020071e-01 6.11529768e-01 3.95422786e-01
3.05548254e-02 1.05099618e+00 1.15284519e-02 1.26708046e-01
-3.97852391e-01 -6.60859585e-01 3.86833012e-01 1.33565354e+00
-3.29927027e-01 1.06311403e-01 -7.86390245e-01 5.96347272e-01
-1.56657732e+00 -1.06667471e+00 4.30286527e-02 2.36415672e+00
1.32125747e+00 6.28609955e-02 4.63204920e-01 6.40210986e-01
5.37498295e-01 -1.19875213e-02 -4.75254297e-01 -9.62073728e-02
-2.37725899e-01 5.74103773e-01 1.91426963e-01 2.37408236e-01
-1.10410357e+00 8.04443836e-01 6.27097368e+00 1.08693206e+00
-1.15942204e+00 1.95330590e-01 2.01581508e-01 -6.78882957e-01
-1.45394519e-01 4.79082055e-02 -4.16754931e-01 3.67832780e-01
1.04772317e+00 2.33303215e-02 9.33511734e-01 6.52583182e-01
1.44563958e-01 4.26081121e-01 -1.37270188e+00 1.29569125e+00
2.17242405e-01 -1.15126646e+00 -1.03245050e-01 -2.13995017e-02
6.34742379e-01 5.49295954e-02 1.99810881e-03 2.69684434e-01
-4.91593927e-02 -1.01222026e+00 1.28379941e+00 4.36501175e-01
1.30148315e+00 -6.06586158e-01 3.19038928e-01 8.85913074e-02
-1.27600622e+00 -3.29494849e-02 1.62008628e-01 -2.66780436e-01
2.38113075e-01 3.25825065e-01 -1.00859249e+00 5.55595756e-01
6.96437061e-01 9.15307820e-01 -6.40557170e-01 9.33930874e-01
-4.01493669e-01 1.05254710e+00 -3.96370113e-01 3.80481571e-01
-2.91856289e-01 -1.40234992e-01 6.61663055e-01 1.12263989e+00
5.82111835e-01 -3.81196231e-01 2.73336411e-01 8.30435872e-01
-2.37371325e-01 3.75413746e-01 -4.28509742e-01 -2.12761357e-01
5.38171589e-01 1.32837152e+00 -4.86350805e-01 -1.88381448e-01
1.39689028e-01 1.04202533e+00 3.51556301e-01 2.37604771e-02
-7.26613581e-01 -4.00830865e-01 5.37831545e-01 1.14654601e-01
1.87373638e-01 -1.24977499e-01 -6.23684406e-01 -9.95554984e-01
2.85859674e-01 -1.06257534e+00 3.21417272e-01 -1.02262914e+00
-1.13881862e+00 6.44725084e-01 -4.56214726e-01 -1.82674026e+00
-6.26563370e-01 -3.30368221e-01 -3.00937295e-01 7.41704047e-01
-8.79892468e-01 -1.20076334e+00 -1.38947353e-01 3.08976442e-01
4.33356553e-01 -2.59810239e-01 1.27695203e+00 6.00677073e-01
-1.44162238e-01 7.19189286e-01 -1.54784378e-02 3.12963754e-01
1.19558084e+00 -1.39676905e+00 2.13210583e-01 3.25696409e-01
1.07164073e+00 3.95023048e-01 6.99926257e-01 -3.74927998e-01
-1.02920270e+00 -9.28941667e-01 6.76957250e-01 -7.57409871e-01
7.78270006e-01 -7.07263291e-01 -7.33558416e-01 7.99264431e-01
-1.61406770e-02 -4.94492888e-01 1.01383257e+00 3.84678215e-01
-4.63818729e-01 -8.48154053e-02 -5.37822604e-01 4.76987004e-01
1.24519384e+00 -9.53517437e-01 -7.21133828e-01 1.75725728e-01
5.85146785e-01 -5.64996839e-01 -1.14863133e+00 3.38305354e-01
9.47169662e-01 -7.88502216e-01 7.86086738e-01 -2.84885585e-01
4.40716386e-01 -5.58760703e-01 -3.80117297e-01 -1.32277977e+00
-1.38652161e-01 -1.00357187e+00 3.09225440e-01 1.59552050e+00
5.69776535e-01 -3.48256007e-02 7.24997044e-01 3.40092704e-02
-4.53956753e-01 -2.71329373e-01 -9.51650918e-01 -8.34406912e-01
-2.73603201e-01 -8.67641509e-01 6.51930153e-01 1.14370596e+00
1.68461204e-01 6.02037072e-01 -5.55260062e-01 -1.12602249e-01
5.34327745e-01 4.29838121e-01 1.01029313e+00 -1.32637215e+00
-8.97500336e-01 -4.96705651e-01 -4.78251129e-01 -6.96756124e-01
6.69942498e-02 -1.33094800e+00 1.68333292e-01 -9.22669530e-01
3.70259546e-02 -5.10872364e-01 -4.15427536e-01 8.01621556e-01
1.86021760e-01 1.09577096e+00 4.26886052e-01 6.45632088e-01
-6.74963117e-01 3.76796365e-01 1.04931843e+00 -1.26862779e-01
-4.36998338e-01 -6.01227349e-03 -6.28757179e-01 6.02447867e-01
8.57189357e-01 -5.55445254e-01 -2.52360314e-01 -3.21653157e-01
4.19876188e-01 6.16307706e-02 3.95473182e-01 -1.37781310e+00
5.98275028e-02 3.01914692e-01 2.23752111e-01 -3.27829182e-01
6.69808030e-01 -4.02157336e-01 6.47158504e-01 -1.31473526e-01
-6.61916852e-01 -8.41045305e-02 2.30883598e-01 2.64183491e-01
-4.29024577e-01 -2.36274779e-01 3.76148999e-01 4.83322032e-02
-2.46403456e-01 -1.30163044e-01 2.29383379e-01 1.89648688e-01
3.43146473e-01 -1.98482215e-01 6.62856624e-02 -4.05308008e-01
-1.02176762e+00 -2.57066846e-01 6.25635445e-01 4.17277545e-01
1.25996917e-01 -1.61900425e+00 -8.21900606e-01 2.83364892e-01
3.14463735e-01 7.08059268e-03 7.48719722e-02 8.23536932e-01
-2.75927842e-01 1.05033115e-01 -1.81589633e-01 -7.30070174e-01
-1.30619860e+00 1.44436657e-01 1.57676518e-01 -1.97343174e-02
-6.66931629e-01 7.33490884e-01 -1.66386530e-01 -7.03597128e-01
2.74442703e-01 -2.51849890e-01 1.05711475e-01 1.67380676e-01
4.09321785e-01 9.61624086e-02 2.58466691e-01 -6.25358701e-01
-1.76043753e-02 6.02575183e-01 3.24644089e-01 -7.18543530e-01
1.30247426e+00 1.42302185e-01 7.55452290e-02 1.04269564e+00
9.02817428e-01 5.40153861e-01 -1.24928665e+00 -5.34617566e-02
1.29036054e-01 -4.72854674e-01 -9.61121693e-02 -1.13895154e+00
-6.42775953e-01 6.65163517e-01 4.20351028e-01 2.18685940e-01
1.21175528e+00 4.02532816e-02 7.20270634e-01 3.41601044e-01
3.79140496e-01 -1.16554785e+00 4.48901812e-03 2.62637496e-01
1.05665410e+00 -8.47828567e-01 -3.11956555e-01 -1.22655720e-01
-6.92974508e-01 8.58162642e-01 1.70915395e-01 1.21625606e-02
2.11616099e-01 2.99307466e-01 4.23043877e-01 6.72900723e-03
-5.63728571e-01 -1.10525712e-01 5.19365132e-01 3.49755764e-01
7.23929703e-01 3.43611568e-01 1.29940689e-01 1.00729609e+00
-1.28541613e+00 7.55436793e-02 3.30178589e-01 5.81729174e-01
-3.47919203e-02 -1.23487818e+00 -5.33421338e-01 4.23282027e-01
-6.07042789e-01 -2.55040824e-01 -6.73309922e-01 4.66942608e-01
2.68307477e-01 8.09147000e-01 1.48221711e-02 -7.20841467e-01
4.67234164e-01 5.74290335e-01 6.66655183e-01 -6.44348562e-01
-8.18912327e-01 4.64769095e-01 3.27191561e-01 -4.12794918e-01
-4.23759788e-01 -8.96039128e-01 -1.15070593e+00 5.19317463e-02
-1.61370918e-01 2.65008807e-01 7.14608431e-01 7.74744570e-01
3.66549432e-01 6.47511244e-01 5.38898051e-01 -1.24109101e+00
-4.74486053e-01 -1.30139518e+00 -8.09592307e-01 8.31740439e-01
2.17950478e-01 -4.19204980e-01 -3.29733729e-01 4.15794641e-01] | [15.830445289611816, 5.348630428314209] |
c7c247f7-06a9-44c7-80f5-afaeb60f8190 | gait-recognition-in-the-wild-a-benchmark-1 | 2205.02692 | null | https://arxiv.org/abs/2205.02692v1 | https://arxiv.org/pdf/2205.02692v1.pdf | Gait Recognition in the Wild: A Benchmark | Gait benchmarks empower the research community to train and evaluate high-performance gait recognition systems. Even though growing efforts have been devoted to cross-view recognition, academia is restricted by current existing databases captured in the controlled environment. In this paper, we contribute a new benchmark for Gait REcognition in the Wild (GREW). The GREW dataset is constructed from natural videos, which contains hundreds of cameras and thousands of hours streams in open systems. With tremendous manual annotations, the GREW consists of 26K identities and 128K sequences with rich attributes for unconstrained gait recognition. Moreover, we add a distractor set of over 233K sequences, making it more suitable for real-world applications. Compared with prevailing predefined cross-view datasets, the GREW has diverse and practical view variations, as well as more natural challenging factors. To the best of our knowledge, this is the first large-scale dataset for gait recognition in the wild. Equipped with this benchmark, we dissect the unconstrained gait recognition problem. Representative appearance-based and model-based methods are explored, and comprehensive baselines are established. Experimental results show (1) The proposed GREW benchmark is necessary for training and evaluating gait recognizer in the wild. (2) For state-of-the-art gait recognition approaches, there is a lot of room for improvement. (3) The GREW benchmark can be used as effective pre-training for controlled gait recognition. Benchmark website is https://www.grew-benchmark.org/. | ['Jie zhou', 'Jiwen Lu', 'Dalong Du', 'Guan Huang', 'Jiankang Deng', 'JunJie Huang', 'Tian Yang', 'Xianda Guo', 'Zheng Zhu'] | 2022-05-05 | gait-recognition-in-the-wild-a-benchmark | http://openaccess.thecvf.com//content/ICCV2021/html/Zhu_Gait_Recognition_in_the_Wild_A_Benchmark_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Zhu_Gait_Recognition_in_the_Wild_A_Benchmark_ICCV_2021_paper.pdf | iccv-2021-1 | ['gait-recognition-in-the-wild', 'gait-recognition'] | ['computer-vision', 'computer-vision'] | [-1.29667297e-01 -7.84648359e-01 -2.47289777e-01 -1.27485722e-01
-5.91086924e-01 -3.31334531e-01 1.86231777e-01 -7.08893359e-01
-2.32647792e-01 4.45668101e-01 1.07362986e-01 2.93703556e-01
4.33581352e-01 -5.74780345e-01 -3.72414291e-01 -1.07694244e+00
-4.37225431e-01 3.98965716e-01 3.42039227e-01 -3.38346899e-01
-1.22815534e-01 2.86888391e-01 -1.82835424e+00 1.47730425e-01
5.98303854e-01 1.12627316e+00 3.28922197e-02 6.16194725e-01
4.87052500e-01 6.47203505e-01 -6.05792582e-01 -6.71064973e-01
3.15655947e-01 -3.00784618e-01 -5.48418939e-01 3.89894724e-01
7.99140394e-01 -6.17191374e-01 -7.35657096e-01 7.05996990e-01
9.73994493e-01 -9.76061895e-02 4.48593587e-01 -1.62421560e+00
-7.72084773e-01 -4.94397134e-02 -5.21487474e-01 2.76051521e-01
6.09023750e-01 7.11239517e-01 7.62802958e-01 -1.02497101e+00
5.62029302e-01 9.18607354e-01 7.75205135e-01 6.87189102e-01
-7.55515456e-01 -7.32822955e-01 3.66440602e-02 5.98771989e-01
-1.23986638e+00 -5.36250353e-01 8.07047367e-01 -3.05695564e-01
9.85976040e-01 2.36641854e-01 1.05441833e+00 1.90502608e+00
6.06833212e-02 1.07617581e+00 7.37381995e-01 -1.53317705e-01
8.90483707e-02 -5.63957870e-01 8.21576174e-03 7.31856704e-01
4.72185999e-01 2.92023104e-02 -7.32108176e-01 -1.56333502e-02
7.80694962e-01 -1.18400499e-01 -4.46817279e-01 -6.53579891e-01
-1.49430740e+00 2.58399546e-01 1.03906617e-01 1.57711804e-01
-9.31694806e-02 3.45270187e-02 8.76166403e-01 4.65172380e-01
7.35764503e-02 -1.00732781e-01 -1.35558203e-01 -6.69042826e-01
-7.82556117e-01 2.25073546e-01 7.12054253e-01 1.08809304e+00
4.39745903e-01 5.62161744e-01 9.22444388e-02 1.08067822e+00
9.79732946e-02 1.09757578e+00 8.59351575e-01 -6.24341607e-01
7.99715817e-01 4.64359760e-01 -1.12333156e-01 -1.16714692e+00
-1.09358355e-01 -6.48132563e-02 -1.10943556e+00 -1.87554896e-01
5.94319940e-01 2.47363940e-01 -9.69053924e-01 1.34926176e+00
1.48682995e-02 4.94729161e-01 -2.21769116e-03 1.02418077e+00
7.57771790e-01 2.06971824e-01 -2.56905168e-01 -1.42095024e-02
1.38172948e+00 -1.04762483e+00 -4.26030904e-01 -2.21843392e-01
6.89100564e-01 -6.29839420e-01 1.28914118e+00 4.20409530e-01
-6.45896733e-01 -5.74618578e-01 -1.04276347e+00 4.05753404e-01
-2.98702806e-01 3.91304046e-01 5.13983786e-01 8.93975973e-01
-9.33184922e-01 4.32036847e-01 -1.17297614e+00 -9.45866704e-01
2.94651508e-01 1.27212450e-01 -8.65140140e-01 -4.63780344e-01
-1.11959434e+00 5.82012177e-01 1.35210156e-01 3.65280420e-01
-8.06992710e-01 -2.50955373e-01 -1.12032902e+00 -4.06923085e-01
4.18633133e-01 -8.15789878e-01 8.59204471e-01 -4.70181972e-01
-1.29517663e+00 1.13737082e+00 5.05144633e-02 -2.25135446e-01
7.69755960e-01 -2.58296520e-01 -9.08174157e-01 1.97004303e-01
3.95008296e-01 1.68298915e-01 1.05546165e+00 -9.73485112e-01
-3.33097070e-01 -4.31339949e-01 -3.23944002e-01 -5.02112918e-02
-4.51947629e-01 4.21277583e-02 -1.04003131e+00 -8.55166972e-01
-1.69475958e-01 -1.08160949e+00 9.94342789e-02 -7.43246451e-02
-3.54315937e-01 2.12619349e-01 1.07733095e+00 -7.21167147e-01
1.28479707e+00 -1.99888992e+00 -1.24150217e-01 -6.20347336e-02
3.24808322e-02 4.84503120e-01 -2.67433077e-01 5.71430624e-01
-1.14448249e-01 -1.36572301e-01 -1.20705985e-01 -3.37841928e-01
1.70066983e-01 4.91430193e-01 -8.20715576e-02 5.58432877e-01
2.74511874e-01 8.23785543e-01 -8.63515675e-01 -5.23466766e-01
4.23111022e-01 1.60656005e-01 -4.67851073e-01 3.27349216e-01
5.27655900e-01 4.87884432e-01 -4.45475429e-01 1.42365825e+00
8.23434532e-01 -1.94219187e-01 -2.34309323e-02 -5.55112302e-01
3.23734909e-01 -4.20660734e-01 -1.31681097e+00 1.75358880e+00
1.02629373e-02 3.12384486e-01 -1.77294955e-01 -1.27224374e+00
1.02790892e+00 1.56103581e-01 8.73915792e-01 -6.76489651e-01
3.66488891e-03 4.08810556e-01 -1.60492346e-01 -8.64761055e-01
4.72091883e-01 3.96469742e-01 -2.26894706e-01 9.14373621e-02
3.98739845e-01 3.13408434e-01 7.02947438e-01 -3.35265025e-02
1.45820093e+00 3.36562306e-01 2.46275261e-01 1.54328883e-01
5.13909757e-01 -1.10379472e-01 8.86028230e-01 5.58936298e-01
-9.19338346e-01 9.62210000e-01 1.29950091e-01 -7.59843051e-01
-1.02987051e+00 -1.33407366e+00 9.19907540e-02 6.93415523e-01
2.48010978e-01 -6.59128070e-01 -5.19906998e-01 -6.21295571e-01
-1.83872469e-02 -3.26141268e-01 -5.95841348e-01 -1.05883859e-01
-9.81487393e-01 -9.46414709e-01 1.11139262e+00 1.00438976e+00
1.16649270e+00 -9.00547385e-01 -6.44411802e-01 3.55257876e-02
-6.80909812e-01 -1.60086775e+00 -9.22629118e-01 -2.85176098e-01
-7.19998956e-01 -1.45498800e+00 -1.13483381e+00 -8.48276973e-01
3.20446491e-01 4.22482699e-01 1.38258004e+00 7.73491561e-02
-4.86037552e-01 8.02659392e-01 -6.09654307e-01 1.68653354e-01
7.76245119e-03 -1.15096651e-01 5.83093405e-01 1.75670981e-01
4.70658064e-01 -7.14372039e-01 -7.79029369e-01 9.44304824e-01
-4.08615112e-01 -2.28525296e-01 6.10050917e-01 1.12570608e+00
4.92174029e-01 -3.28446031e-01 3.55999440e-01 -1.00897424e-01
2.78735489e-01 -2.80809373e-01 -1.69104353e-01 3.75042051e-01
-2.70573974e-01 -2.77179331e-01 6.96890116e-01 -4.99553442e-01
-7.71829903e-01 4.72400784e-02 -1.10546485e-01 -8.69713485e-01
-4.52956438e-01 2.61150002e-01 -4.38479364e-01 -1.56773716e-01
4.15067911e-01 6.56962037e-01 1.23084590e-01 -4.43808526e-01
-6.43444732e-02 4.93390143e-01 8.08648825e-01 -9.63422418e-01
8.13663781e-01 4.38635498e-01 -9.61574167e-02 -1.09986925e+00
-1.23554930e-01 -7.32498765e-01 -8.42694581e-01 -6.34876430e-01
6.25776947e-01 -1.09319854e+00 -5.13361573e-01 1.19495702e+00
-4.70062762e-01 -2.21502498e-01 -1.04663745e-01 5.07754922e-01
-9.00177598e-01 8.12307060e-01 -7.74282157e-01 -4.88716185e-01
-3.99652511e-01 -1.26322687e+00 1.34623706e+00 -1.26970172e-01
-4.54079919e-02 -7.24561870e-01 1.26160040e-01 6.47604227e-01
1.78363293e-01 4.65015650e-01 7.96000957e-02 -3.23666513e-01
-3.23282331e-01 -3.71824056e-01 5.21723442e-02 3.55606347e-01
3.36841196e-01 3.31569493e-01 -8.72274518e-01 -4.89289552e-01
-3.75295699e-01 -6.87577605e-01 8.32293510e-01 8.57730955e-02
8.25263023e-01 2.79960409e-02 -2.68340558e-01 7.81089962e-01
1.11547220e+00 1.18943065e-01 9.85776544e-01 5.96569598e-01
9.28572357e-01 2.05520868e-01 7.81535268e-01 6.00420237e-01
5.00197470e-01 9.71351385e-01 1.42120108e-01 6.50104135e-02
-1.11005925e-01 -1.20425232e-01 7.74124861e-01 1.35407722e+00
-6.49194241e-01 -4.90944609e-02 -1.17599237e+00 5.16932905e-01
-1.91261530e+00 -1.33573365e+00 6.87124804e-02 2.09275746e+00
2.84760237e-01 -1.00204773e-01 5.99471390e-01 3.75263244e-01
6.88480914e-01 2.92587847e-01 -5.67133665e-01 1.66575521e-01
-3.11241359e-01 8.82009938e-02 2.66035974e-01 -4.09200013e-01
-1.51144087e+00 6.73864067e-01 5.83535004e+00 8.89492989e-01
-1.14027095e+00 -2.50764728e-01 2.33404279e-01 5.09852096e-02
7.24637508e-01 -4.77514476e-01 -6.39738381e-01 6.97365820e-01
7.58728623e-01 -1.51945442e-01 -3.89128700e-02 1.03063715e+00
1.29033238e-01 2.13122174e-01 -9.90300000e-01 1.51808131e+00
3.25472236e-01 -9.20351982e-01 1.87786326e-01 2.33838946e-01
3.40464354e-01 -2.60624476e-02 -2.42821779e-02 5.16121268e-01
-7.54949674e-02 -7.46493399e-01 4.39815909e-01 6.10705316e-02
1.01835191e+00 -4.51215386e-01 8.66990864e-01 5.00594592e-03
-2.11481905e+00 9.45553482e-02 -3.76271039e-01 7.40335435e-02
1.50840610e-01 2.58674204e-01 -1.28377974e-01 1.05646598e+00
1.02726007e+00 1.49633169e+00 -7.54879534e-01 1.14457512e+00
1.38700962e-01 7.14883924e-01 -2.59781182e-01 2.70055324e-01
3.40714343e-02 -2.08305195e-01 3.59509349e-01 1.33787453e+00
5.75164616e-01 -2.07849517e-01 5.63614488e-01 2.91013092e-01
3.06369662e-02 3.40574374e-03 -8.47040951e-01 -7.14683682e-02
2.46536463e-01 1.06597471e+00 -7.02499747e-01 -2.71868944e-01
-6.69218779e-01 1.27024233e+00 1.34077370e-01 3.15499187e-01
-1.05775249e+00 -4.43624891e-02 1.02348077e+00 -1.23872273e-01
3.37531269e-01 -2.97292799e-01 2.51548111e-01 -2.05560255e+00
4.82841849e-01 -1.14371634e+00 6.04741156e-01 -5.82512259e-01
-1.68363380e+00 3.91437292e-01 -1.07898355e-01 -2.07014489e+00
-1.97385520e-01 -9.22685444e-01 -6.55944347e-01 3.28442931e-01
-1.15207231e+00 -1.52313900e+00 -6.92926228e-01 1.00480258e+00
5.85127950e-01 -4.60210145e-01 9.62669611e-01 8.34993303e-01
-1.01933849e+00 8.69124770e-01 6.55139461e-02 8.86339247e-01
1.07411206e+00 -1.04477537e+00 8.12868655e-01 1.03520918e+00
2.84508951e-02 3.63222867e-01 3.74404460e-01 -5.38510442e-01
-2.00538087e+00 -1.15493321e+00 2.15201870e-01 -6.21555746e-01
8.26651752e-01 -3.01743835e-01 -8.77000034e-01 5.43290734e-01
-1.51815191e-01 4.82706070e-01 8.25567544e-01 -1.73084855e-01
-5.44967830e-01 -1.96616903e-01 -8.34197819e-01 5.23365021e-01
1.46892405e+00 -2.37812266e-01 -4.19834793e-01 -3.81947346e-02
-2.14222935e-03 -4.84480321e-01 -1.14324832e+00 5.32092869e-01
1.04771066e+00 -9.99390066e-01 1.26649523e+00 -4.60861415e-01
3.57810766e-01 -3.30952227e-01 -4.42518294e-01 -1.34525108e+00
-2.72707164e-01 -2.28715733e-01 -3.79865289e-01 1.28045475e+00
-3.16970766e-01 -4.82730120e-01 1.21269882e+00 4.16556239e-01
-2.18465537e-01 -6.78908765e-01 -9.86016810e-01 -1.46071815e+00
-9.97393951e-02 -4.38243687e-01 4.23108697e-01 9.74922955e-01
1.86288469e-02 -1.90827809e-02 -9.15728450e-01 3.08436323e-02
9.93151784e-01 1.25520408e-01 1.23334038e+00 -9.18118775e-01
-1.90641001e-01 -2.83017009e-01 -1.34058118e+00 -9.25826311e-01
-4.74851467e-02 -5.34179270e-01 -1.83954135e-01 -1.17500925e+00
2.02219054e-01 -7.45872259e-02 -2.33350709e-01 5.43819487e-01
-1.97231576e-01 4.45403963e-01 2.96582878e-01 4.68081325e-01
-7.44440615e-01 8.07833374e-01 1.31646180e+00 -4.39130485e-01
3.13411087e-01 -1.76119477e-01 -1.57634884e-01 5.36262453e-01
9.13899660e-01 8.05365294e-02 -3.07054073e-01 -1.98137686e-01
-6.78665757e-01 -1.78857043e-01 4.42712575e-01 -1.47755158e+00
1.13395259e-01 -1.29024491e-01 3.97946358e-01 -4.50949699e-01
4.88254428e-01 -6.86678469e-01 1.68229118e-01 5.07540524e-01
5.40811002e-01 3.13450933e-01 9.08681229e-02 5.15050232e-01
-5.72763920e-01 4.41870064e-01 5.84472775e-01 -2.00899810e-01
-1.48749685e+00 6.67525113e-01 -3.63997221e-01 4.85009432e-01
9.88169849e-01 -8.84710193e-01 -3.80926639e-01 -2.66488194e-01
-7.29612231e-01 3.19673866e-01 6.26987815e-01 5.94090700e-01
8.84053469e-01 -1.69380438e+00 -7.05119371e-01 4.13019478e-01
7.30610371e-01 -4.91396457e-01 3.61556649e-01 9.46547151e-01
-5.57335436e-01 1.70831025e-01 -7.06292927e-01 -8.94495845e-01
-1.45755529e+00 3.34364265e-01 4.24550831e-01 -4.41916227e-01
-8.59986603e-01 3.19241524e-01 -7.92232752e-02 -4.23649043e-01
1.66124806e-01 -1.71203196e-01 -5.31989522e-02 5.57504818e-02
6.22887909e-01 5.55030644e-01 2.53890634e-01 -1.10820353e+00
-6.53048575e-01 8.37282300e-01 1.41115040e-01 3.95541728e-01
1.08472729e+00 -4.27761376e-02 3.57485801e-01 4.10290748e-01
1.11133051e+00 -5.07876396e-01 -1.14101410e+00 1.03512406e-01
-1.84584975e-01 -8.54806662e-01 -7.31691301e-01 -3.31045479e-01
-1.37344849e+00 8.37528348e-01 7.89140105e-01 -2.32336879e-01
1.29490507e+00 -4.74044651e-01 9.90923166e-01 5.07834435e-01
1.04853666e+00 -1.05193400e+00 5.17314911e-01 5.90615213e-01
8.16622615e-01 -1.40911412e+00 -3.32928956e-01 -4.96683061e-01
-7.33897150e-01 1.12345576e+00 1.11661935e+00 -1.28167361e-01
3.79490942e-01 4.97196645e-01 1.66490465e-01 2.22924761e-02
-5.74159920e-01 -2.57658780e-01 5.06466106e-02 1.10207176e+00
3.70623082e-01 2.42118835e-02 1.31615683e-01 5.21985173e-01
-1.74956903e-01 2.72525668e-01 3.71291667e-01 1.09553599e+00
-8.93248022e-02 -1.11309421e+00 -5.38082063e-01 5.86033762e-01
-1.17322013e-01 2.65290797e-01 -1.19515568e-01 8.88399184e-01
4.96249087e-02 8.95353794e-01 -3.23595583e-01 -8.18089962e-01
6.94055974e-01 1.00767419e-01 6.54275894e-01 -1.06059551e-01
-2.74236411e-01 -1.69602662e-01 3.26698035e-01 -1.05605936e+00
-4.85437065e-01 -7.60497034e-01 -7.13806570e-01 -6.39525831e-01
-6.40824810e-02 -1.75592929e-01 -7.79450610e-02 5.52384555e-01
3.98822546e-01 2.55339056e-01 3.97923380e-01 -1.23404872e+00
-5.52697182e-01 -7.11197615e-01 -8.92385781e-01 8.93503070e-01
7.29790851e-02 -1.10460341e+00 -2.07037732e-01 4.67630118e-01] | [14.303915023803711, 1.4065386056900024] |
19cccf5b-1581-4200-886c-ec7dfda984e2 | o2d2-out-of-distribution-detector-to-capture | 2106.15825 | null | https://arxiv.org/abs/2106.15825v3 | https://arxiv.org/pdf/2106.15825v3.pdf | O2D2: Out-Of-Distribution Detector to Capture Undecidable Trials in Authorship Verification | The PAN 2021 authorship verification (AV) challenge is part of a three-year strategy, moving from a cross-topic/closed-set AV task to a cross-topic/open-set AV task over a collection of fanfiction texts. In this work, we present a novel hybrid neural-probabilistic framework that is designed to tackle the challenges of the 2021 task. Our system is based on our 2020 winning submission, with updates to significantly reduce sensitivities to topical variations and to further improve the system's calibration by means of an uncertainty-adaptation layer. Our framework additionally includes an out-of-distribution detector (O2D2) for defining non-responses. Our proposed system outperformed all other systems that participated in the PAN 2021 AV task. | ['Dorothea Kolossa', 'Robert M. Nickel', 'Benedikt Boenninghoff'] | 2021-06-30 | null | null | null | null | ['authorship-verification'] | ['natural-language-processing'] | [ 1.16791494e-01 1.17284860e-02 8.33898708e-02 -5.03414989e-01
-1.30373871e+00 -7.55222976e-01 1.11343014e+00 1.99040085e-01
-5.39088607e-01 6.95941031e-01 2.33780518e-01 -1.48708194e-01
9.43811331e-03 -3.20437223e-01 -6.78481340e-01 -2.11512119e-01
4.46610063e-01 7.04065323e-01 3.28104943e-01 6.22772202e-02
4.34634775e-01 2.22607672e-01 -1.25655746e+00 5.56771994e-01
6.63425326e-01 8.24317813e-01 -2.17255697e-01 8.46817672e-01
6.01009242e-02 4.32767510e-01 -1.04308987e+00 -8.33486617e-01
1.10456251e-01 6.08535372e-02 -8.60907853e-01 -7.40411758e-01
9.25283134e-01 3.38866487e-02 -7.64262378e-02 8.79122376e-01
8.45956326e-01 7.11276606e-02 9.72505093e-01 -1.37041497e+00
-8.08659434e-01 1.17196095e+00 -5.66312790e-01 9.19067673e-03
3.94460946e-01 -1.76874325e-01 1.25675833e+00 -1.18436551e+00
6.37375176e-01 1.35501158e+00 9.49012876e-01 6.54083610e-01
-1.19875813e+00 -1.03714311e+00 1.67373493e-01 2.50274539e-01
-1.50912821e+00 -5.18543184e-01 7.73163974e-01 -7.01273739e-01
8.76859725e-01 8.96372274e-02 1.22735508e-01 1.97319281e+00
-3.93346127e-04 1.05611396e+00 1.07533801e+00 -6.21474922e-01
1.30985752e-01 4.80856210e-01 6.01129353e-01 -2.11939830e-02
2.94299901e-01 2.13438198e-02 -8.48724961e-01 -4.13303316e-01
-4.51493673e-02 -7.23138213e-01 -3.05599600e-01 -5.29324971e-02
-9.28043306e-01 7.26652384e-01 -1.55438796e-01 3.50976139e-01
-6.08389042e-02 1.77232668e-01 6.45203292e-01 -3.02887727e-02
3.81099254e-01 6.88301206e-01 -6.58140600e-01 -3.21692258e-01
-1.46315730e+00 6.37013733e-01 1.10371673e+00 8.89037907e-01
1.03602618e-01 -1.91399276e-01 -1.00888956e+00 9.74372387e-01
6.43102705e-01 4.25712407e-01 2.73188204e-01 -7.98468709e-01
6.01319790e-01 3.22482526e-01 1.56696677e-01 -8.72585237e-01
-4.33331609e-01 -4.71789986e-01 -6.12166762e-01 -1.73192605e-01
4.92765248e-01 -2.67570585e-01 -8.30744088e-01 1.78856456e+00
2.90851295e-02 1.97081864e-02 -1.91056356e-01 6.72266424e-01
1.08114815e+00 4.56734538e-01 1.80220708e-01 1.88770071e-01
1.38040900e+00 -7.80701637e-01 -7.74953425e-01 2.69346058e-01
3.43063116e-01 -8.65807056e-01 7.41799533e-01 7.47896194e-01
-6.29240036e-01 -6.59300089e-01 -1.11994433e+00 -1.72267810e-01
-5.79878032e-01 3.37074995e-01 8.52784608e-03 1.15260327e+00
-9.44604218e-01 5.95306754e-01 -1.22398473e-01 -1.32005110e-01
5.15650392e-01 2.31957495e-01 -3.32649022e-01 2.18512371e-01
-1.77703726e+00 9.56783056e-01 5.01597226e-01 -5.15915714e-02
-7.30835855e-01 -1.16552401e+00 -6.84119046e-01 -4.43423260e-03
3.07385325e-01 -4.11835819e-01 1.36688673e+00 -4.27838773e-01
-1.77451932e+00 8.79837811e-01 -5.26979044e-02 -5.96364737e-01
8.64971578e-01 -5.82111895e-01 -6.32787287e-01 -2.04614371e-01
2.02685464e-02 5.39074063e-01 7.33659744e-01 -1.07150483e+00
-4.39488322e-01 -7.82725513e-02 -5.73904991e-01 -2.62093127e-01
-3.31956744e-01 2.57041156e-01 -5.90035021e-01 -9.02656913e-01
-6.09942257e-01 -9.27781045e-01 3.55870187e-01 -4.19817001e-01
-6.72462106e-01 -8.68721128e-01 9.16319728e-01 -8.70692015e-01
1.53483212e+00 -1.92959869e+00 -7.95416012e-02 3.39094996e-01
2.08386958e-01 3.77365172e-01 -4.86986302e-02 4.05250400e-01
4.99828681e-02 2.65154272e-01 -1.54108882e-01 -9.65456784e-01
4.79705513e-01 -3.50507468e-01 -3.49184483e-01 3.30206424e-01
1.37625903e-01 6.86471939e-01 -5.69345713e-01 -5.68592906e-01
-2.32973799e-01 5.33673227e-01 -2.23086491e-01 -3.09080835e-02
-2.15895087e-01 -1.14198066e-01 -6.68481514e-02 6.73104227e-01
9.51146007e-01 -1.21566445e-01 9.29497257e-02 5.18313348e-02
-9.22099203e-02 3.56525898e-01 -1.36300182e+00 1.49626410e+00
-7.67726451e-02 7.99692869e-01 -4.15732153e-02 -5.00532806e-01
1.19275749e+00 3.03994805e-01 3.25092673e-01 -2.22954780e-01
3.91254812e-01 3.40212435e-01 -5.02315275e-02 1.35964565e-02
9.59614635e-01 3.17867726e-01 -5.77684402e-01 3.37378740e-01
4.83741373e-01 -3.19905840e-02 1.83913529e-01 3.98287714e-01
7.48859823e-01 3.71984750e-01 5.64828962e-02 -5.83389163e-01
8.42637002e-01 -4.79016453e-01 6.16357803e-01 1.27058029e+00
-4.66152728e-01 1.07070243e+00 7.92345047e-01 4.31726547e-03
-9.97533977e-01 -9.10365403e-01 -5.52127302e-01 7.78064787e-01
-3.27627510e-01 -6.42521739e-01 -6.52300119e-01 -9.14136589e-01
3.00485790e-01 9.63977575e-01 -8.96088302e-01 2.16802239e-01
-2.68307775e-01 -6.12974346e-01 1.27096331e+00 3.98222744e-01
3.01356643e-01 -7.55181849e-01 -2.61389226e-01 1.34823322e-01
-1.00101426e-01 -9.80383575e-01 -7.12149143e-01 1.91012129e-01
-2.94575691e-02 -1.03650081e+00 -1.08468807e+00 -4.40773875e-01
-2.36345515e-01 -5.22671103e-01 9.73766863e-01 -7.31921494e-01
-2.02658307e-02 2.63970912e-01 -2.97551900e-01 -9.31109250e-01
-5.83398581e-01 4.72219527e-01 1.35111466e-01 6.50910959e-02
6.03052258e-01 -1.02269631e-02 -1.07949197e-01 8.10506046e-02
-3.77293944e-01 -4.05673265e-01 3.24917287e-01 1.16230249e+00
2.92388737e-01 -6.42593265e-01 7.40191281e-01 -1.10650170e+00
1.01259089e+00 -3.87611836e-01 -8.07984293e-01 4.54321504e-01
-1.05365324e+00 -2.73192767e-03 3.85769606e-01 -5.82325041e-01
-1.07765019e+00 -3.09174601e-02 -2.13615000e-01 -4.83229369e-01
1.21030085e-01 3.79533947e-01 -4.70388860e-01 1.99755892e-01
6.51473999e-01 -1.40101537e-01 -2.36498609e-01 -6.12924814e-01
4.33558106e-01 1.21118641e+00 6.99323893e-01 -8.48904312e-01
6.30474210e-01 -2.75064409e-01 -1.12536915e-01 -5.52397192e-01
-7.03060448e-01 -4.50402260e-01 -5.31986475e-01 -1.64538637e-01
4.84252930e-01 -9.18339014e-01 -8.13010514e-01 1.00706291e+00
-1.40220678e+00 1.08567218e-03 -1.83236167e-01 3.01722378e-01
3.09507009e-02 6.15173399e-01 -5.89209974e-01 -8.10885251e-01
-7.18353927e-01 -1.00608802e+00 9.54710484e-01 4.87616807e-01
-5.49319446e-01 -7.84069777e-01 6.09858811e-01 4.67581242e-01
4.52176839e-01 -3.66778299e-02 3.69961351e-01 -1.39961946e+00
8.35906342e-02 -2.85021484e-01 -3.00646186e-01 6.39832914e-01
-4.18185472e-01 4.05996591e-01 -1.45428538e+00 -1.47402942e-01
-5.28328657e-01 -3.75974327e-01 1.16152024e+00 4.34164405e-01
1.05466282e+00 -7.15365559e-02 -3.16453129e-01 4.61603761e-01
8.29197586e-01 -2.02535316e-01 4.87279564e-01 5.28507173e-01
5.80386698e-01 6.31508231e-01 2.58985072e-01 5.18443048e-01
6.08373106e-01 8.90004992e-01 -3.55854742e-02 3.59957606e-01
-1.46902502e-01 -3.01979691e-01 3.35919768e-01 4.11263198e-01
3.52336407e-01 -6.40930951e-01 -1.09559214e+00 6.50994003e-01
-1.67789471e+00 -8.57071579e-01 -7.31743336e-01 2.15479589e+00
9.98127341e-01 1.56455040e-01 2.50201762e-01 -4.86990660e-02
7.68317342e-01 9.10304338e-02 -2.57399410e-01 -8.03523183e-01
-5.85368752e-01 2.77692229e-01 3.19807738e-01 4.76172239e-01
-1.45332551e+00 8.49983692e-01 6.87134552e+00 1.12745500e+00
-7.60023892e-01 1.32701293e-01 4.76154655e-01 1.46122128e-01
-2.98997909e-01 -8.68895799e-02 -1.44434941e+00 6.15832269e-01
1.30650914e+00 -2.04108492e-01 -2.90391415e-01 6.65425479e-01
-4.39596117e-01 1.62491817e-02 -1.15703595e+00 8.06216359e-01
4.92306918e-01 -1.28253043e+00 -4.42286171e-02 2.31189355e-02
7.97815084e-01 -5.57716824e-02 1.90661877e-01 7.21423566e-01
3.14738035e-01 -1.03698981e+00 9.04528558e-01 7.46945322e-01
8.32883239e-01 -5.97976029e-01 1.19260430e+00 -2.47654989e-02
-8.04246128e-01 3.68661359e-02 1.90060928e-01 3.53690147e-01
1.53253255e-02 5.92916846e-01 -8.59821737e-01 6.25045538e-01
8.13361824e-01 6.37339532e-01 -7.79193997e-01 9.68719482e-01
-3.50015759e-01 7.00815976e-01 -3.77178341e-01 -3.75314772e-01
-2.18437821e-01 4.08507615e-01 8.19943786e-01 1.64274192e+00
1.70236155e-01 -4.65695053e-01 -2.36886784e-01 1.09634864e+00
-5.82907379e-01 -5.25896326e-02 7.53088854e-04 -6.18387051e-02
7.47132123e-01 1.17803299e+00 -7.95966685e-02 -3.83037686e-01
-5.67000732e-02 9.16565120e-01 1.96761891e-01 3.09689678e-02
-8.08577120e-01 -5.97274184e-01 3.10387224e-01 -2.19617188e-01
4.89774197e-01 2.27562577e-01 -5.31993210e-01 -1.09638822e+00
-1.82137772e-01 -9.14134443e-01 6.87466443e-01 -5.55197358e-01
-1.73150539e+00 4.62965697e-01 1.07595637e-01 -8.68809164e-01
-1.95586637e-01 -7.02780724e-01 -3.83617401e-01 1.32196128e+00
-1.36454499e+00 -1.44443142e+00 9.20991153e-02 1.93618238e-01
1.06867515e-01 -4.97740358e-01 7.97497988e-01 3.87005150e-01
-5.87801993e-01 1.47388542e+00 3.30428839e-01 2.76380509e-01
1.47993827e+00 -1.45796633e+00 4.13509041e-01 9.88443136e-01
-5.25776204e-03 7.22312748e-01 7.34722435e-01 -6.94416702e-01
-8.83345127e-01 -8.70155632e-01 1.43648708e+00 -1.02262831e+00
8.07478487e-01 -6.76795661e-01 -9.59940076e-01 4.02392507e-01
2.27552697e-01 -2.26638824e-01 7.11212039e-01 5.50794601e-01
-8.32138240e-01 -2.13686507e-02 -1.05962694e+00 8.38719457e-02
4.32790160e-01 -6.33145571e-01 -8.30478013e-01 3.49456295e-02
5.14706790e-01 -5.29954314e-01 -1.04262769e+00 4.53839600e-01
9.11586523e-01 -4.24248368e-01 7.67489076e-01 -4.13820207e-01
5.47707975e-01 -2.64981866e-01 -2.48975709e-01 -1.09212756e+00
-3.96075755e-01 -8.68934333e-01 -5.31054616e-01 1.90536010e+00
5.95859766e-01 -3.78164202e-01 5.28150022e-01 5.86799443e-01
-5.48524186e-02 -2.41773412e-01 -1.23787355e+00 -7.20221043e-01
5.49103677e-01 -5.48920274e-01 3.01750183e-01 8.01296473e-01
-1.62012577e-01 4.21335906e-01 -5.74779272e-01 1.08134001e-01
7.74639904e-01 -1.87317938e-01 8.12661767e-01 -1.49787223e+00
-3.93354237e-01 -8.11357319e-01 -1.13789536e-01 -7.21906126e-01
2.16392353e-01 -8.01865995e-01 5.68853468e-02 -1.04457843e+00
4.84745175e-01 -2.15045840e-01 -5.70679367e-01 4.56422597e-01
-3.61231178e-01 2.66062140e-01 2.78819412e-01 2.44171143e-01
-5.64871848e-01 6.32294953e-01 4.31176782e-01 -2.59102046e-01
-3.87059785e-02 7.31222238e-03 -7.40079463e-01 1.46671787e-01
4.91063088e-01 -6.21899307e-01 2.11515784e-01 1.55253708e-02
1.35286525e-01 -3.52914900e-01 2.12673977e-01 -7.55423725e-01
3.14412177e-01 2.81365514e-01 4.85826939e-01 -1.07650173e+00
1.69632003e-01 -3.70079607e-01 3.57516073e-02 1.95642889e-01
-5.79371750e-01 -2.76334167e-01 5.37984729e-01 5.47846317e-01
-2.17023771e-02 -3.47080141e-01 6.92578733e-01 4.78440970e-01
-3.10057133e-01 -1.88294753e-01 -3.36767644e-01 3.12569350e-01
1.01837683e+00 -3.24817747e-02 -7.00046301e-01 -2.03849878e-02
-3.29618633e-01 5.06836534e-01 2.14869991e-01 8.27399075e-01
1.57752424e-01 -9.67186272e-01 -1.10236430e+00 -8.94817151e-03
3.51050377e-01 -4.37897295e-01 5.34040511e-01 7.11057246e-01
-2.06508100e-01 6.99440062e-01 -3.51654626e-02 -6.17521167e-01
-1.35195076e+00 2.11570352e-01 3.38712811e-01 -8.07147086e-01
-3.86137187e-01 1.25691366e+00 -2.52699673e-01 -9.10586178e-01
5.40397286e-01 5.20512648e-02 -5.09093106e-01 2.96102643e-01
6.47170901e-01 5.15430868e-01 8.98176953e-02 -6.29038751e-01
-6.75694942e-01 1.66976079e-01 -3.99230510e-01 -4.50365663e-01
1.13264477e+00 1.82870835e-01 1.97861835e-01 7.19889760e-01
9.62493718e-01 3.64916682e-01 -9.80710685e-01 -2.65276998e-01
2.76307851e-01 3.84706706e-02 1.74877927e-01 -1.43484831e+00
-4.31012601e-01 8.96071255e-01 6.44538999e-01 -2.50481457e-01
5.97129703e-01 -1.57317802e-01 4.88359123e-01 1.00892492e-01
-1.56771123e-01 -1.50102735e+00 -3.26137096e-01 7.31250405e-01
9.07291472e-01 -1.26224446e+00 1.86525837e-01 -3.96104269e-02
-7.37758279e-01 1.29967129e+00 6.05392814e-01 2.56344438e-01
6.84186101e-01 3.21065426e-01 -8.48982111e-02 5.38894907e-02
-8.08771968e-01 4.69226778e-01 8.46064150e-01 5.03773391e-01
4.27343816e-01 9.16639641e-02 -3.61027330e-01 1.29763484e+00
-1.26975462e-01 -3.41970031e-03 6.37803793e-01 4.70822155e-01
1.07604846e-01 -1.32129562e+00 -3.29161733e-01 3.13351095e-01
-4.62641835e-01 -1.38438001e-01 -8.48731697e-01 9.56428528e-01
2.21813753e-01 7.83559680e-01 -1.93827182e-01 -6.43099129e-01
3.84076655e-01 3.70557457e-01 2.85373509e-01 -3.62700254e-01
-1.10409701e+00 -1.48802260e-02 3.06100547e-01 -1.24103978e-01
-2.48922393e-01 -1.19177687e+00 -6.04561687e-01 -2.72008836e-01
-3.74008805e-01 1.47146061e-01 7.45997012e-01 6.94479883e-01
5.46788394e-01 5.54083705e-01 2.50849187e-01 -3.72638613e-01
-8.87104750e-01 -1.54518008e+00 -5.43322206e-01 1.14216976e-01
1.72934443e-01 -6.67546332e-01 -3.55263203e-01 -2.01408744e-01] | [9.590338706970215, 10.562538146972656] |
07fa7dc9-4ff8-4d4f-8cc4-ddc953f76b34 | data-provenance-inference-in-machine-learning | 2211.13416 | null | https://arxiv.org/abs/2211.13416v2 | https://arxiv.org/pdf/2211.13416v2.pdf | Data Origin Inference in Machine Learning | It is a growing direction to utilize unintended memorization in ML models to benefit real-world applications, with recent efforts like user auditing, dataset ownership inference and forgotten data measurement. Standing on the point of ML model development, we introduce a process named data origin inference, to assist ML developers in locating missed or faulty data origin in training set without maintaining strenuous metadata. We formally define the data origin and the data origin inference task in the development of the ML model (mainly neural networks). Then we propose a novel inference strategy combining embedded-space multiple instance classification and shadow training. Diverse use cases cover language, visual and structured data, with various kinds of data origin (e.g. business, county, movie, mobile user, text author). A comprehensive performance analysis of our proposed strategy contains referenced target model layers, available testing data for each origin, and in shadow training, the implementations of feature extraction as well as shadow models. Our best inference accuracy achieves 98.96% in the language use case when the target model is a transformer-based deep neural network. Furthermore, we give a statistical analysis of different kinds of data origin to investigate what kind of origin is probably to be inferred correctly. | ['Xiang-Yang Li', 'Mingxue Xu'] | 2022-11-24 | null | null | null | null | ['inference-attack', 'memorization'] | ['adversarial', 'natural-language-processing'] | [ 6.80166110e-02 1.32880434e-01 -4.53649044e-01 -5.19540966e-01
-4.04789925e-01 -4.48662728e-01 5.32428563e-01 2.98185587e-01
-2.93556303e-01 8.79957080e-01 -2.15295032e-01 -8.89911234e-01
-1.08857766e-01 -9.28646982e-01 -1.10282135e+00 -1.62085012e-01
1.00017786e-01 2.94916123e-01 1.14074640e-01 2.02596992e-01
6.67117715e-01 5.93953848e-01 -1.63722897e+00 9.01442885e-01
8.71913850e-01 1.32476497e+00 -3.26412022e-02 7.66709805e-01
-3.38635027e-01 1.42358172e+00 -1.23541629e+00 -8.27681839e-01
-1.13281153e-01 1.61419421e-01 -8.97939265e-01 -5.60669005e-01
7.36338794e-01 -7.42300749e-01 2.15352982e-01 8.50265384e-01
3.90590727e-01 -5.29006839e-01 4.63837385e-01 -1.94327235e+00
-9.01351988e-01 9.13131058e-01 -1.22500978e-01 1.90885827e-01
2.83965588e-01 1.23429462e-01 7.68493474e-01 -1.17830479e+00
7.58219182e-01 1.00126827e+00 9.63118970e-01 4.66703176e-02
-7.83007145e-01 -8.97817969e-01 -5.16339540e-02 4.93802786e-01
-1.45853114e+00 -5.61560988e-01 3.97059768e-01 -4.75799322e-01
1.09527075e+00 5.87550282e-01 2.05677807e-01 1.52464437e+00
5.26215553e-01 1.16749382e+00 1.02862453e+00 -3.74565512e-01
1.47048026e-01 8.95123839e-01 6.43571854e-01 9.06030118e-01
5.30056238e-01 -2.72008747e-01 -7.02370584e-01 -5.20083547e-01
3.89892519e-01 -1.15993552e-01 -1.37159050e-01 -1.07434638e-01
-9.01240170e-01 4.58739012e-01 8.20882153e-03 1.67572871e-01
-9.91298258e-02 1.44724712e-01 6.03300631e-01 3.28267962e-01
2.02872470e-01 2.15617612e-01 -1.17256796e+00 -2.42356330e-01
-9.81079578e-01 3.68876517e-01 1.02448452e+00 1.33255053e+00
4.69603091e-01 -2.87091918e-03 -1.56204194e-01 4.74066854e-01
3.84307951e-01 5.11800706e-01 5.77904940e-01 -4.19762433e-01
4.58401322e-01 9.69219923e-01 1.19449235e-01 -9.80735540e-01
-2.30439216e-01 -6.06272221e-01 -7.53804266e-01 1.05651766e-01
5.06568789e-01 1.69307172e-01 -4.66090858e-01 1.29503608e+00
3.40635061e-01 3.58760923e-01 4.35485430e-02 -3.17153684e-03
8.55571330e-01 1.61072165e-01 -1.86390311e-01 7.93542564e-02
1.56985319e+00 -5.84828317e-01 -8.69647264e-01 1.76985070e-01
7.69951284e-01 -5.96175790e-01 1.26180923e+00 9.07255828e-01
-6.82712078e-01 -6.90056860e-01 -1.01088274e+00 -3.17878336e-01
-1.15926874e+00 5.77499926e-01 2.73527384e-01 8.90652478e-01
-5.08769274e-01 6.18380070e-01 -5.69885790e-01 2.89059076e-02
6.32215679e-01 1.68937728e-01 -2.44692519e-01 2.75849730e-01
-1.42144060e+00 9.42321658e-01 6.80286586e-01 8.16850588e-02
-7.67056286e-01 -9.91662920e-01 -7.61555135e-01 -1.34310618e-01
3.18251461e-01 -4.01849598e-01 1.35128725e+00 -7.39992201e-01
-8.59622300e-01 6.67545676e-01 -1.75129801e-01 -5.80831349e-01
6.49797022e-01 -4.30954397e-01 -9.83460307e-01 -6.07478201e-01
1.67794004e-02 -8.03330168e-02 1.05427969e+00 -1.28685939e+00
-1.06783807e+00 -3.33334625e-01 5.62601024e-03 -6.14244640e-01
-4.26576704e-01 1.07215513e-02 -9.70115364e-02 -6.42259240e-01
-4.92043763e-01 -2.83673882e-01 6.00296557e-01 9.64347422e-02
-7.89337397e-01 -2.87016362e-01 1.30050480e+00 -1.26970422e+00
1.99867499e+00 -1.97340238e+00 -4.27793890e-01 2.70936459e-01
3.99452895e-01 3.22275430e-01 4.59312528e-01 1.70526862e-01
-1.52932346e-01 3.44507426e-01 -7.19101448e-03 -4.71784174e-01
3.12817812e-01 2.29858086e-01 -7.57320285e-01 2.06376716e-01
1.02606468e-01 7.04784811e-01 -5.80169797e-01 -4.84130740e-01
-4.11989652e-02 2.03649908e-01 -2.88290322e-01 2.93087512e-01
-2.86424845e-01 -2.63259828e-01 -2.58989990e-01 1.17558420e+00
8.76460850e-01 -3.15561593e-01 6.20932430e-02 -5.27391315e-01
-1.95764348e-01 4.00025994e-01 -1.49788392e+00 1.37979329e+00
-6.06837690e-01 6.76317215e-01 -9.35022384e-02 -4.71983910e-01
7.59831011e-01 1.95924610e-01 -1.23078875e-01 -4.85409141e-01
-1.21894129e-01 2.77005434e-01 -2.40964398e-01 -8.94058764e-01
9.61446106e-01 6.23903692e-01 -2.37543717e-01 4.84065533e-01
1.92128271e-02 6.96897089e-01 -2.62281001e-01 1.88990220e-01
1.03336442e+00 4.54467356e-01 4.70162213e-01 2.87232138e-02
6.16547227e-01 -1.18602924e-01 3.79466116e-01 9.37599838e-01
6.91493377e-02 -2.76291519e-02 6.92424834e-01 -5.37534356e-01
-1.02788675e+00 -9.40471113e-01 -6.41918898e-01 1.14681304e+00
-2.08118275e-01 -5.89097977e-01 -6.57701373e-01 -1.06404257e+00
4.23178613e-01 1.25108564e+00 -5.40829539e-01 -3.41885746e-01
-4.35683280e-01 -5.85972250e-01 1.37995672e+00 5.39282322e-01
8.33416581e-01 -1.13503861e+00 -7.33612835e-01 1.22629561e-01
-2.76455075e-01 -9.53033686e-01 2.45376736e-01 2.49411896e-01
-5.14425695e-01 -1.28767419e+00 2.53109783e-01 -2.54972517e-01
3.62364441e-01 -6.05356634e-01 1.33804607e+00 3.15406173e-01
-3.02701294e-01 1.13225430e-01 -1.71822663e-02 -7.68751621e-01
-6.20760381e-01 2.31618494e-01 6.90498203e-02 2.62619881e-03
6.24628723e-01 -1.97972655e-01 -7.71452785e-02 8.84153098e-02
-9.23186004e-01 -9.99934375e-02 5.75585663e-01 8.27694535e-01
2.03809664e-01 1.72981322e-01 3.30144882e-01 -1.37803900e+00
7.35428214e-01 -9.60868418e-01 -5.34623504e-01 7.30318189e-01
-1.03834629e+00 2.30556190e-01 8.73747528e-01 -3.23745281e-01
-8.43012691e-01 -5.15495360e-01 -1.39433399e-01 -2.71418095e-01
-4.10759002e-01 6.10655308e-01 -4.71926749e-01 2.83002913e-01
8.21015537e-01 3.16134691e-01 -1.49176911e-01 -7.80943036e-01
4.54214551e-02 1.26231337e+00 5.65425158e-01 -6.80507421e-01
4.47487622e-01 9.19686407e-02 -4.93473351e-01 -4.67979580e-01
-5.59624970e-01 4.05936642e-03 -6.79497540e-01 -2.33671851e-02
3.49574029e-01 -6.06410563e-01 -1.03050876e+00 9.13732171e-01
-1.44972956e+00 -1.24985307e-01 8.97988398e-03 -1.60152018e-01
-3.28648984e-02 1.65619314e-01 -3.95202041e-01 -1.01561916e+00
-5.42851329e-01 -1.02203989e+00 1.10547566e+00 -3.47740859e-01
-2.57608980e-01 -9.70530272e-01 -4.01779115e-01 4.43394542e-01
3.38416249e-01 2.08561853e-01 1.22969413e+00 -1.13003504e+00
-7.99978733e-01 -3.20589066e-01 -3.46765339e-01 5.13123572e-01
3.75705771e-02 4.82615113e-01 -1.43338537e+00 -5.29423133e-02
-1.14621863e-01 -1.72411025e-01 2.55558729e-01 -1.65300012e-01
1.83675027e+00 -7.19767272e-01 -4.56588000e-01 5.56966543e-01
1.29420388e+00 1.63989916e-01 7.91190505e-01 2.81254977e-01
7.66976237e-01 2.21767515e-01 6.93634272e-01 7.21587181e-01
5.49047291e-01 5.88057578e-01 5.55020154e-01 1.14766285e-01
2.82988697e-01 -4.44106787e-01 4.21926200e-01 3.09203297e-01
2.50255674e-01 -4.72084671e-01 -9.73834872e-01 2.45302990e-01
-1.82410455e+00 -1.11119497e+00 -4.09495771e-01 2.35217142e+00
9.87829149e-01 5.44092894e-01 -1.94817051e-01 5.60273647e-01
3.50217640e-01 -3.31239909e-01 -6.33927166e-01 -4.74141002e-01
3.61418687e-02 2.85237610e-01 6.22973204e-01 3.49080950e-01
-1.04900086e+00 5.26542068e-01 5.91831017e+00 1.05396318e+00
-1.19224870e+00 3.45792532e-01 5.53096890e-01 2.19288349e-01
-3.84835988e-01 -1.20212168e-01 -1.52868724e+00 9.21321213e-01
1.13531232e+00 1.14418954e-01 1.41802415e-01 1.24562895e+00
9.08865854e-02 -2.28501081e-01 -1.39348841e+00 9.64409173e-01
1.99828506e-01 -1.57572520e+00 1.27603009e-01 4.18456048e-02
-1.15099318e-01 -3.27976942e-01 7.17943162e-02 5.66068113e-01
1.11744910e-01 -1.07368875e+00 1.14597404e+00 9.70421970e-01
9.79837477e-01 -6.83338702e-01 8.31966519e-01 6.73491299e-01
-9.37114537e-01 -3.27759624e-01 -8.01355392e-02 3.85976210e-02
-4.86830890e-01 7.34161496e-01 -1.12826550e+00 5.46396911e-01
8.73714566e-01 6.57141626e-01 -1.15584600e+00 5.93750298e-01
-4.82413024e-02 4.89920259e-01 -2.82370090e-01 -1.00659214e-01
-3.99824262e-01 3.66913915e-01 1.21855684e-01 1.22418094e+00
3.22304845e-01 -7.08534479e-01 -2.30926082e-01 1.23653507e+00
-5.58279715e-02 -1.14397824e-01 -7.72021532e-01 2.27955282e-01
9.06723797e-01 1.14576757e+00 -1.14690559e-02 -6.25639021e-01
-2.21316725e-01 9.87659752e-01 3.91500652e-01 1.80300742e-01
-1.24972498e+00 -6.70810938e-01 6.51740789e-01 3.25038195e-01
2.46967413e-02 1.02678947e-01 -5.78976333e-01 -1.13907743e+00
3.97167534e-01 -9.96859550e-01 4.05765027e-01 -9.09951031e-01
-9.01490390e-01 3.96002829e-01 2.39476830e-01 -1.15909970e+00
-2.93195337e-01 -8.85173261e-01 -3.41294587e-01 8.29628587e-01
-1.29319024e+00 -1.41839552e+00 -5.04361808e-01 6.59142613e-01
1.26136690e-01 -5.11892915e-01 9.16913986e-01 7.15305388e-01
-9.28618789e-01 1.27912629e+00 1.20408140e-01 4.15966421e-01
7.10334659e-01 -1.30431390e+00 2.94821262e-01 9.55031812e-01
-7.57698296e-03 1.25841165e+00 4.41544384e-01 -7.13620961e-01
-1.38112950e+00 -1.29255891e+00 1.14045584e+00 -1.01179159e+00
6.39548540e-01 -7.54293799e-01 -9.18889284e-01 1.09496331e+00
-1.03543304e-01 8.81217644e-02 8.79940212e-01 1.64293468e-01
-5.60558259e-01 -2.73042917e-01 -1.53660536e+00 3.80699545e-01
8.16657841e-01 -7.73988247e-01 -4.63125706e-01 2.60047615e-01
5.84990501e-01 -6.53622687e-01 -1.18972278e+00 1.67747304e-01
7.21113801e-01 -1.01702142e+00 9.04697657e-01 -7.53794372e-01
3.70120376e-01 -3.56860608e-01 -1.30324990e-01 -5.54821730e-01
1.24972641e-01 -2.07063779e-01 -9.04575109e-01 1.48978579e+00
6.21529460e-01 -6.26913846e-01 6.49104357e-01 8.64133120e-01
-1.58751771e-01 -7.63565004e-01 -7.78844953e-01 -6.30862713e-01
-2.76846737e-01 -8.72424185e-01 1.08852291e+00 1.22518027e+00
-3.33974004e-01 8.30478966e-02 -4.45846438e-01 3.86970431e-01
5.33251286e-01 9.52539034e-03 1.10244179e+00 -1.26237845e+00
-4.57998663e-01 -1.85258329e-01 -2.48740628e-01 -9.77035105e-01
1.60849895e-02 -8.43451023e-01 -5.36512256e-01 -1.03438866e+00
-2.23508134e-01 -7.64488518e-01 -2.46343553e-01 8.70657802e-01
1.78477392e-01 -3.81536394e-01 -3.31224680e-01 7.35496134e-02
-2.80446380e-01 -8.44326466e-02 6.00452423e-01 -9.94339585e-02
2.13615909e-01 2.66862631e-01 -6.16864562e-01 6.39891744e-01
4.75515723e-01 -4.78520274e-01 -3.89164090e-01 -4.15295959e-01
5.81171095e-01 -2.32346296e-01 8.99383187e-01 -9.40067649e-01
2.68725842e-01 -9.84163880e-02 7.66158938e-01 -1.00873435e+00
5.61078377e-02 -1.29598296e+00 3.65264922e-01 4.91259396e-01
-2.63345987e-01 6.83033168e-02 8.97472724e-02 2.12687045e-01
-4.91708443e-02 -5.72623253e-01 5.51196575e-01 1.53515832e-02
-8.86864722e-01 -6.40777592e-03 -8.80812109e-02 -1.38976917e-01
9.13935363e-01 -5.05998373e-01 -6.31811321e-01 2.13852644e-01
-6.01441920e-01 -1.29256904e-01 2.09177315e-01 4.87598479e-01
6.69708848e-01 -1.24361515e+00 -2.63066232e-01 7.98330307e-01
3.58774900e-01 7.75733888e-02 -2.00789094e-01 6.37965500e-01
-6.12156391e-01 2.51636088e-01 -8.73161256e-02 -4.70769763e-01
-1.28223765e+00 5.07948101e-01 3.86266440e-01 -2.69634396e-01
-1.86399192e-01 7.54037738e-01 -6.05779409e-01 -6.56321526e-01
5.38931847e-01 -1.05372071e+00 -1.94441639e-02 9.40247327e-02
6.69745386e-01 6.05259836e-01 6.00449383e-01 -8.49972591e-02
-2.91268915e-01 -2.82796323e-01 -3.15200061e-01 4.42219406e-01
9.77540016e-01 2.64495492e-01 -2.71148682e-01 9.56052899e-01
1.22705722e+00 1.40170202e-01 -6.03098810e-01 -1.80331931e-01
2.50931084e-01 -4.13524926e-01 -2.49255538e-01 -1.22654140e+00
-7.76664913e-01 7.96257973e-01 6.53651476e-01 3.30627263e-01
9.10951197e-01 -5.05959809e-01 6.59456789e-01 5.08659780e-01
5.97299576e-01 -1.29753387e+00 -3.57364327e-01 4.92327213e-01
9.60096657e-01 -1.10792708e+00 8.99935514e-02 -6.99166209e-02
-2.87901491e-01 1.09853506e+00 9.16415572e-01 4.48501796e-01
9.23036516e-01 6.76691890e-01 -2.39607140e-01 -2.13519454e-01
-8.54184568e-01 6.43570483e-01 9.78545323e-02 5.85763872e-01
3.44753683e-01 -8.85546729e-02 1.45325154e-01 1.02712369e+00
-3.36564779e-01 4.43831325e-01 4.99670893e-01 1.17758846e+00
-7.55576640e-02 -1.08597255e+00 -4.36491251e-01 1.00065386e+00
-5.91995895e-01 -4.71670687e-01 -2.96211123e-01 1.05497479e+00
8.10984313e-01 4.24346566e-01 1.76515326e-01 -5.78681409e-01
3.26300085e-01 4.18421060e-01 -1.78452916e-02 -3.68240952e-01
-9.10771310e-01 -6.85105264e-01 2.75706738e-01 -5.53939462e-01
2.08317861e-01 -4.95625764e-01 -1.21147656e+00 -7.49843359e-01
-4.01155651e-01 -2.92363673e-01 7.84421027e-01 8.52739453e-01
5.94329834e-01 4.98663127e-01 2.96929121e-01 1.66661516e-01
-5.62673867e-01 -1.14124036e+00 -4.55266923e-01 3.35819691e-01
4.02385175e-01 -6.36178672e-01 -2.41874143e-01 1.98422790e-01] | [9.14093017578125, 7.596006393432617] |
fe05dbea-600d-4606-bf5c-74fa6a713dcf | spoofing-and-anti-spoofing-with-wax-figure | 1910.05457 | null | https://arxiv.org/abs/1910.05457v1 | https://arxiv.org/pdf/1910.05457v1.pdf | Spoofing and Anti-Spoofing with Wax Figure Faces | We have witnessed rapid advances in both face presentation attack models and presentation attack detection (PAD) in recent years. Compared to widely studied 2D face presentation attacks (e.g. printed photos and video replays), 3D face presentation attacks are more challenging because face recognition systems (FRS) is more easily confused by the 3D characteristics of materials similar to real faces. Existing 3D face spoofing databases, mostly based on 3D facial masks, are restricted to small data size and suffer from poor authenticity due to the difficulty and expense of mask production. In this work, we introduce a wax figure face database (WFFD) as a novel and super-realistic 3D face presentation attack. This database contains 2300 image pairs (totally 4600) and 745 subjects including both real and wax figure faces with high diversity from online collections. On one hand, our experiments have demonstrated the spoofing potential of WFFD on three popular FRSs. On the other hand, we have developed a multi-feature voting scheme for wax figure face detection (anti-spoofing), which combines three discriminative features at the decision level. The proposed detection method was compared against several face PAD approaches and found to outperform other competing methods. Surprisingly, our fusion-based detection method achieves an Average Classification Error Rate (ACER) of 11.73\% on the WFFD database, which is even better than human-based detection. | ['Zhengquan Xu', 'Shan Jia', 'Xin Li', 'Chuanbo Hu'] | 2019-10-12 | null | null | null | null | ['small-data'] | ['computer-vision'] | [ 4.62574422e-01 -3.36075932e-01 1.81942448e-01 -3.39425765e-02
-4.70882177e-01 -6.71925426e-01 7.33520687e-01 -3.71922404e-01
4.25797999e-02 3.43826562e-01 -9.63614732e-02 -1.81831002e-01
-7.45694637e-02 -4.77444321e-01 -3.16842079e-01 -8.04096758e-01
-3.52863044e-01 -7.45600834e-02 1.85535431e-01 -3.67965639e-01
2.23818839e-01 1.14049709e+00 -1.86719167e+00 3.24127674e-01
3.04128557e-01 1.13684881e+00 -1.90807968e-01 6.06431484e-01
1.85950145e-01 -3.92655693e-02 -1.22016144e+00 -9.01320934e-01
4.59689826e-01 -2.07018569e-01 -1.15280151e-01 8.71166065e-02
9.03141141e-01 -5.59421360e-01 -5.34030437e-01 1.05825210e+00
8.67438436e-01 -4.85490084e-01 6.11012757e-01 -1.68170202e+00
-5.42966247e-01 -5.19526824e-02 -9.98178124e-01 2.42861077e-01
9.24899578e-01 5.30375578e-02 1.48958668e-01 -1.04363286e+00
5.34405053e-01 1.72134113e+00 5.74064612e-01 9.72917974e-01
-9.42222834e-01 -1.41860020e+00 -4.18950260e-01 8.51737931e-02
-1.78505194e+00 -8.35067153e-01 9.61404145e-01 -3.20590496e-01
4.90692228e-01 3.82597208e-01 3.66186559e-01 1.43823719e+00
1.75674960e-01 2.80288547e-01 1.43927038e+00 -5.47718942e-01
-2.50208080e-01 4.26670730e-01 -1.84141114e-01 8.29800367e-01
5.69592714e-01 3.06554019e-01 -6.85069561e-01 -6.02450550e-01
5.96472502e-01 -1.68982551e-01 -5.10003328e-01 2.35370789e-02
-6.20638192e-01 5.79744399e-01 -8.18732083e-02 3.61376077e-01
-6.86849421e-03 -3.35430354e-01 1.73070386e-01 3.98302317e-01
1.84399471e-01 -1.06755346e-01 6.34169057e-02 2.91640550e-01
-8.29918325e-01 2.92812526e-01 8.50204885e-01 6.18091106e-01
2.63350725e-01 2.12134793e-01 1.02020614e-02 8.40768456e-01
5.75553119e-01 1.07344425e+00 2.27460116e-01 -3.17375779e-01
4.70257401e-01 8.79860669e-02 -3.34665477e-02 -1.79511976e+00
2.56035700e-02 -4.05363180e-03 -7.32356250e-01 4.08555627e-01
2.27184966e-01 3.64324637e-02 -8.14139605e-01 1.59464002e+00
3.54219466e-01 1.87776327e-01 -1.15926936e-02 6.18707240e-01
8.96422505e-01 3.62462223e-01 -1.10904820e-01 -4.35559958e-01
1.51105976e+00 -1.14353508e-01 -9.00597811e-01 3.13520700e-01
1.60612404e-01 -1.27627754e+00 5.74612796e-01 5.51175416e-01
-7.48315215e-01 -5.60832679e-01 -1.23664963e+00 6.05721951e-01
-3.91590655e-01 2.28821728e-02 2.26084888e-01 1.95241714e+00
-1.17548931e+00 2.53977835e-01 -1.20463558e-01 -3.78488749e-01
7.28023112e-01 6.60965204e-01 -9.05426979e-01 -1.94947302e-01
-1.10670841e+00 6.64133012e-01 -1.33542389e-01 -1.29624214e-02
-1.03725851e+00 -4.94600773e-01 -6.30145073e-01 -2.39598975e-01
2.17619047e-01 8.26755259e-03 7.00339377e-01 -5.85396230e-01
-1.55848169e+00 1.16711700e+00 -1.64884388e-01 -9.18191150e-02
5.15263796e-01 -1.16806068e-01 -1.24392116e+00 5.03117085e-01
-3.16975921e-01 3.83828759e-01 1.60493600e+00 -1.24964881e+00
7.32404217e-02 -5.86290777e-01 -1.44423291e-01 -4.25537616e-01
-5.89405239e-01 6.76887691e-01 -3.46035995e-02 -6.97862744e-01
-1.15723507e-02 -6.58337057e-01 6.27785504e-01 3.57511818e-01
-2.76197225e-01 -1.42902046e-01 1.69110084e+00 -7.71101177e-01
1.18764162e+00 -2.39522123e+00 -3.99449468e-01 3.66033942e-01
7.82626569e-02 1.10345352e+00 -2.16171458e-01 5.94369709e-01
-3.54617834e-01 4.41423267e-01 9.31842774e-02 -3.53003860e-01
-7.63545707e-02 -1.23925634e-01 -4.20854062e-01 9.77133870e-01
3.68402272e-01 1.88330576e-01 -5.48730671e-01 -5.47862172e-01
2.84152180e-01 8.58945310e-01 -4.63430345e-01 1.44705832e-01
4.26813245e-01 9.63209197e-02 -2.15558127e-01 1.15319026e+00
1.44648778e+00 3.64110857e-01 1.84891373e-02 -2.18615457e-01
2.43993953e-01 -2.36488134e-01 -1.38373947e+00 1.01230264e+00
-3.85416187e-02 7.04435587e-01 2.78607875e-01 -5.19074500e-01
1.22546017e+00 7.35691547e-01 2.36587271e-01 -3.51537496e-01
3.13303202e-01 3.31170678e-01 -6.21464960e-02 -6.47515833e-01
1.20955512e-01 7.02022389e-02 3.57340574e-01 3.59492213e-01
1.53779268e-01 -7.88818076e-02 -1.45410195e-01 -1.63936168e-02
9.33698714e-01 -1.78944767e-01 1.20139644e-01 -3.08572978e-01
9.74722505e-01 -7.45966256e-01 3.01558048e-01 5.12768269e-01
-5.49378872e-01 4.70009446e-01 3.19229692e-01 -2.03175262e-01
-5.67375720e-01 -1.02782750e+00 -3.59411359e-01 2.91443348e-01
1.47820830e-01 -6.57728195e-01 -8.45319808e-01 -8.13700676e-01
2.81343490e-01 1.87572539e-02 -4.39243674e-01 -3.30778733e-02
-6.66261733e-01 -5.17756224e-01 1.28742278e+00 -1.75095960e-01
7.92407870e-01 -5.83670020e-01 -2.42764652e-01 -2.46872202e-01
1.65049791e-01 -1.22540402e+00 -3.21228653e-01 -7.24545002e-01
-2.56351531e-01 -1.39354384e+00 -9.01300311e-01 -7.38292575e-01
6.61212385e-01 6.76736236e-01 5.41264236e-01 3.97707552e-01
-7.00019360e-01 3.51980418e-01 -2.71548063e-01 -5.95024645e-01
-6.39546573e-01 -6.90483510e-01 7.63120353e-01 3.96831006e-01
4.18477148e-01 -2.99882472e-01 -4.73503977e-01 7.36143470e-01
-9.31702375e-01 -7.06535876e-01 3.18138868e-01 4.97436434e-01
-2.28065625e-01 1.95273891e-01 4.87309426e-01 -4.97332901e-01
6.23630643e-01 -2.26817518e-01 -4.79479373e-01 3.24689150e-01
-1.39840499e-01 -5.70057869e-01 1.88548967e-01 -5.98084748e-01
-1.00322318e+00 -2.88025677e-01 -2.64436245e-01 -7.18900740e-01
-4.43335116e-01 -4.20072615e-01 -4.81707066e-01 -8.08700740e-01
6.90165162e-01 1.06348917e-01 4.22648460e-01 -5.26558995e-01
-2.25797951e-01 1.15947819e+00 3.72360826e-01 -2.86647916e-01
1.34107614e+00 4.24015373e-01 2.92200863e-01 -1.43849194e+00
1.24146797e-01 -1.75915584e-01 -2.87085682e-01 -5.36108553e-01
4.50006694e-01 -7.77841091e-01 -1.14684939e+00 1.12594855e+00
-1.20136786e+00 5.39566815e-01 5.55948555e-01 3.64738911e-01
-2.51564709e-03 9.43812251e-01 -5.79029620e-01 -1.30966878e+00
-1.35019481e-01 -1.13481915e+00 1.24305785e+00 6.86043352e-02
-4.94155250e-02 -5.49649596e-01 -2.48728082e-01 3.14046741e-01
4.94592339e-01 7.47558355e-01 6.46513581e-01 -4.53878045e-01
-1.04139142e-01 -6.31580472e-01 -2.85798162e-01 4.36812043e-01
4.94622439e-01 3.60705167e-01 -1.43200183e+00 -6.03729844e-01
3.33262473e-01 -1.36304885e-01 4.63301718e-01 -9.03388411e-02
9.12143528e-01 -1.76668003e-01 -5.11227846e-01 2.91978389e-01
1.26355290e+00 3.54655802e-01 6.48536325e-01 -2.56726772e-01
2.05279037e-01 8.87054861e-01 3.87836754e-01 5.26593745e-01
-4.61443812e-01 9.10365701e-01 3.33746612e-01 1.53723255e-01
-4.13334310e-01 -1.98033661e-01 6.33440733e-01 3.01236451e-01
-5.04757352e-02 -6.25494301e-01 -7.25019574e-01 -2.01990947e-01
-8.74668658e-01 -1.00903022e+00 7.60469958e-02 2.19713569e+00
2.46481702e-01 -5.24788983e-02 1.36692166e-01 7.50330567e-01
1.13598466e+00 2.93450326e-01 5.54581285e-02 -3.71551156e-01
-3.23556632e-01 3.95896316e-01 3.56786638e-01 2.66478240e-01
-1.08885527e+00 6.18860126e-01 6.23680830e+00 9.27644074e-01
-1.14905822e+00 7.58409873e-02 3.96886617e-01 4.90029380e-02
1.19382240e-01 -5.72200298e-01 -1.22109485e+00 5.63060641e-01
7.15240061e-01 1.17915519e-01 1.51343659e-01 4.54832107e-01
-1.63983300e-01 1.12169638e-01 -6.06261611e-01 1.51817954e+00
7.67273724e-01 -9.46316481e-01 2.56265104e-01 5.41026115e-01
3.09995890e-01 -8.18738043e-01 3.84332567e-01 -3.20631385e-01
-1.95855647e-01 -1.09750986e+00 5.01337290e-01 -2.94029191e-02
1.13335907e+00 -7.88740754e-01 6.19731784e-01 1.72233582e-02
-1.14920747e+00 -1.77094683e-01 -3.02121639e-01 2.48828933e-01
1.08702078e-01 3.37526351e-01 -8.56055140e-01 6.52365923e-01
7.29590416e-01 3.22491705e-01 -4.62044567e-01 8.49176764e-01
6.06227927e-02 3.73941958e-01 -5.90880096e-01 2.23560594e-02
-1.08310729e-01 2.52438217e-01 8.12509716e-01 1.01369500e+00
6.12234056e-01 9.83064249e-02 -2.63477594e-01 3.37489992e-01
-1.33505836e-01 -2.01996211e-02 -1.08105397e+00 -1.35796160e-01
6.44037485e-01 1.10926712e+00 -4.82654035e-01 9.17526707e-02
-3.66741776e-01 7.70440936e-01 -3.60078454e-01 8.55472684e-02
-7.21999228e-01 -4.32040602e-01 9.17453825e-01 2.34490827e-01
3.49023074e-01 -2.45533660e-01 5.49739361e-01 -9.25128400e-01
1.03054225e-01 -1.12643778e+00 3.19181174e-01 -4.16585028e-01
-1.19537663e+00 9.24293518e-01 4.00935635e-02 -1.17710090e+00
9.39112622e-03 -9.98787403e-01 -3.10259342e-01 8.58752906e-01
-1.35599685e+00 -1.00919402e+00 -2.76300609e-01 8.66909087e-01
2.32813343e-01 -6.54382348e-01 8.74662936e-01 5.93823671e-01
-6.46407604e-01 1.05766404e+00 -3.11495990e-01 2.12501884e-01
8.13194811e-01 -3.19766939e-01 2.48482794e-01 7.72883058e-01
6.67156056e-02 6.73890769e-01 4.26461965e-01 -7.18403935e-01
-1.73797321e+00 -5.21835029e-01 8.63470733e-01 -2.94459045e-01
2.55990148e-01 -6.29890978e-01 -6.78761780e-01 2.91115083e-02
6.36279881e-02 4.85574231e-02 9.31722045e-01 -6.10311747e-01
-7.79202521e-01 -1.36703730e-01 -1.88384235e+00 4.34107035e-01
1.20278132e+00 -6.76178038e-01 -2.10058168e-01 3.04220855e-01
1.45179376e-01 -5.82881980e-02 -7.00238407e-01 6.27666771e-01
9.07199681e-01 -1.15187883e+00 1.17739093e+00 -2.08767027e-01
-2.15086460e-01 -3.15454334e-01 -5.28551579e-01 -6.44165218e-01
2.84590740e-02 -1.11146414e+00 -1.78857997e-01 1.60824943e+00
-1.95058718e-01 -7.58407533e-01 7.28072226e-01 1.82855502e-01
4.27930623e-01 -2.26611137e-01 -1.02091587e+00 -1.11469865e+00
-4.92257088e-01 -1.07611246e-01 8.45588028e-01 8.35419774e-01
-2.24813670e-01 -3.55312198e-01 -6.50158644e-01 5.36818027e-01
9.54849064e-01 -3.35736752e-01 7.85715520e-01 -1.35406303e+00
-9.68882218e-02 -3.17775160e-01 -9.17754412e-01 -5.09864986e-01
1.44308805e-01 -5.15417457e-01 -5.41703463e-01 -2.63084590e-01
-1.39198765e-01 -2.27844477e-01 8.20405129e-03 1.50182828e-01
1.39413983e-01 8.58814359e-01 3.00039440e-01 1.80972472e-01
3.63715619e-01 2.01762706e-01 1.09556878e+00 -2.62856553e-03
2.12759048e-01 1.80610816e-03 -2.87488043e-01 5.01782298e-01
6.09647810e-01 -3.69649678e-01 -2.07026154e-01 -1.06872067e-01
-4.70642298e-01 1.04032420e-01 4.35804367e-01 -1.20456553e+00
-5.36008775e-02 2.57335186e-01 6.86938763e-01 -4.40964252e-01
6.74017608e-01 -9.24944758e-01 2.58140862e-01 8.12949300e-01
1.26198411e-01 1.11988783e-01 3.19678098e-01 4.87400085e-01
-9.38950852e-02 -1.61078528e-01 8.62257481e-01 4.56063002e-02
-3.58325005e-01 4.22031820e-01 -3.09159964e-01 -5.29817104e-01
1.20405579e+00 -5.83845735e-01 -5.51860631e-01 -1.74819589e-01
-3.08970004e-01 -6.57571018e-01 4.23915684e-01 6.37719154e-01
9.81326342e-01 -1.19514668e+00 -8.02548587e-01 1.02054095e+00
-6.11231364e-02 -9.15039361e-01 1.21519797e-01 3.75591338e-01
-6.10370219e-01 3.99490535e-01 -5.57361782e-01 -5.11686146e-01
-2.08928609e+00 6.46490753e-01 4.27374542e-02 3.13342631e-01
-2.72178471e-01 8.82226229e-01 -6.03416674e-02 9.87414345e-02
2.57708669e-01 4.93259996e-01 -1.51043966e-01 7.37309605e-02
1.08226931e+00 5.14863372e-01 1.91936061e-01 -1.06965303e+00
-6.13995731e-01 7.33094692e-01 -1.41329944e-01 6.71628937e-02
8.00311804e-01 9.48612168e-02 -3.55659891e-03 -4.48172808e-01
1.52731431e+00 3.57848823e-01 -7.45997429e-01 -8.95053223e-02
-2.42695361e-01 -1.30517673e+00 -2.82914132e-01 -3.37948889e-01
-1.07647061e+00 8.58643174e-01 1.05317628e+00 3.68334383e-01
1.04729688e+00 -1.92729205e-01 6.38035297e-01 -1.21181356e-02
4.91069555e-01 -3.95895064e-01 3.16935062e-01 -1.45571351e-01
9.63930070e-01 -1.02110064e+00 1.25727169e-02 -1.09722424e+00
-5.49504980e-02 1.10678720e+00 4.39957559e-01 -2.77741663e-02
9.41713214e-01 3.46671104e-01 6.66154781e-03 -1.77855954e-01
-3.26394290e-01 2.69966871e-01 1.00853629e-01 1.05653536e+00
9.79821682e-02 -2.34079361e-01 1.08190738e-02 1.50605645e-02
-2.02848256e-01 -1.05421662e-01 2.00652689e-01 1.07947087e+00
-1.57776192e-01 -1.25814700e+00 -7.86959708e-01 1.25324190e-01
-8.36655021e-01 4.15810853e-01 -5.64754128e-01 8.44166219e-01
1.62091583e-01 1.29323304e+00 -1.79003686e-01 -6.85185015e-01
1.35167345e-01 5.12008220e-02 7.26248682e-01 -1.89366698e-01
-5.29552341e-01 -1.06014811e-01 -2.26401966e-02 -5.22094011e-01
-5.95806122e-01 -6.47054791e-01 -1.78603515e-01 -7.89098442e-01
-5.76452315e-01 -4.09306213e-02 7.80635357e-01 5.81744850e-01
3.23443949e-01 -1.46306232e-01 1.16642070e+00 -9.21264231e-01
-4.56659108e-01 -7.44394839e-01 -8.42114210e-01 3.93055141e-01
5.46069384e-01 -1.13114989e+00 -5.06384194e-01 -1.68802708e-01] | [13.073365211486816, 1.1427404880523682] |
ed40940e-3ab3-4d27-a20a-db72388bbbf1 | extremely-low-resource-machine-translation | 2105.13065 | null | https://arxiv.org/abs/2105.13065v1 | https://arxiv.org/pdf/2105.13065v1.pdf | Extremely low-resource machine translation for closely related languages | An effective method to improve extremely low-resource neural machine translation is multilingual training, which can be improved by leveraging monolingual data to create synthetic bilingual corpora using the back-translation method. This work focuses on closely related languages from the Uralic language family: from Estonian and Finnish geographical regions. We find that multilingual learning and synthetic corpora increase the translation quality in every language pair for which we have data. We show that transfer learning and fine-tuning are very effective for doing low-resource machine translation and achieve the best results. We collected new parallel data for V\~oro, North and South Saami and present first results of neural machine translation for these languages. | ['Mark Fišel', 'Andre Tättar', 'Maali Tars'] | 2021-05-27 | null | https://aclanthology.org/2021.nodalida-main.5 | https://aclanthology.org/2021.nodalida-main.5.pdf | nodalida-2021-5 | ['low-resource-neural-machine-translation'] | ['natural-language-processing'] | [-5.05985022e-02 -1.38718069e-01 -5.76075017e-01 -4.35130626e-01
-1.49846649e+00 -8.76796603e-01 8.27221096e-01 -6.18405461e-01
-6.71180248e-01 1.59952426e+00 3.34214061e-01 -9.30931568e-01
4.69629496e-01 -6.61628783e-01 -1.03622484e+00 -2.79374897e-01
3.33855689e-01 1.08466315e+00 -3.30067426e-01 -9.27097201e-01
-2.63169501e-02 2.75150865e-01 -5.79541683e-01 6.61374450e-01
1.18255675e+00 -2.61095822e-01 2.68864423e-01 5.67812562e-01
-2.78489709e-01 2.01410607e-01 -4.91386473e-01 -5.49938560e-01
6.26394331e-01 -9.92865264e-01 -1.06463301e+00 -4.32310581e-01
4.63019729e-01 -1.49015367e-01 -8.44061654e-03 7.81366467e-01
7.75621116e-01 -4.19741660e-01 6.28092110e-01 -5.90379834e-01
-7.22514033e-01 1.27696240e+00 -4.87564951e-01 3.16177338e-01
2.11425737e-01 1.36730552e-01 9.92845118e-01 -1.30287504e+00
1.06327093e+00 1.42583191e+00 8.56193483e-01 6.49635911e-01
-1.29678154e+00 -8.81812751e-01 -4.21199232e-01 -1.72631219e-01
-1.35603404e+00 -7.39987016e-01 3.61757636e-01 -2.30563655e-01
1.45936656e+00 1.20942369e-02 4.20225590e-01 1.30170608e+00
3.18344504e-01 7.38185763e-01 1.54567301e+00 -9.60046530e-01
-3.62339169e-01 2.88143456e-01 -5.90532780e-01 4.99281317e-01
2.48699840e-02 3.88830304e-01 -5.34721017e-01 -1.81534719e-02
7.74910390e-01 -7.51443446e-01 9.24605057e-02 2.60407478e-01
-1.69297147e+00 8.97302449e-01 2.56536186e-01 8.22919071e-01
-1.74701720e-01 -3.46633419e-02 4.24349129e-01 1.11675942e+00
7.98324347e-01 8.34663093e-01 -1.04500413e+00 -1.61532745e-01
-8.24992418e-01 3.00084651e-02 7.72425115e-01 1.05327880e+00
8.72056603e-01 4.97197628e-01 2.56575286e-01 1.31642377e+00
-1.32838860e-01 1.00364161e+00 7.03133523e-01 -5.83290577e-01
1.05128038e+00 2.09581032e-02 -1.16484640e-02 -2.73793876e-01
-2.77190298e-01 -2.06101537e-01 -5.43368042e-01 -4.84267287e-02
5.25775254e-01 -6.47850871e-01 -8.43275309e-01 1.78972220e+00
-9.71638598e-03 -5.42446136e-01 5.90729475e-01 4.89700377e-01
3.86673242e-01 8.16266596e-01 -6.78573400e-02 -2.12358892e-01
8.20334911e-01 -1.15059686e+00 -4.15559471e-01 -2.90037602e-01
1.08081543e+00 -1.24698400e+00 1.20236206e+00 1.55417649e-02
-1.29851902e+00 -5.41440547e-01 -9.36134040e-01 3.77556086e-02
-3.67453516e-01 2.65271246e-01 6.07957959e-01 5.75170219e-01
-1.23421633e+00 7.23408937e-01 -7.12595701e-01 -7.22658038e-01
1.47455156e-01 6.42428637e-01 -6.78262830e-01 -1.54417917e-01
-1.42061472e+00 1.34471965e+00 5.19235969e-01 -1.26333892e-01
-7.08345890e-01 -3.81778926e-01 -5.22561371e-01 -4.95477110e-01
-3.33116621e-01 -6.37166440e-01 1.21254086e+00 -1.50175512e+00
-1.61102700e+00 1.05436838e+00 3.33358869e-02 -2.28121325e-01
5.87039053e-01 -7.08661824e-02 -5.19169033e-01 -3.70318949e-01
4.38508987e-01 8.38504672e-01 4.45041567e-01 -1.00913382e+00
-6.34987414e-01 -2.21133769e-01 -4.79325473e-01 4.12194937e-01
-1.70104519e-01 6.51045263e-01 -1.35735646e-01 -8.44164789e-01
-2.15393379e-01 -1.23867309e+00 -2.51822114e-01 -8.56629848e-01
-9.92767140e-02 2.26948243e-02 1.32833645e-01 -1.14366651e+00
5.82160354e-01 -1.52886701e+00 3.72904181e-01 1.95618458e-02
-4.49290156e-01 3.12704593e-01 -6.83390081e-01 6.83239579e-01
-1.64237469e-01 3.51375401e-01 -1.61584586e-01 -3.42318276e-03
-4.50938076e-01 5.94593465e-01 -2.13532552e-01 2.87214786e-01
4.07050848e-01 1.34036171e+00 -8.52867961e-01 -4.72578049e-01
-3.11019897e-01 2.40162849e-01 -2.86945611e-01 -1.54580027e-01
-1.14335403e-01 8.46184492e-01 -6.57843202e-02 5.17037928e-01
3.81549388e-01 4.12921667e-01 3.81062508e-01 4.18379217e-01
-2.30331421e-01 6.87368929e-01 -3.12978119e-01 2.04244685e+00
-8.73974979e-01 6.53048277e-01 -6.84852228e-02 -7.41711676e-01
1.06491661e+00 4.46370095e-01 1.85285628e-01 -1.02469170e+00
7.66693950e-02 1.09605968e+00 4.75590110e-01 -1.98470131e-01
4.47306484e-01 -5.78847706e-01 -2.66372770e-01 7.16529548e-01
3.42831165e-01 -4.18384194e-01 2.49046981e-01 -2.25000307e-01
5.70397258e-01 5.33875346e-01 2.59194285e-01 -7.67056942e-01
3.03721160e-01 5.18377542e-01 5.94739676e-01 4.47111845e-01
2.42487460e-01 3.66429120e-01 7.32784271e-02 -5.29609740e-01
-1.77768624e+00 -1.06756186e+00 1.90781020e-02 1.43643486e+00
-7.12760806e-01 -3.63301523e-02 -7.83164144e-01 -6.32489562e-01
-3.75039667e-01 6.84949756e-01 -2.83352464e-01 1.38143227e-01
-1.36371005e+00 -9.53373373e-01 1.07476079e+00 4.03528780e-01
3.17820221e-01 -1.09343743e+00 4.14618492e-01 2.78121799e-01
-5.85592687e-01 -9.90360022e-01 -4.70049769e-01 2.69566327e-01
-1.22746611e+00 -3.98548484e-01 -1.11183274e+00 -1.28427482e+00
4.69648212e-01 -6.49279207e-02 1.42784953e+00 -2.15330511e-01
2.65914798e-01 -5.47615647e-01 -1.15030587e-01 -4.42398995e-01
-1.17413855e+00 9.28362250e-01 4.32371318e-01 -5.63187301e-01
4.17563796e-01 -6.37742639e-01 2.79504091e-01 3.77562344e-01
-2.57040411e-01 2.82892227e-01 8.90504181e-01 1.18596399e+00
3.59116077e-01 -7.37793684e-01 9.17458832e-01 -1.05751789e+00
7.42297769e-01 -2.95797348e-01 -4.76012111e-01 3.74188393e-01
-5.12105346e-01 3.16488117e-01 9.10105407e-01 -3.69304866e-01
-9.83441591e-01 6.53827339e-02 -2.88542092e-01 1.20961763e-01
1.74953729e-01 3.88895333e-01 3.64635848e-02 -8.49556178e-02
1.30394340e+00 1.55740887e-01 -1.48772895e-01 -4.81366485e-01
5.24703383e-01 8.83728147e-01 4.14495975e-01 -8.34566534e-01
9.42445815e-01 -9.61408541e-02 -2.82902479e-01 -6.90573156e-01
-2.83575356e-01 3.37161869e-03 -1.00099707e+00 1.41323522e-01
6.57863617e-01 -1.29757369e+00 2.09996164e-01 2.52907306e-01
-1.37869287e+00 -7.14916289e-01 -7.35049993e-02 8.44081700e-01
-7.29534388e-01 -2.54860997e-01 -9.08063531e-01 -2.92386562e-02
-5.83772004e-01 -1.15361989e+00 9.68280971e-01 -3.02762777e-01
-3.36319715e-01 -1.24106717e+00 7.10262954e-01 1.94032609e-01
5.81041038e-01 -1.14641622e-01 1.01472819e+00 -7.11847007e-01
-2.36423925e-01 2.60643601e-01 -7.16331303e-02 4.23602372e-01
2.97096699e-01 -3.32259744e-01 -5.05293846e-01 -5.17325461e-01
-3.12547505e-01 -6.02177143e-01 6.36275172e-01 -3.89117114e-02
-1.74043864e-01 -2.87151694e-01 -2.36432657e-01 5.15173733e-01
1.29479980e+00 2.08010271e-01 5.25867879e-01 3.87937099e-01
7.09275186e-01 4.97564882e-01 4.19249743e-01 -4.36742514e-01
2.65447557e-01 7.17729628e-01 -6.09162033e-01 -5.14911532e-01
-3.02708894e-01 -3.50667536e-01 7.51743019e-01 1.57768452e+00
-4.29758370e-01 1.68914005e-01 -1.30666852e+00 8.31342757e-01
-1.62547362e+00 -7.08176792e-01 -9.71595198e-02 1.99448645e+00
1.47747231e+00 -1.75480038e-01 1.38227209e-01 -4.00812358e-01
6.45114541e-01 -3.27689767e-01 -6.52754307e-03 -9.66920912e-01
-4.64193910e-01 7.00867176e-01 8.07137251e-01 9.12617028e-01
-6.52930915e-01 1.88404489e+00 7.84733868e+00 8.91565740e-01
-1.23767364e+00 5.87410152e-01 4.59289342e-01 -6.13093413e-02
-6.11330152e-01 2.02328265e-02 -7.57002711e-01 7.95086399e-02
1.46578407e+00 -2.15990901e-01 9.18105543e-01 3.80597413e-01
2.08360583e-01 3.74647319e-01 -1.25743675e+00 6.53543174e-01
9.49573144e-02 -1.33274853e+00 1.75962463e-01 5.00890650e-02
1.42702448e+00 8.54988039e-01 -1.14905082e-01 5.02573192e-01
7.61444390e-01 -1.04735482e+00 3.42471123e-01 6.07759207e-02
1.20842290e+00 -1.01438963e+00 5.24012268e-01 4.18315202e-01
-6.66510046e-01 4.06509131e-01 -6.03243768e-01 -9.97509956e-02
7.47412890e-02 7.86983296e-02 -1.24337518e+00 5.85565925e-01
3.73585373e-01 5.03750682e-01 -4.15789336e-01 2.61208236e-01
-4.28831458e-01 7.40947723e-01 -2.84715325e-01 6.68488368e-02
4.45255250e-01 -3.94337445e-01 3.04923147e-01 1.38077760e+00
4.59822088e-01 -4.89725143e-01 8.81872028e-02 5.24984121e-01
-3.78342748e-01 6.58429742e-01 -1.20811188e+00 -9.88254249e-02
-8.77573155e-04 7.98883080e-01 -3.75531048e-01 -3.59408170e-01
-3.89871597e-01 1.23390603e+00 5.17794013e-01 3.36179465e-01
-5.72201371e-01 -3.31764013e-01 4.44525063e-01 -1.52866825e-01
-4.20418978e-02 -4.88868982e-01 -4.64418352e-01 -1.42205131e+00
-1.67090580e-01 -1.46051013e+00 -1.36064947e-01 -5.14712036e-01
-1.10726452e+00 1.05243516e+00 -2.20623314e-01 -1.14988029e+00
-9.74478662e-01 -5.69296479e-01 -2.75810182e-01 1.43609464e+00
-1.14945102e+00 -1.59272790e+00 6.86689615e-01 5.92527807e-01
7.55131602e-01 -9.23891127e-01 1.19753385e+00 6.44474566e-01
-1.74355805e-01 8.70541275e-01 5.59369564e-01 2.83080071e-01
1.09498131e+00 -1.06877899e+00 1.14961338e+00 7.93680251e-01
5.28354049e-01 5.44458151e-01 5.01159906e-01 -7.29638457e-01
-1.08838761e+00 -1.06222212e+00 1.60513854e+00 -5.47799587e-01
6.94378257e-01 -5.33524454e-01 -4.46935654e-01 1.11867476e+00
7.94751048e-01 -5.43290138e-01 7.24550962e-01 2.74648875e-01
-3.94072354e-01 1.17274992e-01 -8.51698220e-01 8.21649373e-01
1.04715836e+00 -8.18300843e-01 -6.65392041e-01 6.98372483e-01
7.49891102e-01 -2.09078118e-01 -8.94029200e-01 2.43043259e-01
6.04849100e-01 -2.13207468e-01 7.23753691e-01 -1.10836673e+00
7.01072812e-01 3.31653608e-03 -2.25980580e-01 -2.10641599e+00
-1.77577317e-01 -7.86222935e-01 9.33336377e-01 1.07257712e+00
1.34678721e+00 -6.78555191e-01 4.42755312e-01 -1.89768821e-01
-7.56392479e-02 -2.35435024e-01 -9.25028503e-01 -1.06736565e+00
1.15443122e+00 -1.11284792e-01 4.88457799e-01 1.39250684e+00
5.14927879e-02 1.07711852e+00 -6.68996990e-01 -4.52157319e-01
1.70164868e-01 -3.51599008e-02 9.58951712e-01 -8.20296228e-01
-4.43517625e-01 -2.75417387e-01 1.48596123e-01 -6.34360254e-01
5.07256448e-01 -1.54849112e+00 6.26133755e-02 -1.46349955e+00
6.33405475e-03 -5.99757373e-01 1.14599340e-01 6.65035129e-01
9.25626885e-03 7.16152191e-01 9.12608504e-02 5.04401267e-01
1.20341711e-01 1.68447509e-01 1.46163833e+00 -2.09093049e-01
-3.30792427e-01 -2.32130989e-01 -5.31590521e-01 4.13113594e-01
1.03270853e+00 -7.33452976e-01 -5.45302331e-02 -1.17808831e+00
3.40117812e-01 2.04537719e-01 -6.88827634e-01 -6.46273911e-01
-2.61242181e-01 -2.58573234e-01 3.87622505e-01 -9.57314894e-02
4.15484980e-02 -4.40266222e-01 1.30154788e-01 5.87385356e-01
-3.58253092e-01 7.06188440e-01 3.78117055e-01 -2.42951825e-01
-2.52621531e-01 7.03400671e-02 8.30469310e-01 -5.47075629e-01
-3.12378973e-01 -4.10021320e-02 -5.86499393e-01 1.19511127e-01
6.03609443e-01 2.52746135e-01 -2.06346527e-01 -2.19859317e-01
-5.42033434e-01 2.64875107e-02 5.92680454e-01 5.94092488e-01
1.05832189e-01 -1.63506126e+00 -1.57509458e+00 4.12595928e-01
5.91229908e-02 -6.47532642e-01 -6.56733274e-01 6.74895823e-01
-8.72362614e-01 6.00761354e-01 -8.76908839e-01 -4.56538528e-01
-9.23730075e-01 2.54621297e-01 3.98834020e-01 -5.06294310e-01
-1.02067985e-01 4.99315858e-01 -2.88070381e-01 -1.33656681e+00
-5.88599801e-01 1.28514066e-01 1.21518157e-01 -1.66739956e-01
8.94986391e-02 2.99604293e-02 2.49306098e-01 -9.36950684e-01
-1.20965041e-01 5.72910607e-01 -1.43388540e-01 -8.76589298e-01
1.17419565e+00 5.63725121e-02 -4.47855592e-01 5.54011524e-01
1.25292277e+00 3.92916858e-01 -3.63881946e-01 -1.96984440e-01
3.54241505e-02 -2.55106181e-01 -4.41577524e-01 -1.00226390e+00
-7.86935627e-01 8.05436015e-01 3.10277045e-01 -6.71742201e-01
8.34565520e-01 -1.83340162e-01 8.18058610e-01 7.45770514e-01
8.17942679e-01 -1.21087337e+00 -4.68842685e-01 1.04137874e+00
7.80417979e-01 -1.19057465e+00 -2.47555450e-01 9.56085604e-03
-5.91504157e-01 9.77553368e-01 4.36144471e-01 -1.62269115e-01
1.05519541e-01 3.98574680e-01 6.20168388e-01 4.40216511e-01
-7.49495804e-01 -6.84489682e-02 2.36214586e-02 5.32968700e-01
8.17918718e-01 4.00426358e-01 -9.74195898e-01 -9.23530236e-02
-7.00305045e-01 -1.57214209e-01 4.06712741e-01 6.22753978e-01
-5.14332503e-02 -1.90063930e+00 -4.36512500e-01 1.95292264e-01
-7.58652389e-01 -5.42337775e-01 -7.83755124e-01 9.13556993e-01
3.71327298e-03 7.11569071e-01 -4.10244092e-02 -4.97659117e-01
6.21050708e-02 4.48713422e-01 7.98559964e-01 -6.65783703e-01
-1.02572143e+00 2.98502386e-01 6.52107716e-01 -1.50284871e-01
-2.64251441e-01 -5.67906022e-01 -9.87924218e-01 -4.56576526e-01
-2.67091542e-01 4.10513431e-01 9.36868668e-01 9.98841405e-01
4.47667725e-02 8.83693397e-02 7.69832313e-01 -8.55795979e-01
-4.71999586e-01 -1.35222697e+00 7.50097213e-03 1.14717416e-01
-4.13224623e-02 3.40122543e-02 2.67382525e-02 1.13559291e-01] | [11.588528633117676, 10.331844329833984] |
55f6c0d9-7ea3-447e-a9c0-de8c6c58ea8f | thailmcut-unsupervised-pretraining-for-thai | null | null | https://aclanthology.org/2020.lrec-1.858 | https://aclanthology.org/2020.lrec-1.858.pdf | ThaiLMCut: Unsupervised Pretraining for Thai Word Segmentation | We propose ThaiLMCut, a semi-supervised approach for Thai word segmentation which utilizes a bi-directional character language model (LM) as a way to leverage useful linguistic knowledge from unlabeled data. After the language model is trained on substantial unlabeled corpora, the weights of its embedding and recurrent layers are transferred to a supervised word segmentation model which continues fine-tuning them on a word segmentation task. Our experimental results demonstrate that applying the LM always leads to a performance gain, especially when the amount of labeled data is small. In such cases, the F1 Score increased by up to 2.02{\%}. Even on abig labeled dataset, a small improvement gain can still be obtained. The approach has also shown to be very beneficial for out-of-domain settings with a gain in F1 Score of up to 3.13{\%}. Finally, we show that ThaiLMCut can outperform other open source state-of-the-art models achieving an F1 Score of 98.78{\%} on the standard benchmark, InterBEST2009. | ['Hinrich Sch{\\"u}tze', 'Michael Matuschek', 'Liliana Mamani Sanchez', 'Ivan Bilan', 'Suteera Seeha', 'Johannes Huber'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['thai-word-tokenization'] | ['natural-language-processing'] | [ 2.31034055e-01 4.77729261e-01 -5.69801807e-01 -3.35303307e-01
-1.30889785e+00 -8.05529594e-01 3.01150560e-01 3.53483036e-02
-8.21769118e-01 7.55835176e-01 1.43959522e-01 -4.83388275e-01
4.45900738e-01 -4.61462498e-01 -6.26609743e-01 -5.77743411e-01
2.68614113e-01 6.83015406e-01 4.20849442e-01 -6.44989237e-02
1.72813699e-01 1.16895109e-01 -1.00157380e+00 1.15277991e-01
1.02285707e+00 6.62528694e-01 1.66125089e-01 6.27621531e-01
-5.55211723e-01 3.11611503e-01 -5.50672472e-01 -5.46891809e-01
2.82885760e-01 -1.58225715e-01 -9.86778498e-01 1.49026722e-01
2.80585617e-01 -1.25186518e-01 -4.93697412e-02 8.46931517e-01
3.31296891e-01 1.27287284e-01 6.31593943e-01 -5.07613063e-01
-3.82001251e-01 1.06073463e+00 -7.87423730e-01 8.56624618e-02
-3.63077819e-02 2.01542020e-01 1.45282543e+00 -1.03880990e+00
7.03079343e-01 1.02710414e+00 4.62481469e-01 7.51902282e-01
-1.46971679e+00 -5.60279191e-01 4.14822847e-01 -3.70445013e-01
-9.68762100e-01 -2.83543408e-01 6.39942825e-01 -2.08061874e-01
1.37987506e+00 4.67516072e-02 3.54958445e-01 9.08502162e-01
-2.21946508e-01 1.29377449e+00 1.05050159e+00 -7.18985200e-01
1.09945677e-01 2.08032846e-01 6.84183836e-01 5.72588801e-01
1.26607537e-01 -2.54356116e-01 -4.45185214e-01 1.75768644e-01
2.88558841e-01 -4.97727126e-01 1.02139392e-03 5.51238209e-02
-1.03635561e+00 9.52641666e-01 2.97220588e-01 4.82114106e-01
-1.92576781e-01 9.48514044e-03 4.80639666e-01 5.71380518e-02
7.09516823e-01 6.27264738e-01 -6.92433655e-01 -3.28041255e-01
-1.26281548e+00 -8.29932019e-02 7.43493319e-01 7.67446518e-01
7.22163498e-01 2.47376710e-01 -2.90244788e-01 1.11786473e+00
2.62971312e-01 4.57127482e-01 5.56731403e-01 -8.15415919e-01
7.73005068e-01 6.74625456e-01 2.49467660e-02 -1.70572340e-01
-2.60401964e-01 -4.83585179e-01 -2.81922072e-01 -2.30068147e-01
5.79270184e-01 -4.20875728e-01 -1.48627746e+00 1.80900455e+00
4.27767448e-02 -1.49135858e-01 6.89901188e-02 5.98298788e-01
5.38879812e-01 9.11640882e-01 2.54611939e-01 -1.30989879e-01
1.07383299e+00 -1.31319642e+00 -5.40241659e-01 -8.31624627e-01
7.83829629e-01 -9.41384852e-01 1.43582737e+00 4.25023317e-01
-1.02293658e+00 -3.63858730e-01 -9.75145638e-01 2.06369068e-02
-3.13076913e-01 1.81460917e-01 5.10238171e-01 8.70200932e-01
-9.54860270e-01 4.05777216e-01 -1.00160086e+00 -3.58198196e-01
3.50415498e-01 5.78056812e-01 -1.07097261e-01 -2.09222496e-01
-9.66750801e-01 6.48632288e-01 6.84032023e-01 -4.72390950e-02
-7.40797281e-01 -6.49799943e-01 -8.14774752e-01 -8.20302144e-02
6.21372879e-01 -3.94910127e-02 1.37764955e+00 -8.21115494e-01
-1.55126345e+00 9.90980983e-01 -2.95003533e-01 -5.70377052e-01
5.02059102e-01 -5.41713655e-01 -2.25726917e-01 5.56480736e-02
2.33159840e-01 9.46403205e-01 5.39188981e-01 -1.15847766e+00
-6.31393194e-01 -1.36312306e-01 -1.37368724e-01 1.24157369e-01
-6.31939292e-01 2.71224533e-03 -9.39943194e-01 -7.29822993e-01
-3.77932526e-02 -1.11423767e+00 -3.76236320e-01 -6.06333315e-01
-6.92741454e-01 -2.64958233e-01 5.35526276e-01 -7.39920378e-01
1.44031799e+00 -2.02895498e+00 -3.31799616e-03 2.24926889e-01
-3.20561156e-02 6.80873930e-01 -2.84388185e-01 2.95048505e-01
1.41134575e-01 3.96431565e-01 -6.31020308e-01 -3.99949104e-01
-1.76260229e-02 3.42599541e-01 -1.59069642e-01 2.85035390e-02
3.45136017e-01 1.13251328e+00 -7.43102610e-01 -3.43495488e-01
1.56620234e-01 2.43081361e-01 -4.58529055e-01 1.77757055e-01
-4.54955459e-01 1.27781570e-01 -3.78793448e-01 5.02370059e-01
3.55393708e-01 -2.43419066e-01 3.69569182e-01 3.00488681e-01
-1.52811464e-02 4.96992856e-01 -8.69855046e-01 1.75020444e+00
-5.38921773e-01 6.15199625e-01 -1.84127286e-01 -8.28964591e-01
8.80758345e-01 1.35477096e-01 2.48008937e-01 -5.40368557e-01
1.11394510e-01 4.58223581e-01 1.35009393e-01 -1.57078102e-01
5.92480063e-01 -5.96795557e-03 -3.46360028e-01 6.81033313e-01
1.62074327e-01 -2.02926010e-01 5.19840539e-01 2.99130440e-01
1.07111204e+00 8.15662146e-02 -9.84001160e-02 -3.45676184e-01
3.43304366e-01 2.01784715e-01 4.96232480e-01 6.12241745e-01
-5.43472916e-02 6.24381840e-01 4.41807777e-01 -1.12240307e-01
-9.97194707e-01 -9.95645165e-01 2.22759154e-02 1.40485954e+00
-1.87008902e-01 -3.27716559e-01 -1.11394894e+00 -1.02341068e+00
-1.86339021e-01 9.20805573e-01 -5.27561128e-01 -8.47768411e-02
-7.86410332e-01 -9.07639861e-01 7.62758136e-01 7.61702299e-01
4.93335605e-01 -1.32166255e+00 -3.76426101e-01 4.52864230e-01
-6.44612387e-02 -1.31575239e+00 -5.51140904e-01 5.76495707e-01
-8.68059695e-01 -6.09877288e-01 -1.01779425e+00 -1.05879235e+00
6.20450139e-01 -9.86522138e-02 1.16379988e+00 -1.02759600e-01
1.88466236e-02 -1.65992733e-02 -3.79086763e-01 -1.59722880e-01
-2.31227592e-01 7.21714735e-01 -1.78122416e-01 -2.14705214e-01
6.11283123e-01 -1.05989642e-01 -4.73387569e-01 1.36224344e-01
-7.82400608e-01 -8.23105946e-02 5.94759643e-01 8.02144647e-01
5.85737526e-01 -3.63545626e-01 6.35203958e-01 -1.46157372e+00
5.57149947e-01 -1.26504809e-01 -4.98514980e-01 1.87088296e-01
-8.78870130e-01 2.23565787e-01 5.12328744e-01 -4.71643746e-01
-1.04314244e+00 1.71877712e-01 -3.81882966e-01 1.26964152e-01
-8.94249007e-02 6.28169119e-01 -6.85101002e-02 3.49719882e-01
4.90746289e-01 5.52473217e-02 -4.49833065e-01 -6.96913362e-01
6.25641167e-01 8.34841847e-01 4.34431374e-01 -5.50058007e-01
6.53432310e-01 2.06515014e-01 -7.42064357e-01 -7.54126191e-01
-1.20534635e+00 -7.41432428e-01 -7.34106481e-01 2.67946154e-01
8.85715306e-01 -8.11628938e-01 -1.10973805e-01 3.64853978e-01
-7.82433569e-01 -8.83980811e-01 -3.47175926e-01 2.92944223e-01
-2.85883874e-01 3.29320759e-01 -9.35644031e-01 -7.86057293e-01
-4.51018542e-01 -1.04757440e+00 9.34384167e-01 3.20548683e-01
-4.69728172e-01 -1.11476600e+00 2.40008463e-03 5.79398751e-01
2.32678384e-01 -9.62678157e-03 9.24191535e-01 -1.01837838e+00
-1.71139345e-01 -2.42310911e-01 -2.06886709e-01 5.02981186e-01
1.28935724e-01 -1.16430335e-01 -1.03906381e+00 -3.35849375e-01
-4.64062482e-01 -5.89720547e-01 1.17673206e+00 2.90546626e-01
8.63982499e-01 -9.46444869e-02 -2.67889917e-01 2.20804751e-01
1.23251045e+00 1.65780827e-01 3.34861159e-01 2.42821425e-01
7.54726171e-01 5.03256381e-01 6.70463681e-01 1.42787993e-02
2.04573542e-01 3.94726962e-01 7.86722600e-02 -2.57319868e-01
-2.03137562e-01 -3.52330029e-01 4.53434139e-01 1.11066675e+00
3.40612739e-01 -3.55764300e-01 -1.27216542e+00 9.14897919e-01
-1.66477954e+00 -3.95281583e-01 -4.54626493e-02 2.09251380e+00
1.24236393e+00 7.80835450e-01 1.57567188e-01 1.93988189e-01
6.07081294e-01 2.56442577e-01 -6.71854138e-01 -7.95996606e-01
-1.09776437e-01 5.73066771e-01 6.72187150e-01 6.68797672e-01
-1.06305337e+00 1.79749548e+00 6.29497528e+00 1.15135026e+00
-1.28927016e+00 1.26609728e-01 8.69593740e-01 8.06847438e-02
-4.48482692e-01 -3.12019326e-02 -9.64360893e-01 3.63582224e-01
1.25277364e+00 -9.75164026e-03 3.10282528e-01 6.96013272e-01
7.88399279e-02 -2.84342349e-01 -6.79361880e-01 6.34773493e-01
6.60526603e-02 -1.13153148e+00 -6.71252459e-02 1.64683715e-01
1.03127754e+00 3.65333229e-01 1.12879954e-01 5.42457044e-01
6.00142121e-01 -1.13230801e+00 4.48516160e-01 -2.54287750e-01
1.11872900e+00 -1.01353717e+00 7.66993105e-01 4.40845221e-01
-1.02475595e+00 2.73903370e-01 -3.09692025e-01 1.41557112e-01
2.68407822e-01 6.32342100e-01 -1.00623310e+00 2.98771650e-01
4.55853105e-01 5.46198964e-01 -5.50578952e-01 8.04711103e-01
-6.96148098e-01 1.36288190e+00 -4.14587975e-01 -1.28500596e-01
8.27830553e-01 -4.12448905e-02 2.34077930e-01 1.63426077e+00
-5.86557239e-02 -8.87182727e-02 2.91942269e-01 5.40908992e-01
-4.42005634e-01 3.49793494e-01 -3.40079635e-01 -4.42542136e-01
2.11375788e-01 1.20741332e+00 -1.09264290e+00 -4.43592429e-01
-2.51600474e-01 1.12049985e+00 4.26388949e-01 4.66123253e-01
-6.83340013e-01 -5.36650538e-01 1.92699596e-01 -1.12363860e-01
7.12545156e-01 -2.90701091e-01 -5.39041221e-01 -9.96945918e-01
-8.89544189e-02 -8.34784865e-01 3.47811490e-01 -3.07973146e-01
-9.76815879e-01 7.09997118e-01 -3.14977139e-01 -7.64472187e-01
-4.10802603e-01 -7.77490199e-01 -4.14479524e-01 8.33386838e-01
-1.51225948e+00 -1.26956332e+00 2.87567258e-01 1.51443407e-01
1.03239679e+00 -1.86511621e-01 8.36450577e-01 1.31169796e-01
-7.55858660e-01 8.78083050e-01 2.04042464e-01 4.60626960e-01
5.95757961e-01 -1.47526145e+00 7.19137132e-01 1.09423065e+00
5.91993153e-01 5.17232180e-01 4.97382104e-01 -6.26682341e-01
-1.06610060e+00 -1.10939467e+00 1.13703859e+00 -4.64760423e-01
8.16810369e-01 -6.35908723e-01 -9.64406073e-01 7.78575957e-01
4.03319895e-01 -9.55907553e-02 8.44851434e-01 3.02360386e-01
-4.59408492e-01 1.96442172e-01 -9.76230145e-01 6.55827463e-01
8.21642876e-01 -4.06197190e-01 -7.73066044e-01 2.25952551e-01
9.08670425e-01 -3.37123722e-01 -6.52577043e-01 2.47020289e-01
4.61661100e-01 -6.22984707e-01 6.19055688e-01 -6.69485271e-01
3.36538851e-01 4.34564389e-02 -1.70409426e-01 -1.28233564e+00
-7.15771038e-03 -9.14214253e-01 6.66471422e-02 1.40826523e+00
1.18525875e+00 -5.17505050e-01 1.01023006e+00 6.04167044e-01
-1.13628507e-02 -8.02844822e-01 -7.62592733e-01 -7.58612394e-01
5.01249194e-01 -7.97282159e-01 1.45124391e-01 6.60918951e-01
-4.78055589e-02 6.08527482e-01 -1.49164379e-01 -2.39955291e-01
3.22525442e-01 1.31752431e-01 4.67090636e-01 -9.29229021e-01
-3.04379851e-01 -3.66051883e-01 1.59264237e-01 -1.32924032e+00
5.31929910e-01 -8.97389591e-01 2.58221388e-01 -1.54290342e+00
1.64983481e-01 -5.08717060e-01 -6.10736668e-01 7.67217100e-01
-3.33927244e-01 5.84297240e-01 3.16940337e-01 -6.20787293e-02
-5.77936649e-01 2.16716632e-01 1.00712395e+00 -2.24362746e-01
-4.25154120e-01 5.11106625e-02 -7.52632499e-01 7.30604768e-01
9.73560512e-01 -5.28362811e-01 -3.46112609e-01 -6.97106183e-01
-8.12148973e-02 -2.71724641e-01 -5.11429429e-01 -6.87879741e-01
-6.05755039e-02 -9.79807377e-02 7.58178979e-02 -5.07168889e-01
3.36254776e-01 -5.13688266e-01 -5.50304234e-01 4.48939562e-01
-4.67799366e-01 -1.07773386e-01 4.76558119e-01 3.34216386e-01
-1.80732653e-01 -3.22361171e-01 6.92398131e-01 -6.20333366e-02
-6.95993483e-01 1.16403960e-01 -3.45403880e-01 5.12272656e-01
8.31851006e-01 -2.18798175e-01 -1.91006288e-02 -1.75562605e-01
-6.30958021e-01 3.61775249e-01 2.79063851e-01 4.36309069e-01
2.43964747e-01 -9.21615481e-01 -6.41984820e-01 1.36531547e-01
9.79380533e-02 -2.92149242e-02 -3.08961242e-01 5.93185902e-01
-3.77930939e-01 6.09397411e-01 2.49735445e-01 -5.23856640e-01
-1.05871987e+00 2.40035251e-01 -5.56268133e-02 -6.90149844e-01
-5.00456691e-01 1.09694588e+00 6.33696541e-02 -5.85484147e-01
2.77739614e-01 -4.76682931e-01 -1.12392880e-01 2.24999875e-01
2.44724676e-01 1.69914797e-01 1.42781258e-01 -6.70536518e-01
-4.03444290e-01 7.29047179e-01 -3.85711193e-01 -5.81360221e-01
1.41856444e+00 3.41767855e-02 2.99993187e-01 6.34747207e-01
1.15204382e+00 4.14792985e-01 -1.38138366e+00 -3.10620070e-01
4.09833372e-01 -6.57874867e-02 -9.56359580e-02 -1.17438388e+00
-1.09293699e+00 1.13298225e+00 1.97550356e-01 3.18818800e-02
8.67864549e-01 -5.73205296e-03 1.18354309e+00 4.15851057e-01
1.69030800e-01 -1.33235276e+00 3.98043022e-02 8.51792276e-01
3.87322962e-01 -1.31019986e+00 -2.55595118e-01 -3.62495393e-01
-9.40008581e-01 9.02674854e-01 4.34706360e-01 -2.76092082e-01
4.02202874e-01 3.11668724e-01 4.00125861e-01 1.96250051e-01
-4.55970168e-01 -4.78321999e-01 4.35760826e-01 3.48294586e-01
6.80641353e-01 1.92708492e-01 -4.19973701e-01 6.61736429e-01
-3.10252786e-01 -3.80684465e-01 2.41456658e-01 7.67317653e-01
-6.08963370e-01 -1.36710274e+00 7.52437115e-02 4.57429916e-01
-7.78181493e-01 -3.76803428e-01 -5.47102273e-01 8.45602393e-01
-1.80156395e-01 9.97104883e-01 -2.74723303e-02 -2.10533693e-01
2.62118906e-01 3.70054096e-01 8.41465890e-02 -1.01316237e+00
-7.64859259e-01 5.13464868e-01 2.54974127e-01 -3.70322496e-01
-1.11578546e-01 -4.26075906e-01 -1.61185491e+00 1.36120185e-01
-2.86435992e-01 1.90849200e-01 4.44611073e-01 1.03694677e+00
-6.18401282e-02 3.68708491e-01 4.02326584e-01 -4.64902908e-01
-4.19328660e-01 -9.69480276e-01 -3.76226097e-01 2.21768662e-01
1.08923271e-01 -1.80224314e-01 -4.03575569e-01 8.59053135e-02] | [10.043135643005371, 10.07657241821289] |
7241e7b3-8379-499a-917d-e4ea3f68bdba | mucic-lt-edi-acl2022-hope-speech-detection | null | null | https://aclanthology.org/2022.ltedi-1.20 | https://aclanthology.org/2022.ltedi-1.20.pdf | MUCIC@LT-EDI-ACL2022: Hope Speech Detection using Data Re-Sampling and 1D Conv-LSTM | Spreading positive vibes or hope content on social media may help many people to get motivated in their life. To address Hope Speech detection in YouTube comments, this paper presents the description of the models submitted by our team - MUCIC, to the Hope Speech Detection for Equality, Diversity, and Inclusion (HopeEDI) shared task at Association for Computational Linguistics (ACL) 2022. This shared task consists of texts in five languages, namely: English, Spanish (in Latin scripts), and Tamil, Malayalam, and Kannada (in code-mixed native and Roman scripts) with the aim of classifying the YouTube comment into “Hope”, “Not-Hope” or “Not-Intended” categories. The proposed methodology uses the re-sampling technique to deal with imbalanced data in the corpus and obtained 1st rank for English language with a macro-averaged F1-score of 0.550 and weighted-averaged F1-score of 0.860. The code to reproduce this work is available in GitHub. | ['Grigori Sidorov', 'Hosahalli Shashirekha', 'Fazlourrahman Balouchzahi', 'Anusha Gowda'] | null | null | null | null | ltedi-acl-2022-5 | ['hope-speech-detection'] | ['natural-language-processing'] | [-4.77632761e-01 4.05036807e-01 -7.34310269e-01 -9.21041444e-02
-1.04180288e+00 -3.38908046e-01 7.77896225e-01 5.89685738e-01
-3.12313855e-01 8.74180734e-01 1.00447845e+00 -2.39330411e-01
1.68852538e-01 -3.35611165e-01 -9.61985886e-02 -2.27163523e-01
3.64595443e-01 2.48859569e-01 -2.00674012e-01 -4.40235168e-01
6.11370444e-01 -1.99314058e-01 -1.45543575e+00 6.22778356e-01
9.95381594e-01 3.32533211e-01 -3.07823002e-01 8.69605839e-01
-2.42145032e-01 1.27814925e+00 -4.40403908e-01 -9.36091721e-01
-1.50472328e-01 -7.56285667e-01 -1.02349579e+00 -3.01728874e-01
2.90481448e-01 -3.28947157e-02 -1.77217603e-01 1.22660208e+00
6.38857901e-01 -1.07043557e-01 7.47699142e-01 -1.20203698e+00
-5.66647232e-01 1.23870540e+00 -5.66353023e-01 2.80353516e-01
5.77543259e-01 -5.76027073e-02 1.24832273e+00 -1.16494799e+00
9.35663640e-01 1.37867963e+00 5.71570396e-01 5.97949147e-01
-5.52077949e-01 -5.00945032e-01 -7.28859782e-01 3.16436917e-01
-1.20457494e+00 -8.02795410e-01 7.94857204e-01 -5.15117228e-01
1.10591722e+00 6.26436412e-01 8.78377259e-01 1.31219685e+00
2.69783046e-02 1.09194314e+00 9.01636243e-01 -3.69362563e-01
2.42580220e-01 6.59678221e-01 2.53062904e-01 4.17588651e-01
-1.71081722e-01 -6.49082959e-01 -1.08634973e+00 -4.49029058e-01
-3.47541273e-01 -4.54365969e-01 2.61098612e-02 2.93462425e-01
-9.53588784e-01 1.33801901e+00 -2.08686054e-01 4.34928149e-01
-3.58225077e-01 -4.59408373e-01 8.52817655e-01 3.03184867e-01
1.16871059e+00 -2.36939639e-02 -1.27270475e-01 -9.86353636e-01
-1.16266000e+00 3.33003044e-01 1.01923263e+00 6.82869375e-01
1.98072046e-01 -1.64787039e-01 -2.09499344e-01 1.34779108e+00
6.32313788e-01 4.91468608e-01 9.84340191e-01 -6.61517441e-01
7.50239670e-01 7.39755273e-01 -2.92173848e-02 -9.60609138e-01
-2.94252038e-01 -2.35967949e-01 -7.14173436e-01 -2.58630931e-01
2.94209749e-01 -3.39230776e-01 -1.14776388e-01 1.40120614e+00
3.14738601e-01 -4.85495865e-01 4.95415479e-01 4.88221288e-01
1.34381747e+00 7.65168905e-01 7.30232475e-03 -7.04063952e-01
1.25600755e+00 -1.12849379e+00 -9.53630924e-01 2.29760725e-02
7.70793915e-01 -1.36183560e+00 1.13880551e+00 3.02832693e-01
-1.36569977e+00 1.59114152e-02 -8.55928242e-01 -2.78058767e-01
-2.64183819e-01 1.74319759e-01 7.15318620e-02 1.13126910e+00
-8.58464003e-01 3.63289207e-01 -2.20244884e-01 -7.44043648e-01
6.68591022e-01 -2.70186245e-01 -2.02205569e-01 3.68523836e-01
-1.20384026e+00 5.41457295e-01 -7.17283189e-02 -6.33595645e-01
-5.89992046e-01 -4.42088306e-01 -6.12195134e-01 -2.00785860e-01
-2.39149526e-01 1.56410292e-01 9.51467097e-01 -1.21400928e+00
-1.20654452e+00 1.19088340e+00 -3.74761879e-01 -4.76177871e-01
7.27914572e-01 -1.86911896e-01 -3.35097253e-01 -1.48459479e-01
2.95060158e-01 4.79989111e-01 7.26225078e-01 -5.11378169e-01
-3.61567706e-01 -4.75087851e-01 -1.72944352e-01 3.45814377e-01
-7.94996381e-01 6.85669184e-01 2.69281626e-01 -6.02475286e-01
-2.14775652e-01 -6.25871480e-01 4.26781356e-01 -5.14351249e-01
-5.70212066e-01 -6.36183202e-01 8.35439622e-01 -9.64018047e-01
1.58755422e+00 -2.20418096e+00 -4.97280695e-02 -1.41867325e-01
3.24304491e-01 1.75547197e-01 1.36985511e-01 7.09699512e-01
3.85132730e-02 5.32634139e-01 1.23310074e-01 -4.80301797e-01
-7.32768327e-02 -1.23449311e-01 2.07854100e-02 7.18870759e-01
-1.23000585e-01 5.29325366e-01 -1.05350375e+00 -5.53633809e-01
-2.56775111e-01 5.24224699e-01 -6.13769233e-01 1.08289728e-02
1.16933122e-01 3.39197111e-03 -1.89141139e-01 5.38806915e-01
6.52568698e-01 1.08891763e-01 7.13097155e-02 2.45347679e-01
-5.59549689e-01 6.94420636e-01 -8.93793523e-01 8.91332805e-01
-1.33468881e-01 1.03601050e+00 -1.21164560e-01 -5.71702838e-01
1.23724031e+00 4.03271019e-01 5.05608797e-01 -5.60951114e-01
3.86957496e-01 4.04694200e-01 1.17472537e-01 -7.98176110e-01
9.17841256e-01 2.69450564e-02 8.52044895e-02 6.71799898e-01
-8.93175602e-02 -7.47689903e-02 3.97437513e-01 5.59485495e-01
7.91656256e-01 -4.93891567e-01 6.80392981e-01 -5.14758527e-01
7.87733316e-01 -1.97541177e-01 3.03134263e-01 4.01846081e-01
-7.94260681e-01 5.59040248e-01 1.06375480e+00 -2.48823330e-01
-1.34330213e+00 -4.84377593e-01 -1.67370468e-01 1.20540416e+00
-4.43002343e-01 -8.67557049e-01 -8.59595835e-01 -5.42929471e-01
-4.20134187e-01 1.09815264e+00 -4.58049297e-01 -4.82955277e-02
-1.33508161e-01 -6.47143006e-01 7.85307646e-01 -3.93907279e-01
5.31040788e-01 -1.11225498e+00 -4.66285855e-01 -6.14384748e-02
-7.85195589e-01 -7.76617289e-01 -4.05728489e-01 -2.87112355e-01
-4.36797321e-01 -9.24743772e-01 -9.61995363e-01 -5.42846382e-01
-4.57171574e-02 -1.42908484e-01 9.49322462e-01 -7.88886100e-02
-9.74983349e-02 1.18913084e-01 -8.52117777e-01 -4.29656476e-01
-7.64575243e-01 2.33836621e-01 -1.23839080e-01 1.11436926e-01
7.63665259e-01 -9.62015763e-02 -9.90386903e-02 -3.50969285e-02
-3.74768078e-01 9.70113799e-02 -2.53707647e-01 7.79009342e-01
-1.12300880e-01 -3.51671815e-01 8.46456468e-01 -8.28149080e-01
1.01748073e+00 -1.08622646e+00 1.83834344e-01 -1.10742033e-01
-6.59758925e-01 -5.51117778e-01 3.25771481e-01 -1.07556194e-01
-8.62483501e-01 -2.23861799e-01 -6.18190587e-01 2.05066204e-01
1.81661114e-01 6.51001692e-01 1.09886438e-01 4.66005266e-01
8.51897597e-01 1.14910349e-01 -3.24451411e-03 -3.86883795e-01
1.74899250e-01 1.53312922e+00 -2.10134163e-02 4.30455357e-02
-5.54532483e-02 8.94152522e-02 -7.64650702e-01 -1.27346361e+00
-7.22610116e-01 -7.98946023e-01 -3.96514118e-01 -6.52802348e-01
7.14601159e-01 -1.03694212e+00 -3.21532071e-01 8.02317381e-01
-1.34211385e+00 -6.21180013e-02 -3.60934079e-01 4.03284132e-01
-3.64699543e-01 4.14379746e-01 -5.42872906e-01 -1.27952409e+00
-6.93421841e-01 -8.77097249e-01 6.21600091e-01 2.55828291e-01
-6.90978944e-01 -9.89850819e-01 2.99584210e-01 8.72886717e-01
3.47803324e-01 -2.84907687e-02 7.51697838e-01 -1.25072622e+00
5.21162391e-01 -1.43157527e-01 -1.12352297e-01 4.41923380e-01
-1.99586764e-01 3.33510280e-01 -9.40554202e-01 5.68250865e-02
1.95334330e-01 -8.62620533e-01 5.52897215e-01 1.45052776e-01
4.03096110e-01 -9.25647199e-01 2.53486246e-01 -3.43301833e-01
9.71051097e-01 -3.84753138e-01 8.04517150e-01 2.15305761e-01
3.41189861e-01 6.22751415e-01 6.01823568e-01 1.26772141e+00
7.45089650e-01 4.64343220e-01 4.18098122e-01 4.89611596e-01
-1.86471552e-01 -2.86757946e-01 8.37427497e-01 1.16469252e+00
7.51606598e-02 -4.67289031e-01 -1.05148160e+00 9.21795905e-01
-1.65888321e+00 -1.36553526e+00 -8.45504820e-01 2.07676363e+00
1.02602315e+00 -1.22279562e-02 7.21596301e-01 3.75763834e-01
6.08782828e-01 4.71679389e-01 -2.87377480e-02 -1.05333257e+00
-4.23219025e-01 -3.58401388e-01 2.34227255e-01 9.25497353e-01
-8.56462121e-01 7.21149147e-01 5.35135269e+00 1.14241183e+00
-1.03525591e+00 5.85404992e-01 1.03373086e+00 -1.70517668e-01
-6.05594397e-01 -9.00292620e-02 -7.82815695e-01 5.91570020e-01
1.51877737e+00 -6.38828397e-01 2.74158865e-01 8.48939300e-01
6.58333778e-01 -2.33269721e-01 -4.39504385e-01 1.03805637e+00
6.17099524e-01 -1.09634888e+00 -2.79905915e-01 -7.65227899e-02
8.53072882e-01 4.64595765e-01 2.05492184e-01 3.35575402e-01
-3.56333777e-02 -8.52825046e-01 8.96037400e-01 2.60998517e-01
5.90901494e-01 -6.60778344e-01 8.59398067e-01 8.41294885e-01
-5.54037154e-01 -6.79442734e-02 -4.69885528e-01 -3.08680087e-01
2.85470225e-02 7.68010259e-01 -7.51076877e-01 2.43804306e-02
6.03646636e-01 1.20509636e+00 -6.07860923e-01 6.36321962e-01
3.53922658e-02 9.78979051e-01 -1.92331538e-01 -9.04663920e-01
2.30149135e-01 1.78605825e-01 8.91481400e-01 1.45297873e+00
2.95011759e-01 -3.06613296e-01 -2.29450598e-01 5.07992625e-01
-2.55813301e-01 9.56783116e-01 -4.94088769e-01 -3.55614543e-01
3.44037294e-01 1.24581671e+00 -4.79288042e-01 -4.67953563e-01
-1.97149396e-01 6.63069308e-01 1.57047436e-01 -1.43216148e-01
-6.81817114e-01 -2.34827548e-01 7.43683726e-02 3.12216103e-01
-1.87180713e-01 1.33315071e-01 -5.52417099e-01 -1.20007622e+00
-2.41410986e-01 -1.12984884e+00 2.27164477e-01 -5.24907410e-01
-1.26515543e+00 5.81826508e-01 -1.57196507e-01 -1.05500054e+00
-3.41284066e-01 -1.41549855e-01 -5.26331604e-01 8.55467618e-01
-9.66608524e-01 -8.72331202e-01 -3.07396036e-02 5.56603849e-01
1.11903918e+00 -7.14485705e-01 5.22377372e-01 5.95667899e-01
-2.91872352e-01 7.35058725e-01 2.90961653e-01 4.64428030e-02
7.31902421e-01 -9.46234167e-01 -3.04104984e-02 4.74131942e-01
-6.31383434e-02 2.33664393e-01 9.16267097e-01 -7.15671718e-01
-8.55695248e-01 -8.38936329e-01 2.28658223e+00 -3.19109857e-01
8.12262297e-01 -2.08813787e-01 -4.68604088e-01 1.03332750e-01
6.99767947e-01 -5.50324261e-01 9.10311580e-01 6.91607371e-02
-3.71933132e-01 3.61041158e-01 -1.49723184e+00 4.56812143e-01
5.44481158e-01 -5.52392304e-01 -3.36939454e-01 9.20768499e-01
2.88259268e-01 -1.95307061e-01 -7.62269735e-01 -3.57259482e-01
6.28722072e-01 -1.50792229e+00 3.90931398e-01 -5.83094299e-01
1.12780464e+00 3.72996658e-01 -3.10893178e-01 -1.04689801e+00
1.50414199e-01 -7.10989714e-01 2.01312318e-01 1.56714940e+00
6.84262037e-01 -3.30066353e-01 7.43571401e-01 3.62671286e-01
4.54034507e-02 -5.98906457e-01 -1.42717576e+00 -1.48015231e-01
4.02005970e-01 -5.66582203e-01 -3.87905724e-02 1.04610395e+00
7.72744715e-01 8.17699730e-01 -7.95124829e-01 -7.48216033e-01
3.73185843e-01 -7.06252813e-01 7.02150166e-01 -1.04733288e+00
1.84658587e-01 -5.85018754e-01 -3.16882282e-01 -3.99740279e-01
1.70920283e-01 -1.01578867e+00 -2.58729756e-01 -1.25531268e+00
6.87315464e-01 1.48855774e-02 3.46425742e-01 4.97719139e-01
1.69280380e-01 4.81621325e-01 2.20663935e-01 4.29128319e-01
-6.65796578e-01 6.27822518e-01 7.76201725e-01 -1.74534112e-01
-1.31007597e-01 1.33241221e-01 -8.24361861e-01 6.47530377e-01
8.69104683e-01 -6.66209817e-01 -1.95554525e-01 2.70526558e-01
7.96863377e-01 3.65873687e-02 8.37419182e-02 -7.97531068e-01
2.33212542e-02 1.83581822e-02 -3.82057160e-01 -6.45250201e-01
8.81374329e-02 -2.48063877e-01 -1.47560418e-01 6.68805063e-01
-7.95797288e-01 2.28112176e-01 -1.67985618e-01 -1.89798586e-02
-3.03655773e-01 -7.72806823e-01 9.03581977e-01 -5.42425252e-02
7.04071745e-02 -1.70801386e-01 -7.47662783e-01 6.43525004e-01
7.92241514e-01 5.13700470e-02 -5.78532636e-01 -1.05336118e+00
-5.45540690e-01 -1.99873492e-01 2.77202815e-01 6.47620857e-01
5.25181711e-01 -1.20664787e+00 -1.57044578e+00 1.58768240e-02
1.57513574e-01 -1.00929093e+00 3.84578228e-01 1.25501525e+00
-4.75255609e-01 1.34021237e-01 -6.81575164e-02 -2.87617385e-01
-1.63746130e+00 -7.64838010e-02 9.88766104e-02 -2.31454894e-01
-5.84404245e-02 8.80753994e-01 -6.59158111e-01 -5.82752764e-01
7.64205530e-02 4.08509433e-01 -7.15662181e-01 5.60689747e-01
7.08666027e-01 8.32637250e-01 -1.30497320e-02 -1.37929308e+00
-3.44427705e-01 7.25525990e-02 7.94695988e-02 -1.90151602e-01
1.19371879e+00 -3.21523756e-01 -5.39897919e-01 7.61806726e-01
1.52276778e+00 5.46919286e-01 -2.25902930e-01 2.01644123e-01
-7.20701739e-02 -3.82132232e-01 9.39310789e-02 -8.25281560e-01
-6.80025280e-01 8.23876679e-01 6.84517026e-01 6.63323581e-01
4.29901183e-01 -1.23560186e-02 6.66151285e-01 -9.07661691e-02
-4.40654188e-01 -1.36103928e+00 9.24312100e-02 9.81465518e-01
1.00002360e+00 -1.35309291e+00 2.98963841e-02 -1.18245140e-01
-1.04887569e+00 1.29064274e+00 7.41665810e-02 1.39946014e-01
8.16626906e-01 -9.96072143e-02 1.62772745e-01 -3.70181575e-02
-8.89160037e-01 1.85325861e-01 2.56841213e-01 4.75550473e-01
1.16482329e+00 1.01101778e-01 -1.33003056e+00 5.51324129e-01
-5.11605263e-01 -9.26569477e-02 1.05810606e+00 1.87194288e-01
-6.34575903e-01 -8.42774928e-01 -6.84945881e-02 5.29719591e-01
-9.58086610e-01 -3.43906283e-01 -7.70889878e-01 4.25402552e-01
1.60167310e-02 1.31596994e+00 -9.30599719e-02 -5.08807361e-01
-1.73614427e-01 1.89969257e-01 -3.70951861e-01 -3.88712168e-01
-1.02475500e+00 3.24429870e-01 6.40495181e-01 -8.45057890e-02
-6.71418786e-01 -1.12895262e+00 -8.61786962e-01 -7.51278996e-01
-4.44742478e-02 3.64612430e-01 7.96903968e-01 7.91348696e-01
2.38920867e-01 -5.50764389e-02 6.23809516e-01 -3.54608893e-01
-3.77589792e-01 -1.36186850e+00 -5.79208553e-01 3.57237905e-01
1.70671567e-01 -1.36151671e-01 -7.28616714e-01 -7.25708902e-02] | [8.999302864074707, 10.696272850036621] |
591a2516-b9e9-4aaf-ac59-266331cba605 | zero-shot-cross-lingual-conversational-2 | 2204.04914 | null | https://arxiv.org/abs/2204.04914v1 | https://arxiv.org/pdf/2204.04914v1.pdf | Zero-shot Cross-lingual Conversational Semantic Role Labeling | While conversational semantic role labeling (CSRL) has shown its usefulness on Chinese conversational tasks, it is still under-explored in non-Chinese languages due to the lack of multilingual CSRL annotations for the parser training. To avoid expensive data collection and error-propagation of translation-based methods, we present a simple but effective approach to perform zero-shot cross-lingual CSRL. Our model implicitly learns language-agnostic, conversational structure-aware and semantically rich representations with the hierarchical encoders and elaborately designed pre-training objectives. Experimental results show that our model outperforms all baselines by large margins on two newly collected English CSRL test sets. More importantly, we confirm the usefulness of CSRL to non-Chinese conversational tasks such as the question-in-context rewriting task in English and the multi-turn dialogue response generation tasks in English, German and Japanese by incorporating the CSRL information into the downstream conversation-based models. We believe this finding is significant and will facilitate the research of non-Chinese dialogue tasks which suffer the problems of ellipsis and anaphora. | ['Linqi Song', 'Lianwei Wu', 'Shuqi Liu', 'Kun Xu', 'Haochen Tan', 'Han Wu'] | 2022-04-11 | zero-shot-cross-lingual-conversational | https://aclanthology.org/2022.findings-naacl.20 | https://aclanthology.org/2022.findings-naacl.20.pdf | findings-naacl-2022-7 | ['semantic-role-labeling'] | ['natural-language-processing'] | [ 2.04873249e-01 4.83855039e-01 -2.03680068e-01 -7.63065219e-01
-1.39719164e+00 -7.94474661e-01 7.49720871e-01 -6.58068880e-02
-4.87486959e-01 1.13259041e+00 1.07405174e+00 -4.77405697e-01
3.41649562e-01 -4.11942899e-01 -2.66603738e-01 -3.32825691e-01
9.35517848e-02 7.04733491e-01 2.04697848e-04 -8.70679617e-01
1.13750607e-01 -3.40603255e-02 -9.56637621e-01 9.37295020e-01
9.11397338e-01 2.34271631e-01 4.10793632e-01 5.82205474e-01
-2.70104170e-01 1.36048234e+00 -6.83199406e-01 -6.20767713e-01
-1.31644830e-01 -8.42073798e-01 -1.44833243e+00 2.59138346e-02
-5.24787419e-03 -2.27473199e-01 2.26574093e-02 6.18677676e-01
5.79401791e-01 2.49077320e-01 1.34609073e-01 -8.32020998e-01
-7.71595478e-01 9.28495109e-01 -1.14611007e-01 2.25169167e-01
7.44596422e-01 3.87461632e-02 1.53231895e+00 -8.16531479e-01
9.81780469e-01 1.72552049e+00 5.16732574e-01 1.05367112e+00
-1.04724240e+00 -3.46371025e-01 2.76505530e-01 2.11103871e-01
-8.06670964e-01 -5.85828304e-01 7.27657497e-01 -9.41294283e-02
1.47188318e+00 2.04342678e-01 1.24876797e-01 1.34864628e+00
-9.06280056e-02 1.00073481e+00 1.19906652e+00 -6.96858227e-01
-2.14473814e-01 1.44730300e-01 5.84129170e-02 5.56284487e-01
-3.26038003e-01 -2.56610364e-01 -7.19073474e-01 -1.16568699e-01
3.99365306e-01 -6.68009996e-01 -3.22075814e-01 2.71915048e-01
-1.16592240e+00 1.09782779e+00 7.22134188e-02 6.95742488e-01
-3.46311033e-02 -3.45311798e-02 7.99309075e-01 8.41628432e-01
6.56590462e-01 9.52066720e-01 -7.58028030e-01 -5.40584147e-01
-1.67645231e-01 2.24734098e-01 1.18593907e+00 1.09718275e+00
4.79679465e-01 -2.00258628e-01 -1.90456003e-01 1.32521176e+00
-3.93506773e-02 1.63985997e-01 5.88349760e-01 -1.28932405e+00
9.37001348e-01 6.76478505e-01 2.01428831e-02 -4.78040814e-01
-3.67528051e-01 2.88142432e-02 -2.19469443e-01 -6.55111969e-01
5.51280081e-01 -3.53184193e-01 8.56849253e-02 2.08552289e+00
2.44533792e-01 -3.22286665e-01 7.76240826e-01 9.48714197e-01
1.00983000e+00 6.97690845e-01 2.11433545e-01 -3.32702339e-01
1.61728251e+00 -1.42715561e+00 -9.49277520e-01 -2.85831422e-01
1.09758401e+00 -9.24954236e-01 1.55971384e+00 -2.02186957e-01
-1.17240119e+00 -2.27868080e-01 -6.85635746e-01 -5.33317268e-01
-6.13205284e-02 6.72778413e-02 9.01703417e-01 2.69153386e-01
-9.87329602e-01 1.86125875e-01 -5.47834516e-01 -6.36026204e-01
-9.36327204e-02 -7.54775181e-02 -3.75744253e-01 -1.30355328e-01
-1.79426563e+00 1.34224164e+00 -4.74206544e-02 1.13807514e-01
-5.95227838e-01 -4.16721910e-01 -1.12410295e+00 -1.09082945e-01
4.32004601e-01 -3.27199608e-01 1.81430066e+00 -9.53644454e-01
-1.93128633e+00 1.27963066e+00 -3.81359756e-01 -5.03746152e-01
1.17216580e-01 -2.94626385e-01 -1.80260330e-01 1.43820331e-01
5.17056465e-01 5.05638719e-01 8.38259980e-02 -8.45916152e-01
-6.28765702e-01 -1.68953896e-01 6.33560002e-01 9.13736403e-01
-1.09728239e-02 4.67353791e-01 -2.13871911e-01 -4.69591975e-01
-1.66827932e-01 -1.04699194e+00 -1.48682728e-01 -5.92467606e-01
-1.89773545e-01 -8.56247544e-01 5.04134178e-01 -7.36693263e-01
1.02660108e+00 -1.76388204e+00 2.54311740e-01 -6.91363871e-01
-3.19432884e-01 8.44415128e-02 -3.74631315e-01 9.63577271e-01
2.36816332e-01 9.08983648e-02 -4.00417447e-01 -3.59031498e-01
-2.16927379e-01 4.51231539e-01 -3.03108692e-01 1.22722529e-01
6.19780183e-01 1.00125122e+00 -1.13523853e+00 -4.26177710e-01
-3.46845351e-02 7.58673549e-02 -5.83092511e-01 6.79798365e-01
-4.10296738e-01 8.52856457e-01 -4.71378058e-01 5.15723407e-01
2.95873135e-02 -1.05155885e-01 7.91508973e-01 2.68713772e-01
-2.84192026e-01 1.54007876e+00 -3.46422821e-01 2.07470655e+00
-9.05471027e-01 6.03713930e-01 2.84986466e-01 -9.31242228e-01
6.74949050e-01 7.37932682e-01 3.54194120e-02 -8.59146237e-01
-1.99872792e-01 2.42450714e-01 3.03825468e-01 -6.59746647e-01
4.56146628e-01 -4.24281955e-01 -6.91419363e-01 9.40392375e-01
1.56800792e-01 -2.23296031e-01 1.55137688e-01 2.68783063e-01
1.04672968e+00 2.41960704e-01 5.50659776e-01 -4.71450269e-01
9.84469473e-01 3.46416116e-01 6.17451906e-01 4.61008966e-01
-2.77681589e-01 3.02992404e-01 7.59643555e-01 -3.40703189e-01
-6.63702011e-01 -5.21346807e-01 -1.11327305e-01 1.76484084e+00
-1.77179221e-02 -4.49835837e-01 -9.02467310e-01 -8.71318042e-01
-3.02726567e-01 1.01242805e+00 -2.94322640e-01 1.82872072e-01
-1.30098045e+00 -6.51328743e-01 8.65458131e-01 2.86302269e-01
5.43169498e-01 -1.54203200e+00 -1.18589379e-01 6.32244229e-01
-7.69906998e-01 -1.57020354e+00 -5.08220613e-01 -1.40264342e-02
-6.40617967e-01 -1.23872924e+00 -2.21000224e-01 -1.08801711e+00
2.83351749e-01 3.33465785e-01 1.26526606e+00 5.25532886e-02
8.11777264e-02 3.80196124e-01 -5.70441425e-01 -5.48879355e-02
-8.54411781e-01 4.44768190e-01 -2.95468509e-01 -2.17354834e-01
7.17781901e-01 -3.89813751e-01 -2.03100711e-01 2.05146700e-01
-1.68808758e-01 1.66949630e-01 2.11445481e-01 9.83917534e-01
1.53492605e-02 -7.00844407e-01 1.14299464e+00 -1.38027418e+00
1.22351789e+00 -4.25554633e-01 -2.51867801e-01 3.04503560e-01
-2.88672745e-01 1.53801233e-01 6.30416274e-01 -3.64422426e-02
-1.81650209e+00 -3.10246795e-01 -3.15769285e-01 6.01109803e-01
-1.45156141e-02 5.02143681e-01 -3.18631500e-01 5.01684308e-01
6.49984956e-01 2.96399277e-02 1.59263238e-01 -5.99463105e-01
5.94934583e-01 8.60935152e-01 4.14389491e-01 -9.44625914e-01
2.60819942e-01 5.66513799e-02 -6.87488079e-01 -8.17353129e-01
-1.28507483e+00 -5.72109222e-01 -5.10492444e-01 7.96573609e-02
1.16070306e+00 -1.27320731e+00 -7.05511272e-01 2.03118548e-01
-1.65794110e+00 -5.07980525e-01 1.36236439e-03 2.24109054e-01
-5.25549293e-01 3.70298445e-01 -1.35330188e+00 -6.16139352e-01
-5.42078078e-01 -1.03793275e+00 1.07970619e+00 -1.24914043e-01
-6.86405659e-01 -1.20879185e+00 1.76503539e-01 9.17200923e-01
4.32118505e-01 -1.74462125e-01 1.19455349e+00 -9.50078964e-01
-4.84650552e-01 1.18932217e-01 -8.97250846e-02 3.34767222e-01
2.14610010e-01 -6.47519290e-01 -9.43229556e-01 -1.59880698e-01
3.29753309e-01 -9.10640538e-01 4.17048663e-01 -1.67112648e-01
4.80624974e-01 -6.21890783e-01 5.39913252e-02 2.15580054e-02
9.28754330e-01 2.28078932e-01 2.85084873e-01 2.80106544e-01
5.83140790e-01 1.26109469e+00 7.53867090e-01 -1.42610639e-01
1.05032563e+00 8.04309189e-01 -6.75106421e-02 8.38453323e-02
-2.85989612e-01 -2.64190257e-01 5.98341584e-01 1.41880023e+00
9.26388577e-02 -1.96292520e-01 -6.59308732e-01 6.37246430e-01
-2.00324297e+00 -9.07736719e-01 -2.75605023e-01 1.74093759e+00
1.56258965e+00 -1.98334515e-01 -1.82548895e-01 -5.16893685e-01
7.49211073e-01 3.20128173e-01 -1.35808468e-01 -8.75203490e-01
-2.29103163e-01 1.84641376e-01 -1.89235806e-01 1.08508110e+00
-9.43565667e-01 1.69971728e+00 5.80397129e+00 2.75302470e-01
-8.14843059e-01 6.98494136e-01 4.56844002e-01 2.74392098e-01
-3.49619359e-01 2.47802019e-01 -9.53056574e-01 1.16185211e-01
1.03820634e+00 -1.94907293e-01 5.62850177e-01 8.76403809e-01
2.51812130e-01 -4.07124087e-02 -1.32161188e+00 5.11145890e-01
1.43321604e-01 -1.28985214e+00 -2.66559422e-01 -3.14381897e-01
6.30802035e-01 1.44980073e-01 -7.15318978e-01 7.11162388e-01
5.58472097e-01 -8.78770888e-01 2.16100410e-01 -1.96095422e-01
7.19602287e-01 -4.73645955e-01 8.36640894e-01 3.85090709e-01
-9.36452091e-01 3.01728882e-02 -3.56465906e-01 -4.04137641e-01
3.76253247e-01 -5.36079928e-02 -1.25924683e+00 2.96304941e-01
2.35092640e-01 7.14073718e-01 -1.98490113e-01 -1.50605319e-02
-8.17028940e-01 8.71589541e-01 1.86756149e-01 -2.69832939e-01
4.52899426e-01 -2.21661210e-01 5.80932736e-01 1.55905986e+00
-3.27552974e-01 2.95674890e-01 3.58984202e-01 6.86143935e-01
-3.13557982e-01 3.79742205e-01 -6.73530638e-01 -8.86761546e-02
7.30326831e-01 1.17278349e+00 -2.79334277e-01 -1.79063544e-01
-7.46339679e-01 1.04771328e+00 7.63846576e-01 1.82701528e-01
-3.78911495e-01 -8.10290948e-02 8.23288381e-01 -4.29663397e-02
-2.69877434e-01 -2.95739830e-01 -1.56013221e-01 -1.30691516e+00
-4.71593142e-02 -1.35147130e+00 4.66802865e-01 -4.39437628e-01
-1.60095370e+00 9.18660402e-01 -7.82294199e-02 -8.63265514e-01
-9.59128201e-01 -5.09910941e-01 -7.02824771e-01 9.32165384e-01
-1.68999887e+00 -1.35574985e+00 2.18538955e-01 5.13592243e-01
1.34033358e+00 -3.67610544e-01 1.33944225e+00 2.06548408e-01
-4.75534409e-01 4.73133981e-01 -4.40953791e-01 3.85652214e-01
9.62126017e-01 -1.23742461e+00 5.59077144e-01 6.43844187e-01
-1.49430588e-01 7.59341955e-01 4.77290124e-01 -4.28896844e-01
-1.16259551e+00 -7.62719274e-01 1.87479651e+00 -7.17942774e-01
9.42740679e-01 -7.00492144e-01 -1.10344040e+00 8.46297324e-01
8.16036522e-01 -5.12414277e-01 9.02903795e-01 5.62132895e-01
-1.52945623e-01 1.59349546e-01 -8.16414714e-01 6.12907946e-01
1.10603333e+00 -9.73544359e-01 -1.24863815e+00 5.87962627e-01
1.19696617e+00 -4.44283664e-01 -7.61457324e-01 8.99863169e-02
1.12088226e-01 -6.24119163e-01 5.62440991e-01 -9.13857043e-01
6.84794307e-01 1.52428553e-01 -1.98596731e-01 -1.22647774e+00
-7.93282241e-02 -8.35652173e-01 3.54800165e-01 1.39317393e+00
7.43728697e-01 -5.34115613e-01 3.25590223e-01 6.55676723e-01
-5.37244439e-01 -5.46335578e-01 -1.07228005e+00 -3.62359434e-01
3.51717561e-01 -2.37608090e-01 3.40313226e-01 1.23238945e+00
6.58967137e-01 1.30642998e+00 -3.78322989e-01 -2.85183966e-01
1.02330104e-01 3.55905503e-01 7.62584627e-01 -1.00715005e+00
-3.23014557e-01 -1.62617177e-01 3.80817771e-01 -1.03451884e+00
9.15059566e-01 -1.13570535e+00 2.85221726e-01 -1.42664850e+00
5.87362200e-02 -5.27267039e-01 3.08442265e-01 4.93559122e-01
-3.49609047e-01 -1.70171425e-01 4.33614738e-02 2.69132733e-01
-8.42022002e-01 7.21056759e-01 1.37589526e+00 2.39663422e-01
-2.32882693e-01 3.60625014e-02 -1.08780086e+00 6.24473214e-01
6.24730229e-01 -5.56422293e-01 -5.02610922e-01 -5.93551993e-01
-2.11787540e-02 5.34021199e-01 -2.47029752e-01 -2.50232872e-02
3.79709415e-02 -2.38675967e-01 -5.03435135e-01 -4.88905273e-02
2.90665686e-01 -2.18992963e-01 -6.81292474e-01 1.41331479e-01
-9.67812121e-01 2.20012203e-01 5.24955429e-02 2.92977929e-01
-4.33888316e-01 -3.78862321e-01 5.27569771e-01 -6.15606189e-01
-5.96469402e-01 -3.49894881e-01 -8.78120244e-01 6.99256778e-01
5.92475235e-01 2.53693461e-01 -5.02696574e-01 -8.55905890e-01
-5.15433431e-01 5.21132827e-01 1.13109343e-01 8.15265059e-01
1.54519171e-01 -1.04063988e+00 -1.01706946e+00 3.77905951e-03
2.74857789e-01 4.89179455e-02 -3.94329354e-02 6.16069496e-01
-3.96437645e-01 8.06435764e-01 4.96623553e-02 -3.45567852e-01
-1.18318212e+00 7.41227493e-02 2.95896590e-01 -7.12321758e-01
-6.22365654e-01 8.56654644e-01 2.69739151e-01 -1.11271048e+00
1.56284034e-01 3.48968245e-02 -3.53596568e-01 -5.68666086e-02
4.58292037e-01 -4.00293022e-02 5.43960445e-02 -6.51099801e-01
-3.60790998e-01 -1.12593547e-01 -2.55388349e-01 -3.47253233e-01
1.25518227e+00 -4.72089410e-01 -5.41632593e-01 5.06586373e-01
1.06402576e+00 1.40594155e-01 -1.00245881e+00 -4.46802527e-01
6.67801499e-01 2.48953584e-03 -5.59610903e-01 -8.65199506e-01
-3.08857799e-01 8.70834649e-01 -2.27559879e-01 4.30383459e-02
6.76718950e-01 3.13899428e-01 8.78760755e-01 7.54724026e-01
4.86790836e-01 -1.21992660e+00 2.78971195e-01 1.14022100e+00
1.34589756e+00 -1.30023563e+00 -3.67288560e-01 -6.48728848e-01
-1.22439146e+00 9.36786890e-01 1.07715845e+00 2.21505314e-02
1.77954495e-01 2.09549502e-01 4.76200998e-01 -1.36820748e-01
-1.42548001e+00 -2.51474768e-01 -1.52248502e-01 4.26000953e-01
1.33142698e+00 7.54183233e-02 -9.37968314e-01 6.98273301e-01
-3.90844375e-01 -4.95136291e-01 6.19517565e-01 8.35338414e-01
-1.66817173e-01 -1.65055811e+00 1.12110153e-01 -1.70542181e-01
-6.99112117e-01 -5.60246229e-01 -8.11854959e-01 1.00379717e+00
-4.22868043e-01 1.16843390e+00 9.68619958e-02 1.01091117e-01
3.16782236e-01 5.12365878e-01 2.44841501e-01 -1.28312993e+00
-1.04711854e+00 6.61884621e-02 1.16201854e+00 -5.50006092e-01
-6.56442285e-01 -7.04930902e-01 -1.44107270e+00 -3.49536613e-02
-1.73954815e-01 4.48539644e-01 3.87114793e-01 1.19093907e+00
3.66520286e-01 2.46790484e-01 6.70103192e-01 -4.47653264e-01
-7.29065061e-01 -1.37942374e+00 7.23654032e-02 6.01583719e-01
-2.16152687e-02 -3.17666918e-01 -2.82287091e-01 -1.55347005e-01] | [12.37765884399414, 8.095559120178223] |
5c3f6686-fa5a-4a7d-a434-e779828a50b4 | hagnn-hybrid-aggregation-for-heterogeneous | 2307.01636 | null | https://arxiv.org/abs/2307.01636v1 | https://arxiv.org/pdf/2307.01636v1.pdf | HAGNN: Hybrid Aggregation for Heterogeneous Graph Neural Networks | Heterogeneous graph neural networks (GNNs) have been successful in handling heterogeneous graphs. In existing heterogeneous GNNs, meta-path plays an essential role. However, recent work pointed out that simple homogeneous graph model without meta-path can also achieve comparable results, which calls into question the necessity of meta-path. In this paper, we first present the intrinsic difference about meta-path-based and meta-path-free models, i.e., how to select neighbors for node aggregation. Then, we propose a novel framework to utilize the rich type semantic information in heterogeneous graphs comprehensively, namely HAGNN (Hybrid Aggregation for Heterogeneous GNNs). The core of HAGNN is to leverage the meta-path neighbors and the directly connected neighbors simultaneously for node aggregations. HAGNN divides the overall aggregation process into two phases: meta-path-based intra-type aggregation and meta-path-free inter-type aggregation. During the intra-type aggregation phase, we propose a new data structure called fused meta-path graph and perform structural semantic aware aggregation on it. Finally, we combine the embeddings generated by each phase. Compared with existing heterogeneous GNN models, HAGNN can take full advantage of the heterogeneity in heterogeneous graphs. Extensive experimental results on node classification, node clustering, and link prediction tasks show that HAGNN outperforms the existing modes, demonstrating the effectiveness of HAGNN. | ['Yihua Huang', 'Chunfeng Yuan', 'Hongyang Chen', 'Zhennan Zhu', 'Guanghui Zhu'] | 2023-07-04 | null | null | null | null | ['node-classification', 'link-prediction'] | ['graphs', 'graphs'] | [-3.35733265e-01 4.55525398e-01 -5.24347126e-01 -3.47372927e-02
-2.23296538e-01 -1.62639037e-01 4.28171515e-01 3.91662121e-01
-3.49989161e-03 7.13393331e-01 2.28864267e-01 -1.26437306e-01
-3.39182019e-01 -1.66281891e+00 -3.62577826e-01 -6.77182019e-01
-2.36816660e-01 5.13636291e-01 6.61963522e-01 -3.67153823e-01
-2.70752460e-01 2.58645982e-01 -1.23295021e+00 8.95253122e-02
1.03149104e+00 9.04120803e-01 1.89863220e-01 4.66315061e-01
-5.38542628e-01 8.30908835e-01 -3.21071118e-01 -5.43569028e-01
2.24265307e-01 -2.98196018e-01 -6.16124094e-01 -2.10698143e-01
6.78070560e-02 4.74542305e-02 -5.90549588e-01 1.17288482e+00
5.85976422e-01 1.67629942e-01 3.75046551e-01 -1.77265656e+00
-8.59537542e-01 1.25733995e+00 -3.85702223e-01 4.60832603e-02
1.18837446e-01 -1.02927290e-01 1.37958181e+00 -7.99458385e-01
8.63147199e-01 1.21820700e+00 9.16394949e-01 2.44041085e-01
-8.96661103e-01 -5.11131048e-01 7.16861844e-01 1.76577285e-01
-1.58038199e+00 -1.42910369e-02 1.14605975e+00 -1.58852920e-01
9.08411920e-01 1.09726667e-01 8.87790382e-01 1.00772393e+00
8.04820880e-02 6.38126969e-01 5.05330741e-01 -1.38943583e-01
7.52911493e-02 -1.73859060e-01 4.71977860e-01 9.21345890e-01
6.29567444e-01 -2.91312277e-01 -1.52858332e-01 -2.54269660e-01
6.99949980e-01 2.48842686e-01 -1.73490182e-01 -4.68249917e-01
-1.01495898e+00 9.14188266e-01 9.21027958e-01 5.38335681e-01
-5.48263490e-01 1.16288634e-02 6.57297194e-01 3.08956504e-01
6.63723946e-01 1.28838599e-01 -3.15964013e-01 3.68770510e-01
-5.79022169e-01 -2.01633930e-01 8.60976160e-01 1.16246462e+00
1.07543445e+00 9.16468650e-02 -1.86410591e-01 8.27899694e-01
3.60767663e-01 1.11910008e-01 5.37586391e-01 -6.16171777e-01
6.66514814e-01 1.43947864e+00 -5.86498678e-01 -1.53618526e+00
-6.03624642e-01 -7.05223739e-01 -1.27779877e+00 -4.02940452e-01
-1.48134947e-01 -1.06038168e-01 -8.34850252e-01 1.80762267e+00
5.33381283e-01 4.94371742e-01 3.90900783e-02 5.53741097e-01
1.37405729e+00 5.85757136e-01 2.19654471e-01 2.66157463e-02
9.28090513e-01 -1.43175352e+00 -6.33378386e-01 9.24307555e-02
9.81778502e-01 -2.22791672e-01 7.78192341e-01 -2.89980203e-01
-9.31371748e-01 -5.10789335e-01 -1.10994601e+00 1.75788701e-02
-8.23958337e-01 -3.10430765e-01 8.07157040e-01 4.84925747e-01
-1.39086342e+00 4.60513085e-01 -3.93354744e-01 -5.69391191e-01
1.74600989e-01 3.75608951e-01 -5.65931439e-01 4.18586191e-03
-1.52374613e+00 2.19144821e-01 9.72133815e-01 4.70440611e-02
-4.95865256e-01 -8.18397105e-01 -1.09442329e+00 3.68116468e-01
7.04665005e-01 -1.12087309e+00 6.22485876e-01 -6.98142827e-01
-1.01279223e+00 3.43768358e-01 -6.31561726e-02 -2.29758665e-01
6.96003288e-02 4.81223702e-01 -9.20778215e-01 3.28545421e-01
1.68654546e-01 5.48667252e-01 3.66354793e-01 -1.42144096e+00
-7.01042473e-01 -3.97909880e-01 3.16835910e-01 2.92121679e-01
-7.90613413e-01 -6.00584626e-01 -9.10351694e-01 -5.67640603e-01
3.05499911e-01 -8.89380336e-01 -1.83488280e-01 -3.75672758e-01
-7.51136363e-01 -5.22736967e-01 9.97307003e-01 -2.53911138e-01
1.85861063e+00 -1.86729312e+00 3.12743366e-01 6.76739335e-01
9.86953020e-01 3.12259138e-01 -3.92755389e-01 6.68067217e-01
1.32369146e-01 3.74775678e-01 8.59768465e-02 -3.90206516e-01
1.62369028e-01 2.77290761e-01 3.45103234e-01 1.78532489e-02
-1.77823186e-01 1.24133098e+00 -1.06289113e+00 -8.22023690e-01
1.50451988e-01 3.81407648e-01 -4.99460578e-01 2.88563017e-02
-7.91710094e-02 -1.51678026e-01 -6.62554204e-01 9.72759068e-01
7.33830512e-01 -6.91808641e-01 6.15950704e-01 -4.73413497e-01
3.05779189e-01 -1.14493528e-02 -1.11460459e+00 1.42682052e+00
-3.83893758e-01 1.72173128e-01 1.74248844e-01 -9.06024814e-01
9.29426730e-01 4.07491624e-02 6.34522378e-01 -5.54147482e-01
1.49843276e-01 2.95056254e-01 1.79181714e-02 -2.70153701e-01
7.21960843e-01 3.24078083e-01 1.44620268e-02 3.01460147e-01
2.05312535e-01 5.87684155e-01 4.26051706e-01 7.77898431e-01
1.37731326e+00 -1.59455836e-01 3.02229077e-01 -2.47412264e-01
4.99997824e-01 -1.72172844e-01 6.48443162e-01 6.72726452e-01
-2.87857234e-01 5.34918010e-01 4.74674404e-01 -5.81413865e-01
-6.27231121e-01 -1.02055097e+00 6.35376811e-01 1.18694329e+00
5.39627612e-01 -9.93053138e-01 -6.18909299e-01 -1.11169958e+00
1.52629972e-01 -4.80123563e-03 -8.51393282e-01 -3.59021574e-01
-6.25043213e-01 -9.21187758e-01 4.06675279e-01 6.49682403e-01
7.92624950e-01 -1.07086742e+00 3.82424772e-01 3.60428721e-01
-1.04404449e-01 -1.11918783e+00 -4.34938937e-01 -1.68965250e-01
-7.84853697e-01 -1.17404068e+00 -3.66893440e-01 -8.95737708e-01
6.48288369e-01 5.81721246e-01 1.31852055e+00 6.77966595e-01
2.13946477e-01 2.64492661e-01 -5.94102263e-01 3.05387080e-01
-1.22904882e-01 7.18813121e-01 -2.59782106e-01 3.21648037e-03
3.76484543e-01 -9.75726366e-01 -5.60548544e-01 2.17670351e-01
-8.54181290e-01 3.45424190e-02 7.83273757e-01 6.85580313e-01
4.97482121e-01 3.21375132e-01 6.96622312e-01 -1.43833017e+00
7.79992759e-01 -9.84895349e-01 8.30044076e-02 4.01243746e-01
-7.16833770e-01 -1.06614642e-01 9.28148806e-01 -2.20699072e-01
-6.87483072e-01 -5.17275751e-01 -2.06173941e-01 -3.56867582e-01
2.76569307e-01 9.71228242e-01 -6.96148753e-01 -2.40139425e-01
1.95585221e-01 -1.14969708e-01 -1.88466802e-01 -2.84033895e-01
2.92621195e-01 4.05394644e-01 1.47572353e-01 -5.45434117e-01
8.34284663e-01 4.27871615e-01 1.57588914e-01 -7.11594641e-01
-5.44895470e-01 -3.52150917e-01 -4.74945515e-01 -1.22512460e-01
6.55323625e-01 -7.84317851e-01 -4.39623564e-01 4.55409348e-01
-8.72486174e-01 -3.06314617e-01 -1.38997510e-01 2.34278366e-01
-5.03734574e-02 4.72559005e-01 -7.80461490e-01 -2.17700049e-01
-4.82716173e-01 -1.06201899e+00 8.06704521e-01 2.20990926e-01
1.08091578e-01 -1.53620768e+00 -1.22641221e-01 1.44262224e-01
4.64870125e-01 3.40599030e-01 1.15889323e+00 -8.97895753e-01
-6.61258280e-01 -1.16093613e-01 -6.45442843e-01 -1.24498215e-02
2.95107752e-01 -4.77373647e-03 -5.19358635e-01 -2.72472203e-01
-7.36316502e-01 9.52349156e-02 9.59160924e-01 1.98463857e-01
1.07900250e+00 -3.22789401e-01 -8.19095194e-01 7.90932119e-01
1.57043529e+00 1.41885951e-02 5.13448179e-01 4.02407885e-01
1.40939391e+00 4.59410965e-01 2.48695731e-01 -3.71378125e-03
1.00563204e+00 5.81573665e-01 5.88040233e-01 -2.08922580e-01
-4.58588243e-01 -5.96251667e-01 1.24897197e-01 1.32374728e+00
-1.09900966e-01 -7.67609179e-01 -7.85644114e-01 5.37840068e-01
-2.21817279e+00 -8.00443232e-01 -1.90958098e-01 1.70438254e+00
1.49501517e-01 1.72433093e-01 1.93839237e-01 1.50729425e-03
1.13810611e+00 4.88410592e-01 -5.11829674e-01 -1.41132534e-01
-2.88527489e-01 -2.53882736e-01 2.86325127e-01 4.76576388e-01
-9.45729136e-01 9.13442552e-01 5.01470184e+00 1.08955574e+00
-6.93837047e-01 3.18254262e-01 3.47402275e-01 3.34102511e-01
-8.48598599e-01 1.74133033e-02 -7.92342544e-01 6.94476902e-01
8.28891397e-01 -2.65339047e-01 3.86519618e-02 7.77973831e-01
-1.96665436e-01 4.66866702e-01 -6.58079326e-01 6.99593604e-01
3.41913141e-02 -1.50226045e+00 5.40504694e-01 1.18098654e-01
7.63732433e-01 9.36655104e-02 -2.10999385e-01 6.50426745e-01
6.57427371e-01 -6.30581677e-01 2.88021028e-01 1.21662728e-01
3.76172304e-01 -7.89562702e-01 9.59321737e-01 1.53317124e-01
-2.14137435e+00 -2.08830282e-01 -3.87233615e-01 2.57753253e-01
1.47813842e-01 8.48675013e-01 -4.24336612e-01 1.47635162e+00
5.67679703e-01 1.01142693e+00 -6.89837575e-01 8.22094262e-01
9.25699901e-03 2.63106167e-01 -2.76789993e-01 1.50234178e-02
5.48626840e-01 -4.00932789e-01 5.44460535e-01 1.00097322e+00
5.33213317e-01 -1.62775919e-01 5.97842991e-01 5.51010013e-01
-3.20292979e-01 2.19053760e-01 -6.20232105e-01 -1.43963963e-01
1.02360857e+00 1.46080422e+00 -8.50475907e-01 -4.70454842e-01
-6.90254867e-01 7.74111092e-01 7.45260477e-01 5.25888503e-01
-8.05653274e-01 -6.77192390e-01 4.97591943e-01 1.40234530e-01
2.00970113e-01 -6.90310523e-02 1.02126040e-02 -1.15361571e+00
1.00057065e-01 -6.07896626e-01 8.69259000e-01 -5.40465951e-01
-1.35909760e+00 9.95682776e-01 -5.58754280e-02 -1.02229762e+00
1.78923741e-01 -2.63872921e-01 -8.42215717e-01 5.53942204e-01
-1.52599788e+00 -1.84567750e+00 -7.03381181e-01 6.24131739e-01
2.28941098e-01 -1.62227958e-01 5.44285357e-01 5.21162271e-01
-7.86421061e-01 6.95611179e-01 -2.66754568e-01 3.15526158e-01
2.01732472e-01 -1.15079618e+00 3.54045421e-01 9.10724938e-01
-4.80193049e-02 7.24890649e-01 9.66740921e-02 -1.05173504e+00
-1.31588149e+00 -1.51053870e+00 8.11616719e-01 4.53728177e-02
8.31475616e-01 -2.51271427e-01 -8.81824136e-01 7.51631618e-01
3.30513455e-02 3.81830037e-01 6.10162139e-01 3.35400611e-01
-4.97061402e-01 -2.95112133e-01 -1.16840959e+00 7.47215807e-01
1.63287091e+00 -4.52324718e-01 -1.06501102e-01 1.34577796e-01
1.42390549e+00 6.79681730e-03 -1.21778286e+00 5.92789710e-01
4.00282621e-01 -1.12101686e+00 1.10307920e+00 -3.98463219e-01
2.05116466e-01 -3.98672342e-01 -1.80811867e-01 -1.50584102e+00
-6.83771670e-01 -5.26504159e-01 -4.15114075e-01 1.33531654e+00
4.88451332e-01 -1.07987761e+00 1.00806570e+00 4.33833562e-02
-6.48195446e-01 -8.78112435e-01 -7.00111806e-01 -8.63493681e-01
-1.65238261e-01 -4.83941585e-02 1.21088696e+00 1.07029593e+00
-2.29871035e-01 2.63989419e-01 -9.86258835e-02 3.48025948e-01
6.52946889e-01 2.24415272e-01 6.89704955e-01 -1.62860537e+00
-1.00179806e-01 -5.60486853e-01 -7.78432012e-01 -7.46056080e-01
5.05264103e-01 -1.26622248e+00 -6.00264013e-01 -2.03464746e+00
2.27034912e-01 -8.02397192e-01 -5.55648565e-01 5.03553927e-01
-3.17258507e-01 3.19573462e-01 1.91391900e-01 2.02664226e-01
-9.76855993e-01 5.64581990e-01 1.24871802e+00 -2.54147202e-01
-4.48984832e-01 -4.25041407e-01 -8.85652006e-01 6.20804846e-01
9.21428740e-01 -9.09123197e-02 -8.85264754e-01 -4.50205386e-01
3.98569018e-01 -2.90688753e-01 1.79432184e-01 -1.05122185e+00
4.63767380e-01 1.96450781e-02 5.58952801e-02 -4.96879220e-01
2.61918068e-01 -9.32934821e-01 4.90410596e-01 1.39919236e-01
2.25440925e-03 2.62511730e-01 -2.56027400e-01 7.84854889e-01
-3.79754245e-01 1.83823660e-01 3.77472401e-01 -2.17941061e-01
-1.06777656e+00 8.37885499e-01 1.74810767e-01 1.33401185e-01
1.13855445e+00 -3.74404371e-01 -8.84733260e-01 -1.80179507e-01
-9.41175759e-01 5.02735078e-01 4.96120155e-01 5.55335224e-01
5.43949604e-01 -1.67796683e+00 -3.98438752e-01 2.41362050e-01
2.91500360e-01 -2.51629129e-02 5.21116376e-01 1.10859752e+00
-3.19259465e-01 5.90334982e-02 1.34227395e-01 -3.10037166e-01
-1.13024950e+00 7.17174351e-01 1.24168083e-01 -5.30816257e-01
-7.60554075e-01 8.25405240e-01 1.11643642e-01 -5.10996342e-01
-2.18928587e-02 -6.40131459e-02 -3.00752044e-01 1.50950059e-01
1.88486531e-01 5.48337758e-01 1.53669089e-01 -8.65782082e-01
-2.05959365e-01 5.63898861e-01 -1.39315814e-01 4.94383007e-01
1.12113619e+00 -3.85835826e-01 -5.13781786e-01 3.76503736e-01
1.38084757e+00 -3.50546464e-02 -5.55463433e-01 -3.19830984e-01
-3.60840075e-02 -1.70779452e-01 1.90958977e-02 -4.12404060e-01
-1.64289391e+00 5.36729276e-01 -4.14812714e-02 7.39103794e-01
1.20431185e+00 1.06579371e-01 1.20949721e+00 2.25943610e-01
5.50340176e-01 -9.27064717e-01 -1.26219928e-01 3.71494800e-01
2.91023821e-01 -9.69689608e-01 -1.65350616e-01 -1.06539893e+00
-3.71762127e-01 8.34222376e-01 1.09305012e+00 -5.74887581e-02
1.01989961e+00 3.48019041e-02 -1.72204539e-01 -5.04280210e-01
-6.62404120e-01 -3.88722181e-01 1.84244901e-01 6.96706474e-01
2.97134388e-02 3.41609389e-01 -1.48137361e-01 6.35202229e-01
3.60402279e-02 -2.56650597e-01 3.13995719e-01 8.08697045e-01
-4.94189620e-01 -1.37458551e+00 2.19610393e-01 6.85068548e-01
3.96318547e-02 -1.64370432e-01 -5.34089565e-01 1.00770557e+00
2.41407141e-01 8.91553044e-01 -2.13929527e-02 -9.82209444e-01
1.53888851e-01 -1.08761176e-01 9.03077275e-02 -6.83058977e-01
-6.82056606e-01 -2.27321535e-01 2.38636464e-01 -6.92810416e-01
-3.85591894e-01 9.94517729e-02 -1.25250244e+00 -6.42154217e-01
-2.92479038e-01 5.12363255e-01 1.85486883e-01 5.86182654e-01
8.13160419e-01 7.35820830e-01 5.21260381e-01 -6.72548592e-01
-3.39052756e-03 -8.31316769e-01 -7.79186249e-01 4.30901706e-01
1.20462209e-01 -7.84772217e-01 -7.13177204e-01 -7.65593410e-01] | [7.3465118408203125, 6.357824802398682] |
d72ae7fc-4504-4a54-9cf8-2ac1cb209c60 | automatic-lung-cancer-prediction-from-chest-x | 1808.10858 | null | http://arxiv.org/abs/1808.10858v1 | http://arxiv.org/pdf/1808.10858v1.pdf | Automatic Lung Cancer Prediction from Chest X-ray Images Using Deep Learning Approach | Since, cancer is curable when diagnosed at an early stage, lung cancer
screening plays an important role in preventive care. Although both low dose
computed tomography (LDCT) and computed tomography (CT) scans provide more
medical information than normal chest x-rays, there is very limited access to
these technologies in rural areas. Recently, there is a trend in using
computer-aided diagnosis (CADx) to assist in screening and diagnosing of cancer
from biomedical images. In this study, the 121-layer convolutional neural
network also known as DenseNet-121 by G. Huang et. al., along with the transfer
learning scheme was explored as a means to classify lung cancer using chest
X-ray images. The model was trained on a lung nodules dataset before training
on the lung cancer dataset to alleviate the problem of a small dataset. The
proposed model yields 74.43$\pm$6.01\% of mean accuracy, 74.96$\pm$9.85\% of
mean specificity, and 74.68$\pm$15.33\% of mean sensitivity. The proposed model
also provides a heatmap for identifying the location of the lung nodule. These
findings are promising for further development of chest x-ray-based lung cancer
diagnosis using the deep learning approach. Moreover, these findings solve the
problem of small dataset. | ['Arjaree Thirach', 'Sanparith Marukatat', 'Worawate Ausawalaithong', 'Theerawit Wilaiprasitporn'] | 2018-08-31 | null | null | null | null | ['lung-cancer-diagnosis'] | ['medical'] | [ 8.39288533e-02 1.47197232e-01 -5.88310182e-01 -1.78572968e-01
-7.85152793e-01 -1.69287529e-03 1.63339242e-01 -5.62419221e-02
-4.26188231e-01 7.22461462e-01 -1.70013040e-01 -9.94813681e-01
-8.15560371e-02 -1.34942830e+00 -5.53540289e-01 -6.21843755e-01
-3.81508432e-02 5.31971633e-01 2.03540489e-01 2.61655211e-01
-4.16662663e-01 6.00741982e-01 -8.80524278e-01 3.32312077e-01
7.25343168e-01 1.07722044e+00 5.06831467e-01 6.88504159e-01
-8.28147903e-02 7.89656937e-01 -1.28634036e-01 -2.40571469e-01
2.17404172e-01 -6.81812644e-01 -8.04229856e-01 -6.11394010e-02
2.30664134e-01 -8.28322411e-01 -3.62278402e-01 8.31100404e-01
4.79687929e-01 -2.97530413e-01 7.41154909e-01 -5.08899510e-01
-5.00310659e-01 2.60441065e-01 -6.30873740e-01 4.56881642e-01
-5.11745140e-02 3.81164774e-02 5.00619233e-01 -9.03746068e-01
1.81800976e-01 5.56262314e-01 8.90673339e-01 9.08751667e-01
-3.85724157e-01 -8.76080990e-01 -5.63413203e-01 1.99462287e-02
-1.09931934e+00 2.48490557e-01 1.15646370e-01 -3.35109681e-01
8.06653857e-01 3.76726210e-01 9.78670478e-01 6.88756883e-01
5.28125882e-01 5.85698962e-01 8.83156061e-01 -3.53224993e-01
-4.74918112e-02 1.96190849e-01 -8.13645348e-02 1.26607823e+00
7.41264105e-01 2.91046262e-01 2.53874153e-01 -2.29705453e-01
1.10578430e+00 6.98792696e-01 -2.21554205e-01 1.67302161e-01
-1.00570476e+00 1.00665438e+00 1.18485510e+00 6.75184369e-01
-4.30993497e-01 3.64483058e-01 3.01701307e-01 -1.92296356e-01
2.41085693e-01 3.06612849e-01 -3.44110638e-01 2.93692350e-01
-8.05450737e-01 -9.49712768e-02 5.18455088e-01 5.05122125e-01
2.85687983e-01 1.19220510e-01 -1.95200533e-01 6.76956475e-01
4.35580730e-01 5.27683914e-01 7.56317258e-01 -5.35930634e-01
4.28919554e-01 7.82848537e-01 -3.12558979e-01 -5.30263007e-01
-5.55573642e-01 -6.28899872e-01 -1.45341313e+00 -1.40116632e-01
2.90391922e-01 -2.01403022e-01 -1.16747439e+00 1.17444158e+00
1.24796532e-01 1.96840726e-02 -2.59321332e-01 8.05563509e-01
9.80044544e-01 4.27996814e-01 3.89047682e-01 4.22259122e-02
1.63775325e+00 -7.20632672e-01 -4.80446368e-01 -9.65685844e-02
9.09433842e-01 -3.79240006e-01 8.95962298e-01 -3.79216895e-02
-9.22070980e-01 -5.38507462e-01 -7.39642978e-01 2.00172141e-01
-1.44115806e-01 7.63731897e-01 6.90095663e-01 9.56204891e-01
-8.85764837e-01 2.69789755e-01 -1.17311406e+00 -7.46141195e-01
7.70668030e-01 5.80451071e-01 -5.31844944e-02 -3.53621542e-01
-1.12604415e+00 7.93241382e-01 3.22402269e-01 -1.22221246e-01
-8.51514399e-01 -8.05602193e-01 -4.19414669e-01 2.04138532e-01
3.12649906e-01 -8.95474911e-01 1.33421814e+00 -4.66474891e-01
-1.08894050e+00 8.31163824e-01 1.51125804e-01 -4.64892060e-01
3.94115299e-01 -7.76282921e-02 -2.48704374e-01 4.22302425e-01
2.67934293e-01 7.47250259e-01 2.66550243e-01 -7.18840182e-01
-8.18327487e-01 -3.56222302e-01 -2.61920869e-01 5.85592492e-03
-4.48936373e-01 -1.35017678e-01 -4.99604046e-01 -3.47769648e-01
1.92917913e-01 -9.42704141e-01 -4.97927845e-01 3.13089103e-01
-3.72924805e-01 -1.10866979e-01 9.09196377e-01 -8.56524289e-01
1.09514093e+00 -1.67517376e+00 -7.76786506e-01 2.55171329e-01
3.96114796e-01 5.22600353e-01 5.23338974e-01 3.33289206e-02
-9.56132412e-02 5.47794282e-01 -3.07994634e-01 3.83440912e-01
-5.49838662e-01 5.71760982e-02 4.18133974e-01 3.18114996e-01
2.20442653e-01 1.32432199e+00 -6.64814949e-01 -7.94994235e-01
3.48523617e-01 4.27832633e-01 -3.48881960e-01 2.15128526e-01
1.44503519e-01 5.19421101e-01 -8.64525080e-01 1.04618001e+00
5.19237280e-01 -7.81494617e-01 9.18853730e-02 1.12029903e-01
7.86211193e-02 2.14482751e-02 -4.67777669e-01 1.20895815e+00
-3.48147452e-01 4.01740789e-01 -5.96220084e-02 -8.69253039e-01
6.55026853e-01 7.75687397e-01 7.24297345e-01 -6.20576084e-01
3.35563600e-01 3.15135121e-01 4.24999952e-01 -7.23795831e-01
-2.44179696e-01 -4.83555168e-01 3.56236398e-01 3.96129608e-01
-3.25581133e-01 -1.46548018e-01 -1.50248945e-01 1.66900363e-03
1.52877653e+00 -4.58764791e-01 6.41448736e-01 -2.39548340e-01
4.59465742e-01 2.41965964e-01 1.30740061e-01 6.96867168e-01
-1.48922846e-01 5.40265143e-01 8.45612213e-02 -5.71697891e-01
-8.49273324e-01 -9.27603006e-01 -5.78194737e-01 5.10370791e-01
-2.85077095e-01 1.44190550e-01 -6.27416134e-01 -8.48035693e-01
-9.64179859e-02 2.93407381e-01 -6.23400986e-01 -1.61278725e-01
-7.06024647e-01 -1.02352118e+00 7.56707788e-01 9.82057035e-01
1.17638063e+00 -9.01242137e-01 -8.91636908e-01 1.77536756e-02
1.34291630e-02 -7.17669904e-01 1.18006151e-02 3.24259728e-01
-1.40925908e+00 -1.42740881e+00 -1.20416772e+00 -9.33461785e-01
7.90709078e-01 3.09200257e-01 1.08143270e+00 5.90654492e-01
-6.01110697e-01 1.64177701e-01 -5.21024130e-02 -5.13125896e-01
-4.64008957e-01 2.71801949e-01 -3.48038584e-01 -6.48026824e-01
4.87892836e-01 -2.78756227e-02 -9.95460510e-01 1.74418420e-01
-7.40817487e-01 3.39517207e-03 1.26918316e+00 1.00043070e+00
7.94996738e-01 2.66385257e-01 5.85170984e-01 -1.19044030e+00
5.26878595e-01 -7.91363657e-01 -1.42836556e-01 8.04891735e-02
-5.89155972e-01 -5.63795149e-01 7.35615075e-01 -1.19931616e-01
-1.12138844e+00 1.50893837e-01 -4.85377550e-01 -3.16183656e-01
-2.24632949e-01 6.40515387e-01 2.23595530e-01 -6.75490201e-02
8.41012955e-01 5.04622161e-02 -3.16855252e-01 -3.91184866e-01
-2.99756765e-01 5.63672066e-01 3.09242487e-01 1.19848056e-02
7.77190149e-01 4.02661890e-01 3.36710662e-01 -6.66616976e-01
-7.57870018e-01 -5.14400840e-01 -5.34969032e-01 -1.33747712e-01
1.26484287e+00 -1.08686805e+00 -4.01711851e-01 1.26718625e-01
-6.58426106e-01 -2.06729576e-01 -1.69946805e-01 1.06497908e+00
1.05283316e-02 2.66273767e-01 -6.99912906e-01 -5.38182378e-01
-6.46820962e-01 -8.71402085e-01 8.24036121e-01 2.66399443e-01
-1.52062178e-01 -9.86780286e-01 -1.51606441e-01 4.03200328e-01
5.63417673e-01 3.32299948e-01 1.14991152e+00 -6.30894840e-01
-7.38845050e-01 -8.39446247e-01 -6.31127477e-01 3.26682150e-01
4.75733519e-01 -2.46377811e-01 -8.03662717e-01 -2.01322019e-01
2.81455934e-01 -2.74124444e-01 8.95668864e-01 7.87042677e-01
1.63528013e+00 2.19111051e-02 -8.36857975e-01 6.78935349e-01
1.53730083e+00 7.16130257e-01 4.38738048e-01 1.07772850e-01
7.68108010e-01 -7.79133961e-02 3.49800885e-01 2.51158476e-01
6.49864078e-02 1.04506155e-02 5.68458021e-01 -5.65819681e-01
-1.54098883e-01 -1.37474850e-01 -3.52836162e-01 5.37387609e-01
-2.27875099e-01 -1.94935203e-01 -1.42604327e+00 3.70196104e-01
-1.05869639e+00 -8.04055691e-01 -5.26864529e-01 1.78669512e+00
6.20964348e-01 1.46326929e-01 -7.44166225e-02 1.70439273e-01
6.76165760e-01 -4.21373993e-01 -5.34576237e-01 -9.91886575e-03
5.01341701e-01 7.37505078e-01 6.80681765e-01 -1.40744328e-01
-1.13325143e+00 2.82215595e-01 6.16533518e+00 6.37411118e-01
-1.38337374e+00 2.82610595e-01 1.00056517e+00 2.85918355e-01
1.41405523e-01 -3.68941128e-01 -4.84331578e-01 3.52016777e-01
8.80093157e-01 1.48490116e-01 -2.50084460e-01 1.13270223e+00
1.35762870e-01 -5.04279196e-01 -9.91067350e-01 6.38736606e-01
-3.19783270e-01 -1.32534552e+00 -2.08455697e-01 3.11806232e-01
6.71743870e-01 1.58724397e-01 1.60583556e-01 3.12321424e-01
-7.84734078e-03 -1.27204168e+00 -1.97005153e-01 3.20636928e-01
1.28805006e+00 -6.10636532e-01 1.14501333e+00 6.51545346e-01
-1.31362879e+00 -8.06228295e-02 -4.29664671e-01 4.74802069e-02
-3.43726397e-01 6.32464170e-01 -1.48715615e+00 3.43780786e-01
9.59717453e-01 3.68605286e-01 -6.59102738e-01 9.58034277e-01
-6.36133701e-02 1.07893646e+00 -3.02272230e-01 -2.83826530e-01
1.97004750e-01 8.18743706e-02 -2.28972733e-01 1.11714804e+00
6.31276548e-01 4.41660494e-01 1.11901715e-01 8.62400830e-01
-2.46915027e-01 -2.63432823e-02 -6.52704298e-01 5.54803535e-02
1.84430450e-01 1.17996478e+00 -1.14552891e+00 -5.44560432e-01
-6.60479486e-01 5.65767407e-01 -1.63042635e-01 -2.23387629e-02
-1.07241511e+00 -1.30144179e-01 -1.52597725e-01 4.26578254e-01
4.48850691e-02 2.75110096e-01 -5.72834373e-01 -8.30311239e-01
-3.48496556e-01 -5.18937111e-01 4.08263564e-01 -6.17505908e-01
-9.66473758e-01 5.61809897e-01 5.47326915e-03 -1.19371927e+00
-1.89493343e-01 -7.13067651e-01 -9.98667538e-01 9.37458754e-01
-1.34744167e+00 -9.50542927e-01 -7.05179691e-01 4.51121539e-01
4.66579169e-01 -1.66824132e-01 9.49269235e-01 3.73084694e-01
-6.75504327e-01 3.44738662e-01 2.27133527e-01 3.79002869e-01
2.74416745e-01 -1.26981688e+00 6.90961853e-02 4.89766508e-01
-5.58038294e-01 3.66237134e-01 -1.36986509e-01 -7.95907855e-01
-1.22040081e+00 -1.58833766e+00 7.11620212e-01 -2.59110719e-01
6.10489063e-02 2.13129923e-01 -9.36379313e-01 6.20248914e-01
-3.06101125e-02 2.40688711e-01 8.57053101e-01 -5.31683087e-01
4.22064036e-01 1.15677327e-01 -1.48897457e+00 2.32064560e-01
6.69723868e-01 -3.49097818e-01 -1.98553726e-01 5.36038876e-01
3.27528507e-01 -4.20026004e-01 -1.03870845e+00 6.36251926e-01
3.50453556e-01 -9.19028223e-01 1.06679440e+00 -1.40225798e-01
6.33954227e-01 1.50269598e-01 1.02091759e-01 -6.86109185e-01
-6.27133727e-01 4.57666308e-01 -1.64504088e-02 5.29599369e-01
4.53577697e-01 -4.18983489e-01 1.40066278e+00 5.81834912e-01
-2.50252426e-01 -1.24215126e+00 -8.14970493e-01 -3.76362950e-01
2.95263529e-01 -2.46748134e-01 2.76510030e-01 7.56547868e-01
-4.98382837e-01 -1.00063402e-02 1.15922511e-01 -8.49115103e-02
3.01764637e-01 -3.60838443e-01 3.49278182e-01 -1.19593203e+00
-2.49011934e-01 -2.01465115e-01 -5.92893139e-02 -7.00965106e-01
-4.21479046e-01 -1.02463639e+00 -2.66834110e-01 -1.84568012e+00
5.81590116e-01 -3.62110138e-01 -3.21319640e-01 4.58134830e-01
-3.06118190e-01 3.55375260e-01 -8.38223100e-02 3.75845075e-01
1.81467533e-02 1.02235749e-01 1.66902459e+00 -9.43905935e-02
9.93625745e-02 5.84326565e-01 -5.66422760e-01 8.36107433e-01
1.33035743e+00 -5.82804620e-01 -3.93413961e-01 -3.28355700e-01
-8.46351981e-02 4.79985952e-01 4.66805935e-01 -1.32606089e+00
7.63047859e-02 -9.65737626e-02 1.09931850e+00 -9.04642105e-01
2.72423118e-01 -1.10658002e+00 2.11925507e-01 1.41058648e+00
-1.13571361e-01 -1.32804856e-01 1.46560445e-01 4.81997550e-01
-1.32632256e-01 -3.18807572e-01 1.03209698e+00 -6.31612182e-01
-3.92156094e-01 4.82968062e-01 -6.35742843e-01 -3.35650772e-01
1.15320599e+00 -4.51591790e-01 8.84318352e-03 -7.85289332e-02
-4.10901189e-01 2.06084233e-02 -4.61851843e-02 -3.04180056e-01
7.66305029e-01 -1.34442449e+00 -5.21264076e-01 2.64633685e-01
-1.78175464e-01 3.77744913e-01 2.06405520e-01 1.00095999e+00
-1.02823162e+00 8.70635450e-01 -1.56169549e-01 -6.04415417e-01
-1.15507519e+00 2.64553815e-01 7.39055276e-01 -6.36564493e-01
-5.76318204e-01 9.33660984e-01 3.91587585e-01 -2.22425535e-01
1.43320680e-01 -8.66512537e-01 -1.79997995e-01 -4.83499080e-01
1.15782797e-01 4.98982817e-01 1.82293922e-01 -1.51586711e-01
-3.90240431e-01 4.46165830e-01 -4.59743524e-03 3.11691701e-01
1.10824430e+00 2.21013829e-01 -2.06093118e-02 2.16321703e-02
1.25504410e+00 -4.08400655e-01 -6.15602732e-01 -2.21659499e-03
-3.04702912e-02 -2.83229411e-01 1.97454885e-01 -1.06884181e+00
-1.31716514e+00 1.08501589e+00 1.08937967e+00 1.74773201e-01
1.22811282e+00 1.21145107e-01 8.56620610e-01 7.32837915e-01
-8.05423856e-02 -7.02725649e-01 1.96028128e-01 2.21485108e-01
4.16391701e-01 -1.59176505e+00 1.88810512e-01 -4.26467717e-01
-3.84299368e-01 1.26515412e+00 1.02873135e+00 -1.37720704e-01
9.60492194e-01 1.42409965e-01 3.07276566e-02 -5.04785419e-01
-6.24389291e-01 -1.21281385e-01 2.05816388e-01 5.14995217e-01
9.33485925e-01 2.12019756e-01 -3.58802855e-01 6.15146399e-01
8.28197300e-02 4.06338036e-01 -6.11186726e-03 1.23762202e+00
-7.45090902e-01 -7.48213291e-01 -5.78140140e-01 1.26387095e+00
-9.27376211e-01 -8.94285664e-02 -5.23840427e-01 1.31189501e+00
4.10110682e-01 5.75183928e-01 -1.69490166e-02 -1.67524949e-01
1.17056929e-01 8.37448537e-02 1.58476800e-01 -8.40840876e-01
-6.65864706e-01 3.84955630e-02 -2.18000188e-01 -8.16357210e-02
-3.74045521e-01 -2.55194247e-01 -1.50167465e+00 -2.61901796e-01
-5.76586008e-01 6.82384446e-02 5.36714435e-01 6.36623025e-01
-1.12389266e-01 9.14153576e-01 4.89133716e-01 -3.48066628e-01
-4.30302799e-01 -1.10657239e+00 -6.25620663e-01 -1.75806776e-01
8.45491607e-03 -2.17373013e-01 -3.20713259e-02 -4.32041250e-02] | [15.349769592285156, -2.1978328227996826] |
a27de734-febd-413b-b86f-f725c66d2552 | efficient-large-scale-multi-modal | 1802.02892 | null | http://arxiv.org/abs/1802.02892v1 | http://arxiv.org/pdf/1802.02892v1.pdf | Efficient Large-Scale Multi-Modal Classification | While the incipient internet was largely text-based, the modern digital world
is becoming increasingly multi-modal. Here, we examine multi-modal
classification where one modality is discrete, e.g. text, and the other is
continuous, e.g. visual representations transferred from a convolutional neural
network. In particular, we focus on scenarios where we have to be able to
classify large quantities of data quickly. We investigate various methods for
performing multi-modal fusion and analyze their trade-offs in terms of
classification accuracy and computational efficiency. Our findings indicate
that the inclusion of continuous information improves performance over
text-only on a range of multi-modal classification tasks, even with simple
fusion methods. In addition, we experiment with discretizing the continuous
features in order to speed up and simplify the fusion process even further. Our
results show that fusion with discretized features outperforms text-only
classification, at a fraction of the computational cost of full multi-modal
fusion, with the additional benefit of improved interpretability. | ['T. Mikolov', 'A. Joulin', 'E. Grave', 'D. Kiela'] | 2018-02-06 | null | null | null | null | ['multi-modal-classification'] | ['miscellaneous'] | [ 2.36235067e-01 -3.44231606e-01 -1.72812829e-03 -3.24625313e-01
-1.17085361e+00 -7.88361847e-01 9.50917900e-01 6.32477045e-01
-3.99197251e-01 6.73432887e-01 2.90854931e-01 -3.59007180e-01
-3.14828128e-01 -7.64654577e-01 -4.13921177e-01 -4.40998852e-01
2.58216739e-01 5.51347017e-01 -2.06543818e-01 -6.36718050e-02
6.44087791e-02 4.42560315e-01 -1.93673062e+00 5.68629026e-01
7.87585199e-01 1.39379311e+00 -2.01334879e-01 7.23899961e-01
-2.61410207e-01 4.73989248e-01 -5.17760098e-01 -6.25585616e-01
2.15150669e-01 -1.29268067e-02 -7.66642928e-01 2.99835443e-01
5.92969298e-01 -1.45024016e-01 -1.84666455e-01 9.53766942e-01
4.08286452e-01 2.20734015e-01 7.84970164e-01 -1.58852267e+00
-4.05162692e-01 3.11406374e-01 -5.57755589e-01 3.21907960e-02
6.52284205e-01 -1.57649219e-01 1.04326630e+00 -8.30813229e-01
6.11572087e-01 1.43149030e+00 7.34751642e-01 -2.24378273e-01
-1.34480155e+00 -3.87700170e-01 6.43256828e-02 8.52424651e-02
-1.41881728e+00 -5.16592145e-01 6.50672138e-01 -4.76887017e-01
7.66292870e-01 3.00936878e-01 3.61862600e-01 6.69910431e-01
4.30807769e-01 5.82538486e-01 1.12601447e+00 -4.93406236e-01
1.82772651e-02 5.91466650e-02 9.89707410e-02 7.10914791e-01
3.09128344e-01 -2.45120198e-01 -5.72174132e-01 -3.84423465e-01
3.08666497e-01 1.54938966e-01 -1.05798207e-01 -1.95256010e-01
-1.39970303e+00 8.26942265e-01 3.06764424e-01 4.47524637e-01
-5.27613103e-01 4.94616106e-02 4.76518601e-01 4.85131264e-01
6.77226484e-01 1.48406386e-01 -2.10190535e-01 -1.41411066e-01
-1.18071651e+00 1.07227176e-01 6.67551875e-01 5.34541667e-01
1.02708519e+00 -2.20893934e-01 1.17087118e-01 8.37778807e-01
5.09063676e-02 4.32800293e-01 3.36294174e-01 -1.02747107e+00
6.67641461e-01 6.71736002e-01 2.09767930e-02 -9.55405653e-01
-6.38095796e-01 -1.24465771e-01 -8.55774820e-01 3.32800418e-01
6.98055029e-01 -3.01642239e-01 -1.09438086e+00 1.60985875e+00
9.26867351e-02 -4.99791175e-01 2.35185120e-03 7.01139390e-01
6.19446337e-01 6.58736706e-01 2.71420032e-01 -7.94950053e-02
1.46003175e+00 -5.02318799e-01 -6.78163588e-01 -7.18263984e-02
7.22019136e-01 -9.72127140e-01 8.29977572e-01 4.08809334e-01
-9.53618109e-01 -3.81467879e-01 -9.46215749e-01 -2.50313699e-01
-8.43834043e-01 1.59264490e-01 5.00065029e-01 6.07846081e-01
-1.19081700e+00 4.86468494e-01 -7.59466529e-01 -6.73585534e-01
3.83777976e-01 2.89813608e-01 -7.43090928e-01 -1.76147401e-01
-1.07514942e+00 9.11094248e-01 4.46255237e-01 -3.98502380e-01
-1.41033167e-02 -5.49761653e-01 -9.29806650e-01 2.03910112e-01
2.18433961e-01 -6.79948926e-01 1.04580379e+00 -1.14973819e+00
-8.94662738e-01 6.52452111e-01 -2.74089009e-01 -2.14537054e-01
5.90159595e-01 1.12610951e-01 -4.04902965e-01 3.71552497e-01
1.12316698e-01 8.22273314e-01 7.06176162e-01 -1.30112481e+00
-9.12576437e-01 -4.44411784e-01 4.50453877e-01 5.16825259e-01
-6.63469434e-01 -1.99079648e-01 -2.34809205e-01 -7.05039740e-01
1.52904063e-01 -9.98616040e-01 1.93804175e-01 3.90943855e-01
-1.89847127e-01 -2.26531386e-01 1.19240916e+00 -6.85812533e-01
8.69737506e-01 -2.14178848e+00 1.18888460e-01 2.22022310e-01
4.71292764e-01 -1.21285126e-01 -2.92370822e-02 5.96533775e-01
-7.17055500e-02 3.45914900e-01 -1.58479363e-01 -7.21178651e-01
5.12522459e-02 2.01979011e-01 1.03002349e-02 5.22825599e-01
9.56387445e-02 8.32605243e-01 -4.36658144e-01 -6.73474967e-01
4.61201519e-01 7.09781468e-01 -1.83253646e-01 -4.15501714e-01
7.39383250e-02 9.63040069e-02 -2.18942046e-01 8.82278919e-01
5.16237855e-01 -6.09628022e-01 1.66889012e-03 -4.88057435e-01
2.43180804e-02 -7.71452412e-02 -1.12332380e+00 1.70594656e+00
-7.97760487e-01 9.55704570e-01 2.28058934e-01 -9.92998302e-01
3.66970599e-01 4.66531247e-01 6.51410580e-01 -5.53121746e-01
3.17449510e-01 8.53400007e-02 -1.56577364e-01 -5.17283082e-02
8.88051629e-01 -4.14427698e-01 -1.32919803e-01 6.64144754e-01
2.03025937e-02 6.71155974e-02 3.02976221e-01 3.26051056e-01
8.07238221e-01 -2.49616936e-01 2.86208272e-01 -1.25354469e-01
9.49326903e-02 2.11471707e-01 -2.04319358e-02 6.35716915e-01
7.45530277e-02 7.09330142e-01 3.50404650e-01 -2.30428204e-01
-8.76134813e-01 -8.90062332e-01 -1.21338032e-01 1.07828343e+00
2.65199721e-01 -4.42563295e-01 -3.52387369e-01 -6.90161645e-01
2.87389010e-01 6.16716266e-01 -7.03659892e-01 2.62703020e-02
-6.48429021e-02 -7.05491483e-01 4.79567707e-01 5.08759260e-01
3.94019067e-01 -3.96090806e-01 -6.89577460e-01 2.46700104e-02
-4.81705278e-01 -1.16682112e+00 -9.18041393e-02 2.22034156e-01
-7.26466715e-01 -8.96918595e-01 -7.50676095e-01 -2.07996845e-01
4.55413818e-01 4.22984093e-01 8.83236825e-01 -9.30203274e-02
-4.60160635e-02 7.66223490e-01 -5.50645113e-01 -2.42461964e-01
-1.86938018e-01 1.76864326e-01 -1.77723125e-01 7.07991496e-02
1.75395802e-01 -3.76044631e-01 -3.19916815e-01 1.01961158e-01
-1.35207486e+00 1.51321277e-01 5.57906449e-01 6.56095624e-01
2.34934673e-01 3.40635955e-01 4.81345177e-01 -5.20184100e-01
7.21038640e-01 -6.02892101e-01 -4.68227640e-02 4.70704615e-01
-3.98581684e-01 8.36793408e-02 5.51473618e-01 -4.70274687e-01
-1.01156509e+00 2.44521257e-02 2.63619255e-02 -5.48684776e-01
-2.31662631e-01 1.02759027e+00 7.23270178e-02 -3.54193896e-01
5.70077598e-01 -1.27946669e-02 1.77431002e-01 -3.13590318e-01
4.20540810e-01 8.94847035e-01 2.00667262e-01 -5.84413707e-01
5.45128226e-01 6.23509288e-01 4.70475368e-02 -8.88054132e-01
-5.46022475e-01 -1.73031092e-01 -5.37699521e-01 -4.60366338e-01
6.57771230e-01 -8.38135004e-01 -5.27276635e-01 4.11622822e-01
-8.48520637e-01 8.79824013e-02 -1.37908891e-01 3.92380238e-01
-4.14324462e-01 4.68814373e-01 -5.30960023e-01 -5.16617894e-01
-9.03163403e-02 -1.19378555e+00 1.27764308e+00 4.50804755e-02
-1.70408309e-01 -1.12797856e+00 -4.26236391e-01 4.29447651e-01
4.90556926e-01 2.86197305e-01 1.02179611e+00 -7.87887037e-01
-1.83395252e-01 -5.51591337e-01 -5.50824165e-01 -5.82242869e-02
2.96331972e-01 -9.69081745e-02 -1.01054204e+00 -3.89748454e-01
-4.36789125e-01 -4.43084031e-01 8.61862361e-01 2.57017493e-01
1.06646705e+00 -3.29047702e-02 -3.38440835e-01 2.26468727e-01
1.39034998e+00 6.84304908e-02 3.08560401e-01 2.25065544e-01
5.46594620e-01 6.10767663e-01 6.23853683e-01 4.96970177e-01
7.12947845e-01 5.59082747e-01 2.51621336e-01 -1.97706059e-01
-2.39347190e-01 1.85124144e-01 -1.32191733e-01 4.51850593e-01
-4.58093919e-02 -6.12630486e-01 -1.27165782e+00 4.74175274e-01
-1.86949146e+00 -6.37844920e-01 2.69311100e-01 2.15997386e+00
3.31048816e-01 8.94372612e-02 2.27966905e-01 4.93759692e-01
9.29108262e-01 1.01679534e-01 -4.05304104e-01 -1.78269595e-01
-1.92023054e-01 -1.95411816e-02 4.00964439e-01 4.50343549e-01
-1.21031845e+00 3.47421587e-01 6.59870338e+00 9.00045991e-01
-1.26818430e+00 9.12752822e-02 7.32483387e-01 -1.74116805e-01
-2.60276109e-01 -2.44222671e-01 -3.60392392e-01 3.88211250e-01
1.02043986e+00 -2.83033252e-01 3.71986419e-01 2.69812852e-01
-2.44619757e-01 -4.57938313e-01 -1.09566927e+00 1.15038931e+00
1.38707668e-01 -1.28900993e+00 2.80859202e-01 2.29304954e-01
5.38760960e-01 5.89562953e-02 7.33575895e-02 1.41122505e-01
1.72332197e-01 -8.83205116e-01 8.76785517e-01 5.26517689e-01
8.92227530e-01 -7.85455346e-01 7.02124596e-01 1.95704281e-01
-1.42862034e+00 -1.41969129e-01 3.26038301e-01 -7.06382934e-03
2.58656532e-01 6.71201944e-01 -1.18911922e-01 1.02626348e+00
5.46832979e-01 6.35501981e-01 -8.18548322e-01 9.40498233e-01
6.12512887e-01 1.18195556e-01 -7.49224424e-01 2.25182503e-01
2.23639935e-01 1.07598424e-01 4.64979082e-01 1.17785835e+00
6.87112808e-01 1.93283577e-02 4.31652695e-01 3.34859759e-01
-3.39511670e-02 -4.81061526e-02 -5.82548678e-01 -1.81198686e-01
5.63922524e-01 1.25833142e+00 -1.04218006e+00 -6.06604040e-01
-6.54062271e-01 6.76549375e-01 1.38418317e-01 3.52520883e-01
-6.60130978e-01 -5.33187211e-01 3.28339756e-01 6.00968711e-02
1.14941373e-01 -3.53254408e-01 -5.45911372e-01 -1.23816764e+00
-4.53886874e-02 -6.18745685e-01 5.50232768e-01 -9.88568306e-01
-1.35463595e+00 4.19156820e-01 2.89545625e-01 -1.29088664e+00
-3.77818674e-01 -5.55865645e-01 -7.01401234e-02 8.70359361e-01
-1.24205732e+00 -1.45995283e+00 -3.17293674e-01 6.79918766e-01
3.31962913e-01 1.97350264e-01 6.03918433e-01 5.46801329e-01
-2.64462650e-01 5.13693929e-01 1.78576469e-01 -1.20240614e-01
6.34320617e-01 -1.00590336e+00 -7.31118247e-02 5.49044490e-01
3.49458903e-02 3.45072895e-01 6.89508617e-01 -5.60918629e-01
-1.47020411e+00 -8.74312997e-01 7.29424536e-01 -1.90598279e-01
7.25732684e-01 -1.14655696e-01 -7.18829215e-01 6.09909892e-01
2.52882600e-01 -1.97184458e-02 7.37625360e-01 2.99256742e-01
-4.76029873e-01 3.88687737e-02 -1.36932886e+00 4.43038046e-01
6.23963237e-01 -8.01939845e-01 -4.73822504e-01 2.03487635e-01
3.34838629e-01 -2.03605205e-01 -1.20656407e+00 3.89948756e-01
7.35983729e-01 -7.44766772e-01 9.24607277e-01 -7.76543915e-02
3.49957556e-01 -3.22859526e-01 -5.60155153e-01 -1.38254786e+00
-1.75195664e-01 -1.35990409e-02 5.63091636e-02 1.14269638e+00
3.28256249e-01 -7.89520502e-01 5.84195495e-01 7.78562069e-01
1.44265026e-01 -4.06843126e-01 -1.22674108e+00 -5.81377029e-01
1.87759250e-02 -4.99400914e-01 4.52641368e-01 1.06217420e+00
9.69146192e-02 2.53111959e-01 -1.53744563e-01 -1.99496411e-02
4.29499477e-01 3.52547705e-01 3.95597398e-01 -1.49334013e+00
2.01981977e-01 -5.36260247e-01 -7.07835734e-01 -5.86504221e-01
2.24556532e-02 -8.29156578e-01 -3.94615412e-01 -1.71683812e+00
2.30474055e-01 -3.35887879e-01 -3.19767505e-01 8.31385970e-01
6.42138720e-02 6.14392519e-01 4.98903692e-01 1.86274409e-01
-6.28043175e-01 2.59861797e-01 8.95030677e-01 -2.06553042e-01
2.89539266e-02 -3.99652064e-01 -6.55034423e-01 6.00011766e-01
7.89264321e-01 -1.34923190e-01 -3.19196403e-01 -4.65376824e-01
1.52296126e-01 4.30765361e-01 5.15629709e-01 -1.21918774e+00
3.87341917e-01 6.01374470e-02 6.27529740e-01 -7.23792493e-01
7.77003109e-01 -1.04227221e+00 4.03104186e-01 -1.82926480e-03
-2.59033322e-01 1.93718255e-01 4.13414687e-01 5.78864574e-01
-5.00768840e-01 1.65506139e-01 7.08793283e-01 1.36588335e-01
-4.79868621e-01 -5.61810210e-02 -6.58662617e-01 -3.00867736e-01
9.46011841e-01 -4.33729440e-01 -5.23101807e-01 -7.05001056e-01
-9.48911905e-01 -5.73136434e-02 6.21505678e-01 4.84318674e-01
2.65524656e-01 -1.49382007e+00 -3.98575366e-01 -1.90564431e-02
2.11198926e-01 -5.12748837e-01 1.97533667e-01 9.38267171e-01
-3.21481347e-01 4.33540016e-01 -3.27332139e-01 -7.27007329e-01
-1.38652778e+00 3.82019430e-01 1.42302468e-01 -1.81134522e-01
-4.47254121e-01 2.91680276e-01 1.20977461e-02 -2.49821812e-01
1.06697083e-01 -2.26681337e-01 -5.05948514e-02 6.59223974e-01
2.23811194e-01 4.69328970e-01 4.53226894e-01 -9.52968478e-01
-4.64659512e-01 6.22285545e-01 -5.14291003e-02 -3.97778600e-01
1.10655022e+00 -4.58221734e-01 -7.00497404e-02 6.51416183e-01
1.35081470e+00 -3.62556487e-01 -8.82685184e-01 -3.01546097e-01
-2.40843982e-01 -5.19510210e-01 4.75496113e-01 -9.38541591e-01
-1.10981786e+00 7.06888139e-01 6.47267401e-01 6.93972111e-01
1.35132861e+00 1.46341864e-02 6.08103693e-01 4.34554815e-01
4.33980137e-01 -8.60957026e-01 -3.15709829e-01 3.51089180e-01
6.35860503e-01 -1.35921741e+00 2.55451173e-01 -3.08849394e-01
-5.95327199e-01 1.16279042e+00 1.04065828e-01 1.71613500e-01
8.58465612e-01 1.39582768e-01 -5.35964593e-02 -2.52628207e-01
-7.18401551e-01 -2.99062133e-01 4.52057332e-01 2.50033110e-01
4.23255146e-01 1.04352478e-02 -1.61356822e-01 5.99274561e-02
-1.37630835e-01 -3.99244949e-03 2.63886750e-01 1.26250887e+00
-3.46895576e-01 -8.90187860e-01 -7.12681353e-01 1.05466294e+00
-5.69628596e-01 5.26634455e-02 -3.14116329e-01 9.67694044e-01
4.68727574e-02 1.29233944e+00 3.49906743e-01 -5.29388607e-01
8.61335546e-02 3.24564815e-01 3.64560068e-01 -1.57909200e-01
-4.82458740e-01 1.65640846e-01 2.79976368e-01 -3.51789057e-01
-7.16333032e-01 -9.04010594e-01 -8.85364830e-01 -7.39480734e-01
-3.81181836e-01 7.73400255e-03 8.46175492e-01 1.16076720e+00
4.54704314e-01 4.61825520e-01 2.82614887e-01 -1.24480915e+00
-3.96266431e-02 -9.16148663e-01 -3.72737765e-01 3.85276049e-01
5.97786844e-01 -9.67032015e-01 -3.93273830e-01 1.89794078e-01] | [13.135440826416016, 5.071769714355469] |
2cb312f5-ce14-46c8-be92-cb92908295b6 | impact-of-a-dct-driven-loss-in-attention | 2205.01997 | null | https://arxiv.org/abs/2205.01997v2 | https://arxiv.org/pdf/2205.01997v2.pdf | Attention-based Knowledge Distillation in Multi-attention Tasks: The Impact of a DCT-driven Loss | Knowledge Distillation (KD) is a strategy for the definition of a set of transferability gangways to improve the efficiency of Convolutional Neural Networks. Feature-based Knowledge Distillation is a subfield of KD that relies on intermediate network representations, either unaltered or depth-reduced via maximum activation maps, as the source knowledge. In this paper, we propose and analyse the use of a 2D frequency transform of the activation maps before transferring them. We pose that\textemdash by using global image cues rather than pixel estimates, this strategy enhances knowledge transferability in tasks such as scene recognition, defined by strong spatial and contextual relationships between multiple and varied concepts. To validate the proposed method, an extensive evaluation of the state-of-the-art in scene recognition is presented. Experimental results provide strong evidences that the proposed strategy enables the student network to better focus on the relevant image areas learnt by the teacher network, hence leading to better descriptive features and higher transferred performance than every other state-of-the-art alternative. We publicly release the training and evaluation framework used along this paper at http://www-vpu.eps.uam.es/publications/DCTBasedKDForSceneRecognition. | ['Juan C. SanMiguel', 'Jesús Bescós', 'Marcos Escudero-Viñolo', 'Alejandro López-Cifuentes'] | 2022-05-04 | null | null | null | null | ['scene-recognition'] | ['computer-vision'] | [ 3.59808385e-01 2.39089459e-01 4.48947586e-02 -3.22557837e-01
-2.90003955e-01 -6.17836177e-01 8.90683293e-01 2.04845607e-01
-8.40291142e-01 8.56804311e-01 -2.27543443e-01 -1.97201550e-01
-7.00701296e-01 -1.06289995e+00 -1.01247084e+00 -9.86969829e-01
1.49151832e-01 1.11188255e-01 3.91649634e-01 -2.73689747e-01
2.97130853e-01 8.97349834e-01 -1.91851664e+00 4.92896467e-01
1.00131679e+00 1.00665176e+00 5.15735209e-01 3.97458613e-01
-3.66610378e-01 8.17150533e-01 -6.27450109e-01 -1.33147642e-01
1.30791157e-01 -4.72658485e-01 -9.64397788e-01 -3.06150466e-01
7.42306292e-01 5.36861382e-02 -2.88232118e-01 1.05404043e+00
5.02007663e-01 4.18232560e-01 8.15024555e-01 -8.33853006e-01
-8.66869211e-01 5.79861820e-01 -1.91308185e-01 6.81641698e-01
3.64096910e-01 -2.71061901e-02 5.00817478e-01 -1.12075222e+00
7.63914347e-01 1.00586927e+00 4.81363207e-01 3.55531216e-01
-1.11367738e+00 -5.56166530e-01 1.15123205e-01 8.77033591e-01
-1.51567519e+00 -1.65270522e-01 9.67304468e-01 -4.40843523e-01
9.34368134e-01 2.14361712e-01 9.22517538e-01 8.69544208e-01
-2.15541825e-01 9.30897057e-01 1.51105297e+00 -9.76234794e-01
7.75038749e-02 7.11143553e-01 1.22913823e-01 6.72497988e-01
2.99011230e-01 3.04109186e-01 -9.80101287e-01 6.45047307e-01
9.07023787e-01 -2.83080161e-01 -5.68997562e-01 -6.59579217e-01
-1.08938134e+00 6.62237227e-01 9.36966896e-01 8.32490563e-01
-4.11935121e-01 -5.95998950e-02 4.32368740e-02 3.76325965e-01
2.55193442e-01 6.54775023e-01 -3.60961378e-01 3.68919447e-02
-8.33027303e-01 -1.76893428e-01 5.36952198e-01 6.21362090e-01
1.17363608e+00 2.47775167e-01 -1.33924097e-01 5.25960863e-01
6.11281730e-02 4.38475013e-01 5.83461523e-01 -4.87355411e-01
3.84074658e-01 6.75514340e-01 -4.07001764e-01 -9.63236094e-01
-1.43881679e-01 -5.67610443e-01 -7.03452408e-01 4.12194669e-01
3.70332628e-01 3.39948274e-02 -1.03292298e+00 1.41484797e+00
4.47765797e-01 5.20236671e-01 3.45619023e-01 7.13071823e-01
1.05620015e+00 6.30108297e-01 4.74755131e-02 -1.13243924e-03
8.91543269e-01 -7.98637927e-01 -5.50991237e-01 6.14953786e-02
5.96487045e-01 -4.75313246e-01 7.34225571e-01 6.54796302e-01
-9.45093811e-01 -1.07659757e+00 -1.04752553e+00 6.86122477e-02
-1.23989701e+00 8.60865712e-02 6.21734738e-01 6.40405357e-01
-1.40191472e+00 7.73677468e-01 -4.07814175e-01 -4.80420917e-01
6.97661579e-01 2.93456912e-01 -4.93146032e-01 -2.17754915e-01
-1.29274595e+00 1.29697895e+00 1.09059501e+00 9.36163515e-02
-9.17407155e-01 -1.05773485e+00 -8.23349774e-01 4.34922278e-02
1.04605556e-01 -3.70250046e-01 8.75634909e-01 -1.22766435e+00
-1.53212762e+00 1.00900841e+00 4.08756047e-01 -4.70466882e-01
6.04864299e-01 -3.18280458e-01 -2.83552974e-01 5.22528827e-01
-2.49769479e-01 8.93658221e-01 7.92605758e-01 -1.34317613e+00
-6.97353780e-01 -1.60240740e-01 6.48762807e-02 5.89765966e-01
-6.90037966e-01 -4.35562670e-01 -1.97977245e-01 -5.46029568e-01
-6.92636967e-02 -4.24636185e-01 1.11098357e-01 9.51062795e-03
-1.23472385e-01 -2.56939530e-01 8.18146884e-01 -4.51467156e-01
8.99529517e-01 -2.18361902e+00 3.19497675e-01 4.78594810e-01
4.69719479e-03 5.81361771e-01 -4.24529277e-02 2.25187972e-01
-4.25697029e-01 -2.01228961e-01 -2.20915824e-01 1.78630650e-01
-1.16099574e-01 2.35826790e-01 -1.70394048e-01 4.31765050e-01
1.91793635e-01 9.11254942e-01 -1.01397848e+00 -4.80087757e-01
7.10954368e-01 7.72853076e-01 -2.10648954e-01 1.48302868e-01
5.17597683e-02 4.54921782e-01 -3.19238842e-01 2.14318216e-01
6.64456546e-01 9.30089206e-02 -3.82822682e-03 -2.96318144e-01
-3.99135649e-01 -9.43336487e-02 -1.38966703e+00 1.98230112e+00
-5.82331002e-01 8.68711054e-01 -3.75311077e-01 -1.51092911e+00
1.14497590e+00 1.23388521e-01 2.46873483e-01 -9.15283203e-01
2.01846763e-01 1.69845879e-01 -9.86069664e-02 -4.44740534e-01
3.93938988e-01 1.86655357e-01 3.94125789e-01 8.14401358e-02
6.64972246e-01 -3.36712539e-01 2.65397072e-01 -1.40956566e-01
6.10996544e-01 4.35556710e-01 4.58701819e-01 -5.78771830e-01
7.84145772e-01 -5.81665635e-02 -8.55734274e-02 7.86649346e-01
-2.37060450e-02 1.24203727e-01 1.55524258e-02 -3.13551217e-01
-5.66132426e-01 -1.08767414e+00 -3.87816787e-01 1.12382996e+00
1.17347687e-01 1.31692335e-01 -7.41805911e-01 -6.79085493e-01
2.66512092e-02 7.60888577e-01 -8.82948697e-01 -3.25322151e-01
-4.98800457e-01 -4.00610298e-01 4.81574029e-01 5.91590345e-01
1.06056571e+00 -1.29161775e+00 -9.15227890e-01 4.56944481e-02
3.17652702e-01 -8.75133753e-01 2.47903004e-01 5.17689824e-01
-9.34375465e-01 -1.11053455e+00 -8.92961919e-01 -9.72138584e-01
8.21872890e-01 2.74533898e-01 1.00470603e+00 6.47658631e-02
-5.33656657e-01 7.63061047e-01 -5.44591129e-01 -5.51318407e-01
-1.31178647e-01 4.21561673e-02 -2.09743664e-01 3.58605534e-02
4.30309445e-01 -6.08842790e-01 -5.21332443e-01 -1.02763750e-01
-1.05694878e+00 1.83642045e-01 9.64955449e-01 6.45319283e-01
6.62175238e-01 2.03902349e-01 3.46336275e-01 -8.01183105e-01
5.27649045e-01 -2.66823411e-01 -5.61761320e-01 4.20091420e-01
-6.02216482e-01 2.39506543e-01 4.33229923e-01 -4.49726164e-01
-1.42211485e+00 7.40124658e-02 -9.09164622e-02 -5.06284356e-01
-7.12679267e-01 6.25053287e-01 -6.59174845e-02 -6.97893083e-01
9.55769777e-01 6.74703479e-01 -4.33461756e-01 -3.40864956e-01
6.09050274e-01 1.54635385e-01 6.02491498e-01 -7.05467939e-01
8.78502667e-01 2.72315294e-01 2.72696689e-02 -1.13099265e+00
-7.21074402e-01 -5.72389960e-01 -1.18165398e+00 -4.75706011e-01
9.14423168e-01 -7.37828851e-01 -5.12024581e-01 5.51941872e-01
-1.06366241e+00 -4.45714206e-01 -7.39141643e-01 6.08225465e-01
-4.32959557e-01 -7.28484616e-02 -1.60598457e-01 -6.06591403e-01
5.71343629e-03 -8.56614351e-01 3.77328306e-01 6.22811198e-01
3.95237148e-01 -1.31841481e+00 8.03154930e-02 -8.49263184e-03
4.67849672e-01 3.07010621e-01 8.19834352e-01 -7.64985204e-01
-5.57771087e-01 1.79947227e-01 -2.44348273e-01 5.15833914e-01
-7.47853890e-02 -1.79657027e-01 -1.50192857e+00 2.00893567e-03
-3.57310414e-01 -4.42151517e-01 1.24729514e+00 2.24578410e-01
1.29882717e+00 -6.82200938e-02 -3.17344159e-01 6.38963282e-01
1.71916497e+00 2.19633475e-01 7.33264208e-01 5.38867891e-01
6.61040187e-01 7.10999072e-01 4.32046890e-01 2.30642557e-01
2.03678355e-01 5.86983383e-01 5.09082973e-01 -2.55180866e-01
-5.37770450e-01 -1.73806295e-01 3.19866210e-01 8.30523491e-01
-3.38313550e-01 -2.69693211e-02 -1.09142351e+00 7.91473150e-01
-1.51387477e+00 -9.31345820e-01 2.33672746e-02 2.01623774e+00
1.00380993e+00 1.05710767e-01 -2.92265207e-01 4.45426732e-01
5.64952850e-01 -8.84666666e-03 -2.93517500e-01 -5.47805011e-01
-2.37629771e-01 7.74681687e-01 3.84103268e-01 5.52144766e-01
-1.05369353e+00 1.05279887e+00 5.28956795e+00 1.11672020e+00
-1.14080644e+00 -4.44772132e-02 3.89440119e-01 3.70144576e-01
-1.53711095e-01 -4.91709113e-01 -8.74203563e-01 1.69398286e-03
8.49653840e-01 -7.19390437e-02 1.88384041e-01 8.61494660e-01
-2.25778863e-01 -2.76651263e-01 -1.03935981e+00 9.29556370e-01
2.74659246e-01 -1.53398287e+00 3.11612129e-01 -1.67809531e-01
9.93909895e-01 -2.40816981e-01 3.61078978e-01 4.32147890e-01
1.52538642e-01 -1.06969559e+00 6.83185101e-01 7.79510617e-01
7.38737583e-01 -8.10849607e-01 6.81393206e-01 1.12181492e-01
-1.27452672e+00 4.63338569e-03 -3.67035031e-01 7.04642907e-02
-5.09036481e-01 5.93662918e-01 -1.02791750e+00 9.41509485e-01
7.54439712e-01 7.57419944e-01 -8.87777150e-01 1.21820164e+00
-8.28666925e-01 5.11596441e-01 -1.96924016e-01 -1.71882555e-01
3.59659970e-01 -5.04080206e-02 1.94016039e-01 1.53479266e+00
4.82435137e-01 1.35910302e-01 -6.21923506e-02 8.85291159e-01
4.11624797e-02 1.83096334e-01 -8.41408610e-01 3.19040775e-01
4.45466846e-01 1.14767885e+00 -7.80114233e-01 -3.98853809e-01
-2.38689110e-01 9.02767241e-01 5.18217087e-01 5.27733862e-01
-6.13465905e-01 -6.05472326e-01 2.61406600e-01 -1.92112267e-01
5.73826849e-01 -9.40231830e-02 -1.48069575e-01 -8.51405859e-01
-3.25700551e-01 -4.28998351e-01 5.24718404e-01 -8.98369789e-01
-9.64028955e-01 5.54580152e-01 3.68017375e-01 -9.72324073e-01
-1.27904594e-01 -1.00383651e+00 -7.29090989e-01 9.50714529e-01
-2.06157827e+00 -1.03988969e+00 -7.10766852e-01 1.05682063e+00
3.19809824e-01 -8.43600258e-02 9.17769730e-01 1.50983572e-01
-2.92388380e-01 5.77559471e-01 8.06006789e-02 1.06289163e-01
4.58147794e-01 -1.54337370e+00 -2.49243245e-01 8.04560721e-01
4.87941563e-01 1.63548037e-01 5.38421512e-01 -2.54653513e-01
-1.09623158e+00 -9.64281678e-01 6.02658451e-01 -2.42647469e-01
4.48408157e-01 -1.41956314e-01 -9.39368367e-01 3.73073429e-01
5.36131501e-01 6.08101711e-02 6.60990000e-01 -1.02227107e-01
-2.97351092e-01 -3.47140849e-01 -1.08417225e+00 2.60985315e-01
8.53906214e-01 -5.18395901e-01 -1.02291512e+00 1.60097569e-01
4.87317592e-01 -3.17413509e-01 -9.45707858e-01 2.46292263e-01
2.06758961e-01 -1.14283621e+00 1.08401406e+00 -3.16357970e-01
2.58693963e-01 -2.31782064e-01 1.81726944e-02 -1.60231054e+00
-1.74455419e-01 2.27448009e-02 -1.73565343e-01 1.11663675e+00
3.28228503e-01 -4.98401761e-01 7.47039795e-01 -4.86861318e-02
-2.97822773e-01 -5.94842613e-01 -9.56765771e-01 -8.68206263e-01
2.22322494e-01 -4.78103757e-01 3.44836026e-01 1.40081906e+00
-1.41691253e-01 -6.17120825e-02 3.83060753e-01 1.76322684e-01
5.20050168e-01 2.19864920e-02 5.39319754e-01 -1.25508809e+00
-1.23272732e-01 -6.67791545e-01 -7.24981546e-01 -8.50569785e-01
1.67163417e-01 -1.06504416e+00 -8.21214467e-02 -1.55556810e+00
-2.58975089e-01 -2.95499533e-01 -6.67957544e-01 7.36917496e-01
1.14207320e-01 2.83992499e-01 1.91621765e-01 -1.89371899e-01
-4.75437880e-01 5.98979354e-01 1.48206866e+00 -6.19859956e-02
-1.77140400e-01 -2.98223227e-01 -3.63355190e-01 5.61341941e-01
1.04199958e+00 -2.55423397e-01 -6.54652417e-01 -3.53048831e-01
5.23120910e-02 -4.69391495e-01 5.95473886e-01 -1.48247051e+00
4.30321783e-01 -1.37382925e-01 8.58154297e-01 -4.30598468e-01
3.77306551e-01 -9.45595503e-01 -1.87363699e-01 7.07834840e-01
-2.64118910e-01 -2.77086347e-01 6.69584513e-01 2.15552658e-01
-4.44725215e-01 -4.81265187e-01 8.74349952e-01 -2.58993596e-01
-1.48540390e+00 -4.56242077e-02 -2.33308017e-01 -8.50424170e-02
1.21152627e+00 -5.38868785e-01 -4.75998938e-01 -7.52055123e-02
-7.27070808e-01 -2.55831808e-01 -1.87064916e-01 1.67443559e-01
9.06877339e-01 -1.30612493e+00 -5.34411311e-01 3.60089511e-01
1.61706105e-01 3.96435000e-02 3.86330158e-01 7.65181005e-01
-7.16657043e-01 5.86526155e-01 -7.48069525e-01 -5.91604471e-01
-9.86385107e-01 5.09950876e-01 5.15971124e-01 -8.05496648e-02
-4.78818089e-01 1.24104071e+00 1.30324855e-01 -3.86375993e-01
3.57834518e-01 -3.65992278e-01 -7.23295331e-01 3.62787664e-01
5.52664220e-01 3.26704353e-01 2.57879227e-01 -5.36212027e-01
-3.31087351e-01 6.78906858e-01 1.15325704e-01 3.92446332e-02
1.29829848e+00 1.94596484e-01 1.80903062e-01 2.87132114e-01
1.09129763e+00 -4.06079054e-01 -1.15705383e+00 -2.81149447e-01
5.94416447e-02 -3.56797397e-01 2.67347664e-01 -1.04309714e+00
-1.06356823e+00 1.02395928e+00 9.09971416e-01 1.09071739e-01
1.29028237e+00 -1.40096456e-01 -1.32516697e-01 6.81925535e-01
2.11537868e-01 -1.16524851e+00 3.06913763e-01 5.33972442e-01
9.64219809e-01 -1.09199083e+00 1.95538048e-02 -3.59717935e-01
-4.03210580e-01 1.34392655e+00 9.15577769e-01 -2.72770133e-02
7.65892565e-01 -1.14353374e-01 4.50927056e-02 -1.71147004e-01
-2.48183414e-01 -6.02193475e-01 6.83534741e-01 9.24887419e-01
4.30645764e-01 -1.01380616e-01 8.58004838e-02 2.81919807e-01
-2.81995684e-01 -9.54803750e-02 3.52439344e-01 1.09159327e+00
-7.57170618e-01 -8.28261077e-01 -4.65435326e-01 1.13587432e-01
1.04711540e-01 -2.68593162e-01 -3.58899087e-01 1.20274818e+00
5.96151054e-01 6.47340894e-01 -1.03142753e-01 -2.74276942e-01
5.71003377e-01 1.52637109e-01 8.90929759e-01 -5.96104085e-01
-6.89812124e-01 -4.94483352e-01 -1.87799871e-01 -4.39020604e-01
-8.30814540e-01 -4.17326629e-01 -1.14963746e+00 -1.21263020e-01
-2.51169980e-01 2.32386813e-01 8.28623414e-01 9.56721187e-01
7.46157095e-02 9.17932093e-01 2.71082699e-01 -9.06346798e-01
-1.26469791e-01 -8.24775398e-01 -4.63582337e-01 2.37809315e-01
1.44089490e-01 -8.87610137e-01 -1.72704354e-01 2.57452875e-01] | [9.675445556640625, 2.1573946475982666] |
7cbf9dc4-f319-4c28-8645-d8ad89c79b50 | bert-4ever-evahan-2022-ancient-chinese-word | null | null | https://aclanthology.org/2022.lt4hala-1.22 | https://aclanthology.org/2022.lt4hala-1.22.pdf | BERT 4EVER@EvaHan 2022: Ancient Chinese Word Segmentation and Part-of-Speech Tagging Based on Adversarial Learning and Continual Pre-training | With the development of artificial intelligence (AI) and digital humanities, ancient Chinese resources and language technology have also developed and grown, which have become an increasingly important part to the study of historiography and traditional Chinese culture. In order to promote the research on automatic analysis technology of ancient Chinese, we conduct various experiments on ancient Chinese word segmentation and part-of-speech (POS) tagging tasks for the EvaHan 2022 shared task. We model the word segmentation and POS tagging tasks jointly as a sequence tagging problem. In addition, we perform a series of training strategies based on the provided ancient Chinese pre-trained model to enhance the model performance. Concretely, we employ several augmentation strategies, including continual pre-training, adversarial training, and ensemble learning to alleviate the limited amount of training data and the imbalance between POS labels. Extensive experiments demonstrate that our proposed models achieve considerable performance on ancient Chinese word segmentation and POS tagging tasks. Keywords: ancient Chinese, word segmentation, part-of-speech tagging, adversarial learning, continuing pre-training | ['Ruoyao Ding', 'Yingwen Fu', 'Ziyu Yang', 'Hailin Zhang'] | null | null | null | null | lt4hala-lrec-2022-6 | ['chinese-word-segmentation', 'culture'] | ['natural-language-processing', 'speech'] | [ 2.12282464e-01 -2.79985100e-01 5.38331829e-02 -3.84699076e-01
-6.21375561e-01 -7.12593317e-01 4.30976868e-01 -3.11321050e-01
-9.37171042e-01 5.97511590e-01 3.27496022e-01 -6.72344804e-01
7.54211187e-01 -5.26303947e-01 -1.69517085e-01 -6.57673359e-01
4.64003533e-02 4.50815052e-01 2.54923999e-01 -1.26464829e-01
4.59960580e-01 1.69775873e-01 -6.00809753e-01 3.95974517e-02
1.21294332e+00 6.26619518e-01 4.28241104e-01 6.68951631e-01
-4.34856385e-01 5.62360942e-01 -8.33223462e-01 -4.81211215e-01
1.18912302e-01 -5.05202353e-01 -8.01949918e-01 -7.66169578e-02
-5.43357313e-01 4.61522043e-02 -2.44902924e-01 1.29244554e+00
6.53293669e-01 1.60803005e-01 5.86891115e-01 -4.03139681e-01
-1.11348259e+00 1.07230401e+00 -6.49601817e-01 4.22961235e-01
-4.78641167e-02 -1.75963700e-01 9.27838445e-01 -7.85348117e-01
2.40989894e-01 1.08747423e+00 6.22350693e-01 6.50366008e-01
-5.25594413e-01 -6.85726762e-01 2.24259928e-01 1.00117989e-01
-1.28967142e+00 -1.21829398e-01 6.30542874e-01 -3.06916863e-01
7.25878239e-01 -7.27915317e-02 6.04592741e-01 9.02688742e-01
2.95683712e-01 1.17602098e+00 9.78928566e-01 -7.38520503e-01
1.83412790e-01 -1.04775809e-01 2.33127251e-01 4.42478150e-01
-1.43076360e-01 -7.49439076e-02 2.37966791e-01 1.01921871e-01
7.47765064e-01 -1.18736215e-01 6.08181655e-02 7.02977479e-01
-1.05581319e+00 8.23172867e-01 2.53446009e-02 7.16108143e-01
-1.93572387e-01 -4.98663671e-02 4.93656784e-01 -1.48543529e-02
6.20438278e-01 4.79317635e-01 -7.91311920e-01 -2.66205192e-01
-7.99702466e-01 -1.82850927e-01 7.71511614e-01 1.10789752e+00
5.15167713e-01 3.33701402e-01 3.84432171e-03 1.08333850e+00
3.30220610e-01 7.33715773e-01 1.03419375e+00 -5.74230313e-01
4.18614298e-01 3.70926082e-01 -2.67359108e-01 -4.24717277e-01
-1.57878190e-01 -2.00537905e-01 -7.67413497e-01 -3.76167357e-01
4.18509124e-03 -7.63357699e-01 -1.51310468e+00 1.38978839e+00
5.04156873e-02 1.58368573e-01 2.09434763e-01 5.79107285e-01
3.29754680e-01 1.11345553e+00 7.07708299e-01 -3.16514149e-02
1.57137728e+00 -1.03974724e+00 -8.52198184e-01 -7.13006675e-01
6.29300117e-01 -1.10317230e+00 8.49039495e-01 3.06909442e-01
-7.66866148e-01 -6.45312786e-01 -1.02369487e+00 -1.36469498e-01
-3.85062039e-01 3.06325287e-01 5.44303477e-01 7.86965251e-01
-5.89998722e-01 5.02160132e-01 -1.10082424e+00 -5.82477152e-02
4.49060708e-01 2.86796004e-01 -7.01375380e-02 -1.13717400e-01
-1.46735072e+00 6.53922617e-01 9.99660194e-01 3.92126024e-01
-5.48931479e-01 -3.42298120e-01 -7.59286821e-01 -5.65780289e-02
2.42539927e-01 4.56820354e-02 1.33697021e+00 -1.27258658e+00
-1.39694655e+00 7.02823937e-01 3.05364430e-01 -2.71175951e-01
1.64021343e-01 -7.09307909e-01 -7.89556503e-01 -1.61053509e-01
1.83117643e-01 5.28005958e-01 3.24568123e-01 -8.80339026e-01
-9.16222155e-01 -3.17247212e-01 -6.44537926e-01 9.09456238e-02
-2.57523060e-01 5.07444978e-01 -8.15701962e-01 -9.57171857e-01
-7.80525506e-02 -7.90307701e-01 -7.57519245e-01 -8.17433119e-01
-2.70310879e-01 -2.81166434e-01 7.24198759e-01 -1.38392961e+00
1.36807668e+00 -2.22629189e+00 -6.49479032e-02 2.86966741e-01
-4.70707059e-01 7.55372107e-01 -1.94212273e-01 1.19597785e-01
2.01396599e-01 5.92653513e-01 -7.94362187e-01 -3.55985701e-01
-2.69366026e-01 6.90122962e-01 -1.22090571e-01 1.81815669e-01
3.94331098e-01 8.49293530e-01 -9.31903183e-01 -8.67104411e-01
3.52753326e-02 3.51236433e-01 -1.72064781e-01 2.16759995e-01
-3.54649812e-01 4.69439298e-01 -9.28651035e-01 9.10257518e-01
5.69524825e-01 1.31200850e-01 2.63992697e-01 4.50997859e-01
-1.01715542e-01 1.15578119e-02 -7.71445096e-01 1.77814341e+00
-2.48773605e-01 2.27012813e-01 -2.51004100e-01 -1.12436152e+00
9.61579680e-01 3.65632594e-01 2.55026609e-01 -9.12501156e-01
5.64234614e-01 4.75037456e-01 2.12934792e-01 -6.90448582e-01
6.97844446e-01 -4.54867393e-01 -5.23474276e-01 1.88350841e-01
1.29107311e-02 2.22818270e-01 5.09966984e-02 -8.94507766e-02
9.64251459e-01 4.50906664e-01 3.26926708e-01 -2.00764611e-01
7.87539601e-01 1.45995140e-01 9.83454287e-01 2.07113385e-01
-2.37287134e-01 7.46295929e-01 6.75642043e-02 -1.70783356e-01
-1.26788211e+00 -6.32570982e-01 2.90063079e-02 1.15796375e+00
3.25641222e-02 4.42674896e-03 -1.01143265e+00 -8.29537272e-01
-4.89908278e-01 7.16742277e-01 -4.58918840e-01 4.15978543e-02
-1.22794974e+00 -1.12758493e+00 1.08687854e+00 9.99683499e-01
7.86554873e-01 -1.68383849e+00 -9.07153264e-02 5.73700070e-01
-1.84302226e-01 -1.22351420e+00 -6.82573318e-01 2.66948521e-01
-7.03169107e-01 -7.53680944e-01 -1.17751801e+00 -1.48956466e+00
5.57067633e-01 -2.38664150e-01 7.75201619e-01 1.49786726e-01
-9.62400660e-02 -5.10828905e-02 -9.46654260e-01 -4.28528726e-01
-4.77471441e-01 3.73314381e-01 -3.53571087e-01 -2.41979033e-01
5.79032898e-01 -3.20134342e-01 -4.36266154e-01 1.41838774e-01
-9.32142675e-01 -1.91344485e-01 1.02814806e+00 6.68371499e-01
5.54857075e-01 -1.22466937e-01 5.00391126e-01 -1.23080397e+00
4.51061368e-01 -6.25057459e-01 -3.29282522e-01 4.80135769e-01
-2.88797528e-01 -4.70126793e-03 8.23880196e-01 -5.60671031e-01
-1.63189101e+00 9.32383835e-02 -7.94656098e-01 1.70417890e-01
-1.32405445e-01 8.03413987e-01 -5.19708514e-01 3.06192219e-01
2.77049810e-01 4.12222713e-01 -4.51481491e-01 -8.21263254e-01
5.25979638e-01 1.17357659e+00 8.49416256e-01 -7.06192076e-01
5.95320582e-01 -5.71813360e-02 -5.36087334e-01 -8.35677028e-01
-8.26772988e-01 -5.50919712e-01 -9.25647020e-01 3.20993930e-01
1.49137580e+00 -9.46217597e-01 -1.58842295e-01 7.52010882e-01
-1.10201442e+00 -3.92964810e-01 2.37807721e-01 3.98566961e-01
7.04532266e-02 8.46350372e-01 -8.35464060e-01 -8.14138234e-01
-7.10554183e-01 -8.08263659e-01 7.02262223e-01 7.70931780e-01
1.58248693e-01 -1.14076960e+00 3.13286364e-01 1.51787892e-01
7.86751062e-02 5.95184080e-02 1.03261554e+00 -1.11567307e+00
-1.89228684e-01 -2.57017642e-01 5.29891402e-02 9.24017370e-01
9.04110447e-02 -5.96871488e-02 -5.66802144e-01 7.94826597e-02
-1.66086368e-02 -7.71256387e-02 7.18268454e-01 1.01400934e-01
9.46111858e-01 -1.02378801e-02 -2.93728977e-01 5.12198985e-01
1.34009528e+00 9.46405351e-01 8.10075045e-01 4.64754969e-01
8.09274137e-01 4.17414427e-01 7.26792753e-01 3.04436624e-01
2.09737971e-01 -1.28891110e-01 -2.51801610e-01 -2.59235105e-03
1.30855352e-01 -2.81387717e-01 3.19392681e-01 1.16630638e+00
-2.09712967e-01 -3.42727095e-01 -1.22694600e+00 8.14782083e-01
-1.60100889e+00 -5.77161074e-01 7.26037426e-03 1.68361008e+00
9.82280016e-01 1.02363110e-01 -1.40593410e-01 -9.80997011e-02
9.71428812e-01 7.30179474e-02 -1.88737616e-01 -2.41977274e-01
-1.89940616e-01 7.16751456e-01 6.89616859e-01 2.27139652e-01
-1.17969966e+00 1.62660384e+00 5.82668591e+00 1.03997588e+00
-1.07982481e+00 1.94182277e-01 6.88776553e-01 8.41076076e-01
-3.06119204e-01 2.20139593e-01 -8.44418645e-01 8.16981912e-01
1.02882159e+00 2.20689386e-01 5.49140312e-02 8.80383193e-01
-2.27621675e-01 1.69714674e-01 -5.63188553e-01 3.77476901e-01
1.48089491e-02 -9.13756728e-01 -1.74156874e-01 1.21018905e-02
8.25072646e-01 2.77202845e-01 -2.67304182e-01 5.62339902e-01
6.09820008e-01 -8.87136996e-01 5.86650372e-01 1.89767182e-01
4.70745414e-01 -7.98739076e-01 1.05765212e+00 3.90525162e-01
-1.16799974e+00 9.84913856e-02 -4.49702978e-01 1.11242965e-01
5.84706604e-01 1.99724585e-01 -3.74498785e-01 6.76520705e-01
5.09359539e-01 6.77409112e-01 -6.07434571e-01 8.52827489e-01
-6.20517254e-01 1.16385710e+00 -1.89297557e-01 -2.87263542e-01
5.25820851e-01 -4.64154392e-01 2.06675082e-02 1.40553880e+00
2.67229021e-01 6.66089892e-01 2.44903445e-01 3.85997891e-01
4.01582569e-02 2.84027994e-01 1.46409333e-01 -5.51619232e-01
4.92119014e-01 1.05633867e+00 -1.08716166e+00 -4.37286854e-01
-6.04995251e-01 9.87593234e-01 1.12480953e-01 2.62726873e-01
-8.03558707e-01 -6.58890963e-01 2.23247677e-01 -3.68854851e-01
4.17076051e-01 -4.63221520e-01 -5.72074831e-01 -1.20408785e+00
-3.09455335e-01 -7.75589645e-01 3.40896755e-01 -3.43869269e-01
-1.33131170e+00 6.39348507e-01 -2.41492808e-01 -8.12626183e-01
1.38813138e-01 -6.21824861e-01 -9.85141218e-01 9.26615834e-01
-1.54423833e+00 -1.20818198e+00 3.10837656e-01 2.76336908e-01
7.30376780e-01 -3.78393352e-01 6.56620681e-01 6.05964363e-01
-7.63596952e-01 3.56367201e-01 3.76636475e-01 9.52565074e-01
5.97499549e-01 -1.25307393e+00 9.19984102e-01 1.24846900e+00
1.14551559e-01 4.89444494e-01 3.26930016e-01 -1.02654982e+00
-9.02989268e-01 -1.06109571e+00 8.90063226e-01 -7.09091723e-02
6.23259485e-01 -4.22683656e-01 -1.10994136e+00 1.02372468e+00
3.01475495e-01 -3.54409903e-01 9.37406242e-01 -8.38548914e-02
1.56209916e-01 2.79639065e-01 -6.96040750e-01 5.98718762e-01
6.28512800e-01 -1.43196523e-01 -1.06974185e+00 1.72320336e-01
9.07547057e-01 -1.68247774e-01 -8.37003052e-01 2.15070829e-01
4.95722622e-01 -2.52113312e-01 7.11046100e-01 -4.63452548e-01
3.54727805e-01 -1.25203267e-01 -8.52461159e-02 -1.04264522e+00
-3.25939268e-01 -6.36646807e-01 6.10153556e-01 1.67434263e+00
5.49786925e-01 -3.18081856e-01 8.66157770e-01 4.83927697e-01
-6.78183258e-01 -4.29031312e-01 -6.41059220e-01 -5.11925340e-01
5.32507479e-01 -2.36587316e-01 4.59165365e-01 8.11852157e-01
-1.07842512e-01 5.28964639e-01 -5.45504868e-01 3.64959054e-02
2.18285829e-01 -2.97785968e-01 1.38790488e-01 -9.42148268e-01
-1.92893237e-01 -1.30080819e-01 -4.38232608e-02 -1.28420949e+00
2.48426214e-01 -5.89066029e-01 7.26413429e-01 -1.36498308e+00
-1.85726270e-01 -5.70321679e-01 -4.32303488e-01 6.20887280e-01
-6.60206974e-01 2.64273100e-02 5.78408353e-02 2.69156903e-01
-4.39227313e-01 6.39248073e-01 1.16441619e+00 1.24208033e-01
-2.38500491e-01 5.30683585e-02 -6.45707130e-01 7.01952338e-01
8.02202523e-01 -5.93368292e-01 1.11340716e-01 -6.50725484e-01
-3.96810956e-02 9.96629242e-03 -4.30170923e-01 -7.67723143e-01
1.39035016e-01 -2.15216160e-01 5.18484712e-01 -6.00499034e-01
-1.06497414e-01 -7.10413039e-01 -2.53844917e-01 5.77411771e-01
2.91932654e-02 5.27516045e-02 1.50037661e-01 3.84491980e-01
-2.70974696e-01 -4.99519914e-01 7.21232593e-01 -5.23886323e-01
-1.30422533e+00 1.97280377e-01 -5.94186306e-01 3.46539140e-01
9.51137364e-01 -8.72885957e-02 -1.52387291e-01 2.53646016e-01
-6.18875921e-01 6.03617907e-01 1.93863772e-02 3.83538574e-01
2.70161450e-01 -1.00913954e+00 -6.77898645e-01 1.78397134e-01
-2.55383492e-01 2.65648097e-01 2.47496814e-01 2.77793735e-01
-1.10436380e+00 1.41549945e-01 -3.34192693e-01 -1.36830807e-01
-8.82224917e-01 4.89212602e-01 -9.95816216e-02 -4.42514777e-01
-4.72332120e-01 9.51826632e-01 4.41825688e-02 -5.86091459e-01
6.44885898e-02 -1.45131811e-01 -4.99090999e-01 -1.99340358e-01
3.38062227e-01 2.70159006e-01 -3.79208833e-01 -6.48893952e-01
-3.09886813e-01 7.05787361e-01 -3.70753497e-01 -2.36027747e-01
1.33898032e+00 -1.83871344e-01 -1.17727824e-01 1.53997034e-01
9.25962508e-01 5.81367649e-02 -9.45595622e-01 -2.63971925e-01
4.39804435e-01 3.74894328e-02 -2.42618039e-01 -9.33598042e-01
-9.53583598e-01 1.12476707e+00 2.36451507e-01 -5.51606752e-02
1.05118382e+00 -1.23394758e-01 1.45084059e+00 2.89321423e-01
1.11829929e-01 -1.49154758e+00 -1.03529014e-01 7.88466036e-01
1.32083505e-01 -1.17373240e+00 -1.49826422e-01 -2.67447829e-01
-1.05417454e+00 9.83429253e-01 5.93412578e-01 -3.49112362e-01
8.72488379e-01 3.86252612e-01 5.49032748e-01 3.29955637e-01
-9.25709866e-03 -1.85571700e-01 1.34834111e-01 5.79373300e-01
5.98839104e-01 -1.74486581e-02 -7.65881896e-01 1.21524894e+00
-2.79772878e-01 -1.21428698e-01 2.60933667e-01 1.18021572e+00
-7.82538593e-01 -1.64159942e+00 -2.14539617e-01 6.48319870e-02
-1.17977560e+00 -4.89287823e-01 -2.93262035e-01 8.62876236e-01
3.53310168e-01 8.39218497e-01 5.73085099e-02 -3.92203957e-01
1.12049550e-01 2.05766901e-01 8.25812444e-02 -8.65733266e-01
-7.37924516e-01 5.85736156e-01 -1.24391615e-01 2.00681120e-01
-4.38820779e-01 -5.48538506e-01 -1.72450316e+00 1.16655923e-01
-5.53428769e-01 5.78577757e-01 6.88664556e-01 1.33270884e+00
-1.11545913e-01 5.05118489e-01 4.65770811e-01 -5.31676471e-01
-4.91084188e-01 -1.32630336e+00 -6.58981264e-01 7.10384846e-02
-2.59548336e-01 7.87246898e-02 -1.55362248e-01 4.17596728e-01] | [9.984850883483887, 10.137078285217285] |
fd1c1146-7200-48fa-aedc-3772aa49d2f7 | effect-of-lossy-compression-algorithms-on | 2302.12593 | null | https://arxiv.org/abs/2302.12593v1 | https://arxiv.org/pdf/2302.12593v1.pdf | Effect of Lossy Compression Algorithms on Face Image Quality and Recognition | Lossy face image compression can degrade the image quality and the utility for the purpose of face recognition. This work investigates the effect of lossy image compression on a state-of-the-art face recognition model, and on multiple face image quality assessment models. The analysis is conducted over a range of specific image target sizes. Four compression types are considered, namely JPEG, JPEG 2000, downscaled PNG, and notably the new JPEG XL format. Frontal color images from the ColorFERET database were used in a Region Of Interest (ROI) variant and a portrait variant. We primarily conclude that JPEG XL allows for superior mean and worst case face recognition performance especially at lower target sizes, below approximately 5kB for the ROI variant, while there appears to be no critical advantage among the compression types at higher target sizes. Quality assessments from modern models correlate well overall with the compression effect on face recognition performance. | ['Christoph Busch', 'Juan Tapia', 'Christian Rathgeb', 'Sebastian Schachner', 'Torsten Schlett'] | 2023-02-24 | null | null | null | null | ['face-image-quality', 'image-quality-assessment', 'face-image-quality-assessment'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 5.12064159e-01 -3.30808878e-01 -8.95958766e-02 -4.58255082e-01
-7.46746004e-01 -1.14459746e-01 6.61120296e-01 -1.91513434e-01
-5.03385246e-01 5.26644111e-01 1.57146171e-01 -1.50798738e-01
-3.68889153e-01 -6.32141352e-01 -3.89129072e-01 -6.14853740e-01
-2.35236526e-01 1.02934122e-01 -1.87031478e-01 -1.67889092e-02
4.51625556e-01 9.91079807e-01 -1.97572351e+00 4.71357733e-01
2.94533730e-01 1.20215571e+00 -2.49005169e-01 7.55264997e-01
3.38263288e-02 5.19036353e-01 -8.70764732e-01 -8.67691636e-01
4.79604185e-01 -2.62035698e-01 -5.79787552e-01 1.73492253e-01
1.05906880e+00 -8.44557464e-01 -6.35610580e-01 8.78482223e-01
6.68174028e-01 -1.77454755e-01 4.88523573e-01 -1.32480597e+00
-3.52289796e-01 -1.46948528e-02 -6.82421565e-01 5.41887164e-01
4.66238439e-01 2.89212108e-01 4.18465197e-01 -7.61456788e-01
5.96455932e-01 1.70798719e+00 6.17201269e-01 4.42299932e-01
-1.21807098e+00 -9.32278037e-01 -5.61463177e-01 3.98657560e-01
-1.58286989e+00 -1.22701192e+00 2.86871016e-01 4.76797149e-02
1.18353128e+00 4.12843764e-01 5.31664252e-01 5.71533859e-01
4.04860914e-01 -1.06134437e-01 1.06975198e+00 -4.40369397e-01
1.50596008e-01 -6.80487379e-02 -3.28636408e-01 5.20982206e-01
4.95634884e-01 1.77506939e-01 -1.01588905e+00 -2.69495040e-01
5.92223763e-01 -3.65045995e-01 -1.71666592e-01 2.07098767e-01
-3.57604444e-01 6.26443923e-01 -9.41156968e-02 4.30794477e-01
-2.66044855e-01 2.74050713e-01 4.10769969e-01 6.63463056e-01
4.69091296e-01 -1.66239783e-01 -2.85928312e-04 -3.22324425e-01
-1.53719032e+00 2.37382054e-01 6.89940512e-01 6.80470884e-01
4.78459775e-01 3.94561201e-01 -1.02886342e-01 1.06904447e+00
5.00444055e-01 6.58260584e-01 3.25950980e-01 -1.29177177e+00
5.54995060e-01 1.16438739e-01 -3.18570763e-01 -1.23918986e+00
8.51610154e-02 -8.10878724e-02 -4.39770639e-01 6.52285993e-01
4.60491896e-01 3.57921839e-01 -8.24112117e-01 1.43921065e+00
-4.84997891e-02 -1.17572255e-01 7.82670379e-02 4.52525645e-01
6.45786762e-01 6.40416384e-01 1.78466752e-01 -4.84487712e-01
1.37640822e+00 -3.47166985e-01 -6.43758595e-01 -1.05461076e-01
1.36105642e-01 -1.13663268e+00 8.26594889e-01 5.41329563e-01
-1.39703274e+00 -6.46311522e-01 -1.24799883e+00 1.33653089e-01
-9.22634676e-02 2.93217842e-02 1.08960897e-01 1.58597839e+00
-1.52200532e+00 6.01300895e-01 -5.64205348e-01 -3.79642040e-01
6.76279664e-01 6.14235401e-01 -8.10862541e-01 -7.99350619e-01
-6.52599573e-01 8.30791593e-01 -2.70439144e-02 -3.34301174e-01
-7.53972888e-01 -7.70679355e-01 -6.22453094e-01 3.52188855e-01
-3.24442714e-01 -8.55536237e-02 8.53827417e-01 -1.09947526e+00
-1.09802485e+00 1.00991166e+00 -2.60646254e-01 -5.07204056e-01
3.39323908e-01 9.69499126e-02 -8.23144197e-01 8.53778303e-01
-3.39202225e-01 8.65064144e-01 1.27541006e+00 -1.10008311e+00
-2.88362294e-01 -7.81904697e-01 -2.53438622e-01 1.21924140e-01
-4.88421381e-01 4.59879726e-01 -4.60959435e-01 -5.35329044e-01
-2.53561735e-01 -5.84366143e-01 6.06192350e-01 4.53173906e-01
2.41935089e-01 1.56492159e-01 1.27469277e+00 -9.26073313e-01
1.05162764e+00 -2.26652980e+00 -5.44470847e-01 3.27077478e-01
-1.05984114e-01 3.36218745e-01 -4.24688458e-01 2.73915857e-01
-2.29996383e-01 3.47631663e-01 -8.26388150e-02 -1.73894629e-01
-3.21032286e-01 6.91115633e-02 2.97576994e-01 7.28311598e-01
1.56563997e-01 2.44593561e-01 -7.22909644e-02 -5.83897233e-01
2.06285521e-01 1.05884814e+00 -6.81120336e-01 -4.01068628e-02
2.89543778e-01 -2.38091305e-01 1.62020147e-01 9.11708772e-01
1.25589252e+00 2.11564764e-01 1.77552879e-01 -4.42299217e-01
9.46737528e-02 -1.64798886e-01 -9.21235919e-01 1.09802306e+00
-2.77667969e-01 7.81551540e-01 4.31267530e-01 -4.89848286e-01
9.14474666e-01 4.15720999e-01 4.31601197e-01 -1.03332722e+00
-8.07163790e-02 8.31792653e-02 1.60555035e-01 -2.83651471e-01
5.46225250e-01 -4.05703664e-01 4.90449905e-01 3.95027936e-01
1.83422640e-01 1.85777005e-02 2.40402475e-01 4.38914932e-02
1.05094647e+00 -4.89238828e-01 3.54794934e-02 -3.32989126e-01
5.52365124e-01 -3.77272785e-01 1.65392369e-01 3.51931185e-01
-4.45944458e-01 7.09343493e-01 4.90381807e-01 -1.26248211e-01
-1.14518964e+00 -9.24198627e-01 -3.88514489e-01 8.87349725e-01
-1.29442126e-01 -5.81453323e-01 -8.43212068e-01 -2.03000203e-01
1.29177362e-01 5.49276412e-01 -2.43395105e-01 -3.42789143e-01
-4.51532722e-01 -7.93807745e-01 9.57699418e-01 -1.58146396e-02
1.04252827e+00 -8.93813670e-01 -8.05921257e-01 -3.56858402e-01
1.91382408e-01 -1.05620575e+00 -4.60572362e-01 -5.10314941e-01
-9.80705440e-01 -9.79906380e-01 -6.79611802e-01 -4.00785059e-01
5.86000681e-01 2.35155761e-01 1.11108410e+00 3.73929590e-01
-7.52345502e-01 9.59534168e-01 -2.16596887e-01 4.54686247e-02
-6.49300575e-01 -5.61706364e-01 3.10375728e-02 -4.13045399e-02
3.60791355e-01 -3.53868872e-01 -7.96070576e-01 1.59020931e-01
-1.15403569e+00 -5.73684096e-01 4.52428341e-01 7.02493787e-01
1.51859298e-01 4.33768123e-01 3.27139437e-01 -5.51569283e-01
5.58777213e-01 -2.14107156e-01 -2.97126561e-01 2.27809802e-01
-9.42992806e-01 -1.70326293e-01 3.71452034e-01 1.85281992e-01
-1.33430922e+00 -3.11865389e-01 -2.18936816e-01 -3.40791374e-01
-1.33178771e-01 1.05215430e-01 -3.90617698e-01 -6.18618369e-01
4.82627302e-01 2.47180834e-01 7.01539099e-01 -3.82509470e-01
-1.30765423e-01 7.62077510e-01 5.88002741e-01 -2.96820223e-01
4.85226929e-01 4.14843917e-01 2.37682387e-02 -1.31625867e+00
6.15966499e-01 -7.30256736e-02 -1.97987467e-01 -4.79912013e-01
3.22651416e-01 -7.51152158e-01 -6.24828577e-01 5.87010741e-01
-7.27479935e-01 5.24052195e-02 -2.84985006e-02 3.13278049e-01
-4.32872385e-01 5.94935298e-01 -7.80567229e-01 -7.76842117e-01
-4.51183379e-01 -1.16973829e+00 9.91115510e-01 2.12593257e-01
3.99000980e-02 -5.01725554e-01 -4.38384771e-01 5.77232182e-01
6.45328104e-01 1.04064140e-02 1.12858939e+00 -1.18871681e-01
-3.32905859e-01 -2.81798542e-01 -5.13262749e-01 4.66454089e-01
2.24214315e-01 2.09548086e-01 -1.02980292e+00 -8.99129808e-01
4.30995114e-02 -3.44271451e-01 7.20493972e-01 2.16857910e-01
1.02240443e+00 -4.84655172e-01 6.97860494e-02 5.66681087e-01
1.75616848e+00 5.44312000e-01 1.37893641e+00 9.19836387e-02
-1.42199825e-02 5.67830443e-01 3.69611531e-01 6.04910374e-01
-2.38959730e-01 6.38300478e-01 3.18905234e-01 2.00318843e-01
-5.03051698e-01 1.74235366e-02 5.90597689e-01 3.04301232e-01
6.66887611e-02 -3.37514251e-01 -6.63576126e-01 1.39561117e-01
-6.52040243e-01 -1.20791829e+00 5.39256334e-01 2.49740887e+00
5.75959504e-01 -1.03652999e-01 1.81210250e-01 5.22558153e-01
7.47659147e-01 2.35182479e-01 -2.45096326e-01 -6.19856536e-01
-1.17574163e-01 4.57993835e-01 5.54773092e-01 3.99723083e-01
-5.99793494e-01 4.52813148e-01 7.41248989e+00 9.68696654e-01
-1.21840310e+00 -5.43311052e-02 1.12497008e+00 -5.43894649e-01
3.98789831e-02 -2.70643830e-01 -8.08227718e-01 6.07623696e-01
1.63210773e+00 -3.34347785e-01 4.98576850e-01 4.75736618e-01
7.91006833e-02 -3.99096191e-01 -7.80418277e-01 1.46279693e+00
5.41364074e-01 -1.12707770e+00 4.28674042e-01 4.40582097e-01
1.99959874e-01 -4.07830238e-01 4.14651155e-01 -2.45085210e-01
-5.32499015e-01 -1.38586819e+00 4.74354506e-01 3.15562695e-01
1.58611369e+00 -1.01391423e+00 5.68200946e-01 -4.15728629e-01
-1.01754236e+00 -1.01231650e-01 -5.41038871e-01 4.41802144e-01
-3.26282263e-01 2.57295072e-01 -4.83869880e-01 2.10327268e-01
7.97751009e-01 3.20617050e-01 -8.86745453e-01 8.41568828e-01
7.17599094e-01 5.72581589e-01 -5.38493156e-01 5.99455714e-01
-1.60995156e-01 8.16554949e-02 3.38652134e-01 1.28595972e+00
6.53497159e-01 1.79853126e-01 -6.61329746e-01 2.96984822e-01
-2.74243593e-01 8.56926106e-03 -6.29667342e-01 -1.70168698e-01
7.94083297e-01 8.65687847e-01 -7.34763503e-01 -2.22308144e-01
-5.06014168e-01 8.93494308e-01 -3.58838588e-01 8.99674147e-02
-4.04802591e-01 -2.70807534e-01 7.55444348e-01 3.57740998e-01
5.09118319e-01 -3.86025570e-02 -1.29322261e-01 -5.86813331e-01
-3.35571952e-02 -1.46104312e+00 3.99750918e-01 -5.54710209e-01
-8.06304872e-01 5.29046655e-01 3.31163973e-01 -9.17076647e-01
-1.91709399e-01 -6.11301243e-01 -3.32839191e-01 8.35038781e-01
-1.26752734e+00 -7.78704762e-01 -2.67830938e-01 6.30075693e-01
4.85018015e-01 -6.21860921e-01 8.39783847e-01 5.92693329e-01
-4.04637635e-01 1.20736015e+00 2.14202866e-01 -3.01189214e-01
5.79738319e-01 -5.09100974e-01 4.10378575e-02 6.91453576e-01
-1.82359010e-01 5.64060390e-01 6.88285828e-01 -5.12593389e-01
-1.76612151e+00 -8.16472232e-01 4.92642462e-01 2.09356934e-01
-4.00436312e-01 1.99772671e-01 -7.48301506e-01 8.43897611e-02
4.13647294e-01 -6.47114962e-02 7.75450587e-01 -5.02430916e-01
-6.43543541e-01 -6.52413964e-01 -2.13187838e+00 3.07966977e-01
6.84767306e-01 -7.45199859e-01 2.08883022e-04 2.82924455e-02
-6.75195977e-02 2.32992187e-01 -1.21721065e+00 2.89018303e-01
1.07574749e+00 -1.36111426e+00 1.18438935e+00 2.67424099e-02
5.19320786e-01 4.17677611e-02 -5.10264158e-01 -8.09043288e-01
-1.40267685e-01 -2.94460207e-01 1.65672123e-01 1.41201138e+00
1.14105597e-01 -5.79547882e-01 1.03110099e+00 8.53143156e-01
4.29456383e-01 -3.49250257e-01 -1.53423774e+00 -9.17575598e-01
-1.86231628e-01 -2.58611470e-01 6.07492864e-01 4.18962002e-01
-1.51421860e-01 -1.11004926e-01 -2.15954632e-01 -1.83826968e-01
8.57738614e-01 -3.79535735e-01 4.28968161e-01 -8.53675544e-01
-9.41233486e-02 -4.23176080e-01 -9.93664801e-01 -3.66127044e-01
-2.24761844e-01 -6.03760600e-01 -4.91626978e-01 -9.17807102e-01
2.32893616e-01 -1.27198711e-01 4.31998679e-03 3.36131811e-01
2.95245409e-01 8.34662020e-01 5.53482652e-01 3.27335119e-01
1.26895055e-01 3.00167412e-01 7.60891199e-01 -3.39136153e-01
3.82079899e-01 -3.92621636e-01 -6.23110771e-01 1.75970867e-01
7.97398269e-01 -3.47113580e-01 -5.42923987e-01 -3.57302636e-01
-4.33735788e-01 2.60446340e-01 3.01491410e-01 -1.51148021e+00
-2.11715177e-02 1.43541947e-01 9.00846601e-01 -2.44821131e-01
8.58263731e-01 -9.87488627e-01 5.43032825e-01 7.63630152e-01
-2.43946254e-01 4.11446810e-01 6.65528357e-01 3.64129215e-01
-1.94396809e-01 -2.63979435e-01 1.40659046e+00 -8.25046822e-02
-5.80893934e-01 1.00948088e-01 -3.24173599e-01 -2.22531334e-01
8.40499163e-01 -7.66845882e-01 -5.00687361e-01 -5.92519701e-01
-3.86020482e-01 -5.61410367e-01 7.20462143e-01 1.90855443e-01
1.08634448e+00 -1.19308865e+00 -8.86628509e-01 7.69197226e-01
-1.53403088e-01 -1.20709777e+00 3.93806428e-01 3.93663019e-01
-9.72657382e-01 4.24174011e-01 -8.72288883e-01 -4.21432465e-01
-2.13206720e+00 1.65150538e-01 3.62427592e-01 1.43245071e-01
-3.74760062e-01 6.78730488e-01 -1.51810616e-01 5.05204976e-01
2.29797110e-01 2.98735619e-01 -1.09289307e-02 -7.88466707e-02
1.00610304e+00 8.11083794e-01 3.53593975e-01 -1.04182255e+00
-3.73003155e-01 6.28086090e-01 -3.05168808e-01 -3.00014973e-01
1.03466833e+00 -1.99003607e-01 -2.40299284e-01 -1.90765560e-01
1.56695914e+00 -2.47275889e-01 -9.51162636e-01 4.27572340e-01
-2.96035439e-01 -1.19792008e+00 3.16385508e-01 -7.74764419e-01
-1.39567542e+00 7.91461706e-01 1.47454143e+00 1.18014850e-02
1.64149868e+00 -4.51337337e-01 5.18417478e-01 2.08698725e-03
4.69940424e-01 -8.58736515e-01 -1.60095006e-01 -6.91117346e-02
9.77989316e-01 -7.50660002e-01 5.45649588e-01 -3.77194047e-01
-2.41099507e-01 1.21475792e+00 3.94642234e-01 1.26664028e-01
6.54034436e-01 4.30325538e-01 -2.50988364e-01 -1.59986049e-01
-7.64173269e-01 4.11902010e-01 -1.32763624e-01 7.72417247e-01
5.31361938e-01 -3.32048774e-01 -3.60556573e-01 -1.55322105e-01
-1.65672198e-01 1.10096388e-01 3.59935135e-01 9.73168850e-01
-4.46364790e-01 -1.12353003e+00 -7.67536700e-01 9.06718791e-01
-8.33062768e-01 -1.30476043e-01 -3.98857772e-01 8.73214662e-01
2.16832653e-01 1.14475167e+00 3.02172482e-01 -2.95604676e-01
7.96935856e-02 2.57792741e-01 7.79058993e-01 -1.20061415e-03
-7.89420128e-01 -1.61340132e-01 1.68982416e-01 -6.32025540e-01
-4.51886803e-01 -6.90864563e-01 -6.04659021e-01 -1.07651436e+00
-5.76194562e-02 1.03928156e-01 8.39091778e-01 5.38144886e-01
4.63433355e-01 7.68940002e-02 5.03289759e-01 -6.75272107e-01
-2.91918546e-01 -9.06637847e-01 -8.52397203e-01 2.33479992e-01
2.41011783e-01 -4.32843685e-01 -4.07237828e-01 4.10140753e-01] | [13.062054634094238, 1.0140457153320312] |
ed441225-54b8-46d1-82ba-47234e3b0f67 | machine-made-media-monitoring-the | 2305.09820 | null | https://arxiv.org/abs/2305.09820v1 | https://arxiv.org/pdf/2305.09820v1.pdf | Machine-Made Media: Monitoring the Mobilization of Machine-Generated Articles on Misinformation and Mainstream News Websites | With the increasing popularity of generative large language models (LLMs) like ChatGPT, an increasing number of news websites have begun utilizing them to generate articles. However, not only can these language models produce factually inaccurate articles on reputable websites but disreputable news sites can utilize these LLMs to mass produce misinformation. To begin to understand this phenomenon, we present one of the first large-scale studies of the prevalence of synthetic articles within online news media. To do this, we train a DeBERTa-based synthetic news detector and classify over 12.91 million articles from 3,074 misinformation and mainstream news websites. We find that between January 1, 2022 and April 1, 2023, the relative number of synthetic news articles increased by 79.4% on mainstream websites while increasing by 342% on misinformation sites. Analyzing the impact of the release of ChatGPT using an interrupted-time-series, we show that while its release resulted in a marked increase in synthetic articles on small sites as well as misinformation news websites, there was not a corresponding increase on large mainstream news websites. Finally, using data from the social media platform Reddit, we find that social media users interacted more with synthetic articles in March 2023 relative to January 2022. | ['Zakir Durumeric', 'Hans W. A. Hanley'] | 2023-05-16 | null | null | null | null | ['misinformation'] | ['miscellaneous'] | [-4.67159957e-01 5.54789126e-01 -2.98002571e-01 1.86473131e-01
-1.36138701e+00 -9.13607299e-01 1.26428294e+00 1.94311872e-01
-2.96438247e-01 9.15278316e-01 6.92908704e-01 -4.50388461e-01
5.26193261e-01 -1.08720243e+00 -1.01933539e+00 -1.50509194e-01
1.51991442e-01 6.24981046e-01 3.67825270e-01 -3.18895340e-01
3.45191568e-01 -2.10752890e-01 -8.63413870e-01 3.64314377e-01
8.82814765e-01 3.57826322e-01 -1.12946471e-02 6.29565239e-01
-5.03955245e-01 1.12544322e+00 -1.02122176e+00 -6.58366144e-01
2.18882799e-01 -5.54395199e-01 -2.78983206e-01 -7.00994059e-02
2.71943182e-01 -3.55774105e-01 -6.49477482e-01 1.04168582e+00
2.43183866e-01 -1.45301580e-01 6.76031649e-01 -1.08183146e+00
-8.72218609e-01 1.44609272e+00 -8.17659080e-01 5.02590299e-01
3.92845660e-01 2.40480676e-01 1.07023430e+00 -8.47945631e-01
1.26157165e+00 1.32191908e+00 8.86704028e-01 -5.25104366e-02
-1.14094102e+00 -5.59571207e-01 -1.61704361e-01 -5.75474024e-01
-1.09987462e+00 -2.69213080e-01 3.71943325e-01 -6.75807536e-01
5.80594957e-01 3.35130274e-01 7.29711592e-01 1.50355399e+00
6.83851957e-01 6.55263186e-01 1.07028210e+00 -5.26436940e-02
1.29466891e-01 5.03685534e-01 1.31937727e-01 2.95880646e-01
6.35315061e-01 -3.73559594e-01 -3.58730763e-01 -8.71425390e-01
4.14016277e-01 -2.03794509e-01 2.23331451e-01 8.18849444e-01
-1.21041787e+00 1.29163885e+00 5.80974715e-03 4.56715137e-01
-5.36404431e-01 5.12368754e-02 4.38107461e-01 5.26645064e-01
1.28619790e+00 5.45805037e-01 -1.69998944e-01 -4.13603097e-01
-9.08287764e-01 7.04143524e-01 1.28177130e+00 8.22479963e-01
5.27160525e-01 1.27537400e-01 1.05478190e-01 9.20841694e-01
2.88362980e-01 8.37583303e-01 7.17493117e-01 -9.16923940e-01
7.44886994e-01 2.45722264e-01 4.02505547e-01 -1.45060027e+00
-1.48287460e-01 -6.58088386e-01 -4.63061959e-01 -5.13266861e-01
5.04619300e-01 -4.77456838e-01 -5.13707042e-01 1.34777129e+00
-1.90842509e-01 -2.23014101e-01 -2.98341572e-01 4.52547252e-01
4.27453101e-01 1.24587047e+00 -1.03089437e-01 -3.96533847e-01
9.69030559e-01 -6.99911296e-01 -5.57183325e-01 -3.65404010e-01
4.84064847e-01 -1.02954459e+00 8.28887463e-01 1.67734697e-01
-1.07386887e+00 -6.66934401e-02 -6.07560277e-01 2.47800410e-01
-3.19779992e-01 -6.67908847e-01 1.35709599e-01 2.62763858e-01
-9.36894774e-01 1.76726758e-01 -5.36533177e-01 -5.04782200e-01
3.40677388e-02 -6.86853349e-01 1.27452925e-01 8.42484385e-02
-1.53832293e+00 7.69554853e-01 5.17135449e-02 -6.08187497e-01
-7.95459628e-01 -6.78980112e-01 -2.59592563e-01 -1.67975828e-01
4.43204880e-01 -2.57406682e-01 1.54369199e+00 -8.78690183e-01
-9.03915524e-01 4.78374869e-01 6.82177320e-02 -7.42698848e-01
1.04444575e+00 -1.17120691e-01 -6.31939530e-01 -3.53474580e-02
7.25692332e-01 -1.31067306e-01 5.72943747e-01 -1.22126186e+00
-6.40151978e-01 -7.54628405e-02 -2.36138850e-01 -1.08214021e-01
-2.49125376e-01 3.42210889e-01 -1.31894365e-01 -9.05827999e-01
-1.94355264e-01 -9.85710323e-01 -1.61939234e-01 -6.41137004e-01
-8.74141932e-01 1.93665326e-02 5.64846039e-01 -1.03600800e+00
1.19071603e+00 -1.85866523e+00 -5.53336978e-01 1.83537737e-01
7.15225518e-01 -2.84822732e-01 1.20083824e-01 9.32818949e-01
6.27563357e-01 9.01322842e-01 1.11235440e-01 -1.50516093e-01
7.38182068e-02 -1.66247100e-01 -6.42842591e-01 4.01488781e-01
-3.51451874e-01 8.66370618e-01 -9.22788024e-01 -3.95082608e-02
-3.79458308e-01 1.86144859e-01 -5.79435647e-01 -4.23710763e-01
-5.57640076e-01 2.27052607e-02 -5.70568025e-01 3.49145114e-01
4.98015672e-01 -8.00084472e-01 3.58711556e-02 6.72854245e-01
-6.23321414e-01 7.09967911e-01 -4.50355470e-01 6.74984455e-01
-4.48708475e-01 1.08624458e+00 -4.39768657e-02 -9.75755304e-02
6.71600461e-01 2.23753378e-01 4.94600028e-01 -4.70328718e-01
1.81244478e-01 3.73076469e-01 -1.52345583e-01 -9.64402929e-02
9.52344000e-01 2.00778362e-04 -4.10283715e-01 1.10510409e+00
-2.61066496e-01 -1.91125013e-02 3.15637171e-01 8.31162810e-01
1.23752058e+00 -7.10190475e-01 5.65416254e-02 -1.56290114e-01
-3.67955536e-01 5.01543343e-01 1.72273293e-01 1.30005324e+00
2.25090861e-01 4.73747581e-01 7.83868253e-01 -3.24426413e-01
-1.49721074e+00 -9.67779934e-01 -1.34647498e-02 1.09292364e+00
-3.39542806e-01 -6.48960471e-01 -6.47187173e-01 -4.27116871e-01
2.74426818e-01 1.25621891e+00 -3.77807319e-01 2.57388204e-01
-3.73305976e-01 -1.22328818e+00 5.63290179e-01 -3.43844622e-01
6.18976951e-01 -7.49951124e-01 3.76987636e-01 5.01916111e-01
-6.36889935e-01 -1.09371662e+00 -6.11892045e-01 -5.89275420e-01
-4.14689094e-01 -7.47341573e-01 -8.63872468e-01 -2.02004001e-01
5.52859724e-01 3.72726858e-01 1.12441313e+00 -2.37488061e-01
1.44466653e-01 2.77959108e-01 -4.21146393e-01 -6.21662021e-01
-1.39616919e+00 4.65062469e-01 7.72608370e-02 -8.73544216e-02
1.17387742e-01 -4.23854589e-01 1.59511212e-02 6.97213262e-02
-8.80602360e-01 -1.75983906e-02 4.04876173e-01 3.65438372e-01
-1.33839399e-01 -7.94721097e-02 9.05408621e-01 -1.27002168e+00
1.15229666e+00 -1.44366825e+00 -5.33322871e-01 -2.10316315e-01
-8.27132881e-01 -2.99483985e-01 4.98223603e-01 -5.60849369e-01
-9.88924801e-01 -8.66568863e-01 2.97800172e-02 3.23949665e-01
4.14118499e-01 1.20087552e+00 8.83982420e-01 3.75608683e-01
1.18970537e+00 5.20417333e-01 2.37289984e-02 -5.01032472e-01
1.70534983e-01 9.04676557e-01 5.85647076e-02 7.73465680e-03
1.00674868e+00 2.99750328e-01 -9.46921110e-01 -1.18524098e+00
-7.89364994e-01 -1.99761823e-01 4.07682806e-01 -4.80327755e-01
4.36393291e-01 -1.18478239e+00 -2.69172132e-01 7.40227282e-01
-1.07550859e+00 -4.10335630e-01 3.21785808e-02 3.20257455e-01
-1.74605101e-01 2.05354929e-01 -1.28513002e+00 -7.82404602e-01
-1.87316999e-01 -5.84591627e-01 2.39580020e-01 -1.67737424e-01
-6.07719779e-01 -8.93336952e-01 3.51951331e-01 4.38495427e-01
9.87300754e-01 2.45168537e-01 7.80636609e-01 -1.26401019e+00
-5.67113936e-01 -6.76517129e-01 -1.82249799e-01 1.66355908e-01
2.36035213e-01 4.20209885e-01 -3.28010023e-01 -2.55432595e-02
1.75232843e-01 6.19536005e-02 4.90658849e-01 3.31322998e-01
5.37374355e-02 -1.23310065e+00 -3.48901182e-01 -3.00084442e-01
1.15051365e+00 1.02836631e-01 3.85309488e-01 5.70259631e-01
3.88789475e-01 2.03042358e-01 -2.82475613e-02 7.91410148e-01
5.28068841e-01 2.30038628e-01 -7.56683648e-02 5.20900249e-01
3.11596990e-02 -8.38648021e-01 6.47859991e-01 1.27264452e+00
-2.73852684e-02 -7.16229260e-01 -1.11811256e+00 4.36732769e-01
-1.55445099e+00 -1.29681134e+00 -3.89125347e-01 2.05467463e+00
9.72341239e-01 7.16476023e-01 3.70137602e-01 -6.07170284e-01
1.04249609e+00 5.24236619e-01 -3.88339430e-01 1.09863207e-01
-3.23739529e-01 -7.34322786e-01 8.92422855e-01 6.81522012e-01
-4.70105648e-01 5.51336586e-01 6.94624138e+00 6.44724190e-01
-1.07964134e+00 4.27357435e-01 1.00485229e+00 -1.55423045e-01
-8.53837371e-01 -5.02096340e-02 -8.64387453e-01 1.14848244e+00
1.60571468e+00 -1.22502267e+00 4.40451682e-01 9.94488955e-01
6.62218034e-01 -3.06431592e-01 -9.30811539e-02 5.32489240e-01
1.11343935e-01 -1.81228602e+00 1.65475860e-01 3.86881799e-01
1.19714189e+00 7.42825508e-01 -9.69025344e-02 5.56189358e-01
7.03340709e-01 -4.02105689e-01 1.00128508e+00 6.04712784e-01
3.04801971e-01 -5.93151808e-01 6.20653808e-01 9.74966526e-01
-3.34821761e-01 7.49874189e-02 -2.16064423e-01 -2.36340895e-01
6.27759933e-01 1.23033595e+00 -1.12194276e+00 4.70267422e-02
3.71936649e-01 5.81932425e-01 -5.38152397e-01 7.60066688e-01
1.44198965e-02 1.18538392e+00 -4.41729963e-01 -4.35846269e-01
3.13513964e-01 -1.64020672e-01 1.00758362e+00 1.03876770e+00
4.21767831e-01 -2.68436790e-01 -1.19051956e-01 9.75939155e-01
-7.58549929e-01 7.21617714e-02 -7.64960825e-01 -9.07994628e-01
6.58304989e-01 8.35640013e-01 -6.70941472e-01 -6.83886647e-01
-5.89876652e-01 5.48665583e-01 -7.28738010e-02 3.30396980e-01
-1.02481782e+00 -2.37439424e-01 1.78703398e-01 1.01640773e+00
-4.19095337e-01 -4.20298219e-01 -1.07062533e-01 -1.40094554e+00
-2.44752318e-01 -9.32079911e-01 -2.06462041e-01 -9.10331488e-01
-1.63608837e+00 7.10920215e-01 -3.45609449e-02 -7.67085016e-01
-6.24718547e-01 2.67339915e-01 -5.72224081e-01 7.07425416e-01
-6.00647688e-01 -5.97764373e-01 7.25317225e-02 -1.55932903e-01
6.96253538e-01 -2.58974791e-01 2.14124441e-01 3.54206145e-01
-1.19634777e-01 7.91561976e-02 7.43856311e-01 2.93392777e-01
7.39246786e-01 -8.42760444e-01 1.27403533e+00 5.39509952e-01
4.60330993e-02 7.27497399e-01 1.09510386e+00 -1.33061981e+00
-1.11403000e+00 -1.07972705e+00 1.31622314e+00 -6.76894844e-01
1.34060264e+00 -5.37774742e-01 -6.94257617e-01 1.25459838e+00
3.93918633e-01 -7.18160570e-01 4.44887936e-01 -1.03185862e-01
-2.78643757e-01 4.08072978e-01 -1.02022552e+00 8.48845601e-01
5.96071303e-01 -5.35802245e-01 -6.00549400e-01 1.02890682e+00
8.77958953e-01 -1.20439395e-01 -7.11084425e-01 -3.87298644e-01
4.94420618e-01 -7.11886346e-01 4.55658883e-01 -5.11746168e-01
8.03764999e-01 2.36914814e-01 1.21342123e-01 -1.41845334e+00
-4.06700790e-01 -1.09338892e+00 8.35906938e-02 1.28140044e+00
1.09142518e+00 -1.16815078e+00 6.27765834e-01 7.55362391e-01
3.27034034e-02 -1.14585422e-01 -6.94926679e-01 -7.56884933e-01
2.00991407e-01 -2.62783825e-01 -1.24720350e-01 1.10295820e+00
-9.83543098e-02 2.26653621e-01 -5.43964744e-01 -2.89931148e-01
6.07441485e-01 -2.95105904e-01 9.07525837e-01 -1.12199461e+00
-2.99535543e-01 -2.94254780e-01 3.10366061e-02 -9.07021284e-01
-4.09172475e-01 -7.66393960e-01 -9.47301313e-02 -1.60035205e+00
5.07045984e-01 -3.09405774e-01 4.39331383e-01 9.87543687e-02
1.85437068e-01 1.65559992e-01 4.64733839e-02 8.14912200e-01
-3.91255021e-01 1.29151657e-01 9.18473840e-01 -8.44614133e-02
-1.32861257e-01 1.62190706e-01 -9.58685935e-01 6.76416218e-01
7.41779029e-01 -7.77847826e-01 -1.25422953e-02 -4.28266935e-02
1.02035332e+00 2.04418078e-01 1.43133506e-01 -8.40928495e-01
-5.57716861e-02 -3.41433175e-02 3.50033566e-02 -4.62708622e-01
-6.15360364e-02 -1.24865048e-01 5.91641605e-01 4.24297124e-01
-6.05373740e-01 3.65248978e-01 1.05315577e-02 8.58994186e-01
-1.21156804e-01 -4.69833985e-02 5.70386112e-01 -5.00419855e-01
8.57837573e-02 -2.72332709e-02 -1.55074418e+00 5.91483116e-01
8.58313382e-01 3.04281384e-01 -7.93758750e-01 -1.15912437e+00
-5.66119730e-01 -2.08873123e-01 5.35605788e-01 6.09880805e-01
-2.78178416e-02 -9.95862365e-01 -1.10748982e+00 -2.06540018e-01
-2.48120099e-01 -5.74085295e-01 1.85595185e-01 8.37965608e-01
-7.16771185e-01 4.00199145e-01 1.28962904e-01 -2.14109812e-02
-3.59289825e-01 3.46236259e-01 -1.23383226e-02 -2.08781183e-01
-4.57280695e-01 4.63149667e-01 -1.48501381e-01 -2.22775072e-01
-3.34820122e-01 -2.63756990e-01 4.18077111e-01 3.87094229e-01
7.29727983e-01 5.89916766e-01 -2.48728618e-01 -6.19593859e-01
1.45285755e-01 -2.96556771e-01 -4.09228325e-01 -5.75106323e-01
1.25174546e+00 -2.84861684e-01 -1.80315658e-01 1.01062644e+00
1.25121951e+00 7.87014008e-01 -7.54804611e-01 -8.54420438e-02
1.49540558e-01 -3.33188474e-01 -1.80648535e-01 -7.59517550e-01
-8.09585333e-01 -1.66567326e-01 -5.85592091e-01 1.24270606e+00
-3.75113711e-02 3.55792999e-01 1.16040480e+00 4.39292565e-02
7.39412904e-01 -9.82103944e-01 -3.79841924e-02 6.03716969e-01
1.09595513e+00 -1.03803396e+00 3.40666026e-02 -2.38360211e-01
-6.96697056e-01 7.31030941e-01 1.25434682e-01 -2.91724831e-01
8.38812292e-01 1.52226120e-01 -2.12542806e-02 -2.19519287e-01
-9.70059812e-01 8.31320047e-01 -2.21234292e-01 -2.37841114e-01
3.81557345e-01 1.16761558e-01 -6.46957695e-01 6.93752527e-01
-2.64991999e-01 -1.49642125e-01 1.28520095e+00 7.79200375e-01
-8.03409696e-01 -4.83056813e-01 -3.22982579e-01 1.08733082e+00
-9.73886251e-01 -4.37752008e-01 -5.12921572e-01 8.61045361e-01
-6.57065511e-01 8.85583758e-01 1.64755642e-01 -3.98323774e-01
-1.62241951e-01 1.95389271e-01 -4.88099366e-01 -6.48607194e-01
-7.04195499e-01 1.28292471e-01 5.63639939e-01 -6.31252527e-02
1.66522861e-01 -7.26406038e-01 -8.54976177e-01 -1.00918043e+00
-1.64825127e-01 3.17283928e-01 8.52809787e-01 6.57936811e-01
5.58464706e-01 1.19762057e-02 5.64593792e-01 -4.93028760e-01
-7.17844725e-01 -1.28950226e+00 -6.94730997e-01 8.07268769e-02
1.90117210e-01 -1.21115454e-01 -1.18349814e+00 2.91829929e-02] | [8.351356506347656, 10.123703002929688] |
d6fe86db-b30d-4435-93aa-73e27e21a947 | alignsts-speech-to-singing-conversion-via | 2305.04476 | null | https://arxiv.org/abs/2305.04476v4 | https://arxiv.org/pdf/2305.04476v4.pdf | AlignSTS: Speech-to-Singing Conversion via Cross-Modal Alignment | The speech-to-singing (STS) voice conversion task aims to generate singing samples corresponding to speech recordings while facing a major challenge: the alignment between the target (singing) pitch contour and the source (speech) content is difficult to learn in a text-free situation. This paper proposes AlignSTS, an STS model based on explicit cross-modal alignment, which views speech variance such as pitch and content as different modalities. Inspired by the mechanism of how humans will sing the lyrics to the melody, AlignSTS: 1) adopts a novel rhythm adaptor to predict the target rhythm representation to bridge the modality gap between content and pitch, where the rhythm representation is computed in a simple yet effective way and is quantized into a discrete space; and 2) uses the predicted rhythm representation to re-align the content based on cross-attention and conducts a cross-modal fusion for re-synthesize. Extensive experiments show that AlignSTS achieves superior performance in terms of both objective and subjective metrics. Audio samples are available at https://alignsts.github.io. | ['Zhou Zhao', 'Jinglin Liu', 'Lichao Zhang', 'Rongjie Huang', 'RuiQi Li'] | 2023-05-08 | null | null | null | null | ['voice-conversion', 'voice-conversion'] | ['audio', 'speech'] | [ 2.9452914e-01 -8.7695979e-03 -1.5612048e-01 -3.0916380e-02
-1.2084107e+00 -6.6580737e-01 3.2133305e-01 -4.0986004e-01
2.0740791e-01 2.9443026e-01 8.8002300e-01 2.0797420e-01
2.1444412e-01 -3.5416022e-01 -4.5196438e-01 -6.8984640e-01
3.2571629e-01 3.6295041e-02 -1.5047891e-01 -4.3377748e-01
1.1439169e-02 -2.4705592e-01 -1.6889926e+00 5.0718355e-01
8.4602302e-01 9.7507423e-01 4.1748163e-01 9.0793616e-01
-8.6432829e-02 6.3845164e-01 -6.8443465e-01 -1.9694829e-01
3.5920766e-01 -1.1950810e+00 -5.8593529e-01 -1.0411687e-01
4.2476228e-01 1.4609149e-01 -1.5875603e-01 1.0970552e+00
8.2492411e-01 1.4259891e-01 5.2971160e-01 -1.0605558e+00
-6.9487810e-01 9.2401081e-01 -2.4249020e-01 1.8244728e-02
5.9413338e-01 4.3222634e-03 1.4885328e+00 -9.1212994e-01
3.4266004e-01 1.1863199e+00 6.2551039e-01 6.5388298e-01
-1.2807100e+00 -8.1860298e-01 -1.7591834e-01 2.5601560e-01
-1.3845879e+00 -8.6247802e-01 1.1910756e+00 -3.3606783e-01
5.2724260e-01 5.7106733e-01 7.7704936e-01 1.0646998e+00
-7.7433236e-02 5.9977794e-01 9.0839297e-01 -6.4367568e-01
-1.2939630e-01 -1.7194083e-01 -5.9672248e-01 1.7940700e-01
-7.7717066e-01 2.4935044e-01 -1.1925316e+00 1.6771461e-01
6.9235671e-01 -5.9209919e-01 -5.6769925e-01 1.0251528e-01
-1.5249848e+00 7.1544707e-01 2.1211845e-01 6.2668574e-01
-3.1551245e-01 -7.6146170e-02 4.1616622e-01 4.4445807e-01
1.7256707e-01 6.0437238e-01 -2.9573730e-01 -2.8426966e-01
-1.1027997e+00 9.5813826e-02 6.4881682e-01 5.6612986e-01
2.0940602e-01 5.3362483e-01 -4.8599854e-01 1.2307314e+00
2.6076674e-01 5.6539184e-01 9.6238321e-01 -9.5729238e-01
5.9654039e-01 -5.6785367e-02 -1.1400982e-01 -6.9643492e-01
2.9311130e-02 -4.6855807e-01 -6.8418539e-01 -2.8789928e-02
1.5139382e-01 -8.3634984e-03 -4.9097097e-01 2.1030672e+00
2.6370069e-01 5.1025522e-01 2.5814474e-01 1.0803480e+00
9.5161301e-01 1.1120567e+00 -9.6780248e-02 -5.8576596e-01
1.4165471e+00 -1.1306598e+00 -1.2168169e+00 3.7235390e-02
-1.7562751e-02 -1.1655747e+00 1.6135194e+00 4.2081526e-01
-1.4268779e+00 -1.0933880e+00 -1.1749772e+00 -2.8556371e-01
1.6150647e-01 3.6127162e-01 -1.5710978e-01 3.6643502e-01
-7.6719874e-01 6.3375098e-01 -2.7102473e-01 4.4264317e-02
-1.9025098e-01 -4.1066620e-02 -1.2497073e-01 7.2586542e-01
-1.4257749e+00 5.3862661e-01 4.2447999e-01 -2.5453962e-02
-7.1506965e-01 -7.4151748e-01 -8.4150624e-01 7.7433854e-02
1.5463860e-01 -6.7605990e-01 1.6811117e+00 -1.0466813e+00
-2.0920715e+00 5.2507961e-01 -3.2766247e-01 -2.1091017e-01
1.4612785e-01 -2.7518299e-01 -7.9221505e-01 1.9746953e-01
1.1917750e-01 5.3972644e-01 1.3402448e+00 -1.1519367e+00
-6.7568791e-01 -1.0553297e-01 -6.9901037e-01 8.3186179e-01
-4.0914917e-01 -8.5311912e-02 -3.9039201e-01 -1.2091682e+00
4.0497300e-01 -7.3711818e-01 4.2666662e-01 -5.6161427e-01
-4.2653662e-01 -2.0734099e-01 6.8987739e-01 -9.5252723e-01
1.7038021e+00 -2.2363083e+00 5.7277685e-01 -6.8910278e-02
-3.5861131e-02 1.4186330e-01 -2.9497644e-01 6.1959785e-01
-2.2808212e-01 -2.4284534e-01 -1.8752018e-01 -4.1105020e-01
8.2014918e-02 -2.2869949e-01 -8.8728297e-01 8.6741395e-02
-7.0529999e-03 7.7700585e-01 -8.3081365e-01 -5.0150710e-01
4.8555441e-02 5.5524677e-01 -5.1369423e-01 5.4865158e-01
-8.4001213e-02 8.7670392e-01 6.1470479e-02 4.2870998e-01
2.3729666e-01 2.2800545e-01 2.8836530e-01 -6.6545546e-01
-2.2185871e-01 8.8526773e-01 -1.0259662e+00 1.8844682e+00
-6.7015982e-01 5.1262790e-01 -9.7142845e-02 -5.7250035e-01
1.1790732e+00 8.0386561e-01 3.4033376e-01 -6.4870143e-01
7.6007962e-02 3.3267441e-01 2.1248589e-01 -4.5993423e-01
5.8485562e-01 -7.3819679e-01 -1.7787015e-01 2.3335490e-01
3.0158296e-01 -7.3308951e-01 -1.0966248e-01 -3.7034154e-01
5.1315385e-01 1.3729322e-01 4.0246370e-01 8.2010195e-02
5.7278836e-01 -4.4176614e-01 6.9335938e-01 1.8954691e-01
-6.5662645e-02 9.7720367e-01 2.0793867e-01 7.9837851e-02
-1.0356278e+00 -1.4478571e+00 1.0727537e-01 1.2104673e+00
-4.7246903e-02 -6.1162645e-01 -9.1395855e-01 -1.2131961e-01
-3.0516461e-01 7.6572990e-01 -4.1383672e-01 -1.9721276e-01
-5.6433696e-01 -2.7772048e-03 7.0338726e-01 3.6567780e-01
2.5653085e-01 -1.3979011e+00 -1.2001194e-01 3.2559505e-01
-8.0388248e-01 -8.9748126e-01 -1.2544849e+00 -5.7218544e-02
-4.7299594e-01 -5.3228241e-01 -7.3285145e-01 -1.0621642e+00
-2.6565183e-02 8.3359145e-02 8.3544743e-01 -1.7449275e-01
4.6760660e-02 1.6660120e-01 -4.9999592e-01 -4.3852973e-01
-6.5108329e-01 -3.4698446e-03 3.2017785e-01 2.3445295e-01
-2.1364778e-02 -8.9885730e-01 -5.2964038e-01 1.5046115e-01
-7.0911068e-01 3.0945116e-01 3.1729439e-01 8.4521073e-01
7.7728713e-01 2.2769801e-02 9.8060793e-01 -2.7213776e-01
9.3769670e-01 -2.0705090e-01 -2.5461844e-01 -1.2589378e-02
-2.9221728e-01 -2.8539419e-01 8.1305206e-01 -7.6438946e-01
-8.8858616e-01 -5.8018651e-02 -2.6341400e-01 -7.1167129e-01
1.2184306e-01 4.2578080e-01 -3.7367499e-01 5.3535634e-01
5.6837338e-01 4.1032127e-01 8.9730449e-02 -5.8844489e-01
5.5821270e-01 1.0165555e+00 1.0851363e+00 -5.3560340e-01
9.2730981e-01 -3.2436728e-02 -4.0497217e-01 -8.7207991e-01
-1.0110551e+00 -2.6898152e-01 -4.7877172e-01 -3.1691086e-01
9.3754101e-01 -9.9241549e-01 -6.5121955e-01 3.9293891e-01
-1.0633498e+00 -2.7241609e-01 -7.1347362e-01 7.9812747e-01
-9.9060243e-01 3.7837458e-01 -6.6240960e-01 -7.7127135e-01
-5.6214660e-01 -9.3794143e-01 1.0235298e+00 2.5989988e-01
-5.1960772e-01 -5.8533883e-01 5.1608789e-01 4.3649161e-01
2.7916563e-01 -1.2871197e-01 7.9849899e-01 -3.1489927e-01
-8.9336960e-03 1.8684277e-01 3.3429846e-01 4.6650249e-01
5.2387339e-01 -5.8442794e-02 -1.2976047e+00 -2.0030233e-01
2.4897507e-01 -2.5037661e-01 5.1244986e-01 3.3019751e-01
9.7348487e-01 -4.8889461e-01 4.2449778e-01 5.8732730e-01
7.6077902e-01 3.9027515e-01 5.7164162e-01 -1.8286909e-01
6.3849825e-01 6.8879879e-01 6.1858493e-01 2.3042414e-01
4.6952456e-01 1.0515761e+00 1.0021549e-02 -1.7249582e-02
-7.6833224e-01 -8.9073467e-01 7.3028153e-01 1.9564220e+00
1.2504357e-01 1.9001678e-02 -5.3966492e-01 5.9412998e-01
-1.5962760e+00 -1.3147945e+00 3.3222836e-01 2.3819437e+00
1.4466918e+00 -8.9812875e-02 3.8270667e-01 4.2547733e-01
9.0381676e-01 2.5159866e-01 -5.3116077e-01 -4.1106901e-01
-2.0212185e-01 3.0122629e-01 -3.3587828e-01 5.6858724e-01
-9.0922755e-01 1.0086994e+00 6.1063156e+00 1.1489869e+00
-1.4412556e+00 1.4368543e-01 2.7839935e-01 -2.2126636e-01
-3.6794040e-01 -7.4792139e-02 -4.5716041e-01 5.9026760e-01
1.0070760e+00 -3.6651701e-01 8.7646264e-01 4.9073434e-01
2.9662061e-01 5.3649360e-01 -1.0476701e+00 1.2187467e+00
2.3163864e-01 -1.1986963e+00 4.8394252e-02 -2.7484798e-01
6.5897655e-01 -3.3711439e-01 3.7238857e-01 4.5548782e-01
-3.1317186e-01 -1.0820323e+00 1.3689260e+00 6.5674591e-01
1.0794474e+00 -6.5382558e-01 2.3808894e-01 2.9892996e-01
-1.6626141e+00 -1.4730659e-01 1.4219010e-01 1.1722278e-01
2.5379649e-01 1.3810800e-02 -1.0696127e+00 5.0492549e-01
6.3681304e-01 5.8062923e-01 3.8188100e-02 7.8306031e-01
-3.0770332e-01 8.7079877e-01 2.6173044e-03 3.0068648e-01
-3.2537571e-01 -2.2600219e-01 8.6832380e-01 1.0237825e+00
6.9577163e-01 -1.1440474e-01 1.3615300e-01 9.3826753e-01
-9.9498658e-03 4.7197902e-01 -3.8583705e-01 -3.7087807e-01
8.9004588e-01 1.1492858e+00 -1.6370347e-01 -1.8481332e-01
-9.3151353e-02 8.8691539e-01 -4.1246139e-02 3.1434557e-01
-7.3779643e-01 -4.2485467e-01 6.1214894e-01 6.6052773e-03
4.3649980e-01 -2.4800221e-02 -2.4676847e-01 -9.5076573e-01
-5.0380304e-02 -1.1543118e+00 1.9404578e-01 -1.0658401e+00
-1.4195818e+00 9.8915738e-01 -2.3478605e-01 -1.9089305e+00
-5.4084337e-01 -6.9097862e-02 -8.3708823e-01 1.0563858e+00
-1.2763561e+00 -1.2878121e+00 6.9946148e-02 5.8630741e-01
8.3369952e-01 -3.2761824e-01 9.8043203e-01 2.3612587e-01
-2.6921365e-01 6.6779166e-01 -3.8028002e-01 -4.1671183e-02
9.2096585e-01 -1.2343827e+00 2.8132707e-01 6.1304241e-01
5.2581942e-01 4.0964869e-01 8.2967079e-01 -4.6955487e-01
-1.0632731e+00 -8.7972480e-01 1.1186469e+00 -1.9338879e-01
6.4731139e-01 -2.9538164e-01 -8.8299292e-01 2.0949736e-01
5.3775120e-01 -2.2317591e-01 1.0450952e+00 -9.0494938e-02
-4.5281598e-01 -2.7902710e-01 -6.3274533e-01 7.7359957e-01
7.9622418e-01 -1.1240289e+00 -9.9854040e-01 -9.1018103e-02
9.9253589e-01 -5.4303998e-01 -9.3584716e-01 3.5440901e-01
6.4821881e-01 -8.9805394e-01 7.8603989e-01 -3.5454375e-01
3.6472008e-01 -7.9802221e-01 -5.2941012e-01 -1.6170164e+00
-1.8589118e-01 -1.0398260e+00 -1.7320666e-01 1.4875960e+00
3.3665577e-01 -1.4463952e-01 6.3546926e-02 -3.3414239e-01
-4.0676188e-01 -4.7295448e-01 -1.0871879e+00 -6.6667318e-01
-2.1763969e-02 -4.6427679e-01 9.1888040e-01 1.0323776e+00
3.1943733e-01 8.3141553e-01 -7.3088205e-01 1.0042207e-01
1.8316713e-01 4.8794425e-01 6.7418349e-01 -1.0594479e+00
-5.2333575e-01 -5.8915973e-01 2.7616223e-02 -9.5560861e-01
1.4351864e-01 -1.0572027e+00 2.9596001e-01 -1.0627598e+00
-1.8803334e-01 3.3344943e-02 -4.1772881e-01 3.9382613e-01
-1.1692473e-01 3.8210881e-01 7.6388711e-01 3.5823533e-01
-2.9093838e-01 1.0947219e+00 1.4622928e+00 1.8350404e-02
-6.3423193e-01 9.5121548e-02 -7.1651989e-01 6.7074555e-01
9.6269912e-01 -2.5320080e-01 -4.5507482e-01 -5.7733558e-02
-7.3793516e-02 7.8501314e-01 1.8491704e-02 -1.1635904e+00
1.2163215e-01 -1.4075711e-02 9.2533715e-02 -6.7888165e-01
8.3834374e-01 -4.5745265e-01 1.9512998e-01 1.7969385e-01
-6.0883701e-01 1.7006563e-02 1.2465800e-01 2.2466157e-01
-6.5964818e-01 -6.1532654e-02 7.7652776e-01 1.5783437e-01
-1.9026992e-01 7.0554450e-02 -5.9232131e-02 6.1960384e-02
5.1533657e-01 -1.6314583e-01 -1.8370539e-01 -7.8456873e-01
-8.9378548e-01 -2.2055964e-01 4.6361633e-02 6.7572278e-01
4.9603018e-01 -1.8679683e+00 -8.4378922e-01 3.5495022e-01
7.5547419e-02 -3.9505190e-01 5.6928104e-01 7.5001240e-01
8.1351578e-02 1.6880651e-01 -1.6263270e-01 -4.7431701e-01
-1.2304473e+00 2.6330966e-01 3.0754358e-01 7.1970031e-02
-5.7104295e-01 8.1120563e-01 3.0108491e-01 -4.8159856e-01
2.6273501e-01 -2.0750506e-01 -4.1100791e-01 2.6943722e-01
6.0631937e-01 1.3332927e-01 -1.4567843e-01 -1.0860317e+00
5.1375080e-02 7.5138825e-01 2.8670678e-01 -6.2198770e-01
1.0617683e+00 -4.3896458e-01 2.7093200e-03 1.0204985e+00
1.0577580e+00 6.3030922e-01 -1.2267044e+00 -3.1784388e-01
-2.8742641e-01 -4.2252067e-01 3.8869474e-02 -8.3101922e-01
-8.7577242e-01 9.1706651e-01 4.9855664e-01 4.0126804e-01
1.4110732e+00 4.8563141e-02 1.1148500e+00 -7.7187300e-02
-1.7075548e-01 -1.2470958e+00 4.0585342e-01 6.1425781e-01
1.4389336e+00 -7.5603437e-01 -5.2072012e-01 -2.8953198e-01
-1.0815855e+00 1.1347041e+00 5.0996649e-01 9.2851054e-03
4.6378946e-01 -1.6783332e-02 4.7688913e-01 2.1587059e-01
-8.2485592e-01 -2.3191836e-01 6.6494167e-01 5.2067029e-01
7.1910495e-01 2.1363682e-01 -1.1049831e-01 9.5274258e-01
-1.0989811e+00 -4.1782978e-01 1.4201178e-01 6.0103521e-02
-3.7236497e-01 -1.2072194e+00 -6.3583761e-01 -5.0430186e-02
-3.4413725e-01 -2.7881551e-01 -4.0330040e-01 1.5155363e-01
2.9441428e-01 1.0807787e+00 1.3279676e-02 -7.6476252e-01
3.8881132e-01 2.9791486e-01 4.2371580e-01 -6.0570896e-01
-5.8270085e-01 8.9472961e-01 -1.3260794e-01 -3.1214574e-01
-2.4393484e-01 -7.0103979e-01 -1.2852767e+00 3.0804774e-01
-1.3862054e-01 4.0058857e-01 6.0043532e-01 6.2105548e-01
3.1326321e-01 8.9005601e-01 1.1537627e+00 -9.9830145e-01
-7.1096224e-01 -1.2087697e+00 -7.5612938e-01 4.0119377e-01
6.3096690e-01 -4.6595263e-01 -3.9800248e-01 3.3509848e-01] | [15.56523609161377, 5.963982105255127] |
20da8a9e-ac5c-4397-a13d-ad3e3e87723d | learning-implicit-fields-for-generative-shape | 1812.02822 | null | https://arxiv.org/abs/1812.02822v5 | https://arxiv.org/pdf/1812.02822v5.pdf | Learning Implicit Fields for Generative Shape Modeling | We advocate the use of implicit fields for learning generative models of shapes and introduce an implicit field decoder, called IM-NET, for shape generation, aimed at improving the visual quality of the generated shapes. An implicit field assigns a value to each point in 3D space, so that a shape can be extracted as an iso-surface. IM-NET is trained to perform this assignment by means of a binary classifier. Specifically, it takes a point coordinate, along with a feature vector encoding a shape, and outputs a value which indicates whether the point is outside the shape or not. By replacing conventional decoders by our implicit decoder for representation learning (via IM-AE) and shape generation (via IM-GAN), we demonstrate superior results for tasks such as generative shape modeling, interpolation, and single-view 3D reconstruction, particularly in terms of visual quality. Code and supplementary material are available at https://github.com/czq142857/implicit-decoder. | ['Hao Zhang', 'Zhiqin Chen'] | 2018-12-06 | learning-implicit-fields-for-generative-shape-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Chen_Learning_Implicit_Fields_for_Generative_Shape_Modeling_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Chen_Learning_Implicit_Fields_for_Generative_Shape_Modeling_CVPR_2019_paper.pdf | cvpr-2019-6 | ['single-view-3d-reconstruction', '3d-shape-representation'] | ['computer-vision', 'computer-vision'] | [ 3.51187825e-01 4.46506619e-01 -3.34780174e-03 -3.88017029e-01
-7.61417985e-01 -6.53287530e-01 7.28459597e-01 2.16345955e-02
3.64082873e-01 4.70282674e-01 3.41084540e-01 -2.46386632e-01
4.12483633e-01 -1.31389654e+00 -1.02542925e+00 -4.69545335e-01
2.85380751e-01 5.89636683e-01 -2.46259347e-01 2.11000577e-01
1.29699692e-01 6.91554904e-01 -1.14484298e+00 3.87246281e-01
7.43940949e-01 1.09863460e+00 1.43819809e-01 5.75083077e-01
-1.02343805e-01 3.38390321e-01 -3.20936680e-01 -4.97374505e-01
2.27302268e-01 -2.75180340e-01 -5.77291310e-01 1.63086236e-01
2.02236786e-01 -5.28505623e-01 -2.83958822e-01 7.75202692e-01
4.56986457e-01 -1.39611617e-01 1.26915693e+00 -1.06421530e+00
-9.38712537e-01 1.58830099e-02 -3.01746190e-01 -4.81244594e-01
3.13720942e-01 2.74746239e-01 9.47167277e-01 -1.34607911e+00
7.79594779e-01 1.14394569e+00 6.97756171e-01 6.26942039e-01
-1.53980851e+00 -4.70734090e-01 -2.80066013e-01 -4.87408251e-01
-1.22033644e+00 -6.02634370e-01 9.51773465e-01 -7.59354234e-01
7.47046113e-01 2.06330806e-01 7.91678071e-01 8.77080023e-01
1.07230045e-01 8.21060836e-01 5.99476576e-01 -2.69712090e-01
3.66024464e-01 -2.48406172e-01 -6.23084247e-01 6.84155464e-01
-8.88756663e-02 2.51983613e-01 -2.17537224e-01 -1.75225660e-01
1.41091800e+00 -6.25268519e-02 -1.13973446e-01 -4.72750157e-01
-1.13287485e+00 8.47071826e-01 7.90899277e-01 -1.93459958e-01
-5.16381681e-01 4.62616116e-01 -1.26120262e-02 -1.06629409e-01
7.68377244e-01 2.26403132e-01 -1.19459890e-01 -3.32075777e-03
-6.82625949e-01 4.23212260e-01 5.20846128e-01 1.10634482e+00
9.62179959e-01 2.83489525e-01 -3.76057655e-01 7.32322097e-01
5.99234939e-01 7.90705919e-01 -1.97612241e-01 -1.30733705e+00
3.58018845e-01 7.41276503e-01 2.43002642e-02 -7.50130057e-01
7.82455578e-02 -2.75212884e-01 -8.73732924e-01 6.74895942e-01
8.34378973e-02 -5.20567894e-02 -1.17788732e+00 1.61940527e+00
4.17951763e-01 2.49325067e-01 -1.79529667e-01 7.48625934e-01
1.16784549e+00 7.96753466e-01 -1.79067906e-02 2.95062065e-01
9.79249716e-01 -5.65448701e-01 -3.40017706e-01 -3.00476197e-02
4.38932121e-01 -7.83083022e-01 9.02710855e-01 8.62016752e-02
-1.37171376e+00 -4.71617997e-01 -7.64393210e-01 -3.79724711e-01
-2.08711356e-01 3.93118948e-01 3.87166321e-01 2.38467544e-01
-1.03228796e+00 5.32241344e-01 -9.09838974e-01 1.51944906e-01
7.95667827e-01 2.97715813e-01 -2.64488876e-01 7.31700584e-02
-6.66567802e-01 4.41237807e-01 -2.01383650e-01 -1.06463581e-01
-9.93684947e-01 -8.44899595e-01 -1.17176378e+00 2.46283412e-02
-2.77205855e-01 -1.13010311e+00 1.09807634e+00 -8.61040294e-01
-1.33190620e+00 1.07208097e+00 -4.38298285e-01 -5.27128354e-02
3.70701224e-01 9.26621184e-02 4.22440171e-02 -7.58902729e-03
9.17534083e-02 1.01896119e+00 1.16440272e+00 -1.65949297e+00
-1.02408879e-01 -2.89571106e-01 -4.13731448e-02 1.28668144e-01
3.06158960e-01 -3.82337958e-01 -3.55073065e-01 -9.22036707e-01
2.34252810e-01 -7.50659585e-01 -1.82685360e-01 4.74598765e-01
-7.43802309e-01 -4.85293707e-03 7.97020018e-01 -8.55490804e-01
8.99029434e-01 -2.30214572e+00 2.09292650e-01 4.02103245e-01
3.81268233e-01 9.90300402e-02 -1.32333577e-01 2.55959868e-01
3.96030024e-02 2.06423774e-01 -7.13016033e-01 -6.52208686e-01
-4.09675352e-02 1.36401936e-01 -4.34083164e-01 2.24634871e-01
5.16067147e-01 1.21694815e+00 -7.52514303e-01 -1.05760865e-01
3.51214051e-01 9.16161478e-01 -7.14806676e-01 3.89225185e-01
-6.26931727e-01 7.59351909e-01 -4.66795981e-01 6.47440910e-01
9.40083206e-01 -4.35613036e-01 -7.28552863e-02 -2.85691172e-01
-4.24481481e-02 4.56142545e-01 -9.59626019e-01 1.67242134e+00
-5.00770390e-01 4.86289293e-01 3.13156843e-02 -5.16730130e-01
1.15336907e+00 3.70489985e-01 3.98215711e-01 -5.26307285e-01
9.89444181e-03 1.74879283e-01 -4.45103854e-01 -2.57556625e-02
2.84571290e-01 1.74326710e-02 1.63053870e-01 4.56923753e-01
-1.22831743e-02 -5.66303730e-01 -3.45714301e-01 9.53411981e-02
8.32046986e-01 5.64779520e-01 3.92234512e-02 8.97702947e-02
2.63329774e-01 -3.79313260e-01 2.84145474e-01 1.39094755e-01
4.78150517e-01 1.03565168e+00 3.00547898e-01 -2.77040571e-01
-1.40483844e+00 -1.55929840e+00 -2.70935774e-01 3.79548579e-01
-6.46390319e-02 -4.90025818e-01 -6.03977799e-01 -5.01235366e-01
4.19500589e-01 8.77147257e-01 -7.15572536e-01 -6.48468360e-02
-5.42814136e-01 -4.93513420e-02 2.21088544e-01 6.33904278e-01
3.18674520e-02 -1.32215321e+00 -4.41394061e-01 1.24156944e-01
-8.61616340e-03 -6.99127018e-01 -6.51967287e-01 2.01074909e-02
-1.16344416e+00 -7.78782308e-01 -9.06952500e-01 -7.17589140e-01
1.16246819e+00 -1.25050068e-01 1.10497570e+00 1.54805914e-01
-1.85819983e-01 3.37299645e-01 -1.35097310e-01 -2.19360754e-01
-6.86580241e-01 -6.81044906e-02 -5.38500130e-01 1.63691714e-01
-7.28082731e-02 -7.75505602e-01 -8.32096815e-01 2.49077007e-02
-7.46495485e-01 5.37969589e-01 3.33663613e-01 7.48057902e-01
1.02607512e+00 -6.12435758e-01 4.51557457e-01 -7.20225275e-01
3.27783555e-01 -5.10522127e-01 -4.51359749e-01 -9.34149101e-02
-3.52788985e-01 1.96203172e-01 4.25128073e-01 -2.72620786e-02
-9.13252890e-01 2.57480383e-01 -5.57919621e-01 -5.37687540e-01
-2.41114244e-01 2.43721366e-01 -1.55332893e-01 5.33777401e-02
5.87906122e-01 4.23518836e-01 1.66549161e-01 -6.68839991e-01
5.81173778e-01 5.45058429e-01 4.51003075e-01 -5.80282211e-01
8.25144112e-01 5.15824914e-01 1.93932414e-01 -6.22442067e-01
-5.34306169e-01 3.86247486e-02 -7.51147449e-01 -3.44513386e-01
9.83242631e-01 -9.75826144e-01 -3.97447288e-01 4.35804456e-01
-1.33019412e+00 -6.78759217e-01 -5.34594059e-01 2.91554583e-03
-9.38881516e-01 -2.98870653e-02 -5.27731895e-01 -7.78477430e-01
-5.67292690e-01 -1.14668190e+00 1.48590255e+00 1.16792604e-01
-2.14287505e-01 -1.05370176e+00 1.72455120e-03 9.07652751e-02
3.50548387e-01 5.32536268e-01 1.25142813e+00 7.87023753e-02
-8.40294063e-01 -1.05655536e-01 -1.86860487e-01 3.43157172e-01
2.98992902e-01 1.33329034e-01 -1.15781271e+00 -1.18194886e-01
-4.71514136e-01 -1.73563570e-01 7.35941052e-01 6.93929970e-01
1.45722461e+00 -4.31403875e-01 -4.27883685e-01 1.02641213e+00
1.31100011e+00 2.69546121e-01 8.69586349e-01 -1.85762569e-01
8.69180262e-01 1.90506756e-01 1.34535376e-02 5.90227962e-01
4.51196283e-01 8.31636548e-01 6.33625329e-01 -2.78350770e-01
-5.64381003e-01 -8.01343262e-01 1.03347853e-01 6.46844983e-01
-2.94462919e-01 -1.05214931e-01 -5.93450427e-01 4.88058954e-01
-1.46782565e+00 -7.20095873e-01 -1.31728053e-01 2.31833315e+00
8.51268530e-01 -9.23119038e-02 7.55256265e-02 4.92122248e-02
5.27568698e-01 6.10687993e-02 -6.50465310e-01 -2.83746123e-01
-3.14025059e-02 3.96553874e-01 1.50888965e-01 7.41488755e-01
-8.14075947e-01 7.26048112e-01 5.77317762e+00 5.98840594e-01
-1.38928771e+00 2.53922865e-02 9.88068461e-01 2.40018621e-01
-1.01550114e+00 -1.98817942e-02 -4.49395627e-01 4.97290969e-01
4.82040942e-01 4.05140929e-02 4.93772209e-01 6.96430326e-01
6.98265508e-02 2.17440903e-01 -9.72851515e-01 8.93984437e-01
-2.30932713e-01 -1.71085382e+00 3.21600884e-01 2.09980786e-01
8.47235143e-01 -5.46609685e-02 1.86973527e-01 -2.83752438e-02
3.25373523e-02 -1.21910846e+00 9.40129340e-01 7.40360856e-01
1.47619629e+00 -5.82382143e-01 1.23655610e-01 3.21624666e-01
-1.25971293e+00 4.53613460e-01 -2.02354208e-01 2.57435322e-01
2.46865928e-01 6.74023747e-01 -9.38257575e-01 4.29950237e-01
2.44498223e-01 8.46546948e-01 -2.15382189e-01 1.01667488e+00
-4.66781229e-01 5.62671244e-01 -1.93585798e-01 4.38667893e-01
-1.36672035e-01 -4.14324880e-01 7.16757298e-01 9.92049992e-01
6.61140978e-01 3.11422255e-02 -5.77297099e-02 1.59585130e+00
-2.32437745e-01 -1.26281872e-01 -8.69805753e-01 4.42277687e-03
5.58120251e-01 1.13313103e+00 -5.66357613e-01 -2.83269495e-01
-1.80510759e-01 9.90885496e-01 3.01584929e-01 4.43158299e-01
-5.44432282e-01 -1.87795192e-01 6.36611938e-01 6.34477079e-01
5.14034629e-01 -3.33725005e-01 -9.58769143e-01 -9.23106909e-01
5.87023087e-02 -3.85469198e-01 -3.84139717e-02 -1.20057559e+00
-1.10771501e+00 6.42690063e-01 -2.92801946e-01 -1.45136857e+00
-2.95435399e-01 -5.31542361e-01 -7.38329530e-01 1.12932885e+00
-1.16538370e+00 -1.41006827e+00 -3.69865328e-01 3.62948090e-01
3.09378177e-01 -1.93083622e-02 1.04459929e+00 1.02286592e-01
3.88003662e-02 4.46110189e-01 -4.69693653e-02 2.07563177e-01
2.93371081e-01 -1.20637870e+00 9.66771483e-01 4.38747168e-01
1.56903952e-01 3.77415508e-01 3.01318109e-01 -8.96029532e-01
-1.40247846e+00 -1.36209989e+00 8.00039530e-01 -6.11473441e-01
-1.34936288e-01 -5.03288627e-01 -8.31850231e-01 8.07632148e-01
-1.04982965e-01 1.36839002e-01 3.63315433e-01 -2.98499316e-01
-3.04989189e-01 3.02114397e-01 -1.35246277e+00 4.74020660e-01
1.16604245e+00 -5.31606853e-01 -2.30988413e-01 1.77735776e-01
6.62596881e-01 -7.28648365e-01 -1.04607511e+00 4.11728472e-01
3.48768115e-01 -7.18378901e-01 1.22032666e+00 -3.69730443e-01
9.06711340e-01 -4.69251424e-01 -1.78442091e-01 -1.41697919e+00
-5.90748608e-01 -4.19556320e-01 -6.37095153e-01 9.03885484e-01
5.22669852e-01 -4.46146965e-01 8.18073511e-01 4.12395000e-01
-3.95922929e-01 -1.13833201e+00 -7.97428012e-01 -5.15529394e-01
1.88470066e-01 -4.52537626e-01 9.63545978e-01 7.84002304e-01
-5.40659785e-01 1.82205319e-01 -1.38460249e-01 1.74103901e-02
6.25443697e-01 3.19120228e-01 7.28526056e-01 -1.06962562e+00
-1.82171345e-01 -5.39686680e-01 -3.12300116e-01 -1.38926053e+00
3.68457958e-02 -1.34089792e+00 -1.40712604e-01 -2.00210381e+00
-1.30666882e-01 -7.88510919e-01 2.08780095e-01 6.77215815e-01
8.59500393e-02 5.27160406e-01 1.89314172e-01 1.74207598e-01
4.88822758e-02 8.61406863e-01 1.79632246e+00 -1.00577682e-01
-1.50789425e-01 1.60545751e-01 -5.44429839e-01 4.64803338e-01
6.04936182e-01 -2.52597570e-01 -1.80857912e-01 -7.16174483e-01
2.05369338e-01 4.23742324e-01 6.39773071e-01 -7.12204456e-01
-1.56062350e-01 -8.85951892e-02 8.40178549e-01 -6.37144506e-01
7.05188572e-01 -6.21345341e-01 5.71854413e-01 2.68820763e-01
-1.73869446e-01 -1.23142734e-01 1.55370593e-01 2.22422913e-01
-7.12385550e-02 -1.05437949e-01 6.67619765e-01 -2.03852709e-02
-3.04680943e-01 6.70171916e-01 1.03789397e-01 -3.39664817e-02
7.27154136e-01 -3.79264444e-01 -4.77189682e-02 -5.15363216e-01
-1.00639009e+00 -1.33180857e-01 7.80999184e-01 1.77677304e-01
1.02891707e+00 -1.86604869e+00 -8.34182501e-01 7.12177873e-01
9.31232348e-02 4.58047688e-01 1.78938597e-01 4.05993760e-01
-4.79079723e-01 6.23853467e-02 -6.11140691e-02 -6.86768472e-01
-7.81558037e-01 9.42548215e-02 3.66749913e-01 1.73877299e-01
-1.00199246e+00 7.62565911e-01 5.32967746e-01 -6.64377093e-01
-1.24411620e-01 -3.23274523e-01 1.44384339e-01 -3.83045077e-01
2.30903864e-01 1.83980718e-01 3.97212431e-02 -8.07632446e-01
-1.27961472e-01 7.41263688e-01 5.66352546e-01 -1.88008830e-01
1.38614738e+00 3.81208509e-01 -1.02759898e-02 2.12817237e-01
1.10460162e+00 1.97169542e-01 -1.74138618e+00 -6.33951575e-02
-5.82590699e-01 -5.63950360e-01 1.41538968e-02 -9.59333301e-01
-1.24440169e+00 8.28779697e-01 2.73006439e-01 1.83943976e-02
9.81377184e-01 1.10648409e-01 7.91640282e-01 -3.60793620e-01
3.11919034e-01 -4.28540677e-01 5.85900433e-03 6.42516613e-01
1.40588796e+00 -9.37151134e-01 -2.55805701e-01 -5.84909260e-01
-5.92126429e-01 1.15771782e+00 3.27034414e-01 -3.29748541e-01
9.35746133e-01 5.26361644e-01 -4.10210490e-02 -3.08773726e-01
-5.71344495e-01 9.88491774e-02 7.56177485e-01 7.53238797e-01
5.54013252e-01 4.21614468e-01 1.48587331e-01 4.35925484e-01
-3.21087152e-01 1.60725817e-01 1.04310744e-01 6.65565550e-01
-9.83763337e-02 -1.06581163e+00 -2.34931499e-01 6.39009535e-01
1.22729786e-01 1.27277775e-02 -3.08025807e-01 1.64731175e-01
2.11451590e-01 4.19250399e-01 4.81775343e-01 -3.34025383e-01
2.35901639e-01 -1.30406609e-02 6.93920195e-01 -7.21159160e-01
-2.37242118e-01 1.35589898e-01 -4.26716097e-02 -3.70070755e-01
-2.59278361e-02 -6.04794502e-01 -1.43885732e+00 -2.02071458e-01
8.79508406e-02 -2.42085487e-01 6.32683992e-01 5.85133791e-01
7.14350939e-01 4.16276753e-01 7.41924405e-01 -1.17354035e+00
-2.38972187e-01 -7.94278741e-01 -2.29608238e-01 4.95833069e-01
3.36168051e-01 -7.21104741e-01 -2.41228506e-01 3.97897035e-01] | [8.907564163208008, -3.611281633377075] |
3ea73718-3882-43e3-be60-8bbe1b692bfa | better-combine-them-together-integrating | null | null | https://aclanthology.org/2021.findings-acl.49 | https://aclanthology.org/2021.findings-acl.49.pdf | Better Combine Them Together! Integrating Syntactic Constituency and Dependency Representations for Semantic Role Labeling | null | ['Donghong Ji', 'Fei Li', 'Yafeng Ren', 'Shengqiong Wu', 'Hao Fei'] | null | null | null | null | findings-acl-2021-8 | ['semantic-role-labeling', 'semantic-role-labeling-predicted-predicates'] | ['natural-language-processing', 'natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.452225685119629, 3.6311428546905518] |
0d259593-cec8-4a20-b680-7d18308afaf4 | the-blessings-of-unlabeled-background-in | 2103.13183 | null | https://arxiv.org/abs/2103.13183v2 | https://arxiv.org/pdf/2103.13183v2.pdf | The Blessings of Unlabeled Background in Untrimmed Videos | Weakly-supervised Temporal Action Localization (WTAL) aims to detect the action segments with only video-level action labels in training. The key challenge is how to distinguish the action of interest segments from the background, which is unlabelled even on the video-level. While previous works treat the background as "curses", we consider it as "blessings". Specifically, we first use causal analysis to point out that the common localization errors are due to the unobserved confounder that resides ubiquitously in visual recognition. Then, we propose a Temporal Smoothing PCA-based (TS-PCA) deconfounder, which exploits the unlabelled background to model an observed substitute for the unobserved confounder, to remove the confounding effect. Note that the proposed deconfounder is model-agnostic and non-intrusive, and hence can be applied in any WTAL method without model re-designs. Through extensive experiments on four state-of-the-art WTAL methods, we show that the deconfounder can improve all of them on the public datasets: THUMOS-14 and ActivityNet-1.3. | ['Hanwang Zhang', 'Jianqiang Huang', 'Bing Deng', 'Zhenfang Chen', 'Jingyuan Chen', 'YuAn Liu'] | 2021-03-24 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Liu_The_Blessings_of_Unlabeled_Background_in_Untrimmed_Videos_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Liu_The_Blessings_of_Unlabeled_Background_in_Untrimmed_Videos_CVPR_2021_paper.pdf | cvpr-2021-1 | ['weakly-supervised-temporal-action'] | ['computer-vision'] | [ 6.42201126e-01 3.59342024e-02 -6.58348501e-01 1.78099558e-01
-8.07461262e-01 -3.06316733e-01 5.83712041e-01 -4.80364949e-01
-2.43570119e-01 5.51443934e-01 4.83904481e-01 -4.89911251e-02
1.14295110e-01 -1.31311208e-01 -1.01978588e+00 -1.07742858e+00
-1.16810143e-01 -1.31217510e-01 5.55025339e-01 4.94981587e-01
-1.44157782e-01 -3.56849246e-02 -1.26425827e+00 4.86411542e-01
6.70007944e-01 7.78588951e-01 -1.39515558e-02 5.16417801e-01
3.15221399e-01 1.71561682e+00 -5.16452610e-01 -2.68261820e-01
2.80575663e-01 -6.48238420e-01 -5.62254608e-01 4.55270350e-01
6.71233892e-01 -3.92896324e-01 -6.98238254e-01 1.04893029e+00
1.98263675e-01 1.12577312e-01 5.78284740e-01 -1.65242529e+00
-3.97262305e-01 4.16834533e-01 -9.70408499e-01 3.70581448e-01
1.52165681e-01 3.43213230e-01 7.60537505e-01 -8.13762844e-01
6.16867185e-01 1.44747162e+00 6.56668901e-01 6.57728255e-01
-1.44555748e+00 -6.29457951e-01 6.55440569e-01 5.32776654e-01
-1.16049743e+00 -5.41951835e-01 8.25804174e-01 -7.27045834e-01
5.07742345e-01 2.88101673e-01 4.18323040e-01 1.95319295e+00
2.39524513e-01 1.44994330e+00 1.16299820e+00 -2.77002186e-01
4.04138356e-01 -4.15645927e-01 6.14454858e-02 7.96630442e-01
6.76002726e-02 1.49024367e-01 -8.12301576e-01 -1.99775964e-01
8.47061098e-01 2.46329889e-01 -3.66265476e-01 -6.70234561e-01
-1.41742742e+00 4.01286155e-01 6.75483271e-02 1.20586373e-01
-4.54845458e-01 5.35960555e-01 2.78290004e-01 -1.82452634e-01
6.06239021e-01 -1.97048530e-01 -5.85681736e-01 -3.89933288e-01
-8.27151120e-01 -6.40780553e-02 4.65530425e-01 8.38063359e-01
4.87454742e-01 2.93848459e-02 -5.77377498e-01 6.62780344e-01
2.07321540e-01 3.95173997e-01 4.87483591e-01 -1.11568809e+00
3.61848176e-01 5.04735708e-01 1.88357115e-01 -8.08882892e-01
2.66816299e-02 -4.18287933e-01 -7.74593592e-01 1.68056011e-01
5.84021270e-01 -1.72046393e-01 -1.17710996e+00 1.85801387e+00
4.65249628e-01 1.15690410e+00 -2.51424402e-01 6.99164629e-01
4.93108153e-01 2.11697638e-01 4.65072572e-01 -2.74509907e-01
1.35017490e+00 -1.42078757e+00 -9.02585030e-01 -5.60603619e-01
6.39965594e-01 -3.12161714e-01 1.00242102e+00 4.76981282e-01
-5.98006546e-01 -5.38251460e-01 -7.94798732e-01 1.49340168e-01
1.15968868e-01 3.61530513e-01 5.85009575e-01 4.20933098e-01
-5.24580479e-01 5.88797688e-01 -1.41200042e+00 -3.24805588e-01
7.93273509e-01 -1.54150277e-01 -4.45772201e-01 -9.31073949e-02
-9.40303206e-01 6.58874750e-01 2.12743983e-01 8.94743428e-02
-1.53749776e+00 -6.75315559e-01 -7.22837210e-01 -1.84272438e-01
1.11278188e+00 -5.37757397e-01 1.28976166e+00 -1.46798503e+00
-1.21230090e+00 5.50535381e-01 -7.34234989e-01 -4.79518294e-01
9.14299846e-01 -4.72707629e-01 -4.16460603e-01 1.78247288e-01
3.62859726e-01 2.38663033e-01 1.23188639e+00 -1.30883992e+00
-8.96338165e-01 -1.84293419e-01 -2.49091648e-02 -4.62996354e-03
-7.89466500e-02 -4.07349207e-02 -9.12252903e-01 -8.57407212e-01
1.39867933e-02 -9.08930063e-01 -2.88173914e-01 1.02349922e-01
-4.90685195e-01 -4.28877026e-02 9.37426031e-01 -7.66728818e-01
1.27743888e+00 -2.37625265e+00 -3.52502801e-03 -2.28299364e-01
4.99101400e-01 1.09839968e-01 -2.15814188e-01 7.38792419e-02
-2.79979855e-01 -4.81723920e-02 -1.44168466e-01 -4.99722779e-01
-2.71371696e-02 4.96002257e-01 -1.90467045e-01 7.81869411e-01
2.53461123e-01 9.56492424e-01 -1.29092956e+00 -6.72640264e-01
2.01415986e-01 3.77003163e-01 -3.92321259e-01 5.33721372e-02
-2.10732520e-01 7.14018166e-01 -4.47691470e-01 7.03341782e-01
5.58617413e-01 -2.51380652e-01 1.06208086e-01 -3.10828984e-01
6.21040687e-02 -1.92490537e-02 -1.29456818e+00 1.69831812e+00
9.54831392e-02 5.77061296e-01 -1.55265341e-02 -8.44255447e-01
6.57544136e-02 3.75488013e-01 7.86883175e-01 -4.33013529e-01
-8.71793702e-02 -1.17581740e-01 -2.01379701e-01 -8.34654033e-01
1.00739244e-02 2.06755958e-02 2.43199944e-01 3.95300463e-02
1.89787552e-01 1.00984776e+00 1.74761221e-01 2.58865893e-01
1.77695429e+00 5.66344202e-01 3.14875603e-01 4.58494872e-02
2.96956688e-01 -1.59633785e-01 1.24593031e+00 1.01393366e+00
-7.02269733e-01 5.07706404e-01 8.26368928e-01 -9.22244042e-03
-4.22219366e-01 -1.10085583e+00 2.30258033e-01 1.18497169e+00
1.48896396e-01 -4.29652393e-01 -9.21168447e-01 -1.36295927e+00
-1.14179261e-01 5.62038422e-01 -8.97063613e-01 -2.46049374e-01
-4.21529710e-01 -6.11127973e-01 6.22330070e-01 7.92839348e-01
4.52903211e-01 -8.49279940e-01 -5.05890548e-01 2.49331534e-01
-5.97905040e-01 -1.27069497e+00 -6.44980431e-01 2.42664769e-01
-7.08988130e-01 -1.32612896e+00 -6.16061389e-01 -2.26411626e-01
6.57999694e-01 4.54169989e-01 9.26166892e-01 -1.08947419e-01
-1.42517954e-01 5.95942736e-01 -2.97781587e-01 -3.45433652e-01
-1.82103455e-01 -4.82604206e-01 5.82963636e-04 6.98489785e-01
5.76511145e-01 -4.86019820e-01 -6.43849313e-01 5.31266749e-01
-7.19071984e-01 1.52476415e-01 5.97094893e-01 8.13364506e-01
7.62853801e-01 2.77286619e-01 1.28518060e-01 -8.85400712e-01
-9.99840349e-02 -5.08706510e-01 -2.34095722e-01 2.55019575e-01
-1.76453665e-01 -1.85290556e-02 1.49591550e-01 -1.04711246e+00
-1.08646560e+00 4.91699755e-01 2.09784493e-01 -1.00378346e+00
-4.98333454e-01 2.15069503e-01 -6.15206480e-01 1.36345014e-01
5.18652916e-01 2.39919811e-01 -3.82550895e-01 -7.49745190e-01
3.77611309e-01 5.41403331e-02 6.93970621e-01 -4.24451202e-01
7.04866529e-01 9.00280833e-01 3.03175319e-02 -8.24973047e-01
-1.22800601e+00 -7.25722551e-01 -7.21692502e-01 -3.90123814e-01
9.98068333e-01 -1.16130793e+00 -5.05329132e-01 5.30387461e-01
-1.12815487e+00 -5.29060125e-01 -3.81860554e-01 5.35824180e-01
-5.88396609e-01 6.45164311e-01 -3.56631756e-01 -1.02303386e+00
4.79874730e-01 -9.41958785e-01 1.16537106e+00 -1.60085022e-01
-1.64982408e-01 -8.66548359e-01 1.87360972e-01 5.01113355e-01
-2.10156292e-01 5.44566989e-01 5.03036857e-01 -5.45931160e-01
-5.91953039e-01 -2.08355010e-01 -2.09956663e-03 4.86067563e-01
2.56893307e-01 -8.11629146e-02 -1.26075459e+00 -3.28518078e-02
2.04660952e-01 -3.44477035e-02 1.18271446e+00 5.03086269e-01
1.17310822e+00 -5.13334572e-01 -5.29809296e-01 3.63864303e-01
9.68836367e-01 1.73860848e-01 7.16205060e-01 3.21451761e-02
1.19340277e+00 3.22370648e-01 6.23256385e-01 3.20576817e-01
2.95916080e-01 7.53268003e-01 5.40448070e-01 -4.54443574e-01
-3.03362250e-01 -6.14381611e-01 8.14373016e-01 2.92827487e-01
-3.14506769e-01 -1.95825174e-01 -6.34881675e-01 6.19718611e-01
-2.55353451e+00 -1.27117836e+00 -5.13254344e-01 2.04506302e+00
6.29953384e-01 7.27826431e-02 2.28582859e-01 7.99283162e-02
7.19954550e-01 4.30847347e-01 -6.54403210e-01 5.10818839e-01
-1.03316315e-01 -1.12484902e-01 7.89810359e-01 2.18316793e-01
-1.64010549e+00 9.10033643e-01 5.74774075e+00 1.11249554e+00
-6.80404723e-01 6.48630381e-01 3.99650365e-01 -2.34015003e-01
3.72102618e-01 1.49188042e-01 -5.37318408e-01 6.28024995e-01
7.03664660e-01 2.22961158e-01 2.68326104e-01 9.59869921e-01
6.54474139e-01 -3.25657696e-01 -1.37972558e+00 8.85255516e-01
-3.35610472e-02 -8.80551815e-01 -2.11230084e-01 6.81937635e-02
7.03586936e-01 -1.39998987e-01 -1.96928695e-01 5.03765762e-01
4.17582870e-01 -7.18306243e-01 1.00265157e+00 7.41459250e-01
4.14820403e-01 -3.50756079e-01 3.75373870e-01 4.68101770e-01
-1.26978886e+00 -1.58288404e-01 -9.37411562e-02 -4.28663678e-02
1.93893209e-01 5.95088899e-01 -2.62038022e-01 4.09528583e-01
6.52109802e-01 1.17399013e+00 -6.04430020e-01 9.28111494e-01
-6.89703822e-01 1.11747026e+00 -1.03436960e-02 4.87642556e-01
1.11422807e-01 2.61184722e-01 7.08480835e-01 1.22132039e+00
5.46648866e-03 -1.14794679e-01 3.92324269e-01 6.75038874e-01
-2.62345653e-03 -3.64744872e-01 -3.85278165e-01 -1.17520362e-01
3.83172603e-03 9.73845124e-01 -6.11382484e-01 -4.31857377e-01
-6.75116003e-01 1.35548806e+00 9.73140970e-02 7.55829275e-01
-1.26710534e+00 4.13312167e-01 7.30815768e-01 1.95726082e-01
4.77263361e-01 -4.96318191e-02 5.95867299e-02 -1.69557321e+00
1.52902067e-01 -1.10581791e+00 4.99488801e-01 -6.77762210e-01
-1.28593147e+00 2.48482991e-02 6.56051561e-02 -1.41029942e+00
9.92971659e-02 -4.02373254e-01 -5.02595901e-01 4.14730698e-01
-1.30768144e+00 -1.20524788e+00 -3.15006375e-01 8.05795550e-01
8.37904274e-01 2.57888228e-01 4.37767476e-01 4.03699130e-01
-9.98696804e-01 2.79546887e-01 -9.89656076e-02 2.72377312e-01
7.45036840e-01 -1.25261664e+00 5.97654767e-02 1.41296518e+00
4.08620834e-01 3.60908061e-01 6.43223464e-01 -8.61736894e-01
-1.26713121e+00 -1.48294532e+00 6.95248127e-01 -8.66379917e-01
1.06687093e+00 -5.43589234e-01 -9.18466151e-01 9.20917094e-01
-4.16230895e-02 3.96571070e-01 4.21408594e-01 -1.11452471e-02
-4.14454103e-01 5.89497052e-02 -6.74102366e-01 7.57921576e-01
1.45088279e+00 -5.27532518e-01 -6.47986352e-01 3.48538935e-01
5.50783575e-01 -2.25302964e-01 -3.85849357e-01 3.99181724e-01
5.50837517e-01 -9.55635369e-01 9.59039032e-01 -7.85605967e-01
4.49173331e-01 -5.00907660e-01 -4.70715854e-03 -9.37033772e-01
-4.59562272e-01 -7.50650227e-01 -7.70555854e-01 1.31561923e+00
8.64783153e-02 -2.05774128e-01 8.82223845e-01 6.49278402e-01
1.36547774e-01 -2.93207169e-01 -1.14196432e+00 -1.18829370e+00
-4.14975435e-01 -8.51589262e-01 1.31870896e-01 9.66816306e-01
-2.41064951e-01 2.59525836e-01 -9.83608544e-01 3.80356282e-01
8.35557044e-01 -4.10497487e-01 9.26619291e-01 -1.04633737e+00
-6.42157495e-01 -7.77573884e-02 -3.67775798e-01 -1.28262603e+00
2.39362895e-01 -3.45702201e-01 4.20913965e-01 -1.16718340e+00
4.72128928e-01 2.71328092e-01 -7.38064051e-01 8.56310427e-01
-4.02781546e-01 1.03535488e-01 9.27302241e-02 2.18914166e-01
-1.00579846e+00 6.60064459e-01 1.08652997e+00 -2.67329723e-01
-3.00283581e-02 4.45277281e-02 -5.09662390e-01 1.30367720e+00
2.40160376e-01 -9.60910559e-01 -4.73624945e-01 -1.48358703e-01
-3.30356270e-01 -1.07243508e-02 1.01110423e+00 -9.97006595e-01
8.21328610e-02 -4.64695483e-01 1.93819433e-01 -5.22876322e-01
2.49802321e-01 -1.09024465e+00 2.80053616e-01 3.38876903e-01
-3.34855527e-01 -5.15970409e-01 -3.56202088e-02 1.30127597e+00
-4.28539841e-03 -3.49965245e-02 6.39387965e-01 -2.68549751e-02
-8.63875985e-01 3.56366426e-01 -5.78142643e-01 6.80243745e-02
1.10046506e+00 -1.59863949e-01 -3.31398487e-01 -2.93763995e-01
-7.81462312e-01 1.10082619e-01 2.39741564e-01 3.35788518e-01
2.95016557e-01 -1.35203266e+00 -4.58371758e-01 -7.84301311e-02
2.56175220e-01 -4.35942799e-01 4.41345513e-01 1.45929551e+00
2.45460823e-01 7.31648281e-02 3.90277505e-01 -6.64661050e-01
-1.45358455e+00 8.55135202e-01 2.34722883e-01 -3.66330266e-01
-8.17340672e-01 6.55721247e-01 8.22913110e-01 3.34727317e-01
6.61996961e-01 -5.48700154e-01 5.94751127e-02 4.41549253e-03
6.58727527e-01 5.72642088e-01 -3.71421129e-01 -6.51958466e-01
-5.71162701e-01 1.92836806e-01 2.55724132e-01 4.68530990e-02
9.90267217e-01 -1.89540461e-01 1.95581362e-01 7.00261354e-01
8.90683949e-01 -2.48643756e-02 -2.06670928e+00 -4.07885790e-01
2.55899042e-01 -6.33069694e-01 7.22701922e-02 -8.41217697e-01
-1.09944713e+00 8.75777185e-01 6.16266906e-01 8.99187922e-02
1.14365280e+00 2.36926973e-02 3.80920440e-01 1.17419146e-01
3.39452624e-01 -9.85544324e-01 2.56315470e-01 2.98387766e-01
7.81324387e-01 -1.11809409e+00 -1.06391832e-01 -4.86961931e-01
-8.46763909e-01 5.59714794e-01 5.83555520e-01 -3.29349563e-02
4.70190227e-01 2.90326834e-01 -1.16643764e-01 -7.20800757e-02
-8.08054268e-01 -4.51422006e-01 2.99531251e-01 6.96343184e-01
2.84034922e-03 -1.22746378e-01 -1.43408194e-01 8.42571795e-01
7.86059916e-01 3.31526041e-01 2.72275507e-01 1.05677557e+00
-1.27970532e-01 -7.90342748e-01 -4.14986253e-01 1.48049593e-01
-5.87399423e-01 -2.82525904e-02 -5.29772401e-01 7.62431026e-01
5.33119738e-01 1.16979349e+00 -2.93566912e-01 -4.32278633e-01
4.52182561e-01 1.99818209e-01 3.37528706e-01 -4.00996983e-01
-1.34478331e-01 5.88632822e-01 1.78023860e-01 -1.22516251e+00
-9.13214803e-01 -7.89150059e-01 -9.07321870e-01 1.02258570e-01
-3.06185007e-01 -2.31389880e-01 2.45403871e-02 1.16779602e+00
3.24925065e-01 8.13160181e-01 3.96931678e-01 -7.71082222e-01
-3.06263626e-01 -8.76504779e-01 -6.89667821e-01 5.50572932e-01
5.48566937e-01 -1.14036763e+00 -6.40651584e-01 7.41819799e-01] | [8.539482116699219, 0.6872782111167908] |
b47ebc21-7273-4af0-85c8-c2d8f772732a | loop-closure-detection-using-local-3d-deep | 2111.00440 | null | https://arxiv.org/abs/2111.00440v2 | https://arxiv.org/pdf/2111.00440v2.pdf | Loop closure detection using local 3D deep descriptors | We present a simple yet effective method to address loop closure detection in simultaneous localisation and mapping using local 3D deep descriptors (L3Ds). L3Ds are emerging compact representations of patches extracted from point clouds that are learnt from data using a deep learning algorithm. We propose a novel overlap measure for loop detection by computing the metric error between points that correspond to mutually-nearest-neighbour descriptors after registering the loop candidate point cloud by its estimated relative pose. This novel approach enables us to accurately detect loops and estimate six degrees-of-freedom poses in the case of small overlaps. We compare our L3D-based loop closure approach with recent approaches on LiDAR data and achieve state-of-the-art loop closure detection accuracy. Additionally, we embed our loop closure approach in RESLAM, a recent edge-based SLAM system, and perform the evaluation on real-world RGBD-TUM and synthetic ICL datasets. Our approach enables RESLAM to achieve a better localisation accuracy compared to its original loop closure strategy. Our project page is available at github.com/yiming107/l3d_loop_closure. | ['Yi Wan', 'Qi Qin', 'Fabio Poiesi', 'Yiming Wang', 'Youjie Zhou'] | 2021-10-31 | null | null | null | null | ['loop-closure-detection'] | ['computer-vision'] | [-1.75247520e-01 -1.54407486e-01 -8.27841833e-02 -3.06392282e-01
-1.00385034e+00 -5.73190928e-01 8.49818468e-01 6.16460085e-01
-5.14494181e-01 4.43389267e-01 -1.64886087e-01 -2.05860317e-01
-1.48007661e-01 -8.30138743e-01 -1.11226964e+00 -8.39670449e-02
-4.08105135e-01 7.56092489e-01 3.56756389e-01 -1.38529599e-01
4.13811028e-01 1.02127373e+00 -1.67079866e+00 -2.64477253e-01
8.39797914e-01 1.17209733e+00 1.56094402e-01 3.64092767e-01
2.23223511e-02 1.78182289e-01 -8.09611678e-02 2.88813978e-01
6.03334725e-01 -5.58147356e-02 -4.51234430e-01 -4.47235882e-01
9.73973274e-01 -1.23759568e-01 -2.05583811e-01 6.95632100e-01
5.43500841e-01 -6.60243407e-02 4.92699891e-01 -1.31637049e+00
-8.66923332e-02 -2.60198861e-01 -7.79720843e-01 -8.17542821e-02
7.80992746e-01 -2.57144541e-01 8.92071486e-01 -1.33889580e+00
7.81135142e-01 1.13627815e+00 1.39219296e+00 -1.44397557e-01
-1.70189404e+00 -7.72836864e-01 -3.22887450e-01 2.14261208e-02
-2.01819301e+00 -4.25489962e-01 7.06703961e-01 -4.90877658e-01
1.23282552e+00 1.74279124e-01 8.07151914e-01 5.71681619e-01
5.44399977e-01 4.61901307e-01 1.07669985e+00 -5.21646202e-01
2.50198156e-01 -2.32201457e-01 -2.61561394e-01 1.03213918e+00
3.97891104e-01 4.95616138e-01 -8.55149925e-01 -3.94746631e-01
1.05028951e+00 2.88311783e-02 -2.49873504e-01 -1.30864179e+00
-1.52849126e+00 9.34129059e-01 9.77233052e-01 -2.13641915e-02
-2.22374201e-01 5.04090130e-01 1.22745268e-01 1.72645137e-01
5.40803373e-01 1.73398718e-01 -2.98797369e-01 2.27827355e-02
-1.06175053e+00 4.21883255e-01 6.99100494e-01 1.17902815e+00
1.52767313e+00 -7.00290382e-01 2.42268756e-01 2.02637091e-01
4.98578370e-01 6.75696611e-01 -1.00113131e-01 -8.14395785e-01
2.05121845e-01 8.74613106e-01 1.89881742e-01 -1.12523699e+00
-5.86022615e-01 -3.72382998e-01 -6.61149859e-01 5.29273152e-01
-9.77284312e-02 3.56328130e-01 -6.91603899e-01 1.16157556e+00
4.72086251e-01 2.93197274e-01 -1.36783376e-01 7.75366545e-01
6.60537422e-01 2.46950731e-01 -5.18758357e-01 2.57033288e-01
8.94822299e-01 -5.95909894e-01 -1.55739605e-01 -4.18209493e-01
8.87618601e-01 -7.62643635e-01 6.56365931e-01 1.21905215e-01
-4.07384098e-01 -4.60150868e-01 -1.21242833e+00 -3.05792332e-01
-2.61030972e-01 3.04091126e-01 7.18803823e-01 1.98469624e-01
-1.25928664e+00 5.92441022e-01 -9.46460247e-01 -6.96459115e-01
2.21832141e-01 4.66352016e-01 -8.12821388e-01 6.17149472e-02
-7.31042147e-01 9.25195336e-01 3.90687436e-01 1.20427772e-01
-4.93970245e-01 -5.52614689e-01 -1.20148349e+00 -5.99556983e-01
4.36078906e-02 -6.37588978e-01 8.36141765e-01 -3.34930450e-01
-8.10027659e-01 1.39108968e+00 -4.74763513e-01 -7.56243110e-01
7.28493452e-01 -5.29983580e-01 -2.22418085e-02 9.59069375e-03
4.96633112e-01 8.51436019e-01 4.72904384e-01 -1.41891921e+00
-5.73114753e-01 -5.23869991e-01 -3.02168608e-01 1.52078018e-01
5.58317602e-01 -5.89188039e-01 -2.08667964e-01 -1.56415358e-01
1.01956940e+00 -1.23680270e+00 -2.14672610e-01 6.22968137e-01
-7.35427141e-02 -2.68492401e-01 1.01405740e+00 -2.14972436e-01
7.86855221e-01 -2.08185053e+00 -1.49432316e-01 3.22435558e-01
2.29744405e-01 -8.97601098e-02 5.13314344e-02 6.96488678e-01
1.85825810e-01 -1.59420252e-01 -2.92995602e-01 -5.97834885e-01
9.81566310e-02 3.48222256e-01 -2.19669223e-01 1.00083017e+00
1.08035766e-01 9.00749505e-01 -9.19364989e-01 -5.34608841e-01
7.17610359e-01 6.61806226e-01 -3.70449007e-01 1.64636448e-02
1.61337131e-03 5.28435171e-01 -1.83108762e-01 7.37261832e-01
1.13981926e+00 2.83990085e-01 -3.60931396e-01 -2.42962450e-01
-5.69138527e-01 3.44147712e-01 -1.24498415e+00 2.49059176e+00
-5.10182440e-01 8.54393065e-01 -2.41963472e-02 -4.99014348e-01
1.43632197e+00 -1.95358872e-01 4.28158402e-01 -8.23679149e-01
-1.10137701e-01 6.03022516e-01 -7.02993095e-01 1.42985076e-01
6.75916791e-01 1.75093293e-01 -2.05160394e-01 8.19892585e-02
-3.29663604e-02 -5.00012398e-01 -3.36788327e-01 5.69178946e-02
1.12875378e+00 6.97316587e-01 5.00901222e-01 -4.93171006e-01
4.32466716e-01 3.26639920e-01 3.85893732e-01 8.00107121e-01
-2.54659176e-01 6.99443877e-01 8.74119848e-02 -7.72265315e-01
-8.25110972e-01 -1.26490343e+00 -5.00506878e-01 3.83351237e-01
6.21681333e-01 -6.09104395e-01 -1.62214488e-01 -3.49939764e-01
7.91382372e-01 2.26847291e-01 -5.68501174e-01 6.91045001e-02
-4.05455977e-01 1.46811560e-01 5.76177239e-01 3.37009072e-01
7.14300394e-01 -4.64235097e-01 -1.07869387e+00 1.12138093e-01
-3.85316764e-03 -7.70172715e-01 -7.07581639e-02 4.70709413e-01
-9.19661224e-01 -1.13515663e+00 -1.57609746e-01 -8.66226912e-01
6.24922216e-01 2.57541448e-01 1.02136838e+00 -6.22558743e-02
-3.28306347e-01 4.20351893e-01 -3.12688112e-01 -2.22985238e-01
5.08132018e-03 -5.06956875e-03 3.46807003e-01 -4.72733796e-01
7.15318382e-01 -6.31529570e-01 -7.47828126e-01 2.89084524e-01
-2.75107652e-01 -7.67055387e-03 6.13004565e-01 4.93368089e-01
1.08648002e+00 -5.19016147e-01 -1.89834476e-01 -1.58366174e-01
2.10666806e-01 -1.17772482e-01 -1.04597926e+00 -1.27194792e-01
-6.02317750e-01 2.70240575e-01 -2.78582335e-01 4.38623801e-02
-2.19337583e-01 7.37802386e-01 -2.73514129e-02 -6.41949356e-01
-2.26612493e-01 4.56310004e-01 3.67711008e-01 -8.84229124e-01
8.45127165e-01 1.05983838e-01 9.60617214e-02 -4.83923912e-01
4.51579928e-01 5.81861377e-01 6.13167405e-01 -4.66268420e-01
9.09851670e-01 8.36535275e-01 3.49185020e-01 -7.10193336e-01
-3.01207155e-01 -8.91541839e-01 -1.30306649e+00 -9.09167305e-02
4.16372269e-01 -1.37391472e+00 -7.46696651e-01 3.40849191e-01
-1.36217642e+00 -6.13891333e-02 -2.84308076e-01 5.10468960e-01
-7.48537540e-01 2.97476470e-01 -1.27081633e-01 -6.89500153e-01
-3.37055832e-01 -9.09704924e-01 1.74112165e+00 -5.75965345e-02
-2.10917920e-01 -7.99953878e-01 7.31239021e-01 -2.64330089e-01
1.25995740e-01 9.09500539e-01 1.82918012e-01 -1.42421514e-01
-8.95674169e-01 -4.81654495e-01 -1.85488597e-01 -1.25336036e-01
-1.02111166e-02 -2.92368233e-01 -9.55720723e-01 -5.78720272e-01
-2.77584106e-01 -7.04087988e-02 9.83906507e-01 1.46390036e-01
3.22912723e-01 2.62405604e-01 -8.61224532e-01 9.30624187e-01
1.80399501e+00 -2.39474788e-01 5.60572624e-01 7.65272796e-01
6.89538062e-01 2.23845229e-01 1.04664886e+00 4.43995982e-01
5.48178375e-01 7.67917633e-01 7.84345508e-01 -2.13588700e-01
1.81074023e-01 -7.08729386e-01 1.91152066e-01 4.81074572e-01
1.24294711e-02 3.75172466e-01 -1.29110265e+00 6.10014439e-01
-1.98631716e+00 -3.64351481e-01 -4.25584644e-01 2.58012366e+00
3.35550606e-01 8.48495960e-02 -2.42095739e-01 -7.17643425e-02
5.66606522e-01 1.89378187e-01 -3.44009012e-01 -1.85203478e-01
2.09378661e-03 4.21247482e-01 1.05826950e+00 7.97471464e-01
-1.34385395e+00 1.11709094e+00 5.36975574e+00 4.08985108e-01
-1.10718751e+00 1.37811899e-01 -3.42246920e-01 5.99909201e-02
-1.86806351e-01 5.29810250e-01 -6.80761456e-01 -1.01830952e-01
6.28594398e-01 2.51740795e-02 -2.97485199e-02 8.23046625e-01
6.05579205e-02 -6.78960562e-01 -1.20447445e+00 1.26542020e+00
1.11839771e-02 -1.40694141e+00 -3.19997758e-01 3.11548382e-01
5.89341164e-01 7.07932889e-01 -3.83599252e-01 7.94152170e-02
-1.69951260e-01 -7.53675997e-01 1.01301610e+00 4.24633175e-01
1.08176303e+00 -8.04544687e-01 5.81869483e-01 5.25473297e-01
-1.81322169e+00 2.32730344e-01 -5.29330790e-01 -4.00687397e-01
-1.06820203e-01 6.86466873e-01 -1.10998988e+00 7.08575845e-01
7.26962924e-01 9.21435714e-01 -8.05313945e-01 1.12952197e+00
-2.20476687e-01 -1.29548877e-01 -7.21739948e-01 1.98989108e-01
1.36095271e-01 -2.98462689e-01 5.56663275e-01 1.01037288e+00
5.79551756e-01 -3.76917094e-01 2.64623761e-01 1.15444827e+00
2.58528799e-01 2.42168121e-02 -1.24380589e+00 4.37305033e-01
7.46757984e-01 1.04464102e+00 -7.72903264e-01 1.33824095e-01
-1.13618992e-01 1.05481982e+00 3.22613955e-01 -1.00243777e-01
-5.68286836e-01 -6.52809203e-01 8.10623109e-01 4.22802687e-01
2.08936214e-01 -8.67090166e-01 4.50475849e-02 -1.01732743e+00
2.06293076e-01 -1.74107492e-01 -1.77683890e-01 -8.46476138e-01
-7.05579281e-01 4.57686245e-01 -2.14351311e-01 -1.68016744e+00
-1.69467658e-01 -4.54364479e-01 -2.45252982e-01 8.66310537e-01
-1.50683463e+00 -1.42364609e+00 -6.50987208e-01 4.98215646e-01
-1.16657354e-01 3.69371712e-01 1.05198598e+00 1.36628240e-01
3.56704295e-01 2.14581594e-01 2.61930645e-01 7.98479915e-02
8.27520251e-01 -1.14358747e+00 9.86109197e-01 7.18316019e-01
3.90838563e-01 6.07652009e-01 5.78959405e-01 -8.92528355e-01
-1.62994409e+00 -1.00010991e+00 9.08449531e-01 -7.86183655e-01
5.09553850e-01 -8.68622065e-01 -5.89460254e-01 9.49533224e-01
-3.34628314e-01 4.71358716e-01 9.34719220e-02 -2.58593801e-02
-3.21479648e-01 -2.43757352e-01 -1.26226234e+00 1.28751015e-02
1.41749048e+00 -1.00341058e+00 -7.21432626e-01 2.21594974e-01
5.94971180e-01 -9.34379876e-01 -7.73852825e-01 7.56232798e-01
7.34640598e-01 -1.39076841e+00 1.01609969e+00 5.25536060e-01
-2.56600529e-01 -7.03194499e-01 -2.43145436e-01 -1.08156943e+00
-1.71049193e-01 -5.44955432e-01 1.08358068e-02 9.15840685e-01
-1.49624705e-01 -9.56070781e-01 8.49014223e-01 -2.35078245e-01
-9.69467312e-02 -5.49601018e-01 -1.46929431e+00 -9.52208877e-01
-2.49518648e-01 -5.19283652e-01 4.69538033e-01 7.18583345e-01
-2.68221200e-01 -3.10019962e-02 2.60609947e-02 5.06211519e-01
9.73718464e-01 4.22096252e-01 1.20739901e+00 -1.69682920e+00
4.05950218e-01 1.64651666e-02 -1.25539339e+00 -1.18329513e+00
1.58140048e-01 -1.14557648e+00 2.84390956e-01 -1.72336960e+00
-3.98670912e-01 -7.20321357e-01 -4.60816398e-02 3.07158113e-01
4.98141527e-01 5.04669905e-01 -5.60658798e-02 4.54611778e-01
-6.37809694e-01 6.71091735e-01 5.45007050e-01 1.18384495e-01
-2.68917114e-01 -1.76010400e-01 4.69369115e-03 6.39846683e-01
3.65422487e-01 -5.70507407e-01 9.29256678e-02 -4.27540600e-01
3.26433241e-01 -1.64964139e-01 8.99597168e-01 -1.59418690e+00
4.56117600e-01 3.36561680e-01 4.27497923e-01 -1.36162984e+00
4.08781320e-01 -1.05445540e+00 5.34977973e-01 7.81249523e-01
1.64927974e-01 1.91181138e-01 4.13608789e-01 6.97466373e-01
-1.82840958e-01 2.51621366e-01 4.55759704e-01 -1.29863555e-02
-1.16818595e+00 2.47274518e-01 1.14582837e-01 -3.90581191e-01
1.09149194e+00 -4.74850774e-01 -8.88663530e-02 -1.87422916e-01
-4.22541857e-01 2.33788326e-01 1.21993709e+00 4.68040556e-01
9.02273715e-01 -1.63980138e+00 -5.87861836e-01 7.69212842e-01
8.23339105e-01 4.51098263e-01 -3.38249803e-01 1.06778431e+00
-1.21782088e+00 5.76859891e-01 -1.42698422e-01 -1.41661048e+00
-1.18881047e+00 2.31071934e-01 5.14149010e-01 2.98204124e-01
-7.74075687e-01 9.74907517e-01 -1.77845955e-01 -9.74861383e-01
1.49286970e-01 -7.14245081e-01 5.00577390e-01 -1.13743881e-03
7.14403093e-02 2.36336738e-01 3.53610516e-01 -1.07070065e+00
-9.79732454e-01 1.25826311e+00 5.29907644e-01 1.45935710e-03
1.17698753e+00 -2.07584739e-01 -3.70281607e-01 5.71088791e-01
1.34305525e+00 9.08989310e-02 -1.22016037e+00 -2.82073587e-01
2.31542990e-01 -6.96445525e-01 7.85177574e-02 -1.64417580e-01
-3.57129455e-01 8.37205112e-01 1.30656576e+00 -3.23083997e-01
6.67744637e-01 2.25332722e-01 4.76167351e-01 4.57025409e-01
1.24784362e+00 -6.82469189e-01 -4.20035303e-01 6.37971401e-01
9.52728391e-01 -1.44265091e+00 3.04053843e-01 -1.52240798e-01
5.90621233e-02 1.00767469e+00 1.48059532e-01 -5.03223419e-01
5.70237637e-01 2.51529217e-01 4.22957800e-02 -4.00177538e-01
-1.22359529e-01 -4.33185756e-01 2.37886935e-01 7.86692977e-01
2.06835598e-01 1.83396503e-01 -1.03100307e-01 -3.25796574e-01
-1.51380807e-01 6.31466880e-02 -3.48126963e-02 1.35653543e+00
-7.30329931e-01 -1.05467367e+00 -6.60334885e-01 1.36344492e-01
4.39103037e-01 3.62679586e-02 -5.64180553e-01 1.00549841e+00
3.68552536e-01 4.95915830e-01 4.34112489e-01 -6.29757702e-01
4.99223322e-01 -7.66020864e-02 7.88398564e-01 -5.31700373e-01
-6.26977012e-02 -5.92439435e-02 -2.27461457e-01 -1.05107486e+00
-4.98803079e-01 -6.76386535e-01 -1.40755033e+00 -3.09289753e-01
-4.53828990e-01 1.14879988e-01 1.03138030e+00 6.21766567e-01
7.44552612e-01 -1.91547409e-01 4.09528673e-01 -1.51617980e+00
-1.39162347e-01 -6.53170586e-01 -6.51356161e-01 1.24254860e-02
7.56162703e-01 -9.47196364e-01 -3.16198856e-01 -5.05873203e-01] | [7.442224502563477, -2.2054550647735596] |
09e03723-a0e7-465a-8247-e565e458deeb | face-alignment-in-the-wild-a-survey | 1608.04188 | null | http://arxiv.org/abs/1608.04188v1 | http://arxiv.org/pdf/1608.04188v1.pdf | Face Alignment In-the-Wild: A Survey | Over the last two decades, face alignment or localizing fiducial facial
points has received increasing attention owing to its comprehensive
applications in automatic face analysis. However, such a task has proven
extremely challenging in unconstrained environments due to many confounding
factors, such as pose, occlusions, expression and illumination. While numerous
techniques have been developed to address these challenges, this problem is
still far away from being solved. In this survey, we present an up-to-date
critical review of the existing literatures on face alignment, focusing on
those methods addressing overall difficulties and challenges of this topic
under uncontrolled conditions. Specifically, we categorize existing face
alignment techniques, present detailed descriptions of the prominent algorithms
within each category, and discuss their advantages and disadvantages.
Furthermore, we organize special discussions on the practical aspects of face
alignment in-the-wild, towards the development of a robust face alignment
system. In addition, we show performance statistics of the state of the art,
and conclude this paper with several promising directions for future research. | ['Xin Jin', 'Xiaoyang Tan'] | 2016-08-15 | null | null | null | null | ['robust-face-alignment'] | ['computer-vision'] | [ 7.29047582e-02 -9.95912179e-02 -3.53368431e-01 -6.81196034e-01
-5.01139224e-01 -5.50278544e-01 4.70623851e-01 -5.16263008e-01
-5.39362617e-02 5.43235362e-01 1.28209945e-02 1.42095864e-01
8.94656107e-02 -2.26642061e-02 -1.88327149e-01 -8.00489068e-01
-6.63660392e-02 2.71416932e-01 -5.12394845e-01 -5.72935566e-02
3.06192368e-01 9.53410208e-01 -1.69202948e+00 -3.56659263e-01
4.77333218e-01 8.31289649e-01 -3.58962327e-01 3.55910659e-01
1.63235858e-01 -3.57865095e-02 -7.49634683e-01 -7.77739882e-01
3.04353476e-01 -4.96039450e-01 -6.23946369e-01 3.77566457e-01
1.11517036e+00 -3.53318453e-01 -3.40280589e-03 1.13172853e+00
9.20621634e-01 -4.48882692e-02 4.22554106e-01 -1.53152239e+00
-3.67859960e-01 -1.46974429e-01 -8.67180765e-01 1.01641662e-01
6.17887855e-01 -3.33370835e-01 5.25640786e-01 -9.85397637e-01
4.51732337e-01 1.22166371e+00 7.73945987e-01 1.06545210e+00
-8.80198836e-01 -8.56648326e-01 1.61607787e-01 -1.02672219e-01
-1.57049835e+00 -1.25043261e+00 9.32131231e-01 -2.84529507e-01
6.11012220e-01 4.19429034e-01 6.59886599e-01 7.67487705e-01
1.04387581e-01 4.01190400e-01 1.01134205e+00 -7.45498717e-01
-2.33593285e-01 1.19729415e-01 -1.31128147e-01 9.02297914e-01
2.60133177e-01 -6.25061914e-02 -5.20852625e-01 -5.56133807e-01
8.93080771e-01 -4.24232304e-01 -2.02635989e-01 -4.91429090e-01
-7.83379257e-01 5.67633033e-01 -4.76073138e-02 5.30844748e-01
-7.29021206e-02 -1.00898631e-01 2.46537417e-01 -1.12953372e-01
6.70053303e-01 1.92205638e-01 -1.27572343e-01 6.63449010e-03
-1.11124587e+00 3.79448980e-01 7.58553445e-01 9.86740112e-01
4.21182930e-01 2.28417829e-01 2.02080607e-01 1.04848146e+00
5.20804405e-01 5.94277561e-01 2.37320229e-01 -1.00886321e+00
8.57930481e-02 -5.79407923e-02 -4.08168416e-03 -1.25839245e+00
-3.89363825e-01 1.70098320e-02 -5.57375371e-01 1.60011753e-01
3.48450243e-01 -3.37335497e-01 -5.08068085e-01 1.80515993e+00
6.38081789e-01 2.90280551e-01 -2.94436932e-01 7.63667703e-01
1.13966596e+00 4.24034707e-02 -7.80325457e-02 -7.46058822e-01
1.61982501e+00 -1.03044844e+00 -1.26902139e+00 -2.64428347e-01
6.93761632e-02 -1.36350679e+00 4.14187491e-01 -4.71233279e-02
-1.15254605e+00 -3.93604368e-01 -8.49241376e-01 9.38592199e-03
1.75247863e-02 4.51785743e-01 8.37462723e-01 1.26216710e+00
-1.41915727e+00 2.10248411e-01 -8.13449621e-01 -9.09991264e-01
2.82767177e-01 8.44585359e-01 -6.98180974e-01 3.10777992e-01
-6.98974788e-01 1.07498300e+00 -4.21435922e-01 4.44540322e-01
-1.69069827e-01 -3.44911247e-01 -9.05827820e-01 -3.91773909e-01
2.82476038e-01 -4.64208335e-01 1.43331397e+00 -7.81652629e-01
-1.85138702e+00 1.36219633e+00 -8.78127456e-01 2.98863590e-01
2.55445361e-01 -1.21138319e-01 -6.58043742e-01 -1.33319506e-02
-3.91421355e-02 7.01829255e-01 8.88829648e-01 -1.06384838e+00
-3.77896398e-01 -7.76931703e-01 -1.98156580e-01 4.01600629e-01
-3.63132238e-01 8.69516909e-01 -8.85801435e-01 -6.91786230e-01
1.67869002e-01 -1.20728409e+00 -6.70201480e-02 3.07399750e-01
-7.56047964e-02 -3.05482805e-01 8.17513883e-01 -4.90726322e-01
1.30534959e+00 -2.22802877e+00 -7.64963925e-02 1.27512306e-01
9.24245864e-02 3.23156238e-01 3.64689641e-02 1.49096102e-01
-3.09749782e-01 -8.62291008e-02 1.07588723e-01 -7.09649920e-01
-2.56181091e-01 -1.55051993e-02 -1.24949902e-01 1.01092839e+00
-2.40068629e-01 7.36562610e-01 -6.79841936e-01 -6.53099835e-01
2.42701575e-01 6.56713426e-01 -3.31135541e-01 5.61739504e-02
5.08651555e-01 6.00195110e-01 -3.04743946e-01 1.15067244e+00
9.76759076e-01 2.99382895e-01 3.26217741e-01 -2.49057844e-01
-1.68270603e-01 -4.14471738e-02 -1.22352576e+00 1.22773123e+00
-1.06354520e-01 6.72240436e-01 5.79061627e-01 -7.23159194e-01
9.29934323e-01 7.87803829e-01 9.31285620e-01 -1.96363077e-01
2.94164628e-01 4.19365466e-01 -3.24392086e-03 -3.52460116e-01
5.01972973e-01 -2.86947042e-01 2.75716931e-01 3.75705481e-01
-5.75188100e-02 -1.85774744e-01 4.00286838e-02 -3.23274523e-01
1.44285485e-01 1.49451822e-01 8.36892247e-01 -3.02670300e-01
9.30110574e-01 -4.02838528e-01 5.38488328e-01 1.52854547e-01
-8.69761765e-01 5.51071525e-01 2.26711139e-01 -5.05785465e-01
-4.80542570e-01 -4.89133418e-01 -5.26272714e-01 8.67565453e-01
4.53221053e-03 -5.56751609e-01 -1.23546576e+00 -6.54186785e-01
1.46045657e-02 -8.27652514e-02 -7.61387527e-01 3.52678776e-01
-8.07215273e-01 -8.45006824e-01 5.85806608e-01 3.71895730e-01
3.49093020e-01 -8.21461320e-01 -2.36569479e-01 -9.80171487e-02
-2.95708984e-01 -1.22610283e+00 -6.36431754e-01 -6.31506681e-01
-9.22998786e-01 -1.15780544e+00 -8.29087794e-01 -9.97946501e-01
1.15900147e+00 6.98582470e-01 9.54867065e-01 3.08720142e-01
-3.36290270e-01 6.81304872e-01 1.50211751e-01 -6.04804575e-01
-6.16459511e-02 -1.44007891e-01 6.40139759e-01 5.28712086e-02
6.87624753e-01 -1.70001686e-01 -4.73759562e-01 7.60482967e-01
-3.36907446e-01 -5.30598044e-01 3.83691350e-03 6.00178540e-01
4.15092498e-01 -3.10275346e-01 4.48296070e-01 -7.09301531e-01
4.78319198e-01 -1.51178315e-02 -6.77711785e-01 2.48649642e-01
-3.14389884e-01 -6.16225660e-01 1.28922075e-01 -1.48083538e-01
-1.08096135e+00 1.25269771e-01 -3.95545810e-01 -1.20324917e-01
-1.74556807e-01 -1.01556882e-01 -3.19956720e-01 -9.25661623e-01
4.33920383e-01 -2.59114444e-01 5.61712503e-01 -2.77581215e-01
1.40233144e-01 7.56196201e-01 6.04501784e-01 -6.43205404e-01
7.97078431e-01 7.09150136e-01 1.46461725e-01 -1.22655416e+00
-6.87601626e-01 -5.50513029e-01 -9.17329848e-01 -5.15028775e-01
5.58211207e-01 -5.32856762e-01 -7.73458540e-01 6.05302751e-01
-1.09534860e+00 3.13708425e-01 3.24848801e-01 4.10471141e-01
-5.65763593e-01 6.35789335e-01 -3.03057879e-01 -7.88372278e-01
-3.62791479e-01 -1.74550295e+00 1.14776015e+00 5.89878321e-01
-5.49164057e-01 -1.15189636e+00 8.26832280e-02 5.29223502e-01
4.59354311e-01 2.22690433e-01 2.82426834e-01 -2.29009703e-01
-1.17372647e-01 -4.91384238e-01 2.72186011e-01 2.17107553e-02
6.23704731e-01 4.20081377e-01 -1.26403594e+00 -5.09815574e-01
2.46274889e-01 -1.35419820e-03 -1.00751594e-01 6.71997547e-01
1.05881512e+00 -1.55501753e-01 -7.28986859e-01 7.00787723e-01
8.88741016e-01 3.28469843e-01 4.37751085e-01 1.39990821e-01
3.90053958e-01 8.85764897e-01 8.06765079e-01 1.96043536e-01
1.71351641e-01 1.24851012e+00 1.98258981e-01 -2.16604158e-01
-1.42781392e-01 2.05155984e-01 1.10879712e-01 5.71380854e-01
-6.15562499e-01 2.00771257e-01 -4.87708867e-01 1.28643632e-01
-1.32253420e+00 -9.93435204e-01 5.35618886e-02 2.26120663e+00
5.92685819e-01 -7.33680129e-01 3.68805677e-01 -1.39599759e-02
1.02003384e+00 2.78104216e-01 -1.35831192e-01 -2.97883570e-01
1.93448402e-02 7.61993751e-02 1.03311464e-01 6.59682333e-01
-1.27314854e+00 1.06155813e+00 8.02625751e+00 3.93629193e-01
-1.40702653e+00 -3.71249504e-02 6.75334156e-01 9.54031870e-02
2.25549459e-01 -1.64063379e-01 -1.30441689e+00 2.09742039e-01
4.06548828e-01 -6.03739079e-03 3.87706071e-01 1.01658809e+00
3.71605679e-02 -1.09738462e-01 -1.02756453e+00 1.37134361e+00
7.97573626e-01 -7.76759744e-01 -3.37820768e-01 2.84724146e-01
7.31569648e-01 -4.88244385e-01 3.37785602e-01 -1.59013897e-01
-5.34404039e-01 -1.03925049e+00 3.66922706e-01 -4.39405441e-02
1.08103371e+00 -5.04418790e-01 5.32345355e-01 -1.32377118e-01
-1.30834067e+00 3.25334221e-01 -4.12362576e-01 -4.04186659e-02
1.92691922e-01 3.26720655e-01 -5.51401079e-01 3.91640842e-01
4.74968344e-01 5.63562870e-01 -3.97099882e-01 1.01974201e+00
-1.36506900e-01 1.45727888e-01 -4.19022024e-01 7.77339116e-02
-2.11617023e-01 -4.04320180e-01 4.80539918e-01 8.25136423e-01
4.12248909e-01 3.17588329e-01 -4.79762591e-02 1.73558250e-01
-3.65666188e-02 3.63409281e-01 -7.23183274e-01 1.10575080e-01
5.29827893e-01 1.63198924e+00 -7.05921412e-01 -7.19509833e-03
-8.18334401e-01 7.31584311e-01 2.47390002e-01 2.44031370e-01
-8.52145255e-01 -9.87466723e-02 1.06924224e+00 4.33374718e-02
-2.86326855e-01 -2.51143545e-01 -3.09837073e-01 -1.04240787e+00
1.33425176e-01 -1.11371338e+00 4.76362705e-01 -2.57281542e-01
-7.97186136e-01 7.44465113e-01 2.44088531e-01 -1.20969784e+00
-4.31834400e-01 -6.53410673e-01 -5.33661067e-01 8.72348428e-01
-1.30110383e+00 -1.01077783e+00 -3.91849220e-01 3.99237961e-01
4.21673208e-01 -2.96074539e-01 1.06408525e+00 3.80760968e-01
-8.47777545e-01 1.14142764e+00 -1.46567434e-01 -4.66265157e-03
1.31556046e+00 -5.87089181e-01 3.41523349e-01 6.79264843e-01
-8.46825074e-03 1.21021914e+00 6.34840250e-01 -4.61988121e-01
-1.44290435e+00 -5.39030194e-01 1.23430741e+00 -5.51948547e-01
1.98726997e-01 -2.52756029e-01 -5.77491820e-01 9.25732911e-01
2.30003074e-01 2.81736702e-01 1.07601869e+00 2.76475132e-01
-4.59749289e-02 -2.66010910e-01 -1.48301733e+00 8.61988723e-01
9.51719224e-01 -2.42491439e-01 -2.49852493e-01 3.68572831e-01
-3.58183742e-01 -8.62194955e-01 -8.19619417e-01 4.63739514e-01
1.04337060e+00 -9.91905153e-01 1.02486753e+00 -2.44605824e-01
-3.15798372e-01 -3.86965871e-01 7.68810511e-02 -1.06166565e+00
-4.57454920e-01 -1.08326209e+00 1.71767697e-01 1.43876743e+00
-9.98600423e-02 -8.74882936e-01 9.91147518e-01 8.75688612e-01
3.28243002e-02 -8.24958622e-01 -9.73025620e-01 -3.89614254e-01
-2.47709483e-01 9.38511640e-02 7.49505341e-01 9.92745519e-01
1.97656825e-01 5.88639863e-02 -5.27295113e-01 1.32219031e-01
5.81048310e-01 5.61810285e-02 1.12206376e+00 -1.02057481e+00
2.84742475e-01 -5.43833435e-01 -5.91562390e-01 -8.76017332e-01
5.36654055e-01 -5.20361423e-01 -7.95203075e-02 -1.07380438e+00
2.16509506e-01 -3.39783549e-01 3.22541416e-01 3.94834667e-01
-4.09954816e-01 7.83371747e-01 6.12585768e-02 2.20708057e-01
1.14864232e-02 1.49993673e-01 1.20107639e+00 2.41894916e-01
6.01578280e-02 2.73957014e-01 -8.86200488e-01 9.16111231e-01
8.90394747e-01 -5.11531234e-02 -2.66637981e-01 -3.46012861e-01
-4.36034769e-01 1.20095508e-02 -2.37195075e-01 -7.49216855e-01
2.06219584e-01 -2.68289745e-01 5.83703041e-01 -4.91476983e-01
6.14989460e-01 -8.63815010e-01 2.34771609e-01 2.20694155e-01
2.84203023e-01 4.90403265e-01 3.34937662e-01 -3.64994481e-02
-2.59831190e-01 -3.59404087e-01 1.23274481e+00 5.29663153e-02
-4.72644120e-01 5.05917192e-01 -1.60553470e-01 -3.04960430e-01
1.33587587e+00 -4.84727681e-01 -3.03276516e-02 -4.57562089e-01
-5.15686393e-01 -1.12980701e-01 6.67920232e-01 4.46309477e-01
2.42716104e-01 -1.39944828e+00 -5.19108772e-01 7.00741887e-01
-3.59947160e-02 -3.75021845e-01 2.76435703e-01 9.68421638e-01
-5.20865262e-01 6.25361919e-01 -2.77849078e-01 -6.24217272e-01
-2.13035107e+00 3.65891188e-01 4.85244751e-01 3.72594416e-01
-7.07645640e-02 8.12012970e-01 2.60639668e-01 -4.14046556e-01
2.93319792e-01 4.23485279e-01 -4.76282567e-01 -6.56518489e-02
7.34992445e-01 4.98606443e-01 3.05118799e-01 -1.33586001e+00
-6.73904598e-01 1.14621091e+00 -2.13796515e-02 6.76606372e-02
7.49834120e-01 -2.18483403e-01 -6.65447474e-01 -1.85628325e-01
1.04448307e+00 3.95355642e-01 -7.68967152e-01 6.46279901e-02
-3.31607640e-01 -9.23744857e-01 -3.77596647e-01 -2.11292222e-01
-1.36439407e+00 6.75026238e-01 3.84141624e-01 -3.35779458e-01
1.22902560e+00 -4.80141081e-02 3.40265900e-01 -6.08771071e-02
5.13575375e-01 -8.61301720e-01 -1.07502349e-01 4.63049233e-01
9.78840768e-01 -1.04216576e+00 3.77150804e-01 -1.07151914e+00
-3.15448791e-01 1.24341893e+00 7.82330692e-01 2.35687375e-01
8.67314756e-01 4.80176210e-01 5.08960187e-01 -6.87066242e-02
-5.26428968e-02 2.17419103e-01 4.25630838e-01 7.63318062e-01
1.03639388e+00 -1.63703918e-01 -5.19455254e-01 7.95844495e-02
-5.42229533e-01 -1.89909101e-01 7.74621367e-02 1.03601551e+00
-1.24615915e-01 -1.51021469e+00 -1.02270889e+00 2.16442019e-01
-8.32021356e-01 2.86453933e-01 -4.08365667e-01 9.05019999e-01
-8.83562416e-02 9.40805495e-01 1.15243226e-01 -6.25371113e-02
3.17829967e-01 -5.06416708e-03 9.30249929e-01 -4.70198542e-01
-4.28529918e-01 4.51614618e-01 1.23787746e-02 -5.13711333e-01
-7.93648839e-01 -1.05977142e+00 -7.70443022e-01 -5.43193519e-01
-4.93276894e-01 2.40917578e-01 7.83421218e-01 7.36579835e-01
2.63790548e-01 -3.97969969e-02 6.75204456e-01 -1.21348906e+00
-3.97355258e-01 -8.79098356e-01 -5.59020638e-01 6.47820113e-03
1.88697293e-01 -1.04006505e+00 -2.34536186e-01 4.75979745e-02] | [13.357195854187012, 0.4080319106578827] |
c606b457-c924-40ed-b061-38092335d8e0 | online-obstructive-sleep-apnea-detection | 2110.00660 | null | https://arxiv.org/abs/2110.00660v2 | https://arxiv.org/pdf/2110.00660v2.pdf | Automatic Home-based Screening of Obstructive Sleep Apnea Using Single Channel Electrocardiogram and SPO2 Signals | Obstructive sleep apnea (OSA) is one of the most widespread respiratory diseases today. Complete or relative breathing cessations due to upper airway subsidence during sleep is OSA. It has confirmed potential influence on Covid-19 hospitalization and mortality, and is strongly associated with major comorbidities of severe Covid-19 infection. Un-diagnosed OSA may also lead to a variety of severe physical and mental side-effects. To score OSA severity, nocturnal sleep monitoring is performed under defined protocols and standards called polysomnography (PSG). This method is time-consuming, expensive, and requiring professional sleep technicians. Automatic home-based detection of OSA is welcome and in great demand. It is a fast and effective way for referring OSA suspects to sleep clinics for further monitoring. On-line OSA detection also can be a part of a closed-loop automatic control of the OSA therapeutic/assistive devices. In this paper, several solutions for online OSA detection are introduced and tested on 155 subjects of three different databases. The best combinational solution uses mutual information (MI) analysis for selecting out of ECG and SpO2-based features. Several methods of supervised and unsupervised machine learning are employed to detect apnoeic episodes. To achieve the best performance, the most successful classifiers in four different ternary combination methods are used. The proposed configurations exploit limited use of biological signals, have online working scheme, and exhibit uniform and acceptable performance (over 85%) in all the employed databases. The benefits have not been gathered all together in the previous published methods. | ['Hosna Ghandeharioun'] | 2021-10-01 | null | null | null | null | ['sleep-apnea-detection'] | ['medical'] | [ 5.53040430e-02 -2.06204355e-01 -6.61611557e-03 -1.70191273e-01
-1.84478872e-02 -3.72662634e-01 -2.20645607e-01 1.89354181e-01
-4.89016682e-01 1.14727044e+00 5.89634757e-03 -3.09379995e-01
-6.09434843e-01 -4.95957047e-01 5.30581236e-01 -7.26243019e-01
-1.41511515e-01 7.53056467e-01 2.18890309e-01 1.14633562e-03
6.73716366e-02 6.97870910e-01 -1.71959472e+00 -3.04316550e-01
1.12577844e+00 9.14446473e-01 3.87097836e-01 7.95525789e-01
-1.31009566e-02 2.99993772e-02 -6.62522852e-01 1.56209335e-01
2.03167260e-01 -8.11775386e-01 -3.64393741e-01 1.38696253e-01
-3.95051390e-01 -1.94871742e-02 2.16659203e-01 6.90807819e-01
1.02539492e+00 1.53747007e-01 6.42643034e-01 -1.05021942e+00
4.21004772e-01 2.13131189e-01 3.09104547e-02 7.15354919e-01
4.58914667e-01 4.02849108e-01 5.71520984e-01 -4.85639125e-01
-8.53011310e-02 5.55376291e-01 6.35985136e-01 4.95889544e-01
-1.08980656e+00 -6.26658082e-01 -8.38191330e-01 4.11797374e-01
-1.50492895e+00 -6.80269361e-01 6.75331235e-01 -3.44227821e-01
8.30965400e-01 8.79445910e-01 1.22203624e+00 5.22065163e-01
1.83263153e-01 -3.15385580e-01 1.39479351e+00 -2.32039869e-01
2.93462723e-01 7.03490138e-01 8.74055102e-02 7.51837909e-01
8.08755696e-01 -2.72201747e-01 -1.74841121e-01 -1.84988990e-01
1.16290472e-01 4.88163322e-01 -4.84826088e-01 1.32925794e-01
-7.31987417e-01 4.22599673e-01 -9.43717510e-02 8.02270293e-01
-4.16899264e-01 -5.59884846e-01 2.62460470e-01 2.23817959e-01
-1.99919507e-01 4.72787768e-01 -6.36812568e-01 -5.45928478e-01
-1.01887608e+00 -3.65987271e-01 9.43256795e-01 4.83175159e-01
4.08960223e-01 -2.20573187e-01 1.79148023e-03 6.71967149e-01
6.48892164e-01 5.84051967e-01 1.10394847e+00 -6.13285422e-01
3.29637043e-02 9.28757906e-01 2.13975728e-01 -8.06521118e-01
-8.39763522e-01 -5.16067922e-01 -6.98716164e-01 -3.03298775e-02
3.71421054e-02 -7.76333436e-02 -4.13654178e-01 1.09018290e+00
2.06102774e-01 1.24188483e-01 -1.35463834e-01 6.31647348e-01
7.56181240e-01 2.69539177e-01 4.37391140e-02 -9.21492815e-01
1.53658700e+00 -4.38327789e-01 -1.01334751e+00 2.83840261e-02
1.51237220e-01 -9.05000269e-01 7.97304094e-01 5.33224702e-01
-1.13217902e+00 -4.51533496e-01 -1.21001720e+00 2.63132602e-01
-5.90333819e-01 2.93698370e-01 1.45930514e-01 1.16227138e+00
-9.26781416e-01 9.87009943e-01 -1.02704561e+00 -9.83078539e-01
1.41559497e-01 8.48050416e-01 -5.76461926e-02 4.56613421e-01
-8.18889022e-01 9.45575356e-01 1.72254488e-01 1.84549555e-01
-2.42408484e-01 -2.25261256e-01 -4.40496802e-01 7.13375434e-02
3.83076351e-03 -1.11349761e+00 4.64168549e-01 -2.90546924e-01
-1.44194508e+00 9.33702588e-01 -4.32438761e-01 -4.06512022e-01
4.14701432e-01 -1.44596398e-01 -8.41634333e-01 2.88359582e-01
-1.46411195e-01 -1.07260197e-01 7.54258275e-01 -5.08218706e-01
-3.15996557e-01 -7.33682692e-01 -5.83400667e-01 2.58492172e-01
-5.22019625e-01 7.06486031e-02 1.53309822e-01 -9.74074006e-02
3.44412088e-01 -9.11349356e-01 -8.78673717e-02 -2.71304529e-02
-4.24251914e-01 -4.36161645e-02 7.88555801e-01 -4.93336529e-01
1.61174119e+00 -2.11855912e+00 -6.61037769e-03 1.25841409e-01
3.20617616e-01 5.79749405e-01 8.80600214e-01 5.00372529e-01
1.55128971e-01 2.51445085e-01 -3.89820307e-01 -4.97384369e-01
-8.06591585e-02 5.84824145e-01 4.29785013e-01 6.78961754e-01
-2.63375342e-01 2.68275261e-01 -4.44897234e-01 -8.65360141e-01
6.66520119e-01 4.30448830e-01 -3.22169602e-01 5.86307883e-01
3.35939318e-01 8.31142545e-01 -3.62598270e-01 5.28857708e-01
2.89155602e-01 -1.67685449e-01 1.52408004e-01 -2.98411667e-01
-3.72269154e-01 5.02296686e-01 -1.16882598e+00 1.38603723e+00
-3.06002498e-01 3.26875865e-01 -1.17396690e-01 -9.49727535e-01
8.89960229e-01 8.04056108e-01 6.08825028e-01 -4.20978032e-02
5.87567687e-01 4.73275572e-01 2.89780676e-01 -1.32077968e+00
-2.03467458e-01 -4.43996191e-01 5.93368351e-01 2.27505967e-01
-6.92464784e-02 -9.54686701e-02 3.34809661e-01 -4.29889023e-01
9.91960883e-01 -4.33404952e-01 1.10885489e+00 -4.05919820e-01
1.18268061e+00 -5.34270346e-01 4.94655222e-01 2.84607559e-01
-6.68144464e-01 4.29583669e-01 -4.57809232e-02 -2.77119558e-02
-5.39252698e-01 -9.41797018e-01 -3.59808773e-01 7.87200704e-02
3.16594931e-04 -2.78708428e-01 -2.62181878e-01 -2.14430302e-01
-1.48979187e-01 4.88014340e-01 -1.15086257e-01 -1.06597245e-01
-1.72608212e-01 -8.19648623e-01 4.19739068e-01 -7.61146769e-02
5.13543844e-01 -1.03174162e+00 -9.89985526e-01 2.93731898e-01
1.51995849e-02 -7.63725460e-01 8.46925527e-02 1.60627663e-01
-1.39862514e+00 -1.14948392e+00 -5.42478204e-01 -3.51483971e-01
2.71144211e-01 4.45088483e-02 8.25727582e-01 5.02768695e-01
-9.01435733e-01 2.34111413e-01 -1.36765674e-01 -3.48088861e-01
-3.96400779e-01 9.25468951e-02 5.11535108e-01 1.80825651e-01
5.57455242e-01 -1.30636573e+00 -1.09761691e+00 4.26280022e-01
-2.79290229e-01 -6.61385417e-01 5.70838690e-01 3.32565099e-01
3.01030397e-01 6.16509020e-02 6.03641331e-01 -5.22134125e-01
6.26634061e-01 -5.00148594e-01 -4.06356156e-01 -2.95401633e-01
-1.10788500e+00 -3.65253180e-01 6.71301305e-01 1.73227653e-01
-4.50533390e-01 -2.27543131e-01 -3.67984176e-01 -2.64794469e-01
-6.54092014e-01 -6.06235228e-02 -2.11471096e-01 1.12984873e-01
5.65408647e-01 2.77434558e-01 2.74091631e-01 -6.91489935e-01
-4.61693019e-01 1.01504648e+00 2.63088048e-01 7.92074874e-02
7.76159585e-01 4.18967843e-01 3.23562473e-01 -1.47024560e+00
-5.88745534e-01 -1.04994392e+00 -5.95245779e-01 -6.35534376e-02
1.22062409e+00 -4.75425392e-01 -9.49136257e-01 -7.58226812e-02
-4.80310798e-01 2.75529414e-01 -2.58237123e-01 8.32716703e-01
-2.54313260e-01 4.75547284e-01 -6.09581694e-02 -1.32989025e+00
-6.03604019e-01 -1.05491924e+00 3.04583222e-01 4.63512719e-01
-5.09710848e-01 -9.79593098e-01 9.12223086e-02 4.48739827e-01
7.14936078e-01 1.76193997e-01 6.19272172e-01 -8.63547921e-01
-1.53223783e-01 -2.63632178e-01 1.74038440e-01 5.46164572e-01
8.10342014e-01 8.46988559e-02 -9.81665969e-01 -9.22840312e-02
6.08916998e-01 1.99595377e-01 2.87904501e-01 3.12268049e-01
8.43777716e-01 -2.66292185e-01 -3.63766551e-01 4.07047272e-01
1.53414512e+00 7.89763451e-01 5.07115066e-01 1.52902186e-01
1.89243838e-01 2.78659523e-01 3.89637560e-01 7.15096533e-01
-3.47071253e-02 4.47225064e-01 4.12484080e-01 4.82392423e-02
-3.75412479e-02 3.69293272e-01 1.98580533e-01 1.08444476e+00
-3.73741120e-01 -1.67058781e-01 -4.52888459e-01 3.63607079e-01
-1.14823425e+00 -1.09528720e+00 -5.36759079e-01 2.61912966e+00
6.54851019e-01 1.80342391e-01 4.80432808e-01 7.08934605e-01
5.88918567e-01 -2.76832312e-01 -2.73214638e-01 -2.47270793e-01
1.69502750e-01 6.41375959e-01 2.12919876e-01 3.42079133e-01
-7.69761741e-01 6.39521703e-02 5.70763969e+00 -3.69532965e-02
-1.03790617e+00 3.73320609e-01 7.65597448e-02 -1.81181625e-01
3.04782152e-01 -1.03471771e-01 -7.52062261e-01 9.55531895e-01
1.18563259e+00 5.99598624e-02 3.86191785e-01 5.84157646e-01
8.51185381e-01 -2.95926571e-01 -6.68236375e-01 1.30450594e+00
-9.22352076e-02 -7.79132307e-01 -7.04510093e-01 1.76967025e-01
2.44200349e-01 3.12238038e-02 -2.57765263e-01 -1.78883567e-01
-8.67804468e-01 -7.55736351e-01 -2.34662428e-01 5.92589140e-01
8.73113692e-01 -3.61393005e-01 9.08982456e-01 5.20554148e-02
-1.18395650e+00 -3.60074759e-01 -1.08106337e-01 -2.04031408e-01
2.74808258e-01 5.60057819e-01 -1.02609086e+00 3.02659601e-01
5.96410215e-01 5.52778482e-01 -4.84671414e-01 1.54210067e+00
-1.52475789e-01 8.06441247e-01 -8.09404790e-01 -3.69968027e-01
-3.58194858e-01 -5.23920417e-01 9.06804025e-01 7.17906237e-01
5.51171720e-01 1.95701838e-01 -1.20877735e-01 9.29546416e-01
4.17334259e-01 4.29232746e-01 -7.60425389e-01 -8.35775509e-02
4.45850909e-01 1.27755439e+00 -7.68044293e-01 -4.16109800e-01
-3.12768966e-01 7.67466307e-01 -5.49429238e-01 -1.42480597e-01
-6.61146939e-01 -3.14865410e-01 5.07086337e-01 5.27583480e-01
-7.67050087e-02 7.84800872e-02 -3.21847290e-01 -9.59814906e-01
-3.29925239e-01 -5.98756433e-01 6.06886089e-01 -4.57051814e-01
-8.56555700e-01 6.05861604e-01 -1.72535807e-01 -1.51065838e+00
-4.97346073e-02 -3.06685537e-01 -7.94834316e-01 8.79023850e-01
-1.16650844e+00 -3.62237811e-01 -6.67385042e-01 7.54957676e-01
5.90404570e-01 -1.63133278e-01 1.20902526e+00 6.94003105e-01
-8.45974982e-01 9.55534503e-02 -2.72157609e-01 -6.31867349e-01
6.64704084e-01 -1.35108733e+00 -8.45338643e-01 6.83501720e-01
-3.50849420e-01 8.94670486e-01 1.06036019e+00 -5.06133437e-01
-1.06300402e+00 -6.15385056e-01 1.08283770e+00 -1.73873737e-01
5.03932059e-01 2.14150131e-01 -5.04018545e-01 2.12428525e-01
2.27399975e-01 -2.66983241e-01 1.21823418e+00 -3.03141356e-01
7.05466390e-01 -4.39482510e-01 -1.40845919e+00 2.66313940e-01
6.94159687e-01 -2.75354564e-01 -8.36332202e-01 4.41058367e-01
2.13217661e-01 1.71166033e-01 -1.07159626e+00 3.15753013e-01
5.12525737e-01 -1.62036681e+00 7.49136448e-01 -2.77217746e-01
-2.92267054e-01 -3.70390624e-01 1.34722337e-01 -7.11205602e-01
1.32857487e-01 -1.19221032e+00 1.85446113e-01 1.14345634e+00
3.67902756e-01 -1.00480223e+00 5.31839192e-01 5.19058406e-01
-2.89956391e-01 -6.01680696e-01 -7.48619795e-01 -4.92818326e-01
-9.16426182e-01 -1.99785624e-02 1.83514684e-01 7.28533149e-01
1.03855148e-01 4.66953546e-01 -1.22250251e-01 1.39106512e-01
4.32121754e-01 -1.47729769e-01 5.25051713e-01 -1.79098833e+00
-4.45371985e-01 -3.04237843e-01 -7.19048321e-01 -9.76509973e-03
-6.11104429e-01 -5.95073342e-01 -4.44277972e-01 -1.65585136e+00
-2.84468472e-01 -3.36939782e-01 -5.68089604e-01 2.98298243e-02
5.28554469e-02 3.63397747e-01 -3.15909535e-01 2.52972811e-01
4.13756818e-02 2.48528868e-01 8.33787024e-01 3.50992620e-01
-7.23732173e-01 7.97423244e-01 -1.99528828e-01 8.66878986e-01
1.16575301e+00 -6.59065545e-01 -8.41107368e-01 5.51076531e-01
-2.75645573e-02 2.02946529e-01 -5.14828973e-02 -1.47395861e+00
7.01140165e-02 2.61468083e-01 1.38137117e-01 -5.78837216e-01
7.50903368e-01 -1.28746951e+00 5.58390975e-01 8.99238825e-01
2.84237802e-01 2.32418031e-01 -5.22693098e-02 3.94525111e-01
6.32991344e-02 -5.43280661e-01 1.04633152e+00 -4.13574427e-01
2.45980173e-03 1.22275785e-01 -6.90224469e-01 -5.72878495e-02
9.05930579e-01 -4.62268293e-01 1.97945982e-02 -2.44047537e-01
-1.32313585e+00 -3.13883543e-01 3.37627560e-01 -2.37746224e-01
3.92919272e-01 -6.19336784e-01 -8.75255987e-02 3.81041706e-01
5.96410595e-02 -4.62960333e-01 2.61832178e-01 1.83957422e+00
-7.49676824e-01 5.05222857e-01 -2.77142376e-01 -5.21341026e-01
-1.53723061e+00 4.98093814e-01 5.37641346e-01 -5.26721142e-02
-6.12354398e-01 5.32458782e-01 -6.72354698e-01 3.16184193e-01
2.00150236e-01 -3.48981857e-01 -6.63620412e-01 8.80656689e-02
3.87157589e-01 7.40358174e-01 1.35879591e-01 -3.42872590e-01
-4.77491081e-01 6.69781387e-01 4.91631508e-01 -8.14868808e-02
9.86862242e-01 -5.62899590e-01 -5.19420505e-01 8.23766470e-01
1.00909185e+00 5.38480878e-01 -1.23371258e-01 5.31491876e-01
-3.97656411e-02 -2.59874582e-01 -1.71210364e-01 -7.48263359e-01
-7.31377900e-01 8.77191782e-01 1.15983343e+00 7.77319372e-01
1.33464432e+00 -3.92317325e-01 6.42355740e-01 2.56913155e-01
2.59870380e-01 -7.59032309e-01 -4.17989433e-01 -5.75172007e-01
4.33743209e-01 -1.07614458e+00 1.51140749e-01 -3.79404128e-01
-3.23412180e-01 1.07109010e+00 3.00269097e-01 -4.54606749e-02
1.10752594e+00 -7.64747486e-02 -5.46892453e-03 -3.36047769e-01
-5.02557516e-01 -5.09065986e-01 1.31314903e-01 4.72913593e-01
4.75694656e-01 7.48031959e-02 -1.14742100e+00 4.24507499e-01
-2.41853312e-01 1.51920751e-01 4.44301218e-01 7.61379540e-01
-6.95378482e-01 -1.09764636e+00 -2.77546078e-01 7.88870215e-01
-8.48747492e-01 5.43281399e-02 -4.00189273e-02 6.96079791e-01
3.94330353e-01 1.39865088e+00 -2.35054821e-01 -2.49704033e-01
1.95292115e-01 5.20439148e-01 3.10285836e-01 -5.31212628e-01
-5.51742375e-01 2.17334956e-01 2.91543901e-01 -5.50556242e-01
-7.44835138e-01 -6.36610210e-01 -1.22515881e+00 1.82302654e-01
-3.56244266e-01 6.82499826e-01 9.03980017e-01 1.03223181e+00
3.30292523e-01 5.23286819e-01 6.01080477e-01 -3.93780261e-01
-3.71046513e-01 -1.27508569e+00 -8.63712788e-01 2.44363144e-01
4.38570857e-01 -9.13442910e-01 -1.00447941e+00 -1.09695494e-01] | [13.674422264099121, 3.3322713375091553] |
8ab96aa0-db5c-4ab3-9ec3-e08d1059d856 | refining-generative-process-with | 2211.17091 | null | https://arxiv.org/abs/2211.17091v4 | https://arxiv.org/pdf/2211.17091v4.pdf | Refining Generative Process with Discriminator Guidance in Score-based Diffusion Models | The proposed method, Discriminator Guidance, aims to improve sample generation of pre-trained diffusion models. The approach introduces a discriminator that gives explicit supervision to a denoising sample path whether it is realistic or not. Unlike GANs, our approach does not require joint training of score and discriminator networks. Instead, we train the discriminator after score training, making discriminator training stable and fast to converge. In sample generation, we add an auxiliary term to the pre-trained score to deceive the discriminator. This term corrects the model score to the data score at the optimal discriminator, which implies that the discriminator helps better score estimation in a complementary way. Using our algorithm, we achive state-of-the-art results on ImageNet 256x256 with FID 1.83 and recall 0.64, similar to the validation data's FID (1.68) and recall (0.66). We release the code at https://github.com/alsdudrla10/DG. | ['Il-Chul Moon', 'Wanmo Kang', 'Se Jung Kwon', 'Yeongmin Kim', 'Dongjun Kim'] | 2022-11-28 | null | null | null | null | ['conditional-image-generation'] | ['computer-vision'] | [ 2.32536286e-01 2.52108067e-01 -2.96818227e-01 -4.67472553e-01
-1.09492803e+00 -4.93340552e-01 6.30256832e-01 -2.79116541e-01
-5.22434056e-01 8.19269836e-01 6.47545829e-02 -6.43447563e-02
2.06317872e-01 -7.84408987e-01 -4.63333696e-01 -1.00521290e+00
6.59398288e-02 4.75728780e-01 1.78166673e-01 1.18176481e-02
1.97820634e-01 1.76024228e-01 -8.80438745e-01 3.02383214e-01
1.07819438e+00 1.18402731e+00 9.30046290e-02 1.02428806e+00
3.46041679e-01 8.78443122e-01 -9.88695443e-01 -6.17159426e-01
4.39564377e-01 -8.31098974e-01 -4.51505035e-01 -1.38422564e-01
5.80876648e-01 -6.47563219e-01 -2.88238019e-01 1.44011593e+00
5.26417851e-01 -8.62992778e-02 8.50708961e-01 -1.08121431e+00
-9.18183148e-01 5.67521691e-01 -7.02438831e-01 2.28241950e-01
-2.15621203e-01 4.17159349e-01 7.73123920e-01 -7.04893947e-01
6.11656487e-01 1.00862992e+00 6.09920979e-01 1.01234126e+00
-1.28138554e+00 -1.14252162e+00 -7.47062638e-02 -9.89808068e-02
-1.18543732e+00 -3.87460887e-01 7.74105906e-01 -4.03843492e-01
5.80534458e-01 3.87118338e-03 7.13869035e-01 1.55523133e+00
2.12424859e-01 8.22379529e-01 1.14877534e+00 -4.80358787e-02
3.57940882e-01 1.96091741e-01 -1.81322709e-01 6.67340934e-01
8.12893510e-02 4.25040126e-01 -3.39719027e-01 -5.74991256e-02
1.02964497e+00 -2.92485744e-01 -2.33423531e-01 -7.27244839e-02
-8.32273364e-01 1.13920748e+00 7.75281250e-01 1.92176536e-01
-5.94405890e-01 4.80920345e-01 2.16920421e-01 5.07246196e-01
6.68905795e-01 3.08991849e-01 -2.52040923e-01 -5.75421304e-02
-1.36579645e+00 1.41231090e-01 5.69660008e-01 4.83702153e-01
5.03575444e-01 4.58725035e-01 -4.58933294e-01 8.08934748e-01
2.90579230e-01 7.29309082e-01 4.96615857e-01 -1.33135259e+00
4.38188553e-01 1.50063604e-01 -1.00110069e-01 -5.89053035e-01
7.77130276e-02 -9.66403961e-01 -1.21391642e+00 6.01455331e-01
5.23582518e-01 -3.86159867e-01 -1.27144313e+00 1.70992148e+00
8.33168179e-02 3.00847083e-01 9.61954147e-02 8.48071456e-01
5.42403281e-01 6.01359844e-01 -1.34544866e-02 1.58656284e-01
8.20097089e-01 -1.19145715e+00 -3.77663702e-01 -3.28297168e-01
6.21566892e-01 -7.25981116e-01 1.07149899e+00 8.19362283e-01
-1.16964173e+00 -4.98049676e-01 -1.16178715e+00 3.62389296e-01
6.05834015e-02 3.24039668e-01 4.39398646e-01 7.30724871e-01
-1.29442012e+00 8.30422103e-01 -8.58523965e-01 1.33855894e-01
8.85286808e-01 3.86578292e-01 -1.15873039e-01 -4.11701202e-03
-9.88471389e-01 7.31518090e-01 1.90245017e-01 -2.28724450e-01
-1.37844014e+00 -7.12338924e-01 -5.54332972e-01 -1.21009260e-01
-6.34492040e-02 -6.02090299e-01 1.35448039e+00 -1.38465500e+00
-1.75340736e+00 6.41432226e-01 8.57839435e-02 -7.71121144e-01
8.43952894e-01 -8.26303214e-02 -2.49023870e-01 2.02669621e-01
3.39584835e-02 1.16828871e+00 1.01413274e+00 -1.31503022e+00
-3.54998142e-01 -9.79559422e-02 4.31161299e-02 2.30998565e-02
-2.31249735e-01 -3.17435682e-01 -2.05096856e-01 -9.00949299e-01
-2.91107237e-01 -6.95116222e-01 -2.58420527e-01 1.57266989e-01
-3.56205046e-01 8.04834440e-02 5.33891857e-01 -9.50892925e-01
1.21008646e+00 -2.04680991e+00 -3.50563116e-02 4.24643070e-01
3.44619513e-01 4.03535396e-01 -5.91579676e-01 -3.76730487e-02
1.07316427e-01 6.44717067e-02 -5.82264304e-01 -5.15032947e-01
-2.59784579e-01 3.07691753e-01 2.18072366e-02 4.38674271e-01
5.37634790e-01 8.43196154e-01 -1.03378963e+00 -1.57244429e-01
-4.39930372e-02 8.14096928e-01 -8.54934752e-01 9.52692181e-02
-1.38121128e-01 6.88905656e-01 -2.11987138e-01 4.53667432e-01
8.62190783e-01 -3.00538838e-01 6.51516486e-03 9.37336162e-02
3.19222033e-01 1.87783003e-01 -1.08338726e+00 1.68822551e+00
-5.05906045e-01 6.54412508e-01 7.42105395e-02 -7.60319173e-01
1.17444360e+00 8.22907314e-02 1.01904877e-01 -7.18868077e-01
1.12087250e-01 2.86041319e-01 2.10817739e-01 -1.30223297e-02
3.64980936e-01 -7.79082403e-02 3.01987469e-01 5.42241633e-01
2.26885691e-01 -1.31358668e-01 1.14171743e-01 2.89638579e-01
1.23572707e+00 2.30025351e-02 1.93205327e-02 -1.79394647e-01
4.46974784e-01 -2.17580453e-01 4.54306155e-01 7.89245188e-01
-1.81179702e-01 8.13474000e-01 6.22269630e-01 -6.22740053e-02
-1.16358888e+00 -1.17862487e+00 6.19488396e-02 7.28181422e-01
-1.68465346e-01 -1.31931067e-01 -1.13530970e+00 -1.21096563e+00
-3.11493069e-01 6.89996123e-01 -1.00554216e+00 -2.13983774e-01
-5.22970498e-01 -7.17466712e-01 6.77950621e-01 6.68993354e-01
7.57844865e-01 -9.38585579e-01 -2.65636563e-01 6.10773750e-02
-7.36150146e-03 -4.02864188e-01 -6.78441942e-01 3.00686836e-01
-9.85074043e-01 -8.89652431e-01 -9.94157314e-01 -4.70398247e-01
8.68178666e-01 -1.71682924e-01 1.25843716e+00 3.09706628e-01
2.47490317e-01 -1.48793772e-01 -2.25992858e-01 -1.66189879e-01
-7.02185035e-01 4.08796489e-01 -4.53889489e-01 -2.43147373e-01
1.50002301e-01 -5.67690909e-01 -1.08366060e+00 3.02480072e-01
-8.98187041e-01 -1.46804556e-01 7.57287979e-01 1.11417747e+00
5.42227149e-01 -4.09566522e-01 7.23470628e-01 -8.56063902e-01
7.45998025e-01 -4.54205364e-01 -6.15984678e-01 1.81403924e-02
-9.68300402e-01 2.01611832e-01 6.75625920e-01 -7.12797046e-01
-9.66092825e-01 -1.48224115e-01 -6.11800432e-01 -4.24036413e-01
-1.65149122e-02 1.82672128e-01 1.58814475e-01 -2.66710110e-02
9.90779400e-01 5.16497791e-02 1.09176509e-01 -3.20706189e-01
2.76240915e-01 5.41646957e-01 4.98257875e-01 -3.30478221e-01
6.47676826e-01 4.54597592e-01 -2.55593777e-01 -1.31662339e-01
-8.62201989e-01 -1.99159030e-02 -4.25766289e-01 -1.24179915e-01
6.75714195e-01 -8.64619195e-01 -1.68283671e-01 4.96316969e-01
-1.10958767e+00 -1.00517797e+00 -6.54056430e-01 3.98090333e-01
-4.39572752e-01 -6.33382648e-02 -7.25163400e-01 -6.86432302e-01
-6.12290561e-01 -1.06858492e+00 9.08889115e-01 1.22084670e-01
-2.89977878e-01 -1.19741106e+00 3.02415133e-01 1.39905423e-01
6.88184738e-01 2.01918587e-01 4.99030590e-01 -5.06587982e-01
-4.12310690e-01 -6.78204447e-02 -1.83935240e-01 1.10519910e+00
-2.25312784e-02 -4.36031409e-02 -1.15343094e+00 -4.44106668e-01
7.78572038e-02 -4.42354292e-01 1.19461703e+00 5.17413795e-01
9.96952534e-01 -4.41514641e-01 1.11316539e-01 6.65019989e-01
1.33687568e+00 2.33965337e-01 8.50262523e-01 2.18461618e-01
3.56696367e-01 2.14976341e-01 3.87251586e-01 4.57356840e-01
1.01578474e-01 3.34563166e-01 5.36968887e-01 -1.85271680e-01
-6.41732514e-01 -2.96774894e-01 5.96534789e-01 7.85858452e-01
-2.91362163e-02 -3.38645160e-01 -6.92807794e-01 3.78715664e-01
-1.49894297e+00 -1.06578648e+00 -6.59068599e-02 2.11000133e+00
1.14266539e+00 2.17302486e-01 8.52192417e-02 4.90569174e-02
5.92675388e-01 2.48919919e-01 -7.56366074e-01 -2.71186620e-01
-8.88283625e-02 5.52392960e-01 5.68404913e-01 9.28122342e-01
-8.33460867e-01 8.56893063e-01 6.51920557e+00 1.11854100e+00
-1.18182707e+00 4.27007765e-01 1.08581519e+00 -1.86061695e-01
-4.58866924e-01 -2.09440246e-01 -5.76582730e-01 6.69833541e-01
9.53701138e-01 5.59414774e-02 6.23292148e-01 8.14140499e-01
1.06893048e-01 -9.18972269e-02 -8.24685991e-01 6.79531932e-01
8.67793486e-02 -1.24737835e+00 4.48371843e-02 2.40652636e-01
1.09811318e+00 1.95953190e-01 5.20139396e-01 2.06956968e-01
6.77790940e-01 -1.06737554e+00 7.62900472e-01 4.74121600e-01
7.65076697e-01 -7.88648307e-01 8.64794075e-01 2.09315807e-01
-6.97409630e-01 7.45467171e-02 -4.16178674e-01 1.00150369e-01
5.12911268e-02 6.93573117e-01 -9.31787550e-01 2.36033872e-01
3.91441137e-01 7.44228303e-01 -7.03583956e-01 9.36004579e-01
-6.93935215e-01 8.23204279e-01 -1.51868433e-01 2.96312988e-01
3.96051168e-01 -2.41473496e-01 3.03420663e-01 1.07601166e+00
5.85573375e-01 -4.16834146e-01 -1.50998503e-01 8.28682601e-01
-3.88609201e-01 -2.87745029e-01 -3.14670235e-01 2.99397618e-01
4.46832061e-01 1.17168832e+00 -5.54472208e-01 -5.80349743e-01
-2.27021053e-02 1.19072938e+00 2.84330010e-01 4.67415601e-01
-9.97557282e-01 -1.85948715e-01 6.06098473e-01 1.05591804e-01
4.09386218e-01 -3.55607346e-02 -5.34237146e-01 -9.11920846e-01
-1.47993669e-01 -1.01060295e+00 2.96741664e-01 -7.21041799e-01
-1.41016817e+00 8.18897605e-01 -3.21324229e-01 -1.24328947e+00
-5.87291718e-01 -4.34567600e-01 -7.92274535e-01 9.18100417e-01
-1.58257604e+00 -9.09739733e-01 -4.67815816e-01 5.01087368e-01
4.72266287e-01 -1.08367734e-01 6.35272264e-01 3.95213604e-01
-2.60031044e-01 9.32810545e-01 1.40620261e-01 2.05298752e-01
9.33885574e-01 -1.34192944e+00 4.74813551e-01 8.65708947e-01
7.79262977e-03 3.68859649e-01 5.85264087e-01 -7.31758773e-01
-4.90000725e-01 -1.13952720e+00 5.18953741e-01 -3.92894953e-01
6.49773538e-01 -1.54115573e-01 -9.13836777e-01 5.28948069e-01
4.49912131e-01 -1.98371913e-02 5.97919405e-01 -2.57765681e-01
-4.56257790e-01 -1.49722278e-01 -1.47069621e+00 3.01124454e-01
9.17387068e-01 -3.32493395e-01 -8.60470533e-02 1.13501467e-01
3.54430109e-01 -3.99259031e-01 -8.05399954e-01 1.75178856e-01
5.81515849e-01 -1.07822835e+00 7.92392790e-01 -2.60351896e-01
6.57794297e-01 -1.54651165e-01 -3.96323502e-02 -1.68255091e+00
-2.76269138e-01 -3.45567435e-01 -4.47509795e-01 1.10605896e+00
6.51928842e-01 -6.93730295e-01 1.12597740e+00 1.69133395e-01
-3.16718780e-02 -9.55697358e-01 -8.93820226e-01 -9.80754733e-01
5.33497691e-01 -2.39314392e-01 6.38876915e-01 8.16244900e-01
-7.14882314e-01 1.09205805e-01 -5.06998718e-01 -5.24236485e-02
8.94643426e-01 -4.59682047e-01 6.14447773e-01 -1.17018533e+00
-6.91323459e-01 -5.01492381e-01 -8.32212567e-02 -1.32956100e+00
-9.87256970e-03 -9.92927670e-01 -8.87530893e-02 -1.51520491e+00
1.57143891e-01 -5.92497706e-01 -3.14936101e-01 5.75868189e-01
-1.57981321e-01 6.58712983e-01 1.68799147e-01 3.39579850e-01
-3.90158862e-01 6.49965882e-01 1.55612516e+00 -3.76609057e-01
-2.88218707e-01 -7.99405854e-03 -8.64729822e-01 6.22565448e-01
1.42028725e+00 -9.30165887e-01 -5.48806965e-01 -7.43890584e-01
-4.23822179e-02 -2.62813240e-01 5.56931078e-01 -1.04330719e+00
9.74635258e-02 1.43274814e-01 6.48883820e-01 -2.84650266e-01
3.86379331e-01 -5.06269336e-01 1.11186460e-01 8.81971180e-01
-4.27889198e-01 4.85750549e-02 -1.81986928e-01 2.99807042e-01
-2.82795191e-01 -5.82033157e-01 1.17776036e+00 -1.25137940e-02
-2.50269920e-01 3.93657267e-01 -1.45880327e-01 2.07797408e-01
7.23139882e-01 -1.63328633e-01 -5.37138641e-01 -5.86544394e-01
-6.69020832e-01 4.34840359e-02 5.54813683e-01 6.59926310e-02
7.22062647e-01 -1.46695137e+00 -8.42288017e-01 3.56478989e-01
-2.33852014e-01 -1.74448676e-02 1.33306265e-01 8.62233639e-01
-4.02806967e-01 -2.21000597e-01 -1.57494396e-01 -5.09697378e-01
-9.56155837e-01 4.11933325e-02 4.01168257e-01 -6.40924215e-01
-3.99442136e-01 1.02159059e+00 -4.33223695e-03 -2.52827018e-01
2.35879168e-01 -2.74746239e-01 4.98434491e-02 -2.02135276e-02
6.93315983e-01 3.94404411e-01 -3.46520990e-02 -2.49549225e-01
-6.06351756e-02 3.42053413e-01 -2.05167159e-01 -3.99984956e-01
1.34954166e+00 1.75114766e-01 2.35765234e-01 -6.90223426e-02
1.14007521e+00 -1.64991185e-01 -1.84575129e+00 2.09080558e-02
-2.69970268e-01 -4.98823136e-01 2.45088533e-01 -1.17111838e+00
-1.51343751e+00 8.53863418e-01 9.30907130e-01 7.86360055e-02
1.16974890e+00 -9.41886902e-02 8.35452914e-01 -2.41324143e-03
7.68642798e-02 -9.89802122e-01 5.72920918e-01 5.28123736e-01
9.41322744e-01 -1.23412347e+00 -1.79832190e-01 -6.45774007e-02
-7.93289542e-01 8.00106704e-01 6.09655321e-01 -4.91379082e-01
4.99185413e-01 4.64631140e-01 3.22120070e-01 1.54902879e-02
-7.37005532e-01 7.26030320e-02 1.77217364e-01 8.76076043e-01
3.66006136e-01 -1.12202831e-01 -1.61597133e-01 3.91647756e-01
-3.12246978e-01 1.55060381e-01 3.96614671e-01 6.07831776e-01
-3.48705381e-01 -1.27730846e+00 -1.02834098e-01 6.19131744e-01
-5.42392373e-01 -3.67046148e-01 -2.31853604e-01 5.86513102e-01
1.30945444e-01 9.26620781e-01 1.42556414e-01 -4.73712534e-01
1.91061392e-01 -1.14242822e-01 5.63380122e-01 -3.37101132e-01
-7.83694386e-01 -1.48142958e-02 -1.02116816e-01 -4.99671966e-01
-3.79244119e-01 -3.70839924e-01 -1.01989329e+00 -5.19248188e-01
-3.10654879e-01 1.58310756e-01 5.28148234e-01 4.42180932e-01
2.80852050e-01 5.26901245e-01 6.45771861e-01 -7.45406628e-01
-7.78983831e-01 -1.12492418e+00 -4.54557091e-01 2.98356712e-01
2.90278643e-01 -3.25164914e-01 -5.39743483e-01 -3.62460427e-02] | [11.52497386932373, -0.15977443754673004] |
775ab5cf-e51d-474b-b8fc-d8f245aa2cea | improving-generalization-in-language-model | 2305.17378 | null | https://arxiv.org/abs/2305.17378v1 | https://arxiv.org/pdf/2305.17378v1.pdf | Improving Generalization in Language Model-Based Text-to-SQL Semantic Parsing: Two Simple Semantic Boundary-Based Techniques | Compositional and domain generalization present significant challenges in semantic parsing, even for state-of-the-art semantic parsers based on pre-trained language models (LMs). In this study, we empirically investigate improving an LM's generalization in semantic parsing with two simple techniques: at the token level, we introduce a token preprocessing method to preserve the semantic boundaries of tokens produced by LM tokenizers; at the sequence level, we propose to use special tokens to mark the boundaries of components aligned between input and output. Our experimental results on two text-to-SQL semantic parsing datasets show that our token preprocessing, although simple, can substantially improve the LM performance on both types of generalization, and our component boundary marking method is particularly helpful for compositional generalization. | ['Ziyu Yao', 'Yilun Zhou', 'Bailin Wang', 'Daking Rai'] | 2023-05-27 | null | null | null | null | ['text-to-sql', 'domain-generalization', 'semantic-parsing'] | ['computer-code', 'methodology', 'natural-language-processing'] | [ 4.78852391e-01 5.50050676e-01 -2.39724338e-01 -5.01297951e-01
-7.78806090e-01 -9.62188303e-01 3.37970555e-01 4.65609938e-01
-2.31535167e-01 4.26522702e-01 1.05731241e-01 -7.37165272e-01
4.30547446e-01 -9.96663988e-01 -8.94661009e-01 -3.06553021e-02
1.91155389e-01 2.82274723e-01 6.73355937e-01 -1.56438068e-01
2.85838842e-01 1.67544454e-01 -1.18772054e+00 8.47224176e-01
1.05559814e+00 8.32864642e-01 3.37623835e-01 4.68042850e-01
-1.39414024e+00 8.04702640e-01 -7.66201198e-01 -5.27775764e-01
6.95726499e-02 -5.47066092e-01 -1.50614798e+00 -1.37317032e-01
2.46371269e-01 1.03178218e-01 1.96361661e-01 1.11293805e+00
1.16385050e-01 5.42607717e-02 3.65117222e-01 -1.08847988e+00
-6.95733666e-01 1.33198655e+00 -1.90080926e-01 -3.33020017e-02
3.50931525e-01 -1.72857568e-01 1.24483156e+00 -6.06359661e-01
7.16234982e-01 1.63104713e+00 7.31676340e-01 1.01841962e+00
-1.21947038e+00 -5.08716166e-01 6.48123860e-01 8.45966190e-02
-9.02482390e-01 -2.56937683e-01 6.97220683e-01 -1.87514365e-01
1.33999205e+00 -1.04564317e-01 -9.49808583e-03 7.45207250e-01
-2.62458950e-01 8.60874176e-01 9.71471131e-01 -8.25824201e-01
3.68801147e-01 6.68800697e-02 3.77842873e-01 6.58272624e-01
1.38635293e-01 -4.54151481e-01 -5.27021468e-01 -9.16393325e-02
6.23078227e-01 -5.72168589e-01 2.24545479e-01 -1.43913835e-01
-6.65062487e-01 8.75124693e-01 9.18453485e-02 4.07037497e-01
1.51346311e-01 2.12711334e-01 9.78955388e-01 2.38495529e-01
4.81479555e-01 3.70397985e-01 -8.35560203e-01 -2.81446785e-01
-7.89001048e-01 9.30099264e-02 9.14170504e-01 1.43984127e+00
8.05652440e-01 -1.64124817e-02 3.76125425e-02 1.01537728e+00
-2.28132978e-02 1.96357772e-01 6.93633437e-01 -9.60512996e-01
8.25366080e-01 8.69537175e-01 -2.39182532e-01 -2.45084599e-01
-2.59543121e-01 -1.29043788e-01 -6.01118281e-02 -3.06363791e-01
5.03569067e-01 -7.16674477e-02 -1.04093897e+00 1.96284473e+00
2.06802249e-01 8.65309872e-03 4.88382041e-01 3.06425095e-01
9.10358667e-01 4.72153872e-01 9.19867694e-01 2.01994866e-01
1.66231406e+00 -8.99659932e-01 -4.96213794e-01 -6.46442771e-01
1.32714641e+00 -6.25133872e-01 1.52813339e+00 -1.09941959e-01
-1.24171555e+00 -5.83615005e-01 -9.34073269e-01 -3.16391945e-01
-6.17457807e-01 -1.51098648e-03 5.36720097e-01 8.60105991e-01
-9.24891412e-01 8.09332669e-01 -1.02656782e+00 -5.08444786e-01
2.84957469e-01 8.46457027e-04 -1.59484416e-01 -2.70907730e-01
-1.19771469e+00 7.71913052e-01 7.56317854e-01 -3.37118536e-01
-3.74079913e-01 -7.15098977e-01 -1.19445193e+00 3.52248579e-01
4.94794637e-01 -6.14521146e-01 1.63262331e+00 -1.12904048e+00
-1.53633165e+00 1.09684491e+00 -6.08337224e-01 -4.07365978e-01
2.54619569e-01 -3.82026315e-01 -1.74029872e-01 3.47195089e-01
3.06876212e-01 8.17562759e-01 4.96455669e-01 -1.44040012e+00
-7.87163436e-01 -4.10248965e-01 2.26030558e-01 8.13163165e-03
-3.59540969e-01 4.01268482e-01 -3.76741827e-01 -7.28754282e-01
4.66793150e-01 -5.65053105e-01 -2.18717858e-01 -4.59839791e-01
-3.07127655e-01 -5.83695769e-01 7.25694954e-01 -8.22497010e-01
1.29707730e+00 -2.23779416e+00 -9.31057706e-02 4.06676605e-02
-3.18660408e-01 2.33609840e-01 -1.33790180e-01 5.28322756e-01
-1.41900241e-01 5.54692864e-01 -5.91788530e-01 -6.79988503e-01
1.07694104e-01 5.61441302e-01 -5.42627096e-01 -3.38080496e-01
5.58993280e-01 1.01995754e+00 -9.99220967e-01 -6.57113850e-01
-3.50493602e-02 -1.83012217e-01 -6.67288244e-01 3.22556019e-01
-7.01196730e-01 7.18862414e-02 -4.15705383e-01 6.82779670e-01
5.45811832e-01 6.52683452e-02 7.16083586e-01 1.90929249e-01
1.88437432e-01 1.04873312e+00 -1.07303691e+00 1.97370958e+00
-6.93261981e-01 9.41273868e-02 -8.76702219e-02 -1.29702818e+00
8.66688728e-01 1.97617084e-01 1.88904721e-02 -5.83512247e-01
-2.57217377e-01 4.87489253e-01 -2.55156815e-01 -3.75422984e-01
7.64260709e-01 -3.80159080e-01 -4.35517102e-01 3.94886583e-01
6.32686689e-02 -2.20259860e-01 2.14232564e-01 2.59735018e-01
1.07882392e+00 6.32150948e-01 8.34215060e-02 -3.55022252e-01
5.58213651e-01 3.80153328e-01 9.21866715e-01 7.49200583e-01
1.75296158e-01 4.37039167e-01 8.91285181e-01 -1.00890256e-01
-9.02916491e-01 -1.23746765e+00 2.22670939e-02 1.52037001e+00
2.94657331e-02 -7.60090411e-01 -1.30362713e+00 -1.17931664e+00
-8.53107125e-02 1.10728967e+00 -1.23160630e-01 -2.27752358e-01
-1.24015272e+00 -4.86659318e-01 9.91425157e-01 1.28764963e+00
4.68545079e-01 -1.13502717e+00 -4.23077255e-01 4.95319456e-01
-2.49811262e-01 -1.67143679e+00 -1.08543798e-01 4.54114199e-01
-1.22823930e+00 -9.42878962e-01 -8.21882933e-02 -1.18077636e+00
6.83545649e-01 -3.98232155e-02 1.19634175e+00 3.07049364e-01
1.61700740e-01 7.94520229e-02 -7.75270045e-01 -4.39988405e-01
-8.73297274e-01 2.93815136e-01 -4.58578140e-01 -7.78528392e-01
5.08596480e-01 -4.13692236e-01 -2.31305566e-02 1.68653697e-01
-1.10144496e+00 -7.29051307e-02 1.53499365e-01 6.30449951e-01
3.45599473e-01 -3.29095647e-02 5.79705119e-01 -1.43977392e+00
5.19276977e-01 -1.79547921e-01 -4.72189397e-01 4.16424096e-01
-2.84470558e-01 4.58092988e-01 1.17763758e+00 -2.09524795e-01
-1.34647965e+00 -6.90947771e-02 -3.59760284e-01 2.13069037e-01
-3.82982999e-01 4.12753791e-01 -5.17508686e-01 2.78968304e-01
4.06872481e-01 9.46308821e-02 -2.62611389e-01 -8.82749021e-01
6.77237689e-01 6.03662729e-01 7.96303868e-01 -1.51871896e+00
5.88720381e-01 2.73476720e-01 -5.44976741e-02 -5.56098163e-01
-7.10297108e-01 -3.98385108e-01 -7.49103308e-01 5.69385171e-01
8.36067080e-01 -7.77544618e-01 -2.76180834e-01 4.00049269e-01
-1.28825998e+00 -6.48164690e-01 -4.62736011e-01 -3.88689786e-02
-5.35091639e-01 7.58440852e-01 -9.73024189e-01 -4.22421545e-01
-2.87565768e-01 -9.67237890e-01 1.16230106e+00 3.40708233e-02
-3.80271614e-01 -1.25766027e+00 -3.77090871e-01 1.61778152e-01
2.71783501e-01 -1.35252386e-01 1.84919548e+00 -1.00277543e+00
-3.39746147e-01 2.73423463e-01 -3.33855540e-01 6.27093375e-01
1.17611311e-01 -2.74689674e-01 -9.40201819e-01 8.53766873e-02
-1.72258466e-01 -1.67175040e-01 7.96482503e-01 6.06596097e-02
1.38905585e+00 -1.80359244e-01 -3.43607605e-01 5.99985421e-01
1.38972652e+00 9.06365663e-02 5.89534044e-01 5.70663333e-01
6.78885341e-01 7.28404641e-01 5.72066426e-01 6.55488297e-02
5.89264214e-01 3.45510155e-01 9.86254439e-02 -3.49287689e-02
-3.76485795e-01 -7.55904257e-01 4.88763988e-01 6.34864032e-01
3.00352782e-01 -3.90018448e-02 -8.11616421e-01 5.71448803e-01
-1.77736354e+00 -4.51971352e-01 -3.91357690e-02 2.06264663e+00
1.10673654e+00 3.37913126e-01 -7.09196553e-02 6.59132227e-02
7.68553495e-01 -5.36866523e-02 -2.50295579e-01 -9.92633939e-01
-4.71884981e-02 8.77342701e-01 6.33430004e-01 4.34080124e-01
-8.44761968e-01 1.94326925e+00 6.46919012e+00 8.70647728e-01
-8.70177507e-01 1.53933868e-01 1.91745967e-01 4.43090618e-01
-4.70552951e-01 4.71459955e-01 -1.22067058e+00 3.41600329e-01
9.60949421e-01 4.97093834e-02 1.58478752e-01 1.10074782e+00
-2.81853080e-02 -4.22550738e-02 -1.38690305e+00 3.82890493e-01
-2.78748482e-01 -1.27555346e+00 4.27560002e-01 -5.63508689e-01
4.20563966e-01 -4.50772166e-01 -4.01301086e-01 4.58994180e-01
5.13027430e-01 -8.36409628e-01 8.95326316e-01 -3.53177845e-01
7.08370745e-01 -6.80551767e-01 7.01776266e-01 1.38956651e-01
-1.46406555e+00 -1.85389072e-02 -5.92628002e-01 -5.93549721e-02
3.10412616e-01 3.69144678e-01 -8.39207053e-01 8.01153243e-01
5.02602458e-01 4.32718784e-01 -5.60196400e-01 4.89212632e-01
-8.68199825e-01 8.65667582e-01 -1.23722240e-01 1.37536466e-01
2.84340709e-01 4.48799655e-02 2.35403597e-01 1.62490726e+00
8.56237113e-02 3.34187746e-02 3.56676608e-01 8.14185679e-01
-1.11413613e-01 1.67861998e-01 -1.75308228e-01 -1.47400782e-01
7.05930471e-01 8.75408232e-01 -9.67724442e-01 -6.98666334e-01
-5.58053017e-01 1.02266550e+00 4.85106498e-01 3.00778508e-01
-6.48986578e-01 -5.93292415e-01 7.19287694e-01 -2.80389413e-02
3.98727506e-01 -4.00142282e-01 -7.95170963e-01 -9.82598126e-01
4.32431161e-01 -7.34141707e-01 7.91982770e-01 -5.24653971e-01
-1.16959989e+00 2.84516841e-01 1.60882279e-01 -4.22709674e-01
-3.19931537e-01 -9.07339752e-01 -8.33371222e-01 8.99387598e-01
-1.68875432e+00 -1.36026371e+00 1.48402005e-01 3.36633146e-01
7.18627632e-01 2.34743401e-01 1.01876187e+00 1.06288996e-02
-3.54006022e-01 8.48350286e-01 -2.67156035e-01 4.94774610e-01
4.63305533e-01 -1.43870580e+00 1.06977093e+00 1.11841631e+00
-1.06165297e-02 8.73549700e-01 5.58331192e-01 -6.36140227e-01
-1.14389634e+00 -1.04790998e+00 1.15947056e+00 -2.04861194e-01
6.05368018e-01 -5.27834177e-01 -1.22777164e+00 1.10490513e+00
-1.05133235e-01 -2.96828300e-01 5.02734542e-01 3.92278641e-01
-6.13097966e-01 2.99889028e-01 -1.08455181e+00 3.74527723e-01
1.33670533e+00 -7.71264195e-01 -1.10977316e+00 2.96811908e-01
1.12441635e+00 -4.92051572e-01 -6.98005915e-01 3.25081617e-01
2.26464108e-01 -5.12905598e-01 9.09435928e-01 -9.75555539e-01
5.45000374e-01 -1.38869405e-01 -2.04187080e-01 -1.27891707e+00
5.65721132e-02 -4.04693544e-01 2.66813725e-01 1.80992842e+00
5.43001175e-01 -7.61924088e-01 9.18050826e-01 5.92658341e-01
-6.10524893e-01 -4.11512554e-01 -7.89197624e-01 -1.25241005e+00
6.30281270e-01 -6.80526137e-01 7.91656315e-01 7.33973563e-01
4.29452956e-01 2.95203775e-01 4.87065703e-01 1.38015702e-01
3.42467338e-01 1.46940708e-01 5.41404605e-01 -9.06933308e-01
-3.52972329e-01 -5.39178491e-01 -1.19962186e-01 -1.23116386e+00
7.49582350e-01 -1.15156591e+00 8.32659602e-02 -1.54395747e+00
2.39424091e-02 -7.51140237e-01 -6.72169998e-02 7.81100512e-01
-3.71150911e-01 -3.81293863e-01 2.20324129e-01 -9.96561274e-02
-2.51016945e-01 2.24624231e-01 6.80588186e-01 1.94545295e-02
-2.88522333e-01 -1.49840191e-01 -8.23567450e-01 8.52689922e-01
8.17608535e-01 -6.85467780e-01 -2.20202461e-01 -7.91585326e-01
7.64684798e-03 -2.57462204e-01 4.77339141e-02 -7.73010135e-01
-4.40918319e-02 -1.17092006e-01 -9.00604576e-02 -3.04290354e-01
-8.01741853e-02 -4.19162512e-01 -5.22277355e-01 3.78648877e-01
-4.20332462e-01 2.13546544e-01 6.27096236e-01 1.18977025e-01
-3.15463156e-01 -8.23513031e-01 8.32178771e-01 -5.87482929e-01
-1.10775936e+00 -2.19654053e-01 -2.46753156e-01 6.78117335e-01
6.11747324e-01 -2.50436097e-01 -5.03822684e-01 2.29621276e-01
-6.38205230e-01 2.40718558e-01 6.74657643e-01 4.01641071e-01
3.08662087e-01 -8.23935509e-01 -2.42526278e-01 1.06355272e-01
1.35997146e-01 2.62050003e-01 -1.75446257e-01 1.86568350e-01
-5.63883185e-01 4.07679498e-01 1.16192009e-02 -3.51676106e-01
-1.25701296e+00 5.88415563e-01 1.51246771e-01 -3.28476846e-01
-7.19166696e-01 9.19056296e-01 2.80638397e-01 -6.62318766e-01
2.08167225e-01 -9.10971045e-01 2.19465926e-01 -4.05788660e-01
2.48527348e-01 1.95630163e-01 2.50214487e-01 -2.43715957e-01
-4.54765677e-01 5.69133103e-01 -9.43692848e-02 -6.74372613e-02
1.19626415e+00 -1.00328155e-01 -1.99025780e-01 2.30495706e-01
1.00099111e+00 3.78611907e-02 -7.86679149e-01 -3.52843881e-01
7.39121318e-01 -9.37126353e-02 -4.41481948e-01 -7.04939187e-01
-5.48190594e-01 1.07039618e+00 -1.41243234e-01 2.12795690e-01
1.01548100e+00 3.27648848e-01 1.33629453e+00 3.24007004e-01
2.38834336e-01 -1.48204374e+00 -2.63828158e-01 8.04112434e-01
1.32420480e-01 -8.15019190e-01 -3.88007551e-01 -1.20769560e+00
-5.08539140e-01 1.22422349e+00 7.73860037e-01 1.62959516e-01
2.16295779e-01 4.01668578e-01 6.94542229e-02 6.04477003e-02
-4.44962353e-01 -3.34945589e-01 -2.69059896e-01 5.69037735e-01
6.91495359e-01 1.59988597e-01 -5.47083557e-01 8.43128443e-01
-4.10152078e-01 -2.54291862e-01 3.90000135e-01 1.38313878e+00
-6.83445632e-01 -1.77906108e+00 -2.58012891e-01 -1.28295019e-01
-7.24223435e-01 -5.52657604e-01 -4.80079591e-01 8.23865712e-01
-4.25910624e-03 9.85313237e-01 1.79649144e-01 4.24560625e-03
4.52942491e-01 7.32241869e-01 6.61336899e-01 -1.19870007e+00
-7.81611562e-01 -2.88849026e-01 4.92572904e-01 -5.58936536e-01
-2.03677848e-01 -6.42215073e-01 -2.18972349e+00 -2.45789010e-02
-2.21339986e-02 2.70209074e-01 8.54407907e-01 1.26350105e+00
2.24458754e-01 4.22000855e-01 2.39723608e-01 -2.12118864e-01
-7.69050956e-01 -8.18584442e-01 -4.69627291e-01 7.65843451e-01
-3.45196366e-01 -3.31138015e-01 -2.17922360e-01 3.37072939e-01] | [10.478086471557617, 9.205924987792969] |
5092d14d-d925-405d-8d0e-23a138346578 | emotion-and-sentiment-guided-paraphrasing | 2306.05556 | null | https://arxiv.org/abs/2306.05556v1 | https://arxiv.org/pdf/2306.05556v1.pdf | Emotion and Sentiment Guided Paraphrasing | Paraphrase generation, a.k.a. paraphrasing, is a common and important task in natural language processing. Emotional paraphrasing, which changes the emotion embodied in a piece of text while preserving its meaning, has many potential applications, including moderating online dialogues and preventing cyberbullying. We introduce a new task of fine-grained emotional paraphrasing along emotion gradients, that is, altering the emotional intensities of the paraphrases in fine-grained settings following smooth variations in affective dimensions while preserving the meaning of the original text. We reconstruct several widely used paraphrasing datasets by augmenting the input and target texts with their fine-grained emotion labels. Then, we propose a framework for emotion and sentiment guided paraphrasing by leveraging pre-trained language models for conditioned text generation. Extensive evaluation of the fine-tuned models suggests that including fine-grained emotion labels in the paraphrase task significantly improves the likelihood of obtaining high-quality paraphrases that reflect the desired emotions while achieving consistently better scores in paraphrase metrics such as BLEU, ROUGE, and METEOR. | ['Ameeta Agrawal', 'Justin J. Xie'] | 2023-06-08 | null | null | null | null | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 1.36227876e-01 -1.16221420e-01 -3.42671663e-01 -6.80661321e-01
-8.00845087e-01 -9.02331948e-01 6.06232882e-01 1.90227076e-01
3.17024104e-02 8.03753257e-01 1.04395092e+00 1.52417079e-01
3.01005274e-01 -5.58177769e-01 -6.39581978e-01 -2.80976057e-01
7.73121357e-01 1.50048003e-01 -7.61401355e-01 -6.53420448e-01
5.36571383e-01 2.82162696e-01 -1.33290899e+00 8.62146914e-01
1.12450314e+00 7.75861681e-01 -4.43891376e-01 7.56736934e-01
-2.37305611e-01 8.48025858e-01 -9.47218895e-01 -9.43712175e-01
-2.40937516e-05 -9.56783772e-01 -1.04533589e+00 -1.19760320e-01
4.08078521e-01 9.34549794e-02 -1.83210194e-01 1.11280620e+00
3.60646337e-01 3.60108674e-01 6.71601415e-01 -1.22383022e+00
-1.39405358e+00 3.98553282e-01 -5.19940972e-01 -7.70404711e-02
1.01830423e+00 -4.31720167e-02 1.32053244e+00 -1.10264838e+00
4.44383740e-01 1.54696846e+00 7.52723336e-01 8.73446405e-01
-1.53868008e+00 -7.47500896e-01 -1.79808676e-01 1.44454926e-01
-8.58261645e-01 -4.26794827e-01 1.16442156e+00 -1.11187071e-01
8.41503680e-01 6.13671482e-01 7.76216149e-01 1.55456221e+00
4.67179269e-01 6.30400300e-01 1.22337759e+00 -3.34521472e-01
1.92467764e-01 3.83189380e-01 2.38419741e-01 3.39682132e-01
-2.43489519e-01 -3.06399941e-01 -8.99950564e-01 -5.35068512e-01
1.85705900e-01 6.47686347e-02 -2.32255384e-01 -4.92156036e-02
-1.09400845e+00 1.18980169e+00 5.05161047e-01 1.79521382e-01
-1.48671672e-01 -9.93310511e-02 9.19308603e-01 9.42915142e-01
9.12746072e-01 1.18264449e+00 -2.06602409e-01 -3.35311621e-01
-7.94147313e-01 2.58095354e-01 8.31910133e-01 7.43824482e-01
7.68135190e-01 -3.03672850e-01 -7.66958773e-01 1.60135841e+00
-3.43464464e-01 5.98211646e-01 9.34104264e-01 -1.11691177e+00
4.31591362e-01 7.30542123e-01 4.09553289e-01 -1.08617032e+00
-2.02710051e-02 8.36905539e-02 -9.96859729e-01 -4.33546633e-01
-6.39176220e-02 -3.59029353e-01 -3.79409075e-01 2.02686739e+00
1.03759825e-01 -7.61929229e-02 2.00479239e-01 8.07226539e-01
8.77130806e-01 9.36753511e-01 -9.40140616e-03 -2.21652493e-01
1.17397249e+00 -1.12365854e+00 -6.94102228e-01 -4.82542962e-01
3.12061071e-01 -9.26573098e-01 1.93752003e+00 9.49416831e-02
-1.09424841e+00 -5.74094951e-01 -7.88759589e-01 -3.26885015e-01
-2.55129337e-01 2.63829120e-02 4.91482049e-01 5.63052654e-01
-9.70449686e-01 7.47173429e-01 1.02949657e-01 -4.64189202e-01
6.09857105e-02 -1.86765954e-01 -3.34645212e-01 8.84944052e-02
-1.69557941e+00 9.62415516e-01 -5.90549186e-02 -2.35806592e-02
-4.74044457e-02 -9.55202878e-01 -9.32221293e-01 2.15188786e-01
-1.97810113e-01 -9.15286124e-01 1.35789847e+00 -1.17064643e+00
-1.98349571e+00 1.24462235e+00 -4.03999388e-01 -1.32882863e-01
1.79596424e-01 -3.78206462e-01 -2.27964208e-01 2.78764684e-02
2.18040183e-01 5.30688941e-01 1.22572470e+00 -7.58762598e-01
4.27021533e-02 -2.74844646e-01 1.33016212e-02 5.47726095e-01
-6.23178124e-01 5.41526377e-02 8.41823593e-02 -9.92559433e-01
-4.58623260e-01 -9.63005126e-01 -6.65426403e-02 -1.88803583e-01
-3.24098527e-01 -2.61350811e-01 4.78491873e-01 -6.67232454e-01
1.17743874e+00 -2.14338589e+00 5.15647173e-01 -1.08519346e-01
-9.56209004e-03 4.92474698e-02 -4.02746648e-01 6.65412664e-01
-3.45559835e-01 6.27944767e-02 1.31592155e-01 -2.87981182e-01
2.86359698e-01 -8.81618634e-03 -9.13033247e-01 -9.43545103e-02
1.43982232e-01 1.36375785e+00 -1.16473174e+00 -1.63888410e-01
9.53168124e-02 -4.16423455e-02 -6.38142526e-01 5.29213727e-01
-9.53350887e-02 1.34400159e-01 -3.27009857e-01 1.48874700e-01
3.70851845e-01 -1.26354784e-01 -2.03137934e-01 -2.54401922e-01
4.75346565e-01 2.65457392e-01 -2.34732017e-01 1.74630105e+00
-1.23647249e+00 7.05165803e-01 -6.78533837e-02 -5.94419658e-01
1.08875155e+00 6.19030073e-02 1.93173677e-01 -8.23213220e-01
-1.70019031e-01 -9.87681672e-02 -7.05936193e-01 -3.94103080e-01
1.05024576e+00 -8.40403974e-01 -8.90983939e-01 9.81069744e-01
-3.25882226e-01 -9.54213798e-01 -9.29207355e-02 5.56439817e-01
7.85764158e-01 -4.91420358e-01 4.07448649e-01 -3.42499688e-02
4.08537954e-01 -1.48068771e-01 9.18675810e-02 7.93565035e-01
-2.42508918e-01 4.31293726e-01 5.73433518e-01 -1.02655642e-01
-1.08151031e+00 -1.10661817e+00 3.70490223e-01 1.52411151e+00
1.31011024e-01 -3.75179321e-01 -7.00150609e-01 -4.76525426e-01
7.43148103e-02 9.45410013e-01 -9.86750782e-01 -9.11367416e-01
-1.04111247e-01 -5.34900784e-01 8.02580357e-01 2.49964610e-01
4.32042360e-01 -1.26615775e+00 -7.60341138e-02 -2.61599082e-04
-8.11401546e-01 -8.08559120e-01 -9.54795718e-01 -1.48447827e-01
-5.36222398e-01 -5.57036877e-01 -6.76642358e-01 -7.38902926e-01
4.63786453e-01 3.71427745e-01 1.41340411e+00 -3.40938061e-01
1.12616546e-01 1.94340587e-01 -7.11398423e-01 -1.07285872e-01
-8.31838608e-01 -4.37691621e-03 -3.29786725e-02 9.20930132e-03
4.24215466e-01 -5.16401470e-01 -4.44661707e-01 -1.00824624e-01
-8.56318891e-01 2.36508891e-01 2.59230882e-01 1.30191302e+00
9.02275220e-02 -6.14003599e-01 1.14325488e+00 -9.60947454e-01
1.81366765e+00 -5.58717430e-01 4.78204757e-01 3.81362259e-01
-4.40489531e-01 1.65727958e-01 1.24588561e+00 -4.50976521e-01
-1.12671459e+00 -4.72199321e-01 -3.00639272e-01 -4.92774725e-01
-1.46056488e-02 1.76962629e-01 1.95047200e-01 2.47739613e-01
8.81283939e-01 3.85719955e-01 -5.01804613e-02 -1.36151733e-02
9.28175867e-01 1.02232754e+00 5.62285721e-01 -6.41676068e-01
5.95686674e-01 2.07436591e-01 -5.26727021e-01 -5.58986545e-01
-1.48166263e+00 -4.67834741e-01 -7.71911815e-02 -1.84234202e-01
5.19741952e-01 -8.03782165e-01 -3.79112035e-01 4.10415322e-01
-1.27027559e+00 -1.75231233e-01 -6.78685486e-01 -3.29307675e-01
-6.94449663e-01 5.20496845e-01 -8.79153490e-01 2.29127516e-04
-8.42644870e-01 -5.79704285e-01 1.41343224e+00 3.32419693e-01
-1.04976666e+00 -1.03284979e+00 5.60814738e-01 6.68393850e-01
3.39617252e-01 6.85481727e-02 1.22568440e+00 -5.43217778e-01
5.55544734e-01 -2.02919409e-01 -2.76246339e-01 5.28497696e-01
5.04527867e-01 -2.16500089e-01 -7.26570785e-01 -1.78420022e-01
2.36799583e-01 -1.04608583e+00 6.11495316e-01 -7.61625543e-02
1.11343277e+00 -7.39342988e-01 2.52048522e-01 3.64540279e-01
7.40841568e-01 -3.82932574e-01 4.50016856e-01 -5.75773008e-02
4.38883752e-01 6.11097515e-01 8.83133769e-01 8.32587004e-01
3.63372192e-02 5.72828233e-01 -2.79748946e-01 -1.29008695e-01
1.18271373e-01 -7.57251084e-01 5.94836056e-01 8.29720140e-01
8.47010612e-01 -2.24560853e-02 -1.29210368e-01 4.49679106e-01
-1.78580952e+00 -1.33021271e+00 3.51308346e-01 1.62488687e+00
1.46539783e+00 -1.96904421e-01 2.91568004e-02 -4.58921976e-02
8.22314739e-01 6.33542001e-01 -9.34997797e-01 -1.42365205e+00
3.84849869e-02 4.85423863e-01 -3.82056296e-01 6.79830790e-01
-5.21586597e-01 1.21831715e+00 6.02896357e+00 8.43235195e-01
-1.15135503e+00 -5.92959411e-02 6.68905735e-01 -3.95222217e-01
-6.91558063e-01 -3.20463836e-01 -7.45806620e-02 5.94035566e-01
8.92434418e-01 -8.68860722e-01 8.25724840e-01 8.45624149e-01
6.79277301e-01 1.88794345e-01 -1.25340843e+00 1.20654011e+00
4.43043143e-01 -1.31753278e+00 5.28324068e-01 -7.95415878e-01
1.03108490e+00 -3.47848505e-01 2.45972291e-01 6.27722383e-01
4.68728006e-01 -1.04519343e+00 5.10551035e-01 3.16956848e-01
9.45116282e-01 -9.11594808e-01 4.92550373e-01 1.28186151e-01
-6.98914886e-01 1.17012978e-01 -3.81790876e-01 -3.26944828e-01
-1.63540151e-02 6.00756764e-01 -6.86926961e-01 2.46131629e-01
3.72141570e-01 1.05749857e+00 -5.73671579e-01 2.07821354e-01
-4.32648957e-01 4.98644501e-01 2.48934671e-01 -5.35291195e-01
1.07814945e-01 -3.82100880e-01 4.35241818e-01 1.48934197e+00
1.20200207e-02 7.95749277e-02 -2.52061129e-01 9.35057521e-01
-6.12391770e-01 3.90961170e-01 -6.33345604e-01 -1.46615312e-01
5.64741433e-01 1.20561242e+00 5.85231334e-02 -4.58608001e-01
-8.52130428e-02 1.85015476e+00 6.18928969e-01 4.51777726e-01
-9.68346238e-01 -8.02486479e-01 9.30926800e-01 -5.08211136e-01
-1.83049038e-01 3.90409857e-01 -6.80484414e-01 -1.62343943e+00
-3.38655114e-02 -1.23332357e+00 1.27318859e-01 -1.40972126e+00
-1.94511127e+00 4.77585882e-01 -3.52767050e-01 -9.30475533e-01
-5.43957889e-01 -1.41727641e-01 -9.67863142e-01 9.53100920e-01
-1.04607308e+00 -9.50597882e-01 -6.32768333e-01 6.39031708e-01
9.76275504e-01 1.19894497e-01 9.24960911e-01 -2.48712227e-01
-3.60717297e-01 8.91304612e-01 2.50525713e-01 -1.24326073e-01
1.28236508e+00 -1.20217478e+00 5.47944069e-01 5.23261666e-01
1.19231485e-01 7.51932561e-01 1.14450026e+00 -3.87412965e-01
-1.03198314e+00 -1.29676569e+00 1.15575600e+00 -3.73995930e-01
9.97744143e-01 -4.90688920e-01 -8.52663875e-01 3.77802581e-01
4.20951456e-01 -7.41778553e-01 9.09377098e-01 2.52888858e-01
-7.50702024e-01 -5.11738695e-02 -1.29085982e+00 1.03854501e+00
8.10147464e-01 -1.22656476e+00 -1.09203732e+00 5.69903672e-01
9.15923297e-01 -1.93285704e-01 -7.23912001e-01 -1.11955889e-01
6.14094317e-01 -1.04832768e+00 9.02335465e-01 -1.14553571e+00
1.13572466e+00 4.69785929e-01 4.45089117e-02 -2.21005154e+00
-4.60695624e-01 -8.92392874e-01 2.43875086e-01 1.18654740e+00
2.42575407e-01 -4.62693840e-01 8.89398456e-01 8.56571913e-01
6.83201030e-02 -7.47728705e-01 -6.75753236e-01 -4.15422261e-01
3.79432082e-01 3.11752390e-02 5.26832342e-01 1.10491502e+00
9.10466254e-01 9.77980793e-01 -6.68426931e-01 -4.43706006e-01
1.69706464e-01 6.98271573e-01 9.53291297e-01 -6.27182722e-01
-3.01391393e-01 -4.30584311e-01 1.01814270e-01 -1.15618408e+00
1.01104748e+00 -1.17775238e+00 -3.73472385e-02 -1.30079293e+00
6.90703928e-01 9.44795161e-02 3.20560522e-02 2.86826372e-01
-5.30342281e-01 4.63610649e-01 1.85652122e-01 7.56237209e-02
-4.68981534e-01 1.05681109e+00 1.17047548e+00 -2.81423777e-01
-2.15709344e-01 -9.20642354e-03 -1.22686601e+00 5.07368147e-01
9.24959660e-01 -3.82267326e-01 -5.79516232e-01 -1.15260258e-01
4.90433872e-01 2.04300091e-01 3.26040834e-01 -2.44203523e-01
-2.14092746e-01 -5.78322172e-01 1.56930670e-01 -5.30869626e-02
4.94135439e-01 -4.17365223e-01 -8.96377489e-02 2.46344954e-01
-1.19086802e+00 3.39941680e-01 2.08999604e-01 5.95780671e-01
-4.61186558e-01 -3.65616083e-01 1.16048801e+00 -6.94409087e-02
-3.75428647e-01 -1.71300173e-01 -3.84751290e-01 7.24109888e-01
8.82595956e-01 -2.08612397e-01 -2.84047335e-01 -9.44040179e-01
-5.82581103e-01 3.28346491e-02 7.38375604e-01 9.14162815e-01
7.53643334e-01 -1.55871689e+00 -7.29687154e-01 2.01832846e-01
3.91240418e-01 -7.39160419e-01 3.39029521e-01 -3.16572702e-03
-1.33407041e-02 3.18555504e-01 -4.38587546e-01 -4.73282598e-02
-1.45700812e+00 3.39486569e-01 3.56435239e-01 -5.00741363e-01
-1.18148886e-01 9.02333021e-01 8.78695101e-02 -5.94271302e-01
-1.39397129e-01 -3.73849839e-01 4.59266901e-02 1.04924433e-01
4.01326567e-01 2.19276965e-01 -4.80406769e-02 -3.54922593e-01
-4.09661569e-02 4.99559671e-01 -2.96569109e-01 -1.08771034e-01
1.07279253e+00 -4.55524623e-01 -2.90418863e-01 5.34373939e-01
1.42546546e+00 2.59804010e-01 -6.83383286e-01 -8.53776261e-02
-3.55917752e-01 -5.79224706e-01 -4.28643316e-01 -1.01296031e+00
-4.96122271e-01 7.30289876e-01 -2.45999694e-01 1.26307309e-01
1.15061021e+00 -3.71483751e-02 1.19545662e+00 6.19122267e-01
1.63181737e-01 -1.14197958e+00 8.48085821e-01 6.98182881e-01
1.44413102e+00 -1.03024733e+00 -2.42424190e-01 -8.76303017e-02
-1.28296649e+00 8.22096229e-01 6.33983850e-01 -3.00732523e-01
7.59521797e-02 -2.00383365e-01 2.25371689e-01 7.70797729e-02
-1.07295394e+00 5.96622884e-01 1.69742122e-01 2.94037730e-01
5.11160731e-01 1.09832354e-01 -3.79148990e-01 7.12492228e-01
-6.86176479e-01 -1.13370866e-01 7.23792493e-01 3.22811425e-01
-2.33254269e-01 -9.91514802e-01 -2.27459967e-01 6.61211610e-01
-1.72923371e-01 -3.39021474e-01 -1.39006960e+00 3.39311874e-03
-5.47273636e-01 1.11134505e+00 -2.09603950e-01 -5.13489962e-01
5.32685697e-01 2.63177186e-01 4.21136945e-01 -8.90313327e-01
-1.04122388e+00 -8.61288548e-01 2.43110791e-01 -5.69108248e-01
-8.53217673e-03 -3.87342960e-01 -9.94616508e-01 -6.88665450e-01
-2.56207865e-02 6.02575302e-01 2.16967046e-01 8.80321741e-01
6.77691638e-01 2.68708915e-02 1.31109536e+00 -7.39768922e-01
-1.10506606e+00 -1.02450120e+00 -5.35489559e-01 1.32464039e+00
2.91257203e-01 -1.67949319e-01 -5.75011253e-01 -3.80634740e-02] | [11.627874374389648, 9.340893745422363] |
42173a78-5012-459d-8893-3105449cf046 | time-series-analysis-of-big-data-for | 1907.13016 | null | https://arxiv.org/abs/1907.13016v1 | https://arxiv.org/pdf/1907.13016v1.pdf | Time Series Analysis of Big Data for Electricity Price and Demand to Find Cyber-Attacks part 2: Decomposition Analysis | In this paper, in following of the first part (which ADF tests using ACI evaluation) has conducted, Time Series (TSs) are analyzed using decomposition analysis. In fact, TSs are composed of four components including trend (long term behavior or progression of series), cyclic component (non-periodic fluctuation behavior which are usually long term), seasonal component (periodic fluctuations due to seasonal variations like temperature, weather condition and etc.) and error term. For our case of cyber-attack detection, in this paper, two common ways of TS decomposition are investigated. The first method is additive decomposition and the second is multiplicative method to decompose a TS into its components. After decomposition, the error term is tested using Durbin-Watson and Breusch-Godfrey test to see whether the error follows any predictable pattern, it can be concluded that there is a chance of cyber-attack to the system. | ['Mohammad Rajabdorri', 'Mohsen Rakhshandehroo'] | 2019-07-30 | null | null | null | null | ['cyber-attack-detection'] | ['miscellaneous'] | [-5.47016263e-02 -6.63745344e-01 2.72344381e-01 1.76411256e-01
5.01353033e-02 -9.58500564e-01 6.13275290e-01 2.63177663e-01
1.02421261e-01 9.40873444e-01 2.67559737e-01 -7.25558221e-01
-1.72626331e-01 -7.62311459e-01 -1.93830639e-01 -8.06337953e-01
-2.51229316e-01 -7.19245747e-02 3.45130742e-01 -3.19411933e-01
3.80392611e-01 5.56344330e-01 -1.47315967e+00 -1.15381405e-01
4.75105643e-01 1.02572513e+00 -4.53871042e-01 7.47432888e-01
-1.90792531e-01 5.89748263e-01 -1.24153757e+00 3.79392236e-01
1.56960160e-01 -5.13167441e-01 -2.55386353e-01 -1.35037899e-01
-9.43092287e-01 -2.95034405e-02 2.58842349e-01 9.84924674e-01
8.12913552e-02 2.21374556e-01 8.75516653e-01 -1.56603348e+00
-7.48555586e-02 2.29405627e-01 -7.21782804e-01 5.52855790e-01
5.92987478e-01 1.48392439e-01 2.65248716e-01 -5.36140919e-01
3.97347987e-01 1.23825550e+00 8.12662899e-01 -3.94063264e-01
-1.25951838e+00 -5.69533825e-01 -2.53291149e-02 5.07389158e-02
-1.07114625e+00 3.48393857e-01 8.67805123e-01 -4.34836984e-01
1.09207404e+00 4.54184115e-01 7.10905075e-01 8.55344892e-01
1.22206616e+00 1.14224799e-01 1.65378582e+00 -4.47458148e-01
3.64080399e-01 -6.64203539e-02 4.65896487e-01 -4.64258313e-01
4.30364013e-01 2.19460517e-01 1.94714576e-01 -4.41631913e-01
2.66157091e-01 1.73577696e-01 -5.18839806e-02 6.48273528e-01
-7.39731729e-01 4.59009111e-01 -2.07376182e-01 9.51973259e-01
-9.05401886e-01 -1.49930343e-01 7.48291016e-01 7.77187705e-01
4.39072639e-01 -1.70334056e-02 -6.49438083e-01 -4.15398806e-01
-7.32671440e-01 7.37387761e-02 9.17222798e-01 5.08879364e-01
1.07116476e-01 4.77838248e-01 1.61873639e-01 1.33684248e-01
3.78033131e-01 7.42569029e-01 8.27485919e-01 -6.48203911e-03
-2.05205102e-02 2.98812330e-01 1.87758178e-01 -1.14533246e+00
-4.56706852e-01 -4.85076874e-01 -7.18070805e-01 1.99997500e-01
2.73273557e-01 -4.99937713e-01 -7.33534694e-01 1.39420402e+00
2.89159209e-01 4.43891466e-01 2.24834979e-01 2.03478113e-01
2.83588529e-01 1.30161488e+00 9.78719909e-04 -9.52375591e-01
1.59987593e+00 -1.17833644e-01 -1.37401295e+00 5.41271746e-01
1.39855323e-02 -1.17302680e+00 5.27321875e-01 8.27840686e-01
-4.19645369e-01 -4.36256438e-01 -1.03030097e+00 8.89018476e-01
-6.85340941e-01 -3.56512368e-01 1.07953086e-01 7.89403498e-01
-6.88300431e-01 5.71037948e-01 -7.74586916e-01 -4.70336616e-01
-7.92122483e-01 -1.47456631e-01 3.59630585e-02 3.33252221e-01
-1.41217387e+00 8.03987920e-01 2.66524523e-01 9.42287967e-02
-6.87025011e-01 -5.06161988e-01 -4.14671957e-01 7.61450604e-02
1.90627202e-01 2.94814575e-02 9.68884110e-01 -9.36327159e-01
-1.44385409e+00 8.83427411e-02 -1.89040884e-01 -1.68233171e-01
2.80554831e-01 2.01707333e-01 -1.12063468e+00 -1.29765999e-02
-1.39328897e-01 -9.38683093e-01 1.04516292e+00 -1.04670739e+00
-4.47799027e-01 -5.68870664e-01 -4.50244844e-01 -3.39945287e-01
3.20020877e-03 4.68227476e-01 4.88214701e-01 -8.46407831e-01
1.65251538e-01 -1.02172279e+00 1.97055377e-02 -1.04144108e+00
-2.11996153e-01 -3.96964729e-01 1.09996021e+00 -1.11109829e+00
1.76337361e+00 -2.25841069e+00 -6.37338340e-01 7.58137286e-01
-3.46322477e-01 1.24371603e-01 3.23564738e-01 1.09810853e+00
-7.53921747e-01 3.11024159e-01 -7.56433457e-02 3.60138446e-01
-1.10959932e-01 1.48381546e-01 -6.89505279e-01 5.07976055e-01
2.10500970e-01 -8.76523703e-02 -4.78065938e-01 3.21556240e-01
5.37982062e-02 1.65803924e-01 3.13324511e-01 3.46185714e-02
1.04736522e-01 3.77523839e-01 -3.91432643e-01 5.58864057e-01
1.00294495e+00 5.36582291e-01 4.59668152e-02 -2.10242912e-01
-7.84486771e-01 1.29627362e-01 -1.44034457e+00 4.41455811e-01
-3.18606943e-01 5.76440394e-01 -7.82886967e-02 -1.04398263e+00
1.16347611e+00 8.12448382e-01 6.26633763e-01 -4.66280997e-01
3.41121107e-01 3.89414608e-01 4.66898568e-02 -3.57199758e-01
3.70073229e-01 -2.05554858e-01 7.57292062e-02 5.67848146e-01
-1.35981828e-01 -5.79703040e-02 4.10524100e-01 -2.36367390e-01
1.18114758e+00 -4.39254716e-02 5.11137128e-01 -4.25611705e-01
6.96771920e-01 2.11303964e-01 7.53274441e-01 -6.18251190e-02
-1.19980663e-01 -7.50813866e-03 1.17988122e+00 -1.51185811e-01
-1.05057371e+00 -8.30554962e-01 -1.22493535e-01 2.55288750e-01
-2.51239866e-01 -1.46705046e-01 -3.77131164e-01 -4.93400469e-02
-3.17491740e-02 1.11666381e+00 -5.52339256e-01 -7.98219815e-02
-2.37063840e-01 -1.04281449e+00 3.48597109e-01 2.69666791e-01
3.79814655e-01 -8.47347856e-01 -8.02240908e-01 5.95135033e-01
1.59797832e-01 -5.54290295e-01 -6.29934967e-02 3.94252211e-01
-1.05084884e+00 -1.02392030e+00 -5.63313186e-01 -2.97690108e-02
4.07033652e-01 1.61925629e-01 6.95693374e-01 -2.12018937e-01
-2.90438384e-01 4.21167642e-01 -6.39912009e-01 -8.95338297e-01
-3.65579724e-01 -9.25399542e-01 1.52303904e-01 -8.80566165e-02
3.33004385e-01 -6.39979899e-01 -5.13119698e-01 2.90323764e-01
-1.11153138e+00 -1.11870301e+00 1.34895384e-01 3.83152068e-01
2.49480709e-01 1.08077908e+00 4.94333714e-01 -2.30417773e-01
1.34445953e+00 -8.09538782e-01 -8.44547927e-01 1.98268205e-01
-7.61830688e-01 -3.83479685e-01 6.29005253e-01 -5.11343658e-01
-9.85159874e-01 -5.32561421e-01 2.97776878e-01 -4.29064453e-01
-3.04250926e-01 1.13394177e+00 2.67267227e-01 2.83464551e-01
3.17775488e-01 4.19161201e-01 1.63665041e-01 -3.10956270e-01
-4.33484644e-01 6.31435633e-01 2.10985839e-01 -4.51964080e-01
7.99197316e-01 1.08960062e-01 2.17350766e-01 -9.44839299e-01
2.41877679e-02 -6.02275312e-01 -1.86983764e-01 -7.35965192e-01
9.36618268e-01 -5.34467399e-01 -7.00974107e-01 7.76922584e-01
-1.16126549e+00 2.12039843e-01 1.31260142e-01 7.91109502e-01
6.23162128e-02 2.62655973e-01 -6.20220602e-01 -1.52064538e+00
-3.58086258e-01 -7.39833832e-01 1.93783700e-01 4.22488898e-01
-3.14169496e-01 -1.04960716e+00 4.42912251e-01 -4.52053010e-01
4.86177713e-01 8.43191922e-01 8.40886474e-01 -7.41514146e-01
2.63919264e-01 -7.66972065e-01 3.64448845e-01 6.24469995e-01
1.89797699e-01 1.13031816e+00 -3.91499251e-01 -1.30724370e-01
7.94867575e-01 4.60357964e-01 -1.60100311e-01 5.88538766e-01
3.37610632e-01 -2.34281048e-01 4.55535837e-02 4.30868864e-02
1.79266524e+00 1.21397710e+00 8.55413496e-01 5.67945838e-01
-1.02189839e-01 6.08934700e-01 9.97172534e-01 6.80279851e-01
-1.57282770e-01 7.71181658e-02 1.45298436e-01 1.41689420e-01
6.69939637e-01 2.22922474e-01 8.40516150e-01 7.61867642e-01
-4.37186986e-01 -3.48018199e-01 -1.04513669e+00 2.76590705e-01
-1.62452292e+00 -1.33002818e+00 -1.02978683e+00 2.38222528e+00
1.83970958e-01 3.99333388e-01 4.77655470e-01 8.00130486e-01
6.98200881e-01 -4.70302328e-02 -1.01516865e-01 -1.16420794e+00
-6.69893548e-02 1.12757564e-01 5.05505681e-01 1.06324375e-01
-7.43088841e-01 1.71413139e-01 6.41735363e+00 6.79941773e-01
-1.53963852e+00 -2.08059609e-01 5.46294689e-01 6.61037445e-01
-2.78720498e-01 4.84696142e-02 -2.92260349e-01 8.71436894e-01
1.31308556e+00 -6.43991828e-01 4.80546914e-02 4.37308550e-01
7.07840562e-01 -6.43714190e-01 -2.93395966e-01 5.28049767e-01
-3.78533304e-01 -3.01325500e-01 -5.78001797e-01 2.38549963e-01
6.21000826e-01 -4.31014597e-01 -1.66284069e-01 3.02972645e-01
1.88129529e-01 -4.49716717e-01 6.16557896e-01 7.26830125e-01
7.66238645e-02 -9.30695891e-01 1.22547185e+00 1.95816487e-01
-1.47479177e+00 -1.18591048e-01 -5.05652651e-02 -3.31509531e-01
2.85884768e-01 1.09564388e+00 -2.39174068e-01 1.05496395e+00
6.80560112e-01 2.51704991e-01 -3.99301201e-02 9.10462677e-01
-3.07610631e-01 1.17411101e+00 -6.20306373e-01 -1.43682241e-01
3.45332354e-01 -7.77878046e-01 8.82194400e-01 8.33427608e-01
8.99992406e-01 3.33914131e-01 -8.08733478e-02 5.27121723e-01
1.05513811e+00 9.03319865e-02 -6.37706995e-01 -2.26156473e-01
5.48885286e-01 9.11745548e-01 -9.93919432e-01 -5.05886316e-01
-2.86726177e-01 4.34060007e-01 -1.17841733e+00 6.53708220e-01
-9.68552887e-01 -5.74630916e-01 2.92671144e-01 -1.03211112e-01
1.03287354e-01 -3.63725543e-01 -5.70729613e-01 -8.88654411e-01
1.69834554e-01 -8.95014942e-01 3.99116844e-01 -7.57391930e-01
-1.01377237e+00 5.29837489e-01 3.45593154e-01 -1.51069200e+00
-4.16679084e-01 -3.98475349e-01 -1.49278927e+00 1.09823728e+00
-7.33330786e-01 -6.17182076e-01 -4.04513329e-02 8.01522851e-01
1.57530025e-01 -1.77160770e-01 6.81829929e-01 2.23233134e-01
-8.85898173e-01 -3.06075942e-02 6.57171071e-01 -3.49813461e-01
4.38964844e-01 -1.05053782e+00 -1.71888396e-01 1.07799363e+00
-6.90302074e-01 6.79972231e-01 1.44100809e+00 -1.06782949e+00
-1.11403143e+00 -3.14307362e-01 7.98266113e-01 3.29762787e-01
1.26264644e+00 5.12433648e-02 -8.01679552e-01 3.30103010e-01
4.55891222e-01 -5.97070038e-01 7.05262482e-01 -1.48591921e-02
-1.03003278e-01 -1.86361134e-01 -1.43891871e+00 2.97041237e-01
-2.39147633e-01 -1.53880790e-01 -8.40445042e-01 -4.94078659e-02
3.97413760e-01 3.26981425e-01 -1.36978781e+00 3.88252228e-01
4.46675360e-01 -8.89401734e-01 3.99890631e-01 -3.52039635e-01
1.25442758e-01 -5.51252306e-01 3.71824689e-02 -1.48473310e+00
-2.50510722e-01 -7.80924797e-01 4.27409649e-01 1.49580956e+00
3.27900231e-01 -1.18950939e+00 -5.99843301e-02 5.04870594e-01
-3.76547016e-02 -1.61106423e-01 -8.52361381e-01 -1.17208958e+00
-2.68924206e-01 -3.46038580e-01 7.49495387e-01 1.28139210e+00
3.29262525e-01 -1.32104885e-02 -1.31115705e-01 1.93526864e-01
5.57039022e-01 -6.68783337e-02 6.47170305e-01 -1.16209424e+00
-3.39685910e-04 -3.31687301e-01 -3.93460572e-01 3.78863737e-02
-4.46337312e-01 9.15939733e-02 -6.95220456e-02 -1.20361233e+00
-3.67921233e-01 2.15808451e-01 -4.57241893e-01 7.79729709e-02
-4.21202183e-02 -4.26344752e-01 9.44337621e-02 2.00731739e-01
4.52823669e-01 4.02859956e-01 5.39972186e-01 2.37533450e-01
-3.22107375e-01 3.51586044e-01 1.32449254e-01 4.83650416e-01
9.38440919e-01 -5.60145378e-01 -4.69077438e-01 4.94536638e-01
9.55443904e-02 6.78569913e-01 2.80236937e-02 -1.02044821e+00
4.31775525e-02 -3.63764703e-01 1.74725410e-02 -1.07753503e+00
-2.74259508e-01 -1.26632857e+00 8.95913601e-01 1.09053683e+00
4.19129729e-01 8.56043994e-01 5.56623042e-01 6.11104429e-01
-6.79008126e-01 -3.23718399e-01 3.04232657e-01 3.40819478e-01
-2.57644296e-01 -1.79839730e-01 -1.31905651e+00 -6.89067423e-01
1.35564184e+00 -4.21226621e-01 -2.65188187e-01 -5.65300703e-01
-7.95075297e-01 1.33645967e-01 3.21496092e-02 4.19361182e-02
1.55779541e-01 -1.21379840e+00 -4.83250171e-01 -2.42298186e-01
-4.22376066e-01 -6.69817567e-01 5.00508487e-01 1.32668376e+00
-7.23069251e-01 3.84388685e-01 -4.99265522e-01 -1.85547128e-01
-1.43248510e+00 7.05921948e-01 -6.46735951e-02 -4.34602559e-01
-2.11062416e-01 2.52297312e-01 -3.17550719e-01 2.62381136e-01
-8.33315253e-02 -5.24883687e-01 -5.41499853e-01 6.54234350e-01
5.74398696e-01 7.81785488e-01 -5.04044369e-02 -4.26011384e-01
-4.73666310e-01 6.13686502e-01 3.94048750e-01 -4.08874303e-01
1.46568573e+00 -1.70141496e-02 -7.51932323e-01 1.04497683e+00
1.15726745e+00 4.54899967e-02 -5.97442091e-01 2.77344048e-01
3.70509088e-01 -1.20614715e-01 -1.39359802e-01 -6.60576284e-01
-6.17192328e-01 3.76079082e-01 5.84686458e-01 1.20599747e+00
1.52687395e+00 -8.88948739e-01 5.77246487e-01 -1.92780733e-01
2.15242639e-01 -1.20119286e+00 -3.42536539e-01 7.33741820e-01
9.37960088e-01 -6.11264706e-01 -1.67581171e-01 -2.85706311e-01
-6.57419920e-01 1.53402483e+00 -6.34173006e-02 -5.88323176e-01
1.34263301e+00 5.38256884e-01 -1.33239046e-01 -1.14077911e-01
-1.03911400e+00 3.89255509e-02 4.71353680e-02 4.00292635e-01
5.93768001e-01 1.57814473e-01 -1.34766591e+00 5.13308823e-01
-1.80634469e-01 -6.82227015e-02 8.75402808e-01 1.01878440e+00
-3.25830013e-01 -7.42162347e-01 -1.25289369e+00 2.83262849e-01
-8.33411455e-01 2.48008639e-01 -1.86477989e-01 8.94895792e-01
-1.96469590e-01 1.50836146e+00 2.59952515e-01 -6.03398085e-01
5.96282065e-01 2.00226367e-01 -2.17865244e-01 4.16367111e-04
-8.73883367e-01 7.67951787e-01 1.64404035e-01 -4.85232860e-01
-2.16231272e-01 -1.09087360e+00 -1.05648339e+00 -5.89768112e-01
-3.33338946e-01 4.10447657e-01 1.23920178e+00 8.58218133e-01
-2.42466927e-01 9.17255402e-01 1.05627048e+00 -3.77612382e-01
-4.71246064e-01 -1.24563241e+00 -9.66975391e-01 1.45778075e-01
7.62801617e-02 -4.60082501e-01 -9.76553082e-01 2.40274891e-03] | [6.9873785972595215, 3.262935161590576] |
cdc3c8a8-af82-4298-a9b2-2f3d9bfdeb47 | no-shifted-augmentations-nsa-strong-baselines | null | null | https://openreview.net/forum?id=7VH_ZMpwZXa | https://openreview.net/pdf?id=7VH_ZMpwZXa | No Shifted Augmentations (NSA): strong baselines for self-supervised Anomaly Detection | Unsupervised Anomaly detection (AD) requires building a notion of normalcy, distinguishing in-distribution (ID) and out-of-distribution (OOD) data, using only available ID samples. Recently, large gains were made on this task for the domain of natural images using self-supervised contrastive feature learning as a first step followed by kNN or traditional one-class classifiers for feature scoring.
Learned representations that are non-uniformly distributed on the unit hypersphere have been shown to be beneficial for this task. We go a step further and investigate how the \emph {geometrical compactness} of the ID feature distribution makes isolating and detecting outliers easier, especially in the realistic situation when ID training data is polluted (i.e. ID data contains some OOD data that is used for learning the feature extractor parameters).
We propose novel architectural modifications to the self-supervised feature learning step, that enable such compact ID distributions to be learned. We show that the proposed modifications can be effectively applied to most existing self-supervised learning objectives with large gains in performance. Furthermore, this improved OOD performance is obtained without resorting to tricks such as using strongly augmented ID images (e.g. by 90 degree rotations) as proxies for the unseen OOD data, which imposes overly prescriptive assumptions about ID data and its invariances.
We perform extensive studies on benchmark datasets for one-class OOD detection and show state-of-the-art performance in the presence of pollution in the ID data, and comparable performance otherwise. We also propose and extensively evaluate a novel feature scoring technique based on the angular Mahalanobis distance, and propose a simple and novel technique for feature ensembling during evaluation that enables a big boost in performance at nearly zero run-time cost compared to the standard use of model ensembling or test time augmentations. Code for all models and experiments will be made open-source. | ['Unmesh Kurup', 'Tom Bishop', 'Mohamed Yousef'] | 2021-09-29 | null | null | null | null | ['self-supervised-anomaly-detection', 'supervised-anomaly-detection'] | ['computer-vision', 'computer-vision'] | [ 1.26131281e-01 1.85039595e-01 1.98908269e-01 -3.61310124e-01
-5.17510414e-01 -5.25614917e-01 8.21987033e-01 2.49505103e-01
-3.95646542e-01 3.87440383e-01 -1.91253781e-01 -1.59058377e-01
-3.51898611e-01 -5.60686052e-01 -5.91084421e-01 -1.03692496e+00
-4.47333157e-01 6.34225905e-01 2.25516573e-01 -6.41107932e-02
2.74123430e-01 9.09630418e-01 -2.06656241e+00 -1.92898244e-01
1.00000799e+00 1.23245549e+00 -1.34861007e-01 7.57687986e-01
-1.19736962e-01 3.31318200e-01 -7.48870552e-01 -1.47212800e-02
6.78514659e-01 -3.37293744e-01 -5.51698446e-01 2.07089737e-01
7.24729121e-01 -2.32528076e-01 -3.99674773e-02 1.01847208e+00
5.53577423e-01 2.36402780e-01 1.11918449e+00 -1.41535139e+00
-4.12045091e-01 -2.03478187e-01 -4.35414940e-01 2.71634102e-01
1.65461853e-01 1.29714414e-01 7.63605297e-01 -8.92850518e-01
4.38782483e-01 9.88364577e-01 5.86199701e-01 3.51940930e-01
-1.41663003e+00 -4.03615206e-01 -9.48665291e-02 1.16389833e-01
-1.26910341e+00 -3.51184398e-01 8.23331773e-01 -4.92555827e-01
1.00863099e+00 3.92411917e-01 2.49315441e-01 1.10235023e+00
6.88830912e-02 5.81027269e-01 9.67471540e-01 -6.90514326e-01
4.42569464e-01 3.91945302e-01 1.56895950e-01 6.89610124e-01
4.80571896e-01 1.38058186e-01 -2.06345558e-01 -3.96223515e-01
4.53443587e-01 -3.28017808e-02 -2.12356687e-01 -9.02038872e-01
-8.75815809e-01 9.00936246e-01 2.49860317e-01 1.67890266e-01
-1.09495759e-01 -3.70354474e-01 3.91203612e-01 5.87701321e-01
7.06592917e-01 6.59362972e-01 -4.32919472e-01 -1.06908664e-01
-7.72988319e-01 2.67422684e-02 8.52387905e-01 8.15470874e-01
8.60715210e-01 2.18079627e-01 9.25814286e-02 8.58335376e-01
1.37100428e-01 6.07290566e-01 6.52531326e-01 -6.50615573e-01
1.35374397e-01 6.97335184e-01 3.08698900e-02 -1.03733242e+00
-6.52588487e-01 -4.69783932e-01 -9.23241258e-01 5.52826643e-01
5.75564802e-01 8.35679471e-03 -1.09939277e+00 1.44212365e+00
5.32473266e-01 7.84972087e-02 1.00265130e-01 5.94582319e-01
4.03317958e-01 3.14420521e-01 -5.72349846e-01 -9.31363404e-02
1.02693403e+00 -5.34859538e-01 -3.24642122e-01 2.51998678e-02
9.00121689e-01 -5.53007782e-01 1.16841483e+00 6.45876408e-01
-4.99291986e-01 -3.58186692e-01 -1.23997164e+00 2.69298464e-01
-6.29522085e-01 -8.06931257e-02 4.50429440e-01 8.43764007e-01
-7.69549012e-01 7.96360373e-01 -8.40351224e-01 -4.58543241e-01
5.14250815e-01 4.71816301e-01 -6.51066840e-01 2.17282362e-02
-7.02791333e-01 7.94120252e-01 2.83823758e-01 -6.72707632e-02
-7.45572746e-01 -6.07353449e-01 -1.05898595e+00 -2.04242036e-01
2.89135695e-01 -2.05775797e-01 6.88063741e-01 -1.00987577e+00
-1.32452607e+00 8.97974670e-01 9.10873413e-02 -4.85158354e-01
6.38366699e-01 -3.05978656e-01 -4.98912156e-01 2.39981338e-01
-2.96608340e-02 2.97380179e-01 1.41916811e+00 -1.26162684e+00
-3.19383949e-01 -5.57578325e-01 -2.86012292e-01 -4.24350947e-02
-6.93073153e-01 -1.35758817e-01 -8.05259943e-02 -7.11522043e-01
3.04037035e-01 -1.02450323e+00 5.99702913e-03 2.25971937e-02
-5.01723945e-01 -8.40050951e-02 1.14608824e+00 -2.58498073e-01
9.23521042e-01 -2.26051593e+00 -2.64493644e-01 6.68159068e-01
2.09444210e-01 4.05698597e-01 -9.96651053e-02 1.49977311e-01
-3.84316742e-01 -3.54651138e-02 -5.24936080e-01 -3.10314298e-01
8.59793797e-02 3.16853225e-01 -2.93457806e-01 9.02987778e-01
4.98167217e-01 3.09261113e-01 -7.20426261e-01 -1.60314471e-01
3.11020643e-01 3.26557070e-01 -6.00196183e-01 4.04443979e-01
1.21426471e-01 4.57567573e-01 -1.54048488e-01 6.10985518e-01
7.35380888e-01 2.58805975e-02 -5.13664901e-01 2.19497979e-02
-4.12212163e-02 1.46037135e-02 -1.47592545e+00 1.22511816e+00
-3.62343252e-01 5.21257460e-01 -1.94855794e-01 -1.26909089e+00
1.13997853e+00 2.34003477e-02 4.86504197e-01 -5.45974016e-01
4.24165204e-02 3.02140385e-01 1.24943368e-01 -4.07132149e-01
1.31935194e-01 -2.63506570e-03 1.47753119e-01 3.67964715e-01
3.38770390e-01 -1.36813864e-01 1.56165972e-01 -4.94149141e-02
1.18459368e+00 8.76138359e-02 4.44883049e-01 -6.19418859e-01
5.83951950e-01 -9.76674706e-02 3.28830063e-01 8.17264080e-01
-7.28726014e-02 8.70613158e-01 6.52230978e-01 -3.92138511e-01
-9.85557377e-01 -1.15476215e+00 -7.04131424e-01 8.74100387e-01
-1.51025690e-02 -2.20036238e-01 -6.22519493e-01 -1.03715861e+00
1.50193959e-01 7.87482798e-01 -6.99313998e-01 -4.38860118e-01
-3.88668597e-01 -1.12733090e+00 5.32071531e-01 4.12975758e-01
3.26757997e-01 -8.27391326e-01 -5.48284292e-01 -8.05492699e-02
4.39298421e-01 -9.19110000e-01 -3.74979600e-02 7.83909440e-01
-9.68793511e-01 -1.11519337e+00 -5.84414423e-01 -5.62290668e-01
9.63990986e-01 1.12840440e-02 9.10356283e-01 -8.04940462e-02
-6.27079129e-01 6.47999585e-01 -3.93278450e-01 -5.71502030e-01
-3.07749093e-01 -1.61992237e-01 5.65924227e-01 1.83068156e-01
4.71590400e-01 -6.95300221e-01 -4.33255076e-01 4.95685756e-01
-1.06127608e+00 -6.01126254e-01 4.99636292e-01 1.06127381e+00
4.15486783e-01 1.20490722e-01 5.33580303e-01 -8.43065798e-01
2.12349564e-01 -4.31010842e-01 -6.61503077e-01 -2.63794184e-01
-8.43661964e-01 4.80995774e-01 8.96175206e-01 -3.39419752e-01
-7.82758296e-01 2.51237359e-02 -1.43712059e-01 -5.21784842e-01
-6.55603170e-01 2.70059481e-02 -3.06533784e-01 -3.26907873e-01
1.00395215e+00 2.71044318e-02 3.81292015e-01 -5.86170971e-01
1.29603967e-01 7.57217348e-01 4.59415704e-01 -6.05431795e-01
1.20389771e+00 6.60257578e-01 2.50218600e-01 -1.29826748e+00
-6.57124341e-01 -7.18789279e-01 -8.69921744e-01 2.57256597e-01
5.69222152e-01 -6.05005682e-01 -3.80118728e-01 4.99824643e-01
-5.88156819e-01 -1.73284501e-01 -6.18469775e-01 4.89301950e-01
-5.89255929e-01 6.41182601e-01 -8.03650171e-02 -7.60539114e-01
-1.38545528e-01 -8.30020547e-01 9.43649232e-01 -3.32677849e-02
-4.00254652e-02 -1.09118271e+00 1.75447017e-01 -2.37880588e-01
3.91125888e-01 3.97513598e-01 9.28360045e-01 -1.40941131e+00
-1.14044011e-01 -4.26645219e-01 -1.55568600e-01 9.11503851e-01
2.76091486e-01 -5.59507869e-02 -1.30291092e+00 -4.64216560e-01
2.44932815e-01 -2.16968745e-01 9.21652079e-01 5.55161685e-02
1.14453840e+00 -2.69203901e-01 -6.88319132e-02 8.29962611e-01
1.27171588e+00 -5.41754290e-02 5.18602431e-01 5.73182225e-01
7.17362881e-01 6.46920204e-01 6.69568360e-01 6.10996783e-01
-1.12501122e-01 6.14722908e-01 5.75334489e-01 -2.64563411e-01
4.35386896e-02 1.98369846e-01 4.36766237e-01 4.61954474e-01
1.64479464e-02 -7.24342614e-02 -9.83084798e-01 4.90600586e-01
-1.58154428e+00 -7.20833421e-01 -9.88019928e-02 2.69734740e+00
4.31591481e-01 3.43645066e-01 2.82604814e-01 4.52831477e-01
4.53718841e-01 -7.40104541e-02 -5.55355906e-01 -4.24575895e-01
-1.15796700e-01 4.38250661e-01 5.15719891e-01 3.16218734e-01
-1.51342559e+00 4.09063399e-01 5.59926891e+00 6.58637822e-01
-1.15602493e+00 -2.15707794e-01 4.02696341e-01 5.37573099e-02
6.71360493e-02 -1.64742202e-01 -6.93484128e-01 3.61932844e-01
8.79570127e-01 2.20392585e-01 1.71353191e-01 1.10460782e+00
-5.08480333e-02 -1.73529208e-01 -1.21072400e+00 1.03082323e+00
3.39083672e-01 -6.60592020e-01 -6.02213703e-02 2.87174582e-01
5.78320622e-01 1.19619943e-01 4.09205668e-02 2.80202121e-01
-3.17688845e-02 -8.75953376e-01 3.26500177e-01 3.24014723e-01
7.33847439e-01 -8.21909010e-01 8.66232038e-01 2.96428353e-01
-8.14485848e-01 -1.87450349e-01 -6.09695673e-01 8.96389782e-03
-4.88632202e-01 8.53912592e-01 -8.89732242e-01 6.76909566e-01
7.95709312e-01 6.56603396e-01 -1.01265776e+00 1.07523167e+00
6.63133115e-02 6.33051455e-01 -7.66823649e-01 3.16360950e-01
1.73634022e-01 -3.00340712e-01 9.15994704e-01 1.19904339e+00
3.65413249e-01 -4.92867112e-01 9.11177248e-02 6.45190418e-01
2.33588979e-01 2.30797365e-01 -9.84888017e-01 3.05970758e-01
-3.30434814e-02 1.26552534e+00 -7.54045129e-01 -1.63637832e-01
-4.52392489e-01 1.01658535e+00 1.52264953e-01 2.57861644e-01
-6.84298158e-01 -7.78672695e-01 7.17335522e-01 2.73015469e-01
5.24840653e-01 -1.16692863e-01 -1.17974192e-01 -1.18498516e+00
2.45771900e-01 -7.58200645e-01 5.28296649e-01 -1.76018581e-01
-1.64618015e+00 5.55615842e-01 1.55888975e-01 -1.64536321e+00
-4.27543432e-01 -9.98416841e-01 -6.83400452e-01 6.00864351e-01
-1.50520861e+00 -6.89643919e-01 -4.67923373e-01 6.73240304e-01
9.90769193e-02 -2.41195604e-01 9.90986884e-01 9.47032645e-02
-5.18925905e-01 8.28893363e-01 5.52209377e-01 1.54899299e-01
1.00314796e+00 -1.72601867e+00 1.92383572e-01 8.89842331e-01
2.82210708e-01 4.38772976e-01 7.26017416e-01 -3.77419651e-01
-1.14529288e+00 -1.14236701e+00 2.80703187e-01 -5.95999300e-01
6.10077918e-01 -6.91570103e-01 -1.05113292e+00 4.48093951e-01
-2.77248859e-01 5.70290327e-01 6.14535391e-01 8.69428590e-02
-4.47916657e-01 -2.60946780e-01 -1.39843881e+00 3.32494229e-01
9.68970060e-01 -2.65599281e-01 -7.17956960e-01 3.39722097e-01
2.83688813e-01 -1.17863983e-01 -7.42745876e-01 6.32099509e-01
2.40393087e-01 -1.15527499e+00 9.63875353e-01 -5.60842812e-01
-5.95112741e-02 -4.94368106e-01 -1.74929500e-01 -1.29070854e+00
-1.15635116e-02 -6.01409495e-01 -2.36752823e-01 1.18710434e+00
2.20333934e-01 -1.12307596e+00 6.88884377e-01 3.17242920e-01
-3.15665245e-01 -6.45348132e-01 -1.03541005e+00 -1.28753233e+00
-4.23943736e-02 -4.67697024e-01 1.28842354e-01 9.50420380e-01
-1.07944362e-01 1.19504057e-01 -8.06919560e-02 4.42820907e-01
6.91562414e-01 -1.13749810e-01 9.99203742e-01 -1.62431812e+00
-2.46082127e-01 -2.46798366e-01 -1.06730151e+00 -6.73709214e-01
7.96946362e-02 -9.87602949e-01 -1.21773425e-02 -9.49926913e-01
-2.27965295e-01 -5.22776067e-01 -2.64918476e-01 5.07690132e-01
-6.01778850e-02 3.75151575e-01 -7.19767213e-02 2.61674345e-01
-4.32066739e-01 6.43855810e-01 7.23680496e-01 5.56005910e-02
-3.14592928e-01 3.16511653e-02 -4.17851776e-01 9.85823512e-01
6.59110308e-01 -5.98488510e-01 -3.44950438e-01 2.08280861e-01
-1.23852126e-01 -6.99858844e-01 4.91454601e-01 -1.31573105e+00
-1.03012845e-01 3.09220254e-01 6.41814351e-01 -3.13184083e-01
1.52682468e-01 -1.01874042e+00 -4.67532337e-01 2.26638630e-01
3.13798301e-02 -2.51572505e-02 3.00813466e-01 6.58086300e-01
-1.81529477e-01 -5.08624494e-01 9.34666097e-01 2.24815533e-01
-4.65200335e-01 1.33745581e-01 -1.92038909e-01 2.60358185e-01
1.14899457e+00 -4.70136046e-01 -1.71837240e-01 -2.62501419e-01
-6.60025001e-01 -1.05691820e-01 6.73478365e-01 2.53056675e-01
4.24347669e-01 -1.11383855e+00 -5.72149634e-01 8.35380733e-01
4.70260143e-01 1.34992808e-01 -5.43489121e-02 1.08388627e+00
-4.49994534e-01 1.64785340e-01 -1.85655192e-01 -9.52514648e-01
-1.08312082e+00 6.63929582e-01 2.65982985e-01 -1.33122683e-01
-8.51911485e-01 5.31483650e-01 2.98593998e-01 -5.85164130e-01
2.41520122e-01 -3.46856922e-01 -7.69611001e-02 1.83655202e-01
5.36995888e-01 4.28997368e-01 6.01969540e-01 -5.18259287e-01
-4.29501921e-01 5.46864450e-01 -1.64348245e-01 2.80130655e-01
1.36496162e+00 7.67145380e-02 7.63970912e-02 5.97987592e-01
1.41881788e+00 1.88088030e-01 -1.23943150e+00 -8.54270533e-02
2.04840571e-01 -5.04212558e-01 -3.83611396e-02 -5.05629301e-01
-6.88982189e-01 7.80160904e-01 9.05360162e-01 4.67666477e-01
1.16461265e+00 2.69127805e-02 2.31630459e-01 6.60756409e-01
1.61460295e-01 -1.11463106e+00 2.32080311e-01 5.51796019e-01
8.00607324e-01 -1.45020092e+00 2.13557687e-02 -1.95026979e-01
-4.95176643e-01 1.30941427e+00 5.55612624e-01 -5.14463782e-01
7.31830597e-01 2.06671655e-01 5.71292862e-02 -2.05568090e-01
-2.45938897e-01 -3.48922580e-01 5.47975361e-01 8.48780394e-01
6.30172715e-02 -2.73046881e-01 1.86287478e-01 2.95466334e-01
-1.91243887e-01 -6.90095067e-01 4.83649880e-01 9.30479765e-01
-4.32332516e-01 -9.21723187e-01 -5.58903873e-01 8.92273247e-01
-2.77954340e-01 2.93335109e-03 -2.98361242e-01 1.04885709e+00
5.87214231e-02 6.43890917e-01 3.03800583e-01 -2.59823292e-01
3.54389608e-01 4.23553884e-01 2.89127082e-01 -6.21088266e-01
-2.15912923e-01 -8.85439068e-02 -9.63695645e-02 -6.76326573e-01
-2.18719989e-01 -6.79914892e-01 -1.23497498e+00 1.17515840e-01
-5.97071290e-01 5.04070334e-02 5.94165266e-01 9.21524704e-01
4.71534103e-01 2.63505757e-01 8.88647795e-01 -1.00019670e+00
-7.97425270e-01 -9.76187825e-01 -7.48169124e-01 6.01407945e-01
6.47322953e-01 -1.03980517e+00 -1.06687510e+00 -2.03926727e-01] | [7.7504072189331055, 2.440551996231079] |
a50a970e-879b-4677-9491-2fb7174e7c79 | going-deeper-into-action-recognition-a-survey | 1605.04988 | null | http://arxiv.org/abs/1605.04988v2 | http://arxiv.org/pdf/1605.04988v2.pdf | Going Deeper into Action Recognition: A Survey | Understanding human actions in visual data is tied to advances in
complementary research areas including object recognition, human dynamics,
domain adaptation and semantic segmentation. Over the last decade, human action
analysis evolved from earlier schemes that are often limited to controlled
environments to nowadays advanced solutions that can learn from millions of
videos and apply to almost all daily activities. Given the broad range of
applications from video surveillance to human-computer interaction, scientific
milestones in action recognition are achieved more rapidly, eventually leading
to the demise of what used to be good in a short time. This motivated us to
provide a comprehensive review of the notable steps taken towards recognizing
human actions. To this end, we start our discussion with the pioneering methods
that use handcrafted representations, and then, navigate into the realm of deep
learning based approaches. We aim to remain objective throughout this survey,
touching upon encouraging improvements as well as inevitable fallbacks, in the
hope of raising fresh questions and motivating new research directions for the
reader. | ['Mehrtash Harandi', 'Fatih Porikli', 'Samitha Herath'] | 2016-05-16 | null | null | null | null | ['action-analysis', 'human-dynamics'] | ['computer-vision', 'computer-vision'] | [ 6.44713640e-01 7.20716417e-02 -4.09890831e-01 -2.51521975e-01
-8.90635848e-02 -4.53133345e-01 8.33797693e-01 7.13298768e-02
-5.52154958e-01 5.69203198e-01 3.59410703e-01 1.33388881e-02
-2.04771057e-01 -5.38591743e-01 -2.54741907e-01 -5.83504438e-01
-2.96887249e-01 1.22144699e-01 4.13692474e-01 -2.76059389e-01
4.59048986e-01 7.43466318e-01 -1.75723481e+00 3.04376483e-01
3.44841689e-01 8.19752872e-01 -1.24442443e-01 7.56496429e-01
1.59870666e-02 1.13515782e+00 -4.45820957e-01 -4.70817983e-01
1.01809256e-01 -6.95244431e-01 -1.05531275e+00 5.21898508e-01
4.97319400e-02 -2.12866232e-01 -5.18107295e-01 7.75089443e-01
1.74044043e-01 5.39213479e-01 3.96211773e-01 -1.08307683e+00
-4.52367693e-01 -5.52304741e-03 -4.32836086e-01 4.92766649e-01
4.40623701e-01 1.65309176e-01 9.16954637e-01 -3.44350964e-01
7.47781992e-01 1.06371880e+00 4.92194444e-01 6.91598058e-01
-7.03318655e-01 -4.66671288e-02 5.09585977e-01 7.25010633e-01
-7.85897613e-01 -4.16102231e-01 6.95420980e-01 -6.15385175e-01
1.07602417e+00 1.96458116e-01 9.88487422e-01 1.19224024e+00
-6.01908704e-03 1.25914538e+00 8.60265374e-01 -4.51356858e-01
1.65713087e-01 -3.32809746e-01 5.34513034e-02 6.12545073e-01
1.68463290e-01 1.11577869e-03 -3.59826326e-01 2.35590130e-01
8.20980072e-01 2.65920043e-01 -2.58780599e-01 -6.37204647e-01
-1.41660380e+00 7.99861670e-01 2.24801108e-01 5.89780867e-01
-3.17188978e-01 -9.57394093e-02 6.87734842e-01 9.96767059e-02
1.76965240e-02 2.97325134e-01 -2.35729218e-01 -6.16706550e-01
-7.29958653e-01 4.85727847e-01 6.35743320e-01 4.20931846e-01
4.22685027e-01 9.63913128e-02 7.97773525e-02 5.60470104e-01
-1.30640283e-01 2.30936304e-01 5.43356836e-01 -1.12456715e+00
1.19599411e-02 7.19910443e-01 2.30772823e-01 -9.44343626e-01
-4.73292470e-01 -9.31744799e-02 -7.76944578e-01 4.51090693e-01
7.05868602e-01 -3.77116650e-02 -6.69985950e-01 1.41321957e+00
3.50023806e-01 -1.42251542e-02 -1.50096595e-01 8.55370104e-01
3.12386334e-01 2.85186738e-01 1.94307104e-01 -8.36392045e-02
1.16690111e+00 -9.50976253e-01 -5.96886039e-01 -3.13558310e-01
5.07885695e-01 -4.45898354e-01 8.11901152e-01 5.59345067e-01
-7.36540437e-01 -6.43970251e-01 -7.85612822e-01 -5.60287870e-02
-5.59093773e-01 -2.16787294e-01 9.39934015e-01 4.86440510e-01
-6.71826005e-01 8.38583529e-01 -1.25710821e+00 -9.74590003e-01
6.47090614e-01 2.15860218e-01 -4.65326697e-01 6.88311830e-02
-8.53718400e-01 9.14360940e-01 3.45199049e-01 -2.89547332e-02
-6.45415068e-01 -2.61882067e-01 -7.64712453e-01 -3.99996191e-01
8.45415354e-01 -5.09757936e-01 1.36119962e+00 -1.36510909e+00
-1.56105709e+00 1.07963109e+00 -6.12453669e-02 -8.24916661e-01
7.04852164e-01 -4.21221435e-01 -3.46490413e-01 2.91257560e-01
-1.35715291e-01 5.74933767e-01 6.12652004e-01 -7.71536231e-01
-1.10515165e+00 -4.60795909e-01 2.56273001e-01 2.40510166e-01
-2.24628806e-01 2.98161447e-01 -2.28833467e-01 -6.56308770e-01
-1.21317297e-01 -8.88881981e-01 -3.93740386e-01 -1.73267014e-02
-5.05459607e-02 -3.34472239e-01 7.31519878e-01 -4.09130007e-01
1.05384886e+00 -1.87892497e+00 5.36124647e-01 -4.27849650e-01
6.58947006e-02 5.63909709e-01 1.37159467e-01 5.94288349e-01
-1.48168504e-01 -2.92379767e-01 -3.56781393e-01 -5.36932312e-02
-2.53387112e-02 3.17666948e-01 -3.73862714e-01 6.03883266e-01
1.67204320e-01 1.15860772e+00 -1.15156782e+00 -2.59927481e-01
8.40079069e-01 4.52799022e-01 -2.07660452e-01 -4.82303239e-02
-1.28812805e-01 7.15282917e-01 -6.06026709e-01 6.77536488e-01
-7.45231882e-02 -2.71907389e-01 6.46568462e-02 9.37335044e-02
-2.99386144e-01 6.74365014e-02 -9.62158918e-01 1.47917414e+00
-3.36144795e-03 7.99921095e-01 -5.74534982e-02 -1.55193400e+00
5.30255795e-01 1.44759536e-01 1.02368474e+00 -6.40209556e-01
1.49113998e-01 9.46584046e-02 -7.68970093e-03 -6.79971635e-01
3.97956461e-01 -1.58889085e-01 -1.07554346e-01 3.38415593e-01
-1.42670676e-01 2.46375091e-02 3.79247069e-01 -6.89684451e-02
1.04369569e+00 5.14388859e-01 9.34603095e-01 2.68586069e-01
7.25979328e-01 2.53382325e-01 2.04070836e-01 6.76572025e-01
-7.56975710e-01 3.26253623e-01 3.48706514e-01 -1.00728333e+00
-7.17230558e-01 -8.55176985e-01 8.09341073e-02 1.09549570e+00
4.80171619e-03 -2.42952496e-01 -7.91729271e-01 -7.67894149e-01
-2.76460350e-01 1.88427016e-01 -7.63642430e-01 -8.82191882e-02
-8.31144273e-01 -6.23233855e-01 2.87983954e-01 8.45122695e-01
7.48734713e-01 -1.49150991e+00 -1.15691364e+00 2.22586170e-01
1.16526432e-01 -1.20281136e+00 1.80345580e-01 -3.79814692e-02
-1.02514148e+00 -1.35456789e+00 -1.03817081e+00 -6.08928382e-01
2.21679538e-01 3.56271476e-01 9.61524785e-01 -1.94185004e-02
-5.09359479e-01 7.58983016e-01 -5.89224041e-01 -3.84473652e-01
-1.16839468e-01 -4.61522862e-02 1.93413720e-01 1.04810692e-01
7.98776090e-01 -6.13164425e-01 -5.90503335e-01 2.30875567e-01
-8.52233052e-01 -1.04065917e-01 5.61693549e-01 3.77778560e-01
2.97619939e-01 -7.26189315e-02 2.60190606e-01 -5.09100854e-01
2.14450702e-01 -1.88493460e-01 -2.85325170e-01 2.21911550e-01
-2.49757096e-01 -1.44251332e-01 5.36744833e-01 -3.12634021e-01
-9.04273510e-01 3.24887693e-01 -1.42051503e-01 -1.07639581e-01
-7.37408876e-01 1.50852114e-01 -1.26145054e-02 1.50079932e-02
6.02886319e-01 3.96449357e-01 -1.76900476e-02 -4.75556016e-01
5.64242840e-01 6.25569463e-01 7.12022960e-01 -2.95632452e-01
3.93750012e-01 9.74667668e-01 1.04463480e-01 -1.15863574e+00
-9.99089837e-01 -4.76490468e-01 -1.02782857e+00 -5.05801380e-01
1.10734713e+00 -3.23736638e-01 -8.17399681e-01 6.25894189e-01
-7.91862786e-01 -3.85760665e-01 -4.34529126e-01 3.95875752e-01
-8.72622550e-01 7.33706892e-01 -4.61107165e-01 -7.65098095e-01
2.59434015e-01 -8.58314693e-01 8.71267855e-01 4.92106736e-01
-3.98801893e-01 -1.24885380e+00 2.56535292e-01 6.29127264e-01
4.63890806e-02 4.24862355e-01 3.92862886e-01 -5.32551229e-01
-6.08475924e-01 -4.83458877e-01 4.42079343e-02 4.23920393e-01
4.07784343e-01 -1.00805111e-01 -8.50491941e-01 -4.14409153e-02
2.78311223e-02 -3.84256512e-01 9.23386514e-01 4.52172309e-01
1.03622484e+00 1.41039625e-01 -4.99592423e-01 3.69319409e-01
9.87665474e-01 5.03137887e-01 7.19894469e-01 5.87214828e-01
5.93638897e-01 8.17841172e-01 5.93648195e-01 3.29854846e-01
2.37831995e-01 6.92898691e-01 4.49552774e-01 -3.92879881e-02
-9.89708230e-02 -7.10954964e-02 1.27590477e-01 4.01480854e-01
-7.23981619e-01 -1.13477133e-01 -9.43407595e-01 5.53371489e-01
-1.93351805e+00 -1.34877467e+00 9.60724354e-02 2.19607091e+00
2.80281216e-01 1.73252761e-01 5.75207770e-01 3.12273949e-01
6.05979025e-01 4.67610121e-01 -5.79598665e-01 -2.36186549e-01
3.98699939e-02 9.62975919e-02 2.11306632e-01 2.31706962e-01
-1.45966351e+00 1.01773846e+00 6.62034655e+00 2.94439375e-01
-1.21471798e+00 -2.58830309e-01 4.26097274e-01 8.48501176e-02
3.79429936e-01 -5.78720905e-02 -5.45174718e-01 2.72359729e-01
6.30935848e-01 -7.70399347e-02 3.77501905e-01 8.37676942e-01
3.81200105e-01 -3.09711039e-01 -1.00462413e+00 1.20231998e+00
2.67956853e-01 -1.14884627e+00 -1.26814589e-01 9.75005142e-03
6.39181972e-01 -6.86092228e-02 -9.47150439e-02 2.92006373e-01
1.28647119e-01 -1.01947033e+00 3.57298613e-01 5.46011686e-01
1.39597610e-01 -5.95357895e-01 4.20449674e-01 4.72406358e-01
-1.08997178e+00 -3.14333141e-01 -1.36766240e-01 -7.04877377e-01
3.69323730e-01 1.64244801e-01 -2.54571885e-01 4.46104854e-01
6.75215840e-01 1.15144038e+00 -2.53266424e-01 1.08322990e+00
-2.30608463e-01 3.24766457e-01 6.32983148e-02 -1.11191452e-01
4.43218827e-01 -2.07364768e-01 5.05426466e-01 1.09207904e+00
-1.10503368e-01 3.41474235e-01 2.46723920e-01 2.88158566e-01
2.78640389e-01 -6.41815178e-03 -5.94508886e-01 -5.92467844e-01
-1.87632352e-01 1.13801539e+00 -1.02905893e+00 -3.20745856e-01
-7.85860777e-01 1.23077846e+00 2.27873415e-01 2.39576802e-01
-7.71261990e-01 -3.46904486e-01 8.15313160e-01 2.53585696e-01
3.16381454e-01 -5.66177547e-01 -7.15721622e-02 -1.27934825e+00
-5.76670282e-02 -1.05620205e+00 4.45235819e-01 -4.51136708e-01
-8.60384285e-01 4.07157600e-01 1.63999900e-01 -1.20822728e+00
-3.17769140e-01 -8.57436061e-01 -4.35337991e-01 2.07789138e-01
-1.24906516e+00 -8.30525994e-01 -2.72724241e-01 2.62209058e-01
9.49364185e-01 -7.38954395e-02 7.41603792e-01 1.53295338e-01
-4.26732928e-01 -1.43772721e-01 -2.14247536e-02 4.35205907e-01
4.91113573e-01 -9.62385893e-01 5.08743763e-01 8.78019333e-01
5.10087490e-01 3.80153507e-01 5.89679003e-01 -4.46106106e-01
-1.04726183e+00 -6.82045162e-01 9.05841887e-01 -6.89054668e-01
8.26618791e-01 -7.29900971e-02 -6.58256471e-01 8.18913877e-01
3.84287536e-02 1.48734590e-02 3.69542837e-01 4.45436388e-02
2.88926307e-02 -3.87822441e-03 -6.67517066e-01 7.21552074e-01
1.07572925e+00 -3.05230558e-01 -7.54128695e-01 3.08763474e-01
-1.03112645e-02 -2.62730837e-01 -5.37134528e-01 3.29086035e-01
7.30758369e-01 -1.30014849e+00 9.32584703e-01 -1.02638507e+00
2.61388332e-01 -2.06689492e-01 2.02661045e-02 -7.80051351e-01
-2.36041620e-01 -7.22216964e-01 -1.15744151e-01 6.63635492e-01
-2.75139242e-01 -3.45639110e-01 1.05612266e+00 4.70317543e-01
-4.47732061e-02 -8.47077608e-01 -8.00338566e-01 -6.18915200e-01
-6.37029707e-02 -4.44901437e-01 1.59976065e-01 7.63135374e-01
1.20686434e-01 8.41868892e-02 -4.52678293e-01 -4.48086381e-01
5.87001026e-01 6.39308095e-02 7.58041680e-01 -1.25895154e+00
-2.44500518e-01 -6.44504726e-01 -1.01700258e+00 -1.36066449e+00
2.35567018e-02 -3.79360765e-01 -1.33175060e-01 -1.57921016e+00
2.29954600e-01 4.35402960e-01 -2.68436462e-01 4.89178628e-01
-1.45118564e-01 6.13248765e-01 1.20106831e-01 1.39812693e-01
-9.39679503e-01 2.77639002e-01 1.25816226e+00 -3.12294737e-02
-1.65117964e-01 3.41648221e-01 -6.63528919e-01 1.15449893e+00
8.91685009e-01 -1.17442191e-01 -4.30347532e-01 -3.17508727e-01
1.06324881e-01 -1.38921931e-01 6.08084977e-01 -1.39344513e+00
-4.96154279e-02 -4.33386534e-01 3.60251695e-01 -1.68005168e-01
3.01116586e-01 -6.47636592e-01 -2.41172373e-01 4.63716358e-01
-2.95905143e-01 -1.98742330e-01 -8.32409859e-02 6.38234198e-01
-3.44499439e-01 -7.77912363e-02 1.06677723e+00 -5.47576129e-01
-1.28877294e+00 1.75416768e-01 -5.64185023e-01 -5.54584488e-02
1.44425321e+00 -6.93905652e-01 1.06483363e-01 -5.06878257e-01
-1.01584303e+00 5.74346352e-03 5.60716808e-01 6.88161910e-01
2.02251479e-01 -8.76849413e-01 -3.16577196e-01 -1.11837350e-01
6.17777929e-02 -4.47718114e-01 4.76903319e-01 8.38115454e-01
-5.47462404e-01 7.54463553e-01 -4.77889508e-01 -4.61338580e-01
-1.44984639e+00 6.47664845e-01 4.09694314e-01 -5.76079376e-02
-8.94971669e-01 5.96521020e-01 5.44967689e-02 1.09657019e-01
4.38418746e-01 -3.56079489e-01 -4.08102483e-01 1.10648870e-02
7.94994175e-01 7.48264849e-01 -1.72627062e-01 -9.05138075e-01
-5.49058378e-01 6.94729984e-01 1.21989310e-01 3.73211205e-02
1.35389256e+00 -6.26786053e-02 2.61720628e-01 6.12753212e-01
7.83924401e-01 -2.22140312e-01 -1.53096187e+00 5.23883943e-03
2.77023613e-01 -3.32164466e-01 -2.52897531e-01 -6.26232743e-01
-8.90803516e-01 1.02721798e+00 6.82906628e-01 3.47941041e-01
1.17995119e+00 2.44680136e-01 7.87042141e-01 5.74043751e-01
4.44303364e-01 -1.16296577e+00 3.49983901e-01 5.44117033e-01
4.69758570e-01 -1.50060427e+00 1.61122367e-01 -8.35341439e-02
-8.80764782e-01 1.24110496e+00 3.33121032e-01 -2.10588604e-01
4.06184167e-01 4.78114709e-02 1.88091062e-02 -2.15797618e-01
-3.62528116e-01 -5.54596722e-01 1.32105395e-01 8.69462907e-01
5.43699145e-01 -2.11731032e-01 -3.14474136e-01 1.14836842e-01
2.16141924e-01 4.76229638e-01 2.68417835e-01 1.10559297e+00
-8.72989893e-01 -1.41899168e+00 -2.08182290e-01 2.34892458e-01
-4.93210793e-01 5.11275232e-01 -6.78936541e-01 1.11588848e+00
3.17125201e-01 7.58750677e-01 1.45000786e-01 -2.03876588e-02
3.47484231e-01 2.04815745e-01 8.28469336e-01 -3.93440634e-01
-2.96225667e-01 -3.00577730e-01 -1.47815213e-01 -8.10089827e-01
-9.88145113e-01 -9.74093318e-01 -1.10849285e+00 -1.67635664e-01
2.86419839e-01 -1.55300826e-01 3.61057162e-01 1.23548746e+00
2.83364598e-02 3.24269027e-01 1.08872198e-01 -1.04221952e+00
-2.80959040e-01 -7.08974302e-01 -4.76587474e-01 3.67126375e-01
1.95565149e-01 -7.08423436e-01 -6.95023611e-02 4.30602878e-01] | [8.263738632202148, 0.4836721122264862] |
114c7b83-5b12-4ff1-b0ba-377ebb682cf4 | adacoach-a-virtual-coach-for-training | 2204.12935 | null | https://arxiv.org/abs/2204.12935v1 | https://arxiv.org/pdf/2204.12935v1.pdf | AdaCoach: A Virtual Coach for Training Customer Service Agents | With the development of online business, customer service agents gradually play a crucial role as an interface between the companies and their customers. Most companies spend a lot of time and effort on hiring and training customer service agents. To this end, we propose AdaCoach: A Virtual Coach for Training Customer Service Agents, to promote the ability of newly hired service agents before they get to work. AdaCoach is designed to simulate real customers who seek help and actively initiate the dialogue with the customer service agents. Besides, AdaCoach uses an automated dialogue evaluation model to score the performance of the customer agent in the training process, which can provide necessary assistance when the newly hired customer service agent encounters problems. We apply recent NLP technologies to ensure efficient run-time performance in the deployed system. To the best of our knowledge, this is the first system that trains the customer service agent through human-computer interaction. Until now, the system has already supported more than 500,000 simulation training and cultivated over 1000 qualified customer service agents. | ['Biao Fan', 'Xuelian Li', 'Zujie Wen', 'Dan Liu', 'Haozhou Huang', 'Minghui Yang', 'Shuai Zhu', 'Shuang Peng'] | 2022-04-27 | null | null | null | null | ['dialogue-evaluation'] | ['natural-language-processing'] | [-4.44827855e-01 4.73530799e-01 1.53434038e-01 -6.00304186e-01
-2.26479873e-01 -7.36802042e-01 6.49661481e-01 2.91966349e-01
-5.68506777e-01 5.73312044e-01 -3.92227829e-01 -5.21697581e-01
1.03434317e-01 -8.93217623e-01 4.62233573e-02 -4.62751329e-01
3.13046634e-01 1.77349591e+00 2.48899326e-01 -8.40567112e-01
2.18960211e-01 8.80531311e-01 -1.65645814e+00 4.00994152e-01
1.10010362e+00 4.94567335e-01 5.32015204e-01 6.23294890e-01
-3.07394892e-01 7.45900929e-01 -9.52267587e-01 -4.64909226e-01
-3.46175060e-02 -3.30955684e-01 -1.18976283e+00 4.12225872e-01
-5.91054618e-01 -4.19564396e-01 2.97223836e-01 6.86766207e-01
5.10048628e-01 2.78580338e-01 1.36246324e-01 -1.78421748e+00
3.08146160e-02 9.55119312e-01 1.15217946e-01 -8.23533013e-02
5.73977947e-01 3.71498972e-01 7.08821058e-01 -7.03816354e-01
5.26234269e-01 1.06964695e+00 8.40380713e-02 6.23191535e-01
-6.27339780e-01 -3.86547059e-01 9.18556750e-02 2.84716725e-01
-8.44317496e-01 -2.58476436e-01 4.61564779e-01 -3.83181751e-01
1.11075795e+00 3.21643502e-01 7.61972606e-01 7.99567640e-01
-3.46952230e-01 8.00158203e-01 4.84789968e-01 -5.30433714e-01
3.14521313e-01 9.75797772e-01 1.56355828e-01 5.18062234e-01
-2.24525288e-01 -5.06315172e-01 -1.30696669e-01 -3.08808744e-01
7.36782014e-01 -4.72387485e-02 -1.09771639e-01 -2.08880141e-01
-9.37183797e-01 1.03049684e+00 -4.03435901e-02 6.39859855e-01
-7.29352653e-01 -4.28793162e-01 4.32373315e-01 3.78149420e-01
2.08984986e-02 7.55654752e-01 -3.79970431e-01 -7.48731256e-01
-4.82498765e-01 2.42474586e-01 1.53243029e+00 7.81034172e-01
3.69176716e-01 -1.31127248e-02 2.35350102e-01 7.57672489e-01
4.07434076e-01 2.99362570e-01 4.00569946e-01 -1.15587282e+00
5.72033748e-02 1.10209107e+00 7.20971048e-01 -6.73734963e-01
-2.14797035e-01 -4.80837435e-01 -2.12586537e-01 2.27631420e-01
3.03739667e-01 -1.33108631e-01 -2.78862715e-01 9.12612021e-01
5.06137073e-01 1.41378045e-01 3.28589380e-01 1.13774633e+00
7.24296272e-01 7.56202042e-01 -6.38398007e-02 -4.67204273e-01
1.46028745e+00 -1.37017071e+00 -7.68159389e-01 -3.26324970e-01
6.29033625e-01 -7.93422580e-01 1.28216422e+00 7.82940388e-01
-1.42005014e+00 -5.02490222e-01 -8.10612142e-01 5.88445604e-01
-3.06131780e-01 -1.30067691e-01 9.12710547e-01 5.18731892e-01
-8.05584431e-01 3.65328729e-01 -6.78357065e-01 -3.09756517e-01
-2.98041940e-01 5.53132832e-01 -1.29619002e-01 -2.32706517e-01
-1.35104609e+00 1.02722383e+00 4.80686128e-02 2.40429759e-01
-9.43557382e-01 -1.32529706e-01 -5.88176966e-01 2.26927698e-01
5.61206996e-01 -5.18613815e-01 1.99859238e+00 -9.68961298e-01
-1.67528129e+00 7.19577372e-01 2.13874206e-01 -1.73304960e-01
5.12977183e-01 9.74451974e-02 -5.46914518e-01 5.22466488e-02
-6.28985688e-02 8.28947872e-02 1.13811150e-01 -1.57505989e+00
-1.08520126e+00 -8.77942070e-02 2.15439528e-01 2.94488460e-01
8.46288949e-02 3.34304690e-01 -6.10337019e-01 3.39342594e-01
-1.61021754e-01 -8.33428979e-01 -3.15452933e-01 -1.05687273e+00
1.02303185e-01 -6.15772486e-01 6.18434489e-01 -1.26965895e-01
8.78433943e-01 -1.94876909e+00 -8.59011784e-02 5.40501893e-01
3.97830680e-02 7.53049672e-01 -1.07471012e-01 9.76456642e-01
2.95066148e-01 -1.57668486e-01 2.09548265e-01 -3.05190235e-01
5.39029062e-01 5.37483513e-01 6.64616423e-03 -1.16399437e-01
-7.35468790e-02 6.25693738e-01 -1.09830654e+00 -4.17579800e-01
4.59230691e-02 3.63322884e-01 -5.29686749e-01 7.20943630e-01
-2.85889804e-01 4.46201444e-01 -6.98560596e-01 4.49362606e-01
4.20720667e-01 -3.29971135e-01 5.44326723e-01 5.74821532e-01
-5.23304492e-02 2.49041498e-01 -9.82057273e-01 1.11078978e+00
-6.74524784e-01 1.43284127e-01 7.06794679e-01 -8.58588398e-01
1.14193821e+00 5.76251745e-01 4.82321143e-01 -6.43133402e-01
3.01409662e-01 5.64347744e-01 2.93774724e-01 -6.75250530e-01
7.11372316e-01 1.21616729e-01 -1.80592135e-01 7.83883929e-01
-2.27306187e-01 -2.56649494e-01 4.26329851e-01 4.10020769e-01
8.83825541e-01 -2.87235707e-01 -3.83170806e-02 5.18988788e-01
7.74107516e-01 2.68880367e-01 4.57720876e-01 6.04461491e-01
-2.43510023e-01 -1.37688190e-01 3.56489271e-01 -3.12269181e-01
-8.47762465e-01 -4.74810272e-01 4.04907882e-01 1.33385587e+00
1.02070965e-01 -1.38478905e-01 -8.39445710e-01 -6.86573565e-01
-5.95557950e-02 1.09650338e+00 2.21356094e-01 -3.85805825e-03
-6.02882445e-01 -6.78157713e-03 2.94217706e-01 2.20647708e-01
4.45651114e-01 -1.73658705e+00 -6.77655697e-01 6.91155016e-01
-3.29111844e-01 -1.19254410e+00 -2.99339265e-01 -5.18752560e-02
-2.70892113e-01 -8.51104617e-01 -3.98719728e-01 -1.08242691e+00
7.25501060e-01 1.77533105e-01 1.11334348e+00 6.43173933e-01
1.31191254e-01 4.01227593e-01 -6.17777348e-01 -1.90850720e-01
-9.17109370e-01 1.89088464e-01 2.88972445e-02 -2.28589326e-01
4.91383582e-01 -4.26042497e-01 -4.98156339e-01 1.02381396e+00
-5.61853528e-01 1.46366924e-01 4.04437393e-01 7.41218805e-01
-9.38083008e-02 3.43635261e-01 9.06989694e-01 -1.05836177e+00
1.14264917e+00 -2.49554440e-01 -6.09689832e-01 2.67700881e-01
-8.89685333e-01 -4.59706128e-01 7.24124074e-01 -2.80726582e-01
-1.06099558e+00 -1.92108557e-01 -3.96209717e-01 2.29449049e-01
-4.80058312e-01 7.68089235e-01 -3.23842555e-01 3.34981441e-01
2.43311375e-01 -1.07398536e-02 1.40006721e-01 -3.18716139e-01
-1.42174721e-01 1.08446956e+00 5.46963990e-01 -4.74452794e-01
5.07735252e-01 -2.36338958e-01 -5.54635882e-01 -1.39101833e-01
-4.59505945e-01 -8.41053545e-01 -2.72723019e-01 -5.95884264e-01
3.99907708e-01 -3.94569993e-01 -1.72044849e+00 3.52490455e-01
-1.38147449e+00 -6.38637006e-01 -9.00318474e-02 4.25113857e-01
-3.48935008e-01 3.63584980e-02 -6.76131785e-01 -1.09271324e+00
-2.91655779e-01 -1.40573406e+00 6.29359663e-01 4.91071999e-01
-1.83656439e-01 -7.74190903e-01 -2.33020470e-01 9.47912455e-01
5.58095753e-01 -2.34371945e-01 5.77926278e-01 -1.15346932e+00
-1.62503019e-01 -5.88239133e-01 1.26424864e-01 3.80570680e-01
-1.87727317e-01 4.67904173e-02 -4.72573191e-01 -2.96067655e-01
6.44075871e-03 -2.75635123e-01 -2.21211836e-01 -1.08014189e-01
3.45364213e-01 -2.39151493e-01 -2.90500939e-01 -2.68818766e-01
6.66862607e-01 9.75315988e-01 6.19926989e-01 5.99343777e-01
1.18253089e-01 1.05607677e+00 1.37824535e+00 5.89172006e-01
4.68858451e-01 6.58009112e-01 4.42046106e-01 -3.28400820e-01
8.70680809e-01 -1.25051722e-01 4.34996039e-01 9.77922440e-01
-8.02497491e-02 -4.34882641e-01 -1.27870023e+00 1.83969438e-01
-2.26917648e+00 -7.71883667e-01 -3.02328169e-01 1.71797454e+00
8.18896174e-01 3.41223657e-01 9.77446437e-02 1.49255097e-01
6.51912808e-01 -4.96771425e-01 -3.35085064e-01 -7.83060491e-01
4.96855974e-01 -8.13689977e-02 -1.46010146e-01 7.00827956e-01
-3.27013940e-01 1.07976556e+00 5.31359005e+00 4.84417498e-01
-9.00703669e-01 -3.72917093e-02 4.25521195e-01 2.81340986e-01
-1.62835419e-01 2.29604796e-01 -8.40737760e-01 2.81408578e-01
1.00264025e+00 -2.35733107e-01 7.91994870e-01 1.28105414e+00
6.19958282e-01 -1.48776561e-01 -1.05234897e+00 6.71365380e-01
2.65877042e-02 -1.13024616e+00 -5.26632130e-01 1.21182576e-01
2.53202796e-01 -3.96633744e-01 -4.12037700e-01 6.71111286e-01
5.52432775e-01 -9.29038644e-01 4.02159274e-01 5.87631166e-01
3.99604663e-02 -1.25359535e+00 1.50054693e+00 1.00230861e+00
-6.37058198e-01 -1.48906529e-01 7.01347692e-03 -1.48385242e-01
6.59704804e-01 1.78693205e-01 -1.76800823e+00 3.28661054e-01
3.94585818e-01 -2.68723071e-01 -1.99059367e-01 8.42887163e-01
-1.81455314e-01 3.85281801e-01 1.97456062e-01 -2.20906302e-01
2.85434693e-01 -4.87323284e-01 1.00058042e-01 8.88908923e-01
1.96626797e-01 3.37067753e-01 7.84758925e-01 4.07421231e-01
3.57037485e-01 2.19004989e-01 -1.06069483e-01 -4.12803292e-01
7.58121014e-01 1.53434479e+00 -6.95567966e-01 -4.67447549e-01
1.76821090e-02 8.83039773e-01 -7.81624541e-02 2.80494422e-01
-8.67671072e-01 -7.08718061e-01 3.61102581e-01 2.77913034e-01
-3.23499274e-03 -3.41008425e-01 3.57634395e-01 -3.29834014e-01
-2.37997219e-01 -1.51658118e+00 -1.13322124e-01 -1.06337357e+00
-1.02355278e+00 9.61486399e-01 -4.04562682e-01 -7.24831164e-01
-8.35488796e-01 -4.46640730e-01 -6.66049898e-01 8.77630591e-01
-1.09943569e+00 -9.90064800e-01 -4.98933256e-01 4.40493494e-01
4.42092448e-01 -6.40053332e-01 1.06189966e+00 5.14509320e-01
-4.36890692e-01 1.22274593e-01 -5.81930637e-01 5.29853106e-02
6.64956450e-01 -1.00593460e+00 1.69480398e-01 3.14571559e-01
-4.05576169e-01 8.30826104e-01 9.49638426e-01 -4.92987305e-01
-1.45388091e+00 -3.79727274e-01 1.21649516e+00 -5.41043617e-02
4.46524203e-01 -4.14872020e-01 -8.21287334e-01 4.54965174e-01
4.38359648e-01 -6.96648359e-01 6.67098999e-01 1.06265038e-01
4.96545553e-01 -1.42157257e-01 -1.10164857e+00 3.85457575e-01
3.90588522e-01 -2.99393147e-01 -3.63582283e-01 5.24610162e-01
7.50324667e-01 -4.74013835e-01 -7.62890279e-01 1.46910131e-01
1.13690928e-01 -8.93373251e-01 5.13815463e-01 -4.36883986e-01
-9.90700647e-02 -2.66641825e-01 3.38102698e-01 -1.32530153e+00
3.70296128e-02 -7.45279491e-01 4.31945443e-01 1.48593485e+00
6.23791516e-01 -1.08720398e+00 8.88110876e-01 1.00996459e+00
-4.29319233e-01 -4.93679821e-01 -4.86478597e-01 -4.02457416e-01
-6.07459724e-01 -2.89808124e-01 1.13314033e+00 1.01831913e+00
5.03245056e-01 4.72575992e-01 -1.23152547e-02 1.75329193e-01
-1.16028197e-01 1.73946749e-02 1.10202730e+00 -1.35554838e+00
-6.98563039e-01 -2.66920865e-01 1.76281538e-02 -9.78887379e-01
1.52425110e-01 -5.78413963e-01 3.01113009e-01 -2.02621531e+00
-2.32859299e-01 -6.75681412e-01 2.07789689e-01 5.96625745e-01
4.21258628e-01 -4.23416764e-01 8.65248404e-03 1.70961514e-01
-6.93020165e-01 1.02373771e-01 1.68149972e+00 2.64967680e-01
-5.03474653e-01 6.09960556e-01 -7.77823031e-01 6.76214874e-01
9.76823926e-01 -1.43574938e-01 -5.19600868e-01 -9.98493358e-02
3.54802132e-01 4.99254256e-01 1.52843624e-01 -5.07599294e-01
6.94343686e-01 -4.94798869e-01 -3.55125546e-01 -5.41595459e-01
4.09968942e-01 -1.17134011e+00 3.87503982e-01 5.31001151e-01
-4.24655862e-02 3.58345747e-01 -2.03415066e-01 -2.92134643e-01
-4.91566032e-01 -6.91953838e-01 4.09738630e-01 -1.55841097e-01
-5.24654448e-01 -9.73319486e-02 -9.99549210e-01 -5.77761531e-01
1.35076904e+00 -1.03205889e-01 -5.40611267e-01 -7.84684777e-01
-6.97997630e-01 8.67531061e-01 4.18242902e-01 4.42673303e-02
5.07419109e-01 -8.60869050e-01 -4.80591625e-01 1.64579287e-01
-1.92498237e-01 9.32608545e-02 9.22462791e-02 6.80619180e-01
-7.33468413e-01 4.49844062e-01 -1.31106615e-01 -3.15385282e-01
-1.52419126e+00 3.81039262e-01 3.02498132e-01 -5.37684023e-01
-2.57061809e-01 7.45410860e-01 -5.31896293e-01 -7.82068908e-01
5.05881071e-01 1.31379396e-01 -4.45340186e-01 2.64691263e-01
4.00209039e-01 2.46087089e-01 2.21173391e-01 -5.69711566e-01
-3.06004405e-01 -2.07975626e-01 -2.58785993e-01 -6.13365710e-01
1.48362136e+00 -4.43325862e-02 -2.87671745e-01 2.04701304e-01
4.42653537e-01 2.13222094e-02 -7.98660278e-01 6.24256618e-02
1.19353287e-01 -4.41822082e-01 -2.07425624e-01 -1.17864406e+00
-8.02302897e-01 4.69313264e-01 2.32629538e-01 8.45781386e-01
8.38842213e-01 1.06614195e-01 8.97338271e-01 5.13266325e-01
6.65137172e-01 -1.47310090e+00 2.85391361e-01 6.74443245e-01
8.37136567e-01 -1.14463091e+00 -4.22643036e-01 -5.71006179e-01
-1.26757097e+00 8.16377163e-01 9.32386935e-01 4.06018823e-01
1.24054380e-01 2.07768619e-01 5.05679071e-01 -4.45406705e-01
-1.05182421e+00 -9.24192145e-02 -3.86171222e-01 5.66015780e-01
3.22715044e-01 1.87023759e-01 -3.42816949e-01 1.03737020e+00
-3.71046305e-01 -2.96082832e-02 5.48666835e-01 9.95007992e-01
-7.08794594e-01 -1.75297558e+00 -3.15526813e-01 -1.36915028e-01
-5.91187030e-02 1.45648554e-01 -6.03194654e-01 7.68313169e-01
1.15284659e-01 1.47461426e+00 5.00492081e-02 -1.44391134e-01
1.06352603e+00 1.85243338e-01 -2.75291670e-02 -7.94182777e-01
-1.19188190e+00 2.05000490e-01 6.87610686e-01 -5.38574755e-02
-2.03548834e-01 -4.95206386e-01 -1.87430584e+00 -7.01702356e-01
-4.44327146e-01 1.16707826e+00 9.42662835e-01 6.63353503e-01
1.37914345e-01 5.19943953e-01 1.04873800e+00 -7.71515608e-01
-4.54047561e-01 -8.89631093e-01 -7.09815443e-01 2.09314898e-01
-4.34403449e-01 -5.14231861e-01 -3.34951550e-01 -3.21930051e-01] | [12.905326843261719, 7.977902889251709] |
b57e8d8d-d8a7-497d-952b-53a80c91298f | movement-tracking-by-optical-flow-assisted | 2006.13856 | null | https://arxiv.org/abs/2006.13856v1 | https://arxiv.org/pdf/2006.13856v1.pdf | Movement Tracking by Optical Flow Assisted Inertial Navigation | Robust and accurate six degree-of-freedom tracking on portable devices remains a challenging problem, especially on small hand-held devices such as smartphones. For improved robustness and accuracy, complementary movement information from an IMU and a camera is often fused. Conventional visual-inertial methods fuse information from IMUs with a sparse cloud of feature points tracked by the device camera. We consider a visually dense approach, where the IMU data is fused with the dense optical flow field estimated from the camera data. Learning-based methods applied to the full image frames can leverage visual cues and global consistency of the flow field to improve the flow estimates. We show how a learning-based optical flow model can be combined with conventional inertial navigation, and how ideas from probabilistic deep learning can aid the robustness of the measurement updates. The practical applicability is demonstrated on real-world data acquired by an iPad in a challenging low-texture environment. | ['Arno Solin', 'Lassi Meronen', 'William J. Wilkinson'] | 2020-06-24 | null | null | null | null | ['probabilistic-deep-learning'] | ['computer-vision'] | [-1.01135366e-01 -5.80810905e-01 -3.50376725e-01 2.62045622e-01
-3.17595392e-01 -6.13297701e-01 3.61410141e-01 -4.77655858e-01
-3.81127149e-01 8.98913920e-01 2.73508400e-01 -6.26879707e-02
2.72364020e-02 -2.76869446e-01 -6.88392818e-01 -4.92945820e-01
1.93701684e-02 4.77242135e-02 -5.97477928e-02 1.34239152e-01
2.56909221e-01 4.24188316e-01 -1.33130908e+00 -6.06252730e-01
6.99282944e-01 1.11204541e+00 3.22009683e-01 1.08449543e+00
3.11204940e-01 6.98099673e-01 -2.66047955e-01 1.17386356e-01
4.34476852e-01 2.34884191e-02 -1.78546548e-01 2.65286267e-01
1.06652880e+00 -7.10093915e-01 -8.52176130e-01 9.64042187e-01
5.56493461e-01 3.26465577e-01 1.41034961e-01 -1.01465344e+00
-3.15347880e-01 -5.71529984e-01 -5.44677556e-01 3.92396122e-01
9.64178443e-01 3.37064892e-01 4.88500714e-01 -8.46591115e-01
9.05355394e-01 1.09426761e+00 9.16606605e-01 2.48533756e-01
-9.84743774e-01 -3.28104675e-01 -1.10487364e-01 1.72879204e-01
-1.29921210e+00 -3.96179974e-01 5.92783034e-01 -7.87985802e-01
6.41825676e-01 -1.29148200e-01 8.90290320e-01 1.08347166e+00
5.15757442e-01 5.65092087e-01 6.65709317e-01 -9.37072784e-02
-8.94791633e-03 -2.60842800e-01 -2.60178715e-01 8.48051012e-01
6.04017138e-01 3.23610485e-01 -1.03258240e+00 -3.48048136e-02
1.22890472e+00 5.59620619e-01 -6.09766006e-01 -9.21781957e-01
-1.60719371e+00 5.26780963e-01 6.10502422e-01 -1.17091760e-02
-4.44259226e-01 4.07602459e-01 -5.31062260e-02 -3.39981355e-02
3.83829474e-01 1.50097549e-01 -3.75485033e-01 -7.23425567e-01
-9.21552181e-01 -9.00768191e-02 6.34443820e-01 1.17424536e+00
9.21458721e-01 3.48551065e-01 2.25015178e-01 -4.68951883e-03
8.35092068e-01 1.17339563e+00 4.02139157e-01 -1.26701176e+00
6.90451801e-01 1.23544373e-01 7.77263761e-01 -1.30769598e+00
-5.29932201e-01 -3.29617441e-01 -6.58662021e-01 3.38306129e-01
6.07422590e-01 -5.69055855e-01 -4.64228332e-01 1.44506443e+00
6.22150779e-01 1.02887642e+00 -2.90095031e-01 1.32061434e+00
1.66920796e-01 2.67919391e-01 -5.12936473e-01 -1.77345753e-01
8.32612872e-01 -6.68973744e-01 -9.34157312e-01 -4.21373636e-01
4.65199888e-01 -8.58759284e-01 6.31392837e-01 3.42164814e-01
-7.29593635e-01 -9.29057181e-01 -1.28723037e+00 -8.53830874e-02
-6.49454445e-02 1.41049027e-01 4.51068431e-01 6.74505055e-01
-8.97542834e-01 5.23426473e-01 -1.56323063e+00 -4.08674717e-01
6.76995665e-02 4.25470948e-01 -5.34657478e-01 -2.59409875e-01
-8.65370810e-01 8.23190033e-01 -1.15624905e-01 4.63279516e-01
-4.18803602e-01 -6.94610953e-01 -1.27766359e+00 -3.42456520e-01
1.72961906e-01 -1.26577020e+00 9.78007555e-01 -4.98326987e-01
-1.74779940e+00 2.35000193e-01 -6.53504550e-01 -2.87468761e-01
5.07845581e-01 -9.79085624e-01 -1.70426946e-02 3.29410851e-01
1.34074122e-01 2.20252350e-01 1.32689476e+00 -9.96675909e-01
-6.62678182e-01 -4.50644821e-01 -7.26970611e-03 4.66404200e-01
-2.65292712e-02 -7.37507284e-01 -3.30738157e-01 -3.73604685e-01
3.81181419e-01 -1.39954245e+00 -5.67476004e-02 2.37197936e-01
-2.21272498e-01 5.31406999e-01 9.68191445e-01 -6.00566328e-01
9.08870041e-01 -2.02876592e+00 3.03757459e-01 5.56257665e-02
2.58850992e-01 1.13707200e-01 3.70091796e-01 -2.55305588e-01
3.47362548e-01 -4.46497887e-01 4.38110560e-01 -6.65145218e-01
-3.36767346e-01 2.33241484e-01 -2.92326391e-01 1.00665593e+00
-1.50700703e-01 7.44577944e-01 -1.41987634e+00 -1.87875241e-01
1.01195288e+00 9.38718140e-01 -5.08290291e-01 2.60478944e-01
3.00651640e-01 1.32998919e+00 -5.17193198e-01 6.06323242e-01
5.96104145e-01 -4.32426661e-01 -1.06815316e-01 -4.76786554e-01
-8.16436037e-02 1.79869950e-01 -1.51796973e+00 2.46714830e+00
-6.61449313e-01 8.62962067e-01 1.51368022e-01 -3.28571141e-01
4.93625700e-01 1.76939517e-01 7.66316950e-01 -6.95038915e-01
1.36763141e-01 -1.10728983e-02 -4.24451917e-01 -2.33334869e-01
7.96459079e-01 1.04968481e-01 1.45981207e-01 2.79307544e-01
2.06865922e-01 -7.83538297e-02 -1.97739065e-01 1.27006978e-01
1.03140736e+00 8.72662663e-01 4.35000211e-01 2.62386769e-01
4.58312899e-01 -2.33513132e-01 6.72045588e-01 9.07506824e-01
-4.66544807e-01 8.86306107e-01 -3.57460260e-01 -5.01691043e-01
-8.72829854e-01 -1.18463254e+00 -1.38198525e-01 2.97681332e-01
5.51513314e-01 -4.79357898e-01 -1.81224987e-01 -4.06824380e-01
2.62408733e-01 -2.38057539e-01 -2.74963289e-01 -2.78716069e-02
-6.28719866e-01 -3.48588437e-01 -1.92152113e-01 4.18238580e-01
4.96394634e-01 -1.34904832e-01 -7.67148674e-01 2.58320689e-01
-3.53981018e-01 -1.51110971e+00 -6.19530201e-01 -5.20038426e-01
-1.10221970e+00 -1.08240247e+00 -7.84963310e-01 -9.96072441e-02
5.37767172e-01 8.82650197e-01 7.46943295e-01 -9.00558233e-02
-2.61785872e-02 1.08570719e+00 -2.41764840e-02 8.45995545e-02
2.36438960e-01 -2.76682764e-01 7.80337214e-01 2.93178707e-01
-7.17302337e-02 -5.53814769e-01 -7.82642424e-01 3.82463783e-01
-2.94012666e-01 -1.98384836e-01 1.51225049e-02 7.54464447e-01
5.93134224e-01 -4.90196586e-01 -2.82544613e-01 -1.01054072e-01
7.10800961e-02 -4.09875929e-01 -8.21748912e-01 -2.53588080e-01
-3.45106632e-01 -3.89610641e-02 2.69081086e-01 -4.48251724e-01
-9.98055935e-01 5.91946244e-01 2.29986638e-01 -8.94208729e-01
1.50735587e-01 3.19611311e-01 1.16746061e-01 -5.05908191e-01
5.57815790e-01 -1.78818494e-01 1.95720732e-01 -3.22826684e-01
5.83486676e-01 3.43098491e-01 7.42003560e-01 -2.03977674e-01
7.39275813e-01 9.28288281e-01 3.96629035e-01 -9.75362420e-01
-6.77065611e-01 -8.67376685e-01 -1.00443006e+00 -4.37279850e-01
7.54779160e-01 -1.43378639e+00 -9.79762375e-01 3.95906061e-01
-1.17246330e+00 -9.58223939e-02 -1.99771121e-01 1.19346070e+00
-7.82980621e-01 6.95312142e-01 -4.96219903e-01 -6.67084455e-01
-1.16542298e-02 -1.20210171e+00 1.44428098e+00 6.36756301e-01
-2.03296289e-01 -1.37262082e+00 5.20727456e-01 1.27206460e-01
2.75183529e-01 1.46355376e-01 -4.17299658e-01 3.16816211e-01
-9.72648144e-01 -4.87491667e-01 1.93432234e-02 1.25147730e-01
4.48452502e-01 -2.38092601e-01 -8.63623857e-01 -5.82219124e-01
2.67732769e-01 2.79884543e-02 5.00670969e-01 8.49587619e-01
3.60614747e-01 -6.63378239e-02 -5.12352526e-01 1.16038203e+00
1.43686759e+00 -1.90512121e-01 3.83394271e-01 3.64605635e-01
1.43786478e+00 -8.04302916e-02 5.83047628e-01 6.48989975e-01
5.95543563e-01 8.74406397e-01 4.85248476e-01 2.13889375e-01
-2.58171380e-01 -3.31097245e-01 5.18191874e-01 8.44297826e-01
-2.94220537e-01 1.20671496e-01 -6.34066105e-01 1.72001630e-01
-2.16159701e+00 -7.93099701e-01 5.40649556e-02 2.60125709e+00
3.18687201e-01 -3.64989974e-02 -1.55373052e-01 -9.29839462e-02
6.86418295e-01 3.59082609e-01 -6.10357940e-01 3.52254838e-01
4.72137257e-02 -1.95038050e-01 7.15102553e-01 1.18193781e+00
-1.15878975e+00 6.63399756e-01 6.34165192e+00 -1.34019963e-02
-1.34570813e+00 1.64331898e-01 -5.51969633e-02 -3.82490814e-01
1.53116614e-01 1.19159944e-01 -1.06972373e+00 6.68349624e-01
9.82016623e-01 1.27765805e-01 3.17966759e-01 6.54116869e-01
3.25750023e-01 -5.77707052e-01 -9.43339884e-01 1.47614181e+00
1.95938855e-01 -1.50909352e+00 -5.85603237e-01 4.10388887e-01
8.76838088e-01 3.66071045e-01 8.86263028e-02 -1.11428246e-01
2.23372623e-01 -4.66575086e-01 4.72244501e-01 1.05542624e+00
6.18645966e-01 -3.11781049e-01 5.87562084e-01 1.82253376e-01
-1.34239686e+00 1.22795187e-01 -2.38945857e-01 -5.70069790e-01
6.80168688e-01 7.64297664e-01 -5.99827588e-01 6.51713371e-01
7.74919987e-01 1.59134984e+00 -4.47228462e-01 1.32512188e+00
-2.17502430e-01 9.40579772e-02 -6.24957979e-01 3.77212495e-01
-1.03057086e-01 -4.75693166e-01 9.35196400e-01 6.67661607e-01
6.81040168e-01 -3.24444860e-01 4.30129826e-01 5.29502392e-01
3.89418483e-01 -5.40071607e-01 -1.08713830e+00 4.10711765e-01
1.40741527e-01 1.28545392e+00 -3.32414299e-01 -3.14810544e-01
-6.91412508e-01 1.08153296e+00 5.53763211e-02 5.78465402e-01
-5.55828035e-01 -1.30253866e-01 1.13695836e+00 6.45274622e-03
3.51428688e-01 -1.12670171e+00 -1.50205150e-01 -1.88766253e+00
1.45462602e-01 -2.36587137e-01 1.06932648e-01 -1.13918173e+00
-1.03659844e+00 1.97569609e-01 -4.26078439e-01 -1.99628258e+00
-6.81408226e-01 -6.77933872e-01 -2.87299275e-01 9.54352081e-01
-1.43955314e+00 -7.91277766e-01 -6.79902792e-01 8.80673647e-01
7.24861845e-02 1.76836587e-02 4.53255415e-01 2.56164521e-01
-3.39230359e-01 -3.41792554e-02 3.10363889e-01 1.67683333e-01
8.34967375e-01 -1.20432651e+00 3.48771125e-01 1.31014621e+00
4.24167901e-01 8.62827122e-01 6.30792916e-01 -7.61120796e-01
-2.03050661e+00 -6.43505394e-01 4.75880682e-01 -1.05743229e+00
8.14046323e-01 -1.71115145e-01 -4.64661062e-01 8.36793005e-01
-1.25068739e-01 8.92161608e-01 2.70084411e-01 -7.03730062e-02
9.63096842e-02 -2.00313404e-01 -7.92968273e-01 3.71849477e-01
9.84735906e-01 -9.59207892e-01 -3.65440786e-01 2.26989910e-01
4.00879741e-01 -1.13780284e+00 -7.44534850e-01 -3.30599435e-02
8.29737067e-01 -6.98200762e-01 1.34365535e+00 -3.41894001e-01
-3.85629863e-01 -6.65372729e-01 -1.63189888e-01 -1.35934222e+00
-3.09049726e-01 -1.16687131e+00 -1.04116702e+00 5.77583492e-01
-4.24152523e-01 -5.87664604e-01 9.42520082e-01 5.77520192e-01
3.09048951e-01 6.33409396e-02 -1.23434186e+00 -5.91177046e-01
-6.38754904e-01 -7.12701738e-01 8.04344267e-02 5.87986529e-01
-6.24708496e-02 4.14067119e-01 -8.83094311e-01 4.83906090e-01
9.18734193e-01 -1.59694165e-01 1.17253172e+00 -1.14977598e+00
-2.94278562e-01 2.22062141e-01 -9.32400107e-01 -1.77068019e+00
1.42005637e-01 -3.74619246e-01 2.56027328e-03 -1.36478007e+00
-4.99691516e-01 -2.50553429e-01 1.82191618e-02 -2.07799137e-01
-3.64495516e-01 4.22459245e-01 4.60108399e-01 3.32195073e-01
-6.84188902e-01 5.54076791e-01 1.20715547e+00 4.54058163e-02
-3.45783621e-01 2.35839397e-01 -1.83188036e-01 7.92908967e-01
2.35583708e-01 -3.90601069e-01 -3.41054916e-01 -7.44539499e-01
4.30103451e-01 2.97637492e-01 5.45775473e-01 -1.34873974e+00
6.15246892e-01 1.59264565e-01 8.49783301e-01 -4.13032234e-01
4.89676595e-01 -9.42348480e-01 1.00441411e-01 4.15572703e-01
4.15663302e-01 3.39404225e-01 6.97059706e-02 1.01677144e+00
-2.11564720e-01 2.03098938e-01 3.45481247e-01 3.14734913e-02
-8.88531327e-01 5.95325112e-01 -1.69253439e-01 -3.10314074e-02
6.26048386e-01 -5.11991739e-01 -2.06344441e-01 -7.83261299e-01
-8.16079140e-01 -9.56985876e-02 6.70419633e-01 5.39388061e-01
6.05977297e-01 -1.55039656e+00 -7.69370347e-02 5.47418296e-01
-4.77653109e-02 -1.22072371e-02 2.43032917e-01 1.26036882e+00
-5.25286138e-01 5.54685831e-01 -2.94166654e-01 -1.38008535e+00
-7.37386584e-01 3.15804899e-01 4.97077405e-01 -3.61146405e-03
-7.61621833e-01 4.19695795e-01 -1.03627350e-02 -1.54466435e-01
1.09969877e-01 -5.47079206e-01 1.25238588e-02 -3.83667462e-02
8.03380549e-01 6.50261462e-01 3.54682580e-02 -1.14215899e+00
-4.31172550e-01 1.22273505e+00 5.22145867e-01 -3.78177375e-01
7.47381747e-01 -8.71996403e-01 3.69396299e-01 4.08730656e-01
1.39129078e+00 2.45749950e-01 -2.17905831e+00 -3.97865623e-01
-2.84043670e-01 -1.04515052e+00 2.59267837e-01 -3.63336876e-02
-9.77344573e-01 9.88027513e-01 9.95826662e-01 -1.65745303e-01
7.34146357e-01 -6.15588486e-01 6.72479272e-01 3.86497408e-01
7.70430028e-01 -7.00054944e-01 4.00149152e-02 7.73928285e-01
4.19797003e-01 -1.59703743e+00 3.85450423e-01 -5.78697771e-02
-2.73136497e-01 1.20405662e+00 2.39680186e-01 -3.35247636e-01
9.56482172e-01 1.22138299e-01 1.14990294e-01 3.65975380e-01
-1.69734776e-01 -2.80750751e-01 5.26089668e-01 9.85995412e-01
5.07933125e-02 -2.99585223e-01 3.57402027e-01 -1.23047039e-01
1.21524937e-01 2.02428579e-01 5.99884629e-01 1.30447674e+00
-1.38944492e-01 -8.25537622e-01 -6.30802929e-01 -7.71180866e-03
-2.35325411e-01 1.57667741e-01 2.79619157e-01 6.77987635e-01
-7.09368438e-02 8.49846601e-01 2.91560680e-01 -4.00253206e-01
1.60869330e-01 -1.81876346e-01 8.15576613e-01 -3.40127975e-01
-9.45382938e-02 1.69411644e-01 -2.12314591e-01 -1.44748402e+00
-8.41895163e-01 -9.20149505e-01 -9.95692968e-01 -3.52318376e-01
-1.37018953e-02 -2.32425302e-01 8.00584137e-01 1.09155667e+00
7.54501581e-01 1.92353442e-01 5.02519846e-01 -1.69401944e+00
6.15987275e-03 -5.74866414e-01 -4.80918258e-01 2.54213482e-01
1.21340394e+00 -1.00986910e+00 -4.83116508e-01 2.65796423e-01] | [7.869621753692627, -2.0535430908203125] |
ce8df833-ffee-472b-873b-7ebf571a49ac | sign-scalable-inception-graph-neural-networks | 2004.11198 | null | https://arxiv.org/abs/2004.11198v3 | https://arxiv.org/pdf/2004.11198v3.pdf | SIGN: Scalable Inception Graph Neural Networks | Graph representation learning has recently been applied to a broad spectrum of problems ranging from computer graphics and chemistry to high energy physics and social media. The popularity of graph neural networks has sparked interest, both in academia and in industry, in developing methods that scale to very large graphs such as Facebook or Twitter social networks. In most of these approaches, the computational cost is alleviated by a sampling strategy retaining a subset of node neighbors or subgraphs at training time. In this paper we propose a new, efficient and scalable graph deep learning architecture which sidesteps the need for graph sampling by using graph convolutional filters of different size that are amenable to efficient precomputation, allowing extremely fast training and inference. Our architecture allows using different local graph operators (e.g. motif-induced adjacency matrices or Personalized Page Rank diffusion matrix) to best suit the task at hand. We conduct extensive experimental evaluation on various open benchmarks and show that our approach is competitive with other state-of-the-art architectures, while requiring a fraction of the training and inference time. Moreover, we obtain state-of-the-art results on ogbn-papers100M, the largest public graph dataset, with over 110 million nodes and 1.5 billion edges. | ['Davide Eynard', 'Federico Monti', 'Fabrizio Frasca', 'Ben Chamberlain', 'Michael Bronstein', 'Emanuele Rossi'] | 2020-04-23 | null | null | null | null | ['graph-sampling'] | ['graphs'] | [ 2.75004338e-02 2.72191525e-01 -3.31592351e-01 -2.37581015e-01
-3.13642889e-01 -5.27242184e-01 6.23033881e-01 4.22999024e-01
-4.46347594e-01 6.89680278e-01 6.82987347e-02 -6.77028954e-01
-1.29614621e-01 -1.41742861e+00 -9.03778076e-01 -4.86192256e-01
-5.58013320e-01 6.68610275e-01 3.67677271e-01 -1.72068805e-01
5.38885184e-02 6.14480853e-01 -1.19603324e+00 7.92120770e-02
6.96131349e-01 7.42700100e-01 5.49727492e-02 6.54691517e-01
-2.76438206e-01 6.91041529e-01 -1.72487110e-01 -5.91554761e-01
2.07470343e-01 -8.85638520e-02 -8.29610527e-01 -1.16906933e-01
8.03167224e-01 -1.50905952e-01 -9.00667965e-01 1.13083375e+00
5.42280018e-01 3.10882866e-01 3.94242018e-01 -1.00009167e+00
-7.45793700e-01 9.49083924e-01 -6.32759571e-01 1.40001699e-01
1.94501743e-01 7.50136226e-02 1.33812547e+00 -6.32015407e-01
8.40755343e-01 1.28348351e+00 7.92937160e-01 1.72414422e-01
-1.52023804e+00 -6.19124174e-01 3.66895080e-01 2.31696233e-01
-1.36748040e+00 1.17468310e-03 8.03058207e-01 -2.52083808e-01
1.13697481e+00 1.85544565e-01 7.92411685e-01 9.72122431e-01
1.48332641e-01 4.90406215e-01 8.48033667e-01 -1.65369228e-01
1.44476831e-01 -3.70539814e-01 2.17375264e-01 1.16321468e+00
6.66021705e-01 -1.88437760e-01 -1.82610542e-01 -1.92099109e-01
8.23967874e-01 3.13580520e-02 -2.07385689e-01 -4.52379286e-01
-8.84686410e-01 1.14664876e+00 9.76075947e-01 2.06600890e-01
-2.84209102e-01 6.35886014e-01 5.75298786e-01 3.13469410e-01
6.19516671e-01 3.66006106e-01 -1.23838685e-01 2.20383078e-01
-7.82317758e-01 1.76178589e-01 1.00871944e+00 8.26470196e-01
1.18327320e+00 -1.86929498e-02 2.45706197e-02 7.61808455e-01
1.73189893e-01 1.93004876e-01 1.13171771e-01 -5.77231348e-01
4.33486402e-01 9.53167498e-01 -4.22437161e-01 -1.31866133e+00
-6.27162278e-01 -5.45555413e-01 -1.35659492e+00 -3.74060087e-02
3.11992913e-01 5.94333336e-02 -1.09372222e+00 1.58239639e+00
3.77481997e-01 4.32158023e-01 -4.63489443e-01 6.90431297e-01
1.00950956e+00 6.06520712e-01 -1.27313212e-01 2.97243178e-01
1.25417209e+00 -9.48746562e-01 -1.46476030e-01 -3.27857673e-01
7.68538356e-01 -2.56254137e-01 1.07156074e+00 2.97031075e-01
-7.93079078e-01 -3.05679470e-01 -9.11075652e-01 -2.89365053e-01
-6.15949929e-01 -2.85258025e-01 1.26736128e+00 5.73112905e-01
-1.26845968e+00 1.09789634e+00 -8.27500880e-01 -4.94509518e-01
6.66520536e-01 5.55491686e-01 -5.14693558e-01 -2.77520716e-01
-1.03857601e+00 3.17254931e-01 3.21344674e-01 4.11704108e-02
-6.17827952e-01 -6.57615125e-01 -8.66856813e-01 4.36915785e-01
4.71947789e-01 -7.99365819e-01 7.27577686e-01 -6.49136543e-01
-1.32721066e+00 7.54679978e-01 1.99066415e-01 -7.27535605e-01
3.16189528e-01 3.87466233e-03 -2.96222717e-01 1.20614253e-01
1.58186723e-03 4.90445346e-01 4.19904977e-01 -6.44071102e-01
-9.59720984e-02 -3.06691110e-01 3.12645495e-01 -2.94620935e-02
-6.41407073e-01 -2.42174909e-01 -7.61865199e-01 -4.52056289e-01
-1.11267626e-01 -1.11488402e+00 -5.55855155e-01 -6.28948808e-02
-5.45748532e-01 -3.58477175e-01 5.78385115e-01 -3.48495007e-01
1.25636911e+00 -1.79704571e+00 2.38932103e-01 5.55030167e-01
8.01889837e-01 1.83861241e-01 -2.98427016e-01 7.35418200e-01
4.88836169e-02 2.81471372e-01 -1.88090503e-01 -2.15812072e-01
2.23696958e-02 1.73114866e-01 -1.24358316e-03 5.86574912e-01
3.10446098e-02 1.07692897e+00 -9.55322742e-01 -2.38060758e-01
1.97046697e-01 5.35267770e-01 -8.04048121e-01 7.79170543e-02
-3.98998618e-01 9.47471559e-02 -4.52836901e-01 3.03667188e-01
6.07517838e-01 -1.09787619e+00 5.86152792e-01 -9.51705799e-02
2.03678429e-01 6.17443025e-01 -1.19908369e+00 1.70149589e+00
-4.52445835e-01 5.85977793e-01 9.89631712e-02 -1.19772029e+00
8.50753427e-01 -2.26412341e-01 4.93819386e-01 -4.82278973e-01
1.48939207e-01 6.30273372e-02 1.97268531e-01 7.48927752e-03
4.27474499e-01 3.77984047e-01 8.92456099e-02 5.10561347e-01
7.44218379e-02 1.28787234e-01 6.35759234e-01 5.91381490e-01
1.58695209e+00 -2.83529907e-01 1.29282236e-01 -4.84579027e-01
2.91676342e-01 -2.41252348e-01 1.56459063e-01 8.91019583e-01
2.67349213e-01 2.82010555e-01 7.25150406e-01 -8.79260957e-01
-8.44139159e-01 -8.11248958e-01 2.68393338e-01 1.15221167e+00
-7.61099458e-02 -7.92427599e-01 -5.20536542e-01 -6.42733276e-01
2.08403885e-01 1.52049482e-01 -5.95741093e-01 -2.23098189e-01
-7.62926817e-01 -8.53693485e-01 2.72427738e-01 4.70493168e-01
4.06867117e-01 -1.03606486e+00 1.07019216e-01 4.07274216e-01
4.01609629e-01 -1.20278442e+00 -4.16571289e-01 3.55285630e-02
-9.05582845e-01 -1.14984035e+00 -3.36487055e-01 -8.08471322e-01
6.89072311e-01 3.91209930e-01 1.51332283e+00 4.16688085e-01
-3.84046644e-01 8.33806247e-02 -1.73027873e-01 5.50900139e-02
-2.62190729e-01 8.19802463e-01 -2.53671318e-01 -2.51841247e-02
3.68841849e-02 -1.02068102e+00 -6.53686047e-01 -1.36428803e-01
-8.14347804e-01 1.57843232e-01 6.76113784e-01 8.33622813e-01
5.23932457e-01 -5.59093878e-02 3.90041918e-01 -1.54541600e+00
6.35125756e-01 -5.80317557e-01 -9.61583734e-01 -8.24209005e-02
-7.34965384e-01 2.35429630e-01 8.48265469e-01 -2.96052486e-01
-2.95243770e-01 -3.79551984e-02 -1.94184422e-01 -2.87225246e-01
2.36266300e-01 9.61008787e-01 7.49804303e-02 -4.55549300e-01
6.43346786e-01 4.95731831e-03 3.52964588e-02 -4.53066170e-01
5.28193772e-01 1.17601037e-01 2.54206300e-01 -4.33972955e-01
7.71830678e-01 4.91324782e-01 4.95895684e-01 -9.34432805e-01
-6.54807150e-01 -3.63146305e-01 -2.12628886e-01 -1.62115142e-01
5.50225973e-01 -8.73551428e-01 -9.57932115e-01 4.11436439e-01
-9.59248781e-01 -5.24564922e-01 4.46492769e-02 1.96426332e-01
-1.46137387e-01 5.52765191e-01 -9.09131348e-01 -2.31738955e-01
-6.45802617e-01 -1.14283574e+00 1.09762180e+00 9.64341033e-03
2.10771058e-02 -1.17453265e+00 8.39979723e-02 -2.65499894e-02
5.28574407e-01 4.79854792e-01 1.09012377e+00 -6.00298464e-01
-1.14211226e+00 -3.24350268e-01 -5.37697971e-01 6.94513619e-02
5.67647107e-02 -1.61165595e-01 -5.90799093e-01 -7.12829590e-01
-9.03973341e-01 -2.69923329e-01 1.30305743e+00 2.98559785e-01
1.59925103e+00 -3.37379217e-01 -5.92253149e-01 8.91346633e-01
1.47686851e+00 -5.91852903e-01 6.21959329e-01 8.94994009e-03
1.27249837e+00 1.17333226e-01 -1.48013607e-01 1.62296191e-01
4.50806916e-01 4.28685129e-01 6.35129869e-01 -3.08324069e-01
-1.84791297e-01 -4.09393013e-01 4.60288078e-02 8.20968986e-01
-1.10889882e-01 -5.69290340e-01 -8.97465169e-01 3.82261187e-01
-1.85513306e+00 -8.24684978e-01 -1.96835682e-01 2.16358304e+00
5.30352592e-01 3.94933820e-01 1.63831830e-01 -9.35370922e-02
7.46832192e-01 6.13011479e-01 -5.67926109e-01 -2.33096540e-01
6.05065227e-02 5.95727980e-01 8.55460763e-01 6.15970969e-01
-1.08245063e+00 1.05640876e+00 5.93850899e+00 7.48008490e-01
-1.13835669e+00 -1.92529693e-01 7.57357061e-01 1.64246202e-01
-5.35128653e-01 9.39858779e-02 -5.62853456e-01 3.53760809e-01
9.78737533e-01 -1.60276026e-01 9.32646871e-01 8.84653032e-01
-2.34892666e-01 2.61998892e-01 -9.97354567e-01 9.53450799e-01
-2.56457478e-01 -2.01198411e+00 1.66297123e-01 4.07680809e-01
7.19873250e-01 6.43367052e-01 -2.20692217e-01 3.30310374e-01
7.78660059e-01 -1.19508171e+00 1.66758791e-01 2.96261907e-01
9.20370400e-01 -6.83942854e-01 3.90465945e-01 8.08608904e-02
-1.54022658e+00 1.11054339e-01 -5.35244584e-01 -2.54278600e-01
-1.20528445e-01 9.79736924e-01 -9.28669333e-01 6.18285954e-01
6.90035939e-01 8.42635572e-01 -6.09249771e-01 8.95108461e-01
-1.15338281e-01 7.43635118e-01 -5.64408422e-01 -3.95490974e-01
3.45616043e-01 -4.28521663e-01 3.82839292e-01 1.13226831e+00
2.44956031e-01 -2.89904863e-01 4.65658456e-01 8.53184640e-01
-8.91413152e-01 2.29918301e-01 -8.68955851e-01 -5.50756037e-01
5.70110083e-01 1.48922372e+00 -9.58391190e-01 -3.72270882e-01
-6.40680134e-01 7.98488915e-01 1.09224367e+00 2.71687925e-01
-6.49464071e-01 -5.45103550e-01 6.38977706e-01 4.26061332e-01
4.46707308e-01 -4.75494444e-01 2.35678285e-01 -1.06973755e+00
-1.01737060e-01 -7.08231449e-01 4.91147786e-01 -2.60364920e-01
-1.47103882e+00 6.49331152e-01 -3.32487166e-01 -5.82602143e-01
-6.35033846e-02 -9.24520135e-01 -7.84241736e-01 7.65651166e-01
-1.44849265e+00 -1.02665985e+00 -4.92460936e-01 3.89416069e-01
-1.30160302e-01 -3.41154486e-02 7.10694432e-01 6.64347529e-01
-5.28742850e-01 4.51370180e-01 1.15535043e-01 3.60587418e-01
4.05819774e-01 -1.41936719e+00 1.35393643e+00 6.82312071e-01
3.87371868e-01 5.91779709e-01 3.86685431e-01 -6.64930522e-01
-1.96750069e+00 -1.33037233e+00 6.48594975e-01 6.32034661e-03
9.90889668e-01 -8.50084484e-01 -9.23502326e-01 7.65370131e-01
5.70559502e-03 5.33996522e-01 3.02740186e-01 5.39976835e-01
-4.99250174e-01 -2.99476445e-01 -8.04466546e-01 6.86473012e-01
1.43697047e+00 -4.90188241e-01 3.48893434e-01 7.31707633e-01
8.58610868e-01 -4.70488101e-01 -9.23558772e-01 2.71360725e-01
3.55629891e-01 -8.69086683e-01 8.87916863e-01 -7.56104887e-01
2.03145564e-01 -5.20446524e-02 1.46276653e-01 -1.27243686e+00
-7.64047682e-01 -8.47318053e-01 -3.37604493e-01 8.74927282e-01
3.90075564e-01 -9.16665852e-01 1.19990802e+00 1.01599887e-01
-1.52351141e-01 -1.07871878e+00 -7.18915224e-01 -5.56963265e-01
-6.34174645e-02 -1.84371412e-01 8.11395288e-01 9.13578808e-01
-3.42617422e-01 7.60176539e-01 -3.39804649e-01 1.11254700e-01
5.89381993e-01 4.76212949e-01 1.06364083e+00 -1.52525556e+00
-5.23418963e-01 -6.68876588e-01 -6.44812763e-01 -1.22383380e+00
2.77272731e-01 -1.45915091e+00 -4.94922757e-01 -1.93181741e+00
1.04424894e-01 -3.73951733e-01 -2.03967869e-01 5.06344259e-01
-2.38836780e-01 3.87506932e-01 -8.89398679e-02 -2.20405504e-01
-6.66051507e-01 4.53880787e-01 1.23179197e+00 -3.30744922e-01
3.06350980e-02 -2.87570208e-01 -7.86668599e-01 4.03147757e-01
6.55642927e-01 -3.76204699e-01 -3.86023432e-01 -4.98944163e-01
7.03156352e-01 -1.88652247e-01 4.61307317e-01 -9.45252538e-01
1.60553023e-01 1.29601985e-01 1.69518769e-01 -3.33001047e-01
1.88998774e-01 -4.58884150e-01 1.74088389e-01 4.72712040e-01
-1.13380000e-01 1.88324466e-01 1.38906226e-01 8.18360567e-01
8.18026438e-02 1.53896391e-01 7.57666647e-01 -1.76255614e-01
-5.33586204e-01 9.28347111e-01 3.02004274e-02 1.77105099e-01
5.45153201e-01 1.05162054e-01 -4.15000439e-01 -4.09363270e-01
-4.76266980e-01 2.82160282e-01 5.18127441e-01 1.88470885e-01
3.24101806e-01 -1.32005894e+00 -7.93754935e-01 1.03410602e-01
-2.40997486e-02 9.73736644e-02 1.24964096e-01 6.35503948e-01
-7.40707636e-01 1.11488655e-01 3.06352992e-02 -5.00854313e-01
-9.21832979e-01 5.17237842e-01 2.31942385e-01 -6.72406614e-01
-9.69271660e-01 8.86160612e-01 4.70080264e-02 -4.67468977e-01
4.01599854e-02 -3.14868152e-01 2.23197177e-01 -2.92843819e-01
1.79786801e-01 3.91072333e-01 2.35185698e-01 -2.93052197e-01
-3.04144531e-01 2.18449309e-01 -3.62268001e-01 6.48828804e-01
1.55966258e+00 3.22914988e-01 -4.13276315e-01 1.11937687e-01
1.34556091e+00 9.58083048e-02 -9.58316684e-01 -3.09784055e-01
-4.24230173e-02 -3.02437425e-01 1.64612100e-01 -3.60149860e-01
-1.44124579e+00 6.78809643e-01 1.36937812e-01 6.38500631e-01
8.41639280e-01 1.12218626e-01 1.02856243e+00 8.41328681e-01
4.95792598e-01 -7.46757448e-01 -4.33475636e-02 5.89884043e-01
6.23803735e-01 -1.22264111e+00 3.90723228e-01 -5.86495161e-01
-8.28004442e-03 9.69666898e-01 4.07581151e-01 -7.66989708e-01
7.90064573e-01 1.47232324e-01 -5.74996591e-01 -5.48307061e-01
-7.40439832e-01 -1.49208799e-01 4.18466985e-01 2.55286157e-01
6.22240841e-01 2.69248515e-01 -2.56877720e-01 1.92341775e-01
-1.67567283e-01 -3.20195138e-01 3.56186032e-01 5.20761371e-01
-4.65297043e-01 -1.29529619e+00 3.17335784e-01 1.06238925e+00
-2.18835667e-01 -3.58686268e-01 -5.06273508e-01 8.60025287e-01
-4.39025521e-01 4.87414867e-01 9.60794240e-02 -2.55504906e-01
4.22510765e-02 -2.69725084e-01 5.26591718e-01 -7.53124774e-01
-4.65686589e-01 -2.72037745e-01 3.97096306e-01 -8.11139047e-01
-1.23977780e-01 -1.86148226e-01 -1.19966042e+00 -7.39923179e-01
-2.88749665e-01 -2.49382425e-02 3.89491916e-01 5.13154805e-01
7.50825405e-01 7.29185343e-01 4.11552668e-01 -9.39048409e-01
-3.76309097e-01 -9.46323752e-01 -6.21026158e-01 4.37112540e-01
1.22088335e-01 -5.50160527e-01 -1.57885864e-01 -6.30280018e-01] | [6.953188419342041, 6.139644145965576] |
5ef3cd4b-5e44-444a-867d-6a159515cefe | stochastic-mpc-for-energy-hubs-using-data | 2304.12438 | null | https://arxiv.org/abs/2304.12438v1 | https://arxiv.org/pdf/2304.12438v1.pdf | Stochastic MPC for energy hubs using data driven demand forecasting | Energy hubs convert and distribute energy resources by combining different energy inputs through multiple conversion and storage components. The optimal operation of the energy hub exploits its flexibility to increase the energy efficiency and reduce the operational costs. However, uncertainties in the demand present challenges to energy hub optimization. In this paper, we propose a stochastic MPC controller to minimize energy costs using chance constraints for the uncertain electricity and thermal demands. Historical data is used to build a demand prediction model based on Gaussian processes to generate a forecast of the future electricity and heat demands. The stochastic optimization problem is solved via the Scenario Approach by sampling multi-step demand trajectories from the derived prediction model. The performance of the proposed predictor and of the stochastic controller is verified on a simulated energy hub model and demand data from a real building. | ['John Lygeros', 'Philipp Heer', 'Jonas Mehr', 'Varsha Behrunani', 'Francesco Micheli'] | 2023-04-24 | null | null | null | null | ['stochastic-optimization'] | ['methodology'] | [-3.53804827e-01 -1.07788727e-01 1.39556363e-01 -8.00883584e-03
-3.82662565e-01 -7.39936352e-01 5.08488238e-01 1.07348226e-01
3.66193593e-01 1.11417234e+00 1.48052517e-02 -9.15420800e-02
-5.33023119e-01 -1.16862357e+00 -3.32837254e-01 -1.08883607e+00
2.22717181e-01 6.46732867e-01 -4.04138446e-01 1.19269595e-01
-1.03580534e-01 5.67133427e-01 -1.35934210e+00 -6.25019610e-01
1.08981872e+00 1.09081149e+00 7.22131729e-01 5.27776837e-01
1.34702861e-01 1.63138300e-01 -4.36897039e-01 2.35631809e-01
4.79936451e-01 -3.33657950e-01 -1.24230623e-01 -1.67629480e-01
-1.17810643e+00 -3.89257133e-01 2.15516225e-01 8.78079832e-01
6.39742970e-01 8.56782377e-01 7.79555619e-01 -1.31275761e+00
-4.58049357e-01 4.00466204e-01 1.25929505e-01 -1.08869731e-01
-2.42636085e-01 2.70812184e-01 5.29883921e-01 -4.80601132e-01
-8.27805325e-02 8.91460657e-01 5.30115068e-02 2.39381075e-01
-1.36884761e+00 -6.80500194e-02 -5.08577982e-03 1.56580165e-01
-1.51881254e+00 -3.79074067e-02 8.77159178e-01 -3.87683183e-01
1.18143320e+00 3.88588399e-01 1.07412779e+00 6.01974070e-01
4.07885939e-01 4.92532998e-01 1.25984657e+00 -2.71573812e-01
8.89121771e-01 1.37045592e-01 -4.48134124e-01 -1.84671804e-01
1.78318202e-01 4.57883030e-01 2.11350247e-01 -1.39165908e-01
1.95951030e-01 2.04708368e-01 6.69553643e-03 1.18657224e-01
-4.99047548e-01 6.12728119e-01 1.94222182e-01 1.76001430e-01
-1.06641030e+00 1.15430407e-01 -1.48591906e-01 -2.74962217e-01
4.31660831e-01 2.61270523e-01 -5.89199603e-01 -1.16712704e-01
-1.14454019e+00 -5.04270121e-02 1.06816816e+00 1.25258565e+00
2.04679459e-01 7.07827866e-01 -4.76225346e-01 1.00793786e-01
4.49044973e-01 1.46014643e+00 2.85926610e-01 -8.09333563e-01
-1.52536696e-02 1.25429928e-01 7.56846786e-01 -5.94282568e-01
-4.83639956e-01 -4.28136647e-01 -9.66465235e-01 2.77088821e-01
-1.73293024e-01 -8.50500405e-01 -7.87805855e-01 1.38732517e+00
1.10441111e-01 -2.31714826e-02 -1.23244859e-01 7.44831562e-01
-3.45448226e-01 1.31559134e+00 9.46378931e-02 -7.10569322e-01
8.91648412e-01 -6.43558562e-01 -1.03301525e+00 4.77752984e-01
-1.44132018e-01 -5.77563643e-01 1.38167381e-01 1.13754734e-01
-1.22608423e+00 -3.81372631e-01 -6.21805251e-01 3.72023761e-01
-6.73559427e-01 5.05700111e-01 -2.43361279e-01 3.58357459e-01
-8.74910772e-01 6.79066241e-01 -1.18734539e+00 -1.06442168e-01
-2.25272357e-01 1.85049370e-01 7.48157263e-01 6.53569460e-01
-1.16819465e+00 1.30159485e+00 7.58484304e-01 6.42938137e-01
-1.01965296e+00 -8.29446673e-01 -3.02908331e-01 6.71390653e-01
2.20304430e-01 -9.06842947e-01 1.14555526e+00 -2.43355528e-01
-2.18284082e+00 -6.43337965e-01 -1.57368928e-01 -3.66000414e-01
6.74696624e-01 1.24004133e-01 -3.54704618e-01 -1.84418976e-01
-3.06343406e-01 -2.97086090e-01 6.32420361e-01 -9.53224897e-01
-7.49939024e-01 -3.02711688e-02 -7.67259181e-01 1.82887584e-01
1.26783594e-01 -3.84128362e-01 5.03509343e-01 -3.00278574e-01
-5.00237405e-01 -1.16038346e+00 -3.83445233e-01 -9.51848507e-01
-8.22048306e-01 -2.53828853e-01 6.99914753e-01 -1.08699906e+00
9.94857669e-01 -1.66266215e+00 3.88918698e-01 5.91869891e-01
-7.35997856e-01 -3.05234671e-01 5.86122990e-01 9.57135141e-01
-1.04426481e-02 2.12185726e-01 -7.67985508e-02 -3.40414107e-01
4.42097813e-01 4.95326728e-01 -4.64244336e-01 2.30502471e-01
-3.91718037e-02 7.27748632e-01 -9.04513597e-01 2.29437619e-01
7.47281015e-01 4.98467714e-01 9.51999277e-02 3.17966640e-01
-6.94287002e-01 3.69911581e-01 -6.98236227e-01 3.44506800e-01
6.34971678e-01 1.63227320e-01 1.10817514e-01 -1.36255890e-01
-6.66007936e-01 -3.30404669e-01 -1.32234943e+00 1.07525563e+00
-1.15152991e+00 3.19945514e-01 3.60336602e-01 -8.89944851e-01
7.93383002e-01 3.35089982e-01 7.15799510e-01 -1.13176726e-01
2.64809877e-01 3.30623269e-01 -3.06544065e-01 -1.69276580e-01
5.33869863e-01 -5.21309555e-01 2.28207588e-01 2.37382919e-01
-6.55746460e-02 -6.74836278e-01 2.67513242e-04 -5.17279804e-01
5.61298192e-01 2.02448033e-02 1.02031559e-01 -6.28055453e-01
4.38417852e-01 -1.94847107e-01 6.97488487e-01 1.59367584e-02
2.69865483e-01 -7.94995204e-03 1.46642718e-02 5.96038960e-02
-1.21623433e+00 -1.10317755e+00 1.25794798e-01 2.76949733e-01
1.86976120e-02 4.01610196e-01 -2.60841161e-01 7.62048736e-02
2.95786113e-01 2.13707876e+00 -2.53108919e-01 -1.27025381e-01
-2.38128439e-01 -1.00183535e+00 -5.46948791e-01 3.06591362e-01
8.32000077e-02 -3.68645310e-01 -8.46620977e-01 5.05372524e-01
4.13270406e-02 -6.00749314e-01 -1.65886313e-01 4.09586340e-01
-5.38152874e-01 -5.16383350e-01 -6.04457021e-01 -1.52942702e-01
8.30627739e-01 -3.06320816e-01 8.13025773e-01 -4.53942120e-01
-2.58045569e-02 4.67956454e-01 7.97724798e-02 -8.84252667e-01
-3.39266479e-01 7.74312839e-02 4.59020734e-01 9.19588096e-03
-3.45779926e-01 -6.62070334e-01 -6.98610961e-01 2.00969845e-01
-5.49846172e-01 2.28797421e-01 2.13330597e-01 7.17337966e-01
8.28007281e-01 1.03174174e+00 6.23590171e-01 -1.54956862e-01
8.73816252e-01 -7.00290382e-01 -1.36931384e+00 6.48676991e-01
-1.14281774e+00 2.89538622e-01 1.06623030e+00 -6.66749403e-02
-1.42196774e+00 2.73984313e-01 4.87420201e-01 -5.39555430e-01
4.97412942e-02 2.89543509e-01 -2.92855442e-01 4.60366696e-01
-6.52653098e-01 5.12418389e-01 -3.02487165e-01 -6.33249879e-01
3.85433853e-01 2.53859758e-01 3.22442710e-01 -7.68904686e-01
1.25825596e+00 -2.80849077e-03 7.54129887e-01 -5.94216585e-01
-6.71314225e-02 -5.64746298e-02 -3.50719810e-01 -2.58892655e-01
8.85955632e-01 -8.32422853e-01 -1.17914104e+00 3.08559000e-01
-1.02180874e+00 -3.25103045e-01 -7.76175797e-01 5.37828326e-01
-9.68435705e-01 -3.72446179e-01 -1.03941210e-01 -1.63317347e+00
-4.71184611e-01 -1.14560413e+00 5.42862236e-01 7.14350700e-01
2.32201084e-01 -1.23620760e+00 1.73396632e-01 -5.28610110e-01
8.88951182e-01 7.17678607e-01 8.27514172e-01 -2.54333705e-01
-8.98131728e-01 -1.77520543e-01 5.16769290e-01 6.85236156e-01
3.92819613e-01 4.53252316e-01 -4.69629854e-01 -3.46359253e-01
3.29594523e-01 4.12328839e-01 -2.04965286e-02 6.62828207e-01
1.00951838e+00 -7.32412100e-01 -3.25035632e-01 4.35045034e-01
2.16168189e+00 7.14195073e-01 -3.46866250e-02 -8.34080018e-03
1.59839675e-01 3.48386496e-01 3.15440446e-01 9.56637800e-01
3.63398254e-01 3.69990170e-01 5.83202600e-01 4.81215179e-01
7.59317160e-01 -4.50275213e-01 1.88680857e-01 7.94447958e-01
-3.57527018e-01 -4.88902390e-01 -8.92950356e-01 7.19912112e-01
-1.73558438e+00 -1.23574066e+00 -1.17952399e-01 2.28336596e+00
5.32370687e-01 -3.74716938e-01 -1.58625990e-01 -3.08155924e-01
5.96423030e-01 -2.28434175e-01 -5.93407035e-01 -7.23737478e-01
-2.22468656e-02 6.40657311e-03 1.14085400e+00 5.37344992e-01
-4.54701334e-01 3.56535502e-02 5.86858416e+00 7.88154602e-01
-8.55423450e-01 -6.91804364e-02 8.64798188e-01 -3.79345268e-01
-4.44792986e-01 2.18735933e-02 -7.69429028e-01 1.12353706e+00
1.56893885e+00 -1.22953248e+00 1.12031317e+00 9.38261569e-01
1.24230933e+00 -3.98573756e-01 -9.66284871e-01 5.71244478e-01
-6.00036323e-01 -1.06345177e+00 -5.19382536e-01 1.63184643e-01
1.34974015e+00 9.87152755e-02 -3.99707211e-03 9.27339047e-02
8.96531999e-01 -5.90174317e-01 7.89073288e-01 1.42920530e+00
1.68860242e-01 -1.13736236e+00 8.96910369e-01 8.28569353e-01
-1.20653594e+00 -5.39157510e-01 -6.39034212e-02 1.69171840e-01
9.86168444e-01 6.44951046e-01 -8.07848036e-01 9.75276709e-01
3.42097610e-01 9.92523730e-02 -3.72223486e-03 9.60179627e-01
-4.04750317e-01 4.60553467e-01 -8.85095656e-01 -3.53919506e-01
-5.38278706e-02 -9.04407084e-01 5.60224473e-01 8.30935359e-01
1.06006074e+00 2.64258474e-01 1.87244833e-01 1.24536157e+00
3.84072185e-01 -2.35413060e-01 -2.76674986e-01 -1.09553829e-01
7.69117892e-01 1.38679481e+00 -4.30578530e-01 -1.57556355e-01
-1.47505984e-01 7.52605617e-01 -7.18394697e-01 8.06422234e-01
-9.19519484e-01 -3.17163646e-01 6.43020153e-01 -3.17401767e-01
2.55079240e-01 -5.45921564e-01 -5.35759628e-01 -6.99561417e-01
-1.11455724e-01 2.28611290e-01 -1.28603533e-01 -8.00559938e-01
-1.34041834e+00 1.37470573e-01 3.49035382e-01 -1.05480623e+00
-8.82740796e-01 -3.58390868e-01 -1.08004200e+00 1.76245522e+00
-1.50535190e+00 -7.12637365e-01 -1.41068446e-02 3.09261143e-01
6.78898692e-01 1.98573694e-02 7.99901605e-01 -3.07085365e-01
-1.05307734e+00 -5.03267944e-01 1.33794546e+00 -4.54784811e-01
-1.01196125e-01 -1.51966727e+00 -1.56115880e-02 8.33159447e-01
-8.52230370e-01 1.79724589e-01 1.08327532e+00 -7.12702751e-01
-1.66946864e+00 -1.16182816e+00 5.38890600e-01 5.91579080e-03
8.50886762e-01 -8.08684304e-02 -5.06344914e-01 3.03914458e-01
7.60724962e-01 -1.71872973e-01 4.15519565e-01 -7.60911584e-01
7.67587543e-01 -2.15007663e-01 -1.54545450e+00 1.32974446e-01
7.20791817e-02 -2.86460459e-01 -2.94574022e-01 3.24135959e-01
5.72886884e-01 -1.05060481e-01 -1.15062845e+00 1.62031844e-01
2.62637496e-01 -5.43196872e-02 5.31888127e-01 -2.39892781e-01
-5.22302389e-02 -2.66245365e-01 -2.90722609e-01 -2.36463141e+00
-3.24070841e-01 -1.03581691e+00 -5.56142211e-01 1.65440500e+00
3.23414057e-01 -7.05103278e-01 1.11293159e-01 1.55313885e+00
-5.84365577e-02 -5.23875654e-01 -1.44535041e+00 -1.16044533e+00
1.44175172e-01 -4.40785252e-02 9.52406764e-01 5.07465839e-01
-5.74936233e-02 -1.90983400e-01 -7.20160529e-02 7.20891654e-01
1.00158370e+00 3.33212644e-01 1.53046042e-01 -7.87795603e-01
-8.73379335e-02 -4.65257645e-01 3.56582344e-01 -2.62338459e-01
1.40830368e-01 -5.65430105e-01 4.03715581e-01 -1.92496800e+00
-2.51019120e-01 -1.39241610e-02 -3.29262316e-01 -2.77104997e-03
8.07045251e-02 -8.89675558e-01 6.16755903e-01 -7.25718886e-02
2.54768699e-01 1.47056270e+00 6.90049112e-01 -9.63405222e-02
-3.43874305e-01 6.82262957e-01 1.03373580e-01 5.01976132e-01
1.24659979e+00 -1.94016695e-01 -7.84528553e-01 -1.80193305e-01
1.01456471e-01 4.17444348e-01 1.24287114e-01 -1.04169512e+00
2.91706145e-01 -8.76297653e-01 3.44812572e-01 -1.04201353e+00
5.70376925e-02 -1.63326228e+00 1.27805376e+00 5.91775477e-01
2.19448842e-02 3.06056261e-01 1.72568932e-01 8.27885330e-01
9.63829979e-02 -4.28296216e-02 4.68969792e-01 5.67444898e-02
-2.81057477e-01 1.31464675e-01 -5.86728632e-01 -5.23802638e-01
1.55983138e+00 3.82932782e-01 -1.92956641e-01 -3.38995099e-01
-1.12208796e+00 9.69384968e-01 2.53233761e-01 2.60181934e-01
2.20526725e-01 -1.23147297e+00 -4.92778718e-01 -1.34741232e-01
-8.83479297e-01 -5.56239523e-02 1.75097734e-01 6.10930920e-01
-2.47622743e-01 6.30902648e-01 1.28986430e-03 -2.88381368e-01
-4.18323398e-01 8.68121386e-01 7.25458503e-01 -2.07711041e-01
-5.48108518e-02 2.31159106e-02 -5.59806705e-01 2.93967854e-02
-3.53776008e-01 -7.43460894e-01 3.27834398e-01 3.76537353e-01
2.10101917e-01 1.12755322e+00 -1.44566625e-01 -3.61938864e-01
-1.93336159e-01 2.07228661e-01 1.08197153e+00 -2.48997852e-01
1.56276023e+00 -7.00192630e-01 -2.24875018e-01 7.16770768e-01
9.30869162e-01 -2.10918993e-01 -1.35170329e+00 2.99106508e-01
4.15708590e-03 -1.23903260e-01 3.94420773e-01 -1.18686843e+00
-1.10013318e+00 4.07030314e-01 5.43190420e-01 8.04280102e-01
1.40861082e+00 -6.37930512e-01 4.68905807e-01 1.30248249e-01
5.77488422e-01 -1.62879062e+00 -9.42039907e-01 3.79040360e-01
9.19872105e-01 -7.30578363e-01 -2.13923216e-01 1.29765168e-01
-4.95773464e-01 1.20202982e+00 9.30144414e-02 -1.26276761e-01
1.13333130e+00 3.66975397e-01 -7.39219546e-01 4.58909720e-01
-8.71216595e-01 -1.33928865e-01 -1.41655087e-01 2.98354477e-01
-1.94943279e-01 8.72753739e-01 -4.19973761e-01 8.49275231e-01
-5.74086867e-02 2.80353904e-01 5.46044290e-01 7.20268130e-01
-1.84237286e-01 -7.40068197e-01 -5.35347342e-01 1.76548362e-01
-1.45981148e-01 1.58817261e-01 4.92460221e-01 3.70524794e-01
1.58881530e-01 1.10538125e+00 1.05577722e-01 -1.06358886e-01
7.06044912e-01 5.09097397e-01 -2.67428815e-01 -3.27773422e-01
-2.59150594e-01 7.77940899e-02 -1.30678639e-01 -1.92164719e-01
9.27731246e-02 -8.17905009e-01 -1.35813141e+00 -4.90981579e-01
-4.70807016e-01 5.44679880e-01 1.27958000e+00 8.11116040e-01
4.77285355e-01 7.99706638e-01 1.55943680e+00 -9.52234924e-01
-1.21096218e+00 -7.79766381e-01 -9.27447498e-01 -3.65935057e-01
1.88714534e-01 -5.38654566e-01 -6.82741344e-01 -1.39894620e-01] | [5.6777729988098145, 2.523050546646118] |
1960aa36-86ca-4504-a81b-dd1741f258f7 | a-spatial-compositional-model-scm-for-linear | 1509.09243 | null | http://arxiv.org/abs/1509.09243v1 | http://arxiv.org/pdf/1509.09243v1.pdf | A spatial compositional model (SCM) for linear unmixing and endmember uncertainty estimation | The normal compositional model (NCM) has been extensively used in
hyperspectral unmixing. However, most of the previous research has focused on
estimation of endmembers and/or their variability. Also, little work has
employed spatial information in NCM. In this paper, we show that NCM can be
used for calculating the uncertainty of the estimated endmembers with spatial
priors incorporated for better unmixing. This results in a spatial
compositional model (SCM) which features (i) spatial priors that force
neighboring abundances to be similar based on their pixel similarity and (ii) a
posterior that is obtained from a likelihood model which does not assume pixel
independence. The resulting algorithm turns out to be easy to implement and
efficient to run. We compared SCM with current state-of-the-art algorithms on
synthetic and real images. The results show that SCM can in the main provide
more accurate endmembers and abundances. Moreover, the estimated uncertainty
can serve as a prediction of endmember error under certain conditions. | ['Yuan Zhou', 'Anand Rangarajan', 'Paul Gader'] | 2015-09-30 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 5.40059745e-01 -4.78307188e-01 -2.32797787e-02 -1.74794406e-01
-3.43218744e-01 -4.84304488e-01 8.24318349e-01 7.85349160e-02
-1.12215057e-01 1.00093746e+00 1.83204964e-01 -3.82184625e-01
-1.55700058e-01 -9.12129521e-01 -5.22058308e-01 -1.19401371e+00
-1.01423357e-02 4.02701706e-01 1.08317472e-01 1.82652175e-02
3.31261801e-03 5.41403532e-01 -1.95712173e+00 1.16992183e-01
1.26851785e+00 7.82045484e-01 4.99890774e-01 3.26257795e-01
-9.15236771e-02 6.25015855e-01 -3.90842348e-01 2.72596598e-01
4.29964870e-01 -6.99180245e-01 -4.88701940e-01 2.79572606e-01
4.41384435e-01 -8.69263336e-02 2.33706921e-01 1.42222166e+00
2.08639741e-01 2.81235367e-01 1.15983295e+00 -9.55780029e-01
-2.20497936e-01 8.50021303e-01 -7.90917039e-01 -3.55725884e-01
-2.28676260e-01 -6.60949424e-02 5.11848509e-01 -7.02723622e-01
2.64617950e-01 1.20658541e+00 7.15882540e-01 -1.26892090e-01
-1.57428801e+00 -6.22394085e-01 -5.89747354e-02 -2.04477068e-02
-1.62837005e+00 -2.61207253e-01 5.46535194e-01 -6.50577009e-01
4.40054834e-01 6.42021954e-01 8.61461580e-01 4.41698015e-01
1.92026347e-02 4.99313533e-01 1.74966156e+00 -6.98255301e-01
3.52776200e-01 1.24575324e-01 1.27809867e-01 1.58083737e-01
6.12377644e-01 3.22970897e-01 -1.99297264e-01 -2.71227449e-01
5.98317921e-01 8.51020515e-02 -5.87792039e-01 -4.45258170e-01
-1.23381746e+00 8.65759313e-01 5.54876328e-01 2.54504293e-01
-6.64607167e-01 -6.25302047e-02 -1.79982230e-01 8.57686251e-02
7.79811502e-01 2.59385914e-01 -4.92650531e-02 5.62505484e-01
-1.55364656e+00 2.18551338e-01 6.11892939e-01 4.21968609e-01
1.16736817e+00 1.60544649e-01 -1.39218200e-05 1.07599092e+00
7.98908055e-01 1.12237823e+00 3.18234593e-01 -9.17177975e-01
6.34075899e-04 4.54199910e-01 3.68026376e-01 -9.19300735e-01
-1.54522687e-01 -4.23129350e-01 -1.00507903e+00 6.20421469e-01
3.62336069e-01 -7.79486373e-02 -1.12311459e+00 1.47033668e+00
1.09336317e-01 3.42379987e-01 6.04946911e-02 8.60230446e-01
4.18201059e-01 1.08615839e+00 7.10253119e-02 -5.53262055e-01
9.97381210e-01 -7.03320026e-01 -8.52119923e-01 -2.86926657e-01
2.59130865e-01 -8.26348245e-01 5.34391761e-01 2.96499491e-01
-6.51428938e-01 -4.32075381e-01 -1.22800088e+00 4.80707675e-01
-5.37549555e-01 2.36017093e-01 4.17158514e-01 9.34287488e-01
-9.83519316e-01 7.66401768e-01 -8.31922710e-01 -4.92213309e-01
-3.12745497e-02 1.85887203e-01 -1.19290419e-01 4.16750647e-03
-9.45597589e-01 9.83877182e-01 7.48683095e-01 2.85368681e-01
-8.80552590e-01 -5.82796514e-01 -7.50683486e-01 -1.47579446e-01
-1.53374102e-03 -5.05487382e-01 8.75970364e-01 -1.43218291e+00
-1.52383447e+00 6.50307834e-01 -3.07481825e-01 -3.67901534e-01
4.18794841e-01 1.22798525e-01 -4.97202635e-01 1.17267124e-01
1.14596419e-01 6.81209803e-01 9.67843831e-01 -1.81670380e+00
-4.34695661e-01 -4.06046003e-01 -4.96311963e-01 3.11292887e-01
-1.94568932e-01 -1.94619089e-01 2.62626678e-01 -8.00427079e-01
5.70640981e-01 -9.17602420e-01 -1.53576955e-01 2.77600624e-02
-2.10501060e-01 3.74148577e-01 6.33057952e-01 -8.00602674e-01
1.17179954e+00 -2.06111312e+00 1.21938661e-01 6.98612511e-01
-2.30323792e-01 3.74100506e-01 5.85219264e-02 4.72244322e-01
-1.87441558e-01 2.28491742e-02 -9.72895563e-01 -6.34971038e-02
-1.67439103e-01 2.17927665e-01 -3.19913149e-01 8.32988322e-01
-1.07814044e-01 4.02868390e-01 -9.67261910e-01 -2.99474150e-01
7.69664943e-01 5.71404457e-01 3.96253355e-02 2.76131146e-02
-2.07914621e-01 4.32335764e-01 2.51814067e-01 4.44598109e-01
1.29558039e+00 1.71129368e-02 3.72594357e-01 -1.09054849e-01
-6.09279990e-01 -8.26861337e-02 -1.54558456e+00 1.22170293e+00
-2.18617320e-01 4.53121930e-01 4.29034263e-01 -8.15011621e-01
8.70321810e-01 2.32269257e-01 1.99568212e-01 1.62227191e-02
3.00802570e-02 5.53587019e-01 6.88360631e-02 -6.89015239e-02
4.25381809e-01 -4.93646234e-01 7.65559733e-01 4.42592055e-01
-1.15636364e-01 -4.56623524e-01 3.01272482e-01 -2.99478531e-01
9.18093771e-02 2.36579999e-01 5.97545326e-01 -1.09294224e+00
7.33278096e-01 1.87036961e-01 1.83580011e-01 7.55595028e-01
1.30956098e-01 5.54783046e-01 -2.10039675e-01 -2.97309011e-02
-9.09745276e-01 -1.13308787e+00 -5.49395680e-01 6.71111763e-01
2.33905002e-01 2.30238110e-01 -6.69743598e-01 1.75929353e-01
-3.98865938e-02 9.19512570e-01 -4.06777918e-01 3.49912077e-01
-3.46651562e-02 -1.39508617e+00 2.08673269e-01 1.48885056e-01
6.53745353e-01 -8.77595186e-01 -3.58862340e-01 1.12971939e-01
-2.74271041e-01 -3.55843782e-01 -1.35422945e-01 2.20893621e-01
-1.23536837e+00 -9.69512820e-01 -9.64677453e-01 -4.52082425e-01
7.06270337e-01 5.70433497e-01 8.77610326e-01 -2.95402914e-01
1.69163615e-01 -7.22700208e-02 -4.47516263e-01 -5.95003903e-01
-6.52619362e-01 -2.16886401e-01 -1.66198313e-02 2.69127071e-01
4.17945892e-01 -6.38903797e-01 -4.55544859e-01 5.11682034e-01
-1.33163440e+00 2.13319242e-01 5.46655953e-01 5.91164351e-01
6.47731364e-01 2.68154979e-01 2.90296495e-01 -9.03233051e-01
3.26311916e-01 -5.08284509e-01 -8.25876355e-01 3.13929945e-01
-8.85100067e-01 5.62016889e-02 1.65249690e-01 -3.34793776e-01
-1.35981309e+00 1.58954188e-01 8.21060985e-02 -1.07717745e-01
-3.85071605e-01 7.93954968e-01 -9.21742171e-02 -2.09131055e-02
8.73448670e-01 3.43926400e-01 2.34480321e-01 -4.29575145e-01
2.39661410e-01 9.11042869e-01 3.70745510e-01 -5.14884293e-01
7.49157071e-01 8.81735623e-01 1.55549854e-01 -1.40519226e+00
-5.07711589e-01 -8.31116021e-01 -6.51539326e-01 -2.55389720e-01
7.12476254e-01 -1.15781987e+00 -1.18051842e-01 8.24341714e-01
-7.29232788e-01 -3.75826091e-01 -1.25992894e-01 7.56251454e-01
-2.60632187e-01 7.25863516e-01 -2.72213817e-01 -1.40310144e+00
-2.09053174e-01 -1.06751227e+00 7.56354213e-01 1.10617854e-01
3.34432833e-02 -1.11997771e+00 1.53354406e-01 1.27613232e-01
4.44591463e-01 4.01072681e-01 7.49037385e-01 5.37159443e-02
-3.36426586e-01 1.06696062e-01 -3.55124921e-01 6.14878833e-01
3.19916397e-01 1.86708182e-01 -1.22960865e+00 -2.44681746e-01
8.34622234e-02 2.16596901e-01 1.29889023e+00 9.47215378e-01
7.51363099e-01 -2.04431891e-01 -2.44410753e-01 5.48036695e-01
1.74024582e+00 2.54300702e-02 9.15320456e-01 2.33193904e-01
4.28866476e-01 7.72553504e-01 4.71960783e-01 3.48540962e-01
-2.79566228e-01 2.99309611e-01 6.87174022e-01 -1.71401799e-01
-7.26513118e-02 -4.91318852e-02 3.76444727e-01 7.00359821e-01
-3.20326477e-01 -2.27010086e-01 -8.07838261e-01 6.83819771e-01
-2.10264182e+00 -1.22526431e+00 -7.92907596e-01 2.57805610e+00
8.67744207e-01 -5.88252962e-01 1.21118091e-02 2.72819638e-01
1.09978354e+00 2.90250272e-01 -1.67789698e-01 1.60186186e-01
-4.97272313e-01 9.89099070e-02 9.77205336e-01 8.27968597e-01
-1.30192780e+00 5.82408190e-01 6.67172766e+00 7.90009260e-01
-9.79688287e-01 7.70661533e-02 4.51190591e-01 3.24520200e-01
-4.26289976e-01 1.56535104e-01 -3.80758852e-01 4.48410690e-01
6.98639333e-01 2.04495296e-01 6.21603727e-01 4.85774875e-01
4.40249234e-01 -5.33800423e-01 -4.47921872e-01 8.99682641e-01
7.55076259e-02 -8.33698094e-01 1.40509859e-01 1.66956142e-01
1.40735602e+00 7.94325024e-02 -1.36997536e-01 -2.61439323e-01
3.34376127e-01 -9.68852580e-01 8.43844354e-01 6.92915022e-01
6.70753241e-01 -7.07730651e-01 1.08663297e+00 4.53396052e-01
-1.12093449e+00 2.47360626e-03 -7.31048882e-01 -3.77973795e-01
5.10512181e-02 1.07694864e+00 -5.69967985e-01 7.58486509e-01
4.76311654e-01 6.47681653e-01 -2.42983326e-01 1.22263467e+00
-4.44758296e-01 7.22921431e-01 -6.18443549e-01 8.04525688e-02
2.35245213e-01 -9.61527348e-01 4.70866799e-01 1.17586517e+00
8.37640762e-01 -1.65715083e-01 9.66137052e-02 1.10019934e+00
2.92281806e-01 3.75845999e-01 -4.59918708e-01 -2.58183777e-02
4.56543893e-01 1.04392254e+00 -8.46446335e-01 -4.29356366e-01
-1.60369039e-01 7.90806234e-01 -3.80338609e-01 5.39921820e-01
-4.15239841e-01 -1.67156868e-02 6.46745622e-01 2.44005948e-01
1.75978467e-01 -5.54369465e-02 -3.21201682e-01 -1.14100611e+00
-3.91790301e-01 -7.96585441e-01 2.51122177e-01 -9.43837881e-01
-1.37039757e+00 3.03043365e-01 3.63124341e-01 -1.15645993e+00
2.94710905e-03 -1.06207967e+00 -4.53308702e-01 1.29368687e+00
-1.58739519e+00 -1.14204252e+00 -3.99750412e-01 2.57780552e-01
1.50444165e-01 1.21204391e-01 9.92428005e-01 -7.40150437e-02
-3.43825996e-01 -3.90377551e-01 7.75194824e-01 -2.48963356e-01
6.98216796e-01 -1.38816500e+00 -3.78519684e-01 1.21821213e+00
3.94144319e-02 6.11732423e-01 9.23323154e-01 -8.75904739e-01
-6.45866752e-01 -1.34628403e+00 6.95153832e-01 -5.29275388e-02
5.69033444e-01 2.27003857e-01 -7.09036231e-01 3.34446579e-01
2.67599583e-01 -3.36507887e-01 9.09824908e-01 -3.20591591e-02
-1.12566456e-01 -2.08867013e-01 -1.14342332e+00 4.05501962e-01
3.86067599e-01 -3.42996687e-01 -4.44812030e-01 3.06062281e-01
1.26561135e-01 9.03906599e-02 -7.27068603e-01 5.30119598e-01
7.43815124e-01 -1.18135154e+00 9.21229601e-01 6.02161586e-02
2.38219410e-01 -9.89572048e-01 -4.03470457e-01 -1.49281669e+00
-4.46652859e-01 -1.22996554e-01 6.70824274e-02 9.75668788e-01
5.25415182e-01 -6.83834553e-01 1.96940556e-01 1.90554023e-01
1.56368185e-02 1.39714360e-01 -5.82717597e-01 -9.23513174e-01
4.04952466e-03 -3.28311294e-01 6.86202586e-01 1.00986707e+00
-2.55668163e-01 -4.50897142e-02 -5.88706672e-01 6.18669033e-01
7.62815356e-01 2.86074072e-01 3.58292222e-01 -1.76067209e+00
-1.28789485e-01 -5.25403380e-01 1.61072914e-03 -6.70757890e-01
1.66456580e-01 -7.98324645e-01 3.80090266e-01 -1.61353493e+00
3.69977295e-01 -2.81515032e-01 -2.91484654e-01 3.41279835e-01
-3.43618512e-01 5.70305884e-01 -5.75217344e-02 5.09669781e-01
2.08030611e-01 3.73134911e-01 8.85941863e-01 -4.75493878e-01
-3.88158470e-01 -1.69038326e-02 -4.03002053e-01 6.99655473e-01
1.04700184e+00 -4.37094331e-01 -3.86822313e-01 -1.58261821e-01
2.68025696e-01 -3.37626576e-01 4.56353515e-01 -1.09030652e+00
-2.16457590e-01 -3.58955055e-01 3.24946940e-01 -8.26108217e-01
3.00394237e-01 -1.09234524e+00 8.10433686e-01 5.15661240e-01
1.08092234e-01 -5.38184345e-01 2.32171372e-01 4.86730248e-01
-2.70696849e-01 -7.81197011e-01 1.01058388e+00 -2.99178243e-01
-5.98954201e-01 2.58121733e-02 -6.34206951e-01 -7.25643694e-01
7.01399088e-01 -1.95725113e-01 -1.57508358e-01 -3.61084998e-01
-4.59274292e-01 -1.51559278e-01 8.36961210e-01 -2.91537732e-01
2.09595531e-01 -1.20514548e+00 -1.01975036e+00 9.12812725e-02
2.19162330e-01 -2.18325198e-01 1.79206416e-01 8.52002382e-01
-1.06975496e+00 2.89265841e-01 -1.55179158e-01 -7.08708823e-01
-1.14307761e+00 6.19129002e-01 4.86404896e-01 -1.70919765e-02
3.53090130e-02 4.99339938e-01 8.62704366e-02 -6.21044695e-01
-2.28948131e-01 -2.64161259e-01 -2.72768736e-01 2.45436236e-01
6.88851953e-01 6.25649214e-01 -1.87356681e-01 -9.21971202e-01
-1.17357910e-01 4.17985618e-01 7.55736589e-01 -2.94742793e-01
1.23801613e+00 -2.18823954e-01 -9.69379961e-01 7.37329423e-01
6.98135734e-01 2.67350320e-02 -1.17443275e+00 1.07596731e-02
-2.44208932e-01 -5.62307417e-01 4.83984053e-01 -8.40807199e-01
-6.97657824e-01 8.79674733e-01 9.12692249e-01 3.26340228e-01
1.25702381e+00 -5.37646055e-01 -1.16290182e-01 2.12854728e-01
2.67077446e-01 -9.96641695e-01 -7.69701421e-01 2.68547118e-01
7.86732435e-01 -1.33065951e+00 3.45672905e-01 -6.83004498e-01
-2.18544111e-01 1.12293637e+00 -4.51994454e-03 1.62115037e-01
9.37049925e-01 -2.14177277e-02 3.41046721e-01 7.23888054e-02
2.07765535e-01 -5.89060962e-01 4.57253337e-01 7.65435398e-01
6.20275021e-01 6.14850283e-01 -4.79176432e-01 -3.58953439e-02
8.38096440e-02 -1.80438399e-01 2.49570787e-01 5.15918851e-01
-9.18559372e-01 -1.16192973e+00 -1.07397199e+00 5.85356772e-01
-1.24428876e-01 -4.08462912e-01 -2.64109582e-01 3.98968816e-01
4.01416540e-01 1.28290844e+00 -1.29880965e-01 8.23150277e-02
-2.30880827e-01 2.91570097e-01 3.03942561e-01 -5.34627795e-01
-1.40132353e-01 3.64091486e-01 -1.29103959e-01 8.75942875e-03
-1.13917637e+00 -7.73546696e-01 -6.63408637e-01 -3.57093483e-01
-7.09523380e-01 3.34590197e-01 7.38857090e-01 7.34103978e-01
-1.77621216e-01 2.53784776e-01 2.96117812e-01 -1.11733305e+00
-4.97020274e-01 -1.58374023e+00 -1.26458538e+00 2.56315321e-01
7.36935735e-02 -5.49774110e-01 -5.80293179e-01 1.43151715e-01] | [10.027214050292969, -2.0305604934692383] |
097f4cbe-b9d6-4e92-8a13-4952374968da | data-level-recombination-and-lightweight | 2009.05102 | null | https://arxiv.org/abs/2009.05102v1 | https://arxiv.org/pdf/2009.05102v1.pdf | Data-Level Recombination and Lightweight Fusion Scheme for RGB-D Salient Object Detection | Existing RGB-D salient object detection methods treat depth information as an independent component to complement its RGB part, and widely follow the bi-stream parallel network architecture. To selectively fuse the CNNs features extracted from both RGB and depth as a final result, the state-of-the-art (SOTA) bi-stream networks usually consist of two independent subbranches; i.e., one subbranch is used for RGB saliency and the other aims for depth saliency. However, its depth saliency is persistently inferior to the RGB saliency because the RGB component is intrinsically more informative than the depth component. The bi-stream architecture easily biases its subsequent fusion procedure to the RGB subbranch, leading to a performance bottleneck. In this paper, we propose a novel data-level recombination strategy to fuse RGB with D (depth) before deep feature extraction, where we cyclically convert the original 4-dimensional RGB-D into \textbf{D}GB, R\textbf{D}B and RG\textbf{D}. Then, a newly lightweight designed triple-stream network is applied over these novel formulated data to achieve an optimal channel-wise complementary fusion status between the RGB and D, achieving a new SOTA performance. | ['Xuehao Wang', 'Shuai Li', 'Chenglizhao Chen', 'Hong Qin', 'Aimin Hao', 'Yuming Fang'] | 2020-08-07 | null | null | null | null | ['rgb-d-salient-object-detection'] | ['computer-vision'] | [ 3.91427368e-01 1.40434101e-01 -6.85448349e-02 -3.07824999e-01
-7.01382935e-01 -9.46642682e-02 4.11856383e-01 -4.52273376e-02
-3.10895383e-01 4.22960669e-01 1.46989033e-01 1.26655838e-02
-3.37889232e-02 -9.18882608e-01 -5.74470758e-01 -1.08958042e+00
3.93271118e-01 -6.25577495e-02 8.58695328e-01 -3.66991878e-01
6.48275465e-02 5.71457982e-01 -1.97029698e+00 2.84309387e-01
8.22407663e-01 1.70496798e+00 3.55562568e-01 4.02299613e-01
-5.01245081e-01 6.26268625e-01 -4.95761812e-01 -1.72250107e-01
4.98632014e-01 -4.57200855e-01 -3.62716496e-01 -9.11519974e-02
2.65267879e-01 -4.88009840e-01 -4.96244282e-01 1.17644536e+00
6.87291384e-01 -1.30618870e-01 1.87815413e-01 -1.44626164e+00
-5.17103434e-01 3.61975342e-01 -9.45730805e-01 1.61689699e-01
1.21134156e-02 2.71636665e-01 8.67298126e-01 -7.85124183e-01
2.57001579e-01 1.25978768e+00 2.14114383e-01 4.66370404e-01
-7.30028570e-01 -7.23358035e-01 4.87937033e-01 2.07315192e-01
-1.09826553e+00 -1.80605009e-01 1.42924452e+00 1.30662307e-01
7.58274019e-01 2.69460887e-01 1.10832107e+00 7.91774094e-01
-5.14359735e-02 1.18484902e+00 1.06621778e+00 -1.61441281e-01
2.02776492e-01 -1.48379579e-01 2.02441830e-02 6.69484556e-01
2.69127309e-01 1.16444580e-01 -8.39067876e-01 3.86082232e-01
9.04279888e-01 2.14301273e-01 -3.12908143e-01 -4.98152882e-01
-1.17727578e+00 5.18546343e-01 1.07696891e+00 1.26040041e-01
-4.58248138e-01 2.32707977e-01 1.70877874e-01 -9.81408060e-02
3.15826118e-01 6.26736134e-02 -5.25940001e-01 8.29395875e-02
-9.30698574e-01 1.56982124e-01 9.82770845e-02 1.02566576e+00
1.14634609e+00 -1.00949220e-02 -2.43419588e-01 4.64052320e-01
5.45429647e-01 5.57658851e-01 4.91344661e-01 -7.22071946e-01
5.61863720e-01 1.16990852e+00 -5.20966277e-02 -7.91101336e-01
-6.61160290e-01 -6.14501834e-01 -8.80974293e-01 3.06133211e-01
2.20505416e-01 1.03782021e-01 -1.25535440e+00 1.55862319e+00
5.79345942e-01 -1.09031096e-01 1.61271244e-01 1.36851954e+00
1.32550645e+00 5.99426627e-01 -8.01847056e-02 -1.38280138e-01
1.36605299e+00 -9.78738308e-01 -3.70367587e-01 -3.76504511e-01
2.72059411e-01 -5.46548605e-01 1.11139584e+00 2.78472722e-01
-1.13402271e+00 -5.60075998e-01 -1.32037902e+00 -6.77292824e-01
-2.92259574e-01 1.75907478e-01 8.01663935e-01 3.30054998e-01
-9.41703796e-01 1.39126405e-01 -5.89048743e-01 3.96279320e-02
6.37008786e-01 3.39528471e-01 -1.15233719e-01 -6.49103895e-02
-1.19052100e+00 5.52089453e-01 4.06541407e-01 4.03046489e-01
-8.04042578e-01 -4.80101705e-01 -8.52124333e-01 -6.02532004e-04
4.38297451e-01 -5.75623810e-01 9.53001142e-01 -9.54027951e-01
-1.47636330e+00 8.14337492e-01 -1.55773610e-01 3.04935477e-03
3.02059174e-01 -3.01836208e-02 -2.49183446e-01 3.25744629e-01
8.78136680e-02 8.80861461e-01 1.09309268e+00 -1.41069257e+00
-1.20994532e+00 -7.82183230e-01 2.81904101e-01 4.67774540e-01
-4.52712119e-01 -3.43576789e-01 -7.01323926e-01 -6.25436962e-01
7.75891840e-01 -3.73823822e-01 9.85922217e-02 2.09373504e-01
-4.71229851e-01 -1.77870184e-01 9.18280005e-01 -2.89788157e-01
1.12344313e+00 -2.18324089e+00 4.73041773e-01 1.07622631e-01
6.19962335e-01 2.00187460e-01 6.53499588e-02 -3.19469541e-01
-2.62659416e-02 -2.38740623e-01 -2.74258077e-01 -4.08445567e-01
-3.00987720e-01 1.57001466e-01 -2.22162932e-01 4.97519672e-01
5.14833093e-01 1.00064981e+00 -1.12348235e+00 -6.59267783e-01
3.17284703e-01 4.22231585e-01 -3.45426917e-01 8.78077596e-02
-1.68744460e-01 2.06579000e-01 -6.84410572e-01 1.08387053e+00
9.89805937e-01 -1.75769702e-01 -3.07074904e-01 -5.94507754e-01
-4.43209410e-01 2.41179019e-01 -8.16771448e-01 2.00941777e+00
-1.41231611e-01 4.47207868e-01 1.90519944e-01 -7.73443699e-01
1.30268323e+00 -2.58403301e-01 6.79069161e-01 -1.19865465e+00
3.77128839e-01 3.62522125e-01 -7.37157464e-02 -1.74925745e-01
5.89877665e-01 -1.01072147e-01 -1.38184711e-01 1.72752187e-01
-4.91191261e-02 -3.92457217e-01 -3.36281419e-01 1.59767680e-02
8.35775197e-01 2.63473272e-01 3.22446367e-03 -5.76862320e-02
5.93082726e-01 3.12430207e-02 7.78385937e-01 3.57013166e-01
-4.44879979e-01 8.47954750e-01 7.19529986e-01 -2.48474911e-01
-8.13488305e-01 -1.13159060e+00 1.73387066e-01 8.64407659e-01
9.54821169e-01 -1.22390710e-01 -4.55668598e-01 -7.93042123e-01
-5.95567236e-03 4.37454790e-01 -7.17299879e-01 -5.59539676e-01
-6.56672895e-01 -5.76759875e-01 3.50586504e-01 5.73068380e-01
1.01647770e+00 -9.13571060e-01 -1.16855299e+00 1.28714129e-01
1.36358961e-01 -6.18196130e-01 -2.17305899e-01 6.55246973e-01
-8.10015857e-01 -9.56306219e-01 -9.65474665e-01 -7.13313341e-01
4.55459595e-01 7.38217056e-01 8.37940037e-01 2.19061859e-02
-5.45625538e-02 -9.29615349e-02 -4.55669940e-01 -5.35890758e-01
3.12087178e-01 1.44649476e-01 -3.88503909e-01 1.83762293e-02
3.77449006e-01 -4.55057085e-01 -1.12617183e+00 1.58177659e-01
-1.08816981e+00 5.03735125e-01 8.03325772e-01 6.57387257e-01
7.83552885e-01 -1.38692841e-01 4.75354403e-01 -9.18992534e-02
-8.85269418e-03 -3.35840791e-01 -5.74194372e-01 1.95900962e-01
-2.65773803e-01 -2.23231725e-02 4.67558682e-01 -1.27999216e-01
-1.02717805e+00 1.90814778e-01 -1.10932469e-01 -7.88218379e-01
1.43510163e-01 9.89486426e-02 -5.82415104e-01 -1.37533313e-02
2.33560115e-01 4.52386588e-01 3.14274840e-02 -3.60008895e-01
3.95660877e-01 6.98720276e-01 6.57822371e-01 -2.49058217e-01
6.11283958e-01 6.94619656e-01 1.05286136e-01 -4.96656656e-01
-8.14165175e-01 -3.98639500e-01 -4.34222162e-01 -4.14291769e-01
7.71455646e-01 -9.65208292e-01 -6.59526289e-01 9.15638208e-01
-1.15547049e+00 -2.74603546e-01 -5.55064499e-01 2.47768119e-01
-3.11160922e-01 7.38480538e-02 -3.37566823e-01 -6.92688465e-01
-4.09204453e-01 -1.35744596e+00 1.59530425e+00 7.25151598e-01
3.85313600e-01 -3.10251981e-01 -4.89209563e-01 5.19732982e-02
3.10708851e-01 2.01891616e-01 9.10310090e-01 -8.42200518e-02
-7.45055258e-01 -1.11510446e-02 -9.62396204e-01 1.63747877e-01
2.64026552e-01 -8.65218788e-02 -1.11920893e+00 1.75036900e-02
1.15444437e-01 -1.74410686e-01 1.07572639e+00 1.71984658e-01
1.12536466e+00 4.93157767e-02 -1.71385199e-01 8.62036943e-01
1.42728209e+00 1.10055134e-01 5.40086150e-01 5.61268985e-01
1.05639577e+00 4.70996618e-01 7.58071721e-01 4.83628213e-01
6.75138414e-01 5.42205334e-01 9.83785808e-01 -4.50974047e-01
-4.41144794e-01 -2.07983449e-01 1.73114717e-01 5.38990676e-01
1.44381434e-01 -1.87679097e-01 -7.96847284e-01 4.80883241e-01
-1.70818245e+00 -4.39044476e-01 -1.70395896e-01 2.09177065e+00
8.52118134e-01 3.26026380e-01 1.54941708e-01 3.30796272e-01
6.52787328e-01 3.87085974e-01 -9.14850533e-01 -1.36495335e-02
-5.74225783e-01 1.62833497e-01 6.97469831e-01 -7.34603032e-02
-9.28523123e-01 7.86555171e-01 4.91827059e+00 9.63416100e-01
-1.51951849e+00 1.64785206e-01 5.88836432e-01 -6.14396691e-01
-6.93411648e-01 -3.97954881e-02 -7.46165335e-01 5.16870022e-01
1.35711670e-01 1.80079907e-01 1.22331470e-01 7.54136324e-01
-7.07295164e-02 -4.88193393e-01 -7.84922659e-01 1.21802390e+00
-1.84720289e-02 -1.14711022e+00 1.45173118e-01 -5.06633408e-02
6.41858459e-01 -5.49793020e-02 1.80348679e-01 1.41416177e-01
-4.84546237e-02 -6.03100479e-01 1.33686244e+00 3.28821689e-01
9.22924578e-01 -7.90655732e-01 6.15227818e-01 1.88419417e-01
-1.37874746e+00 -3.86937529e-01 -3.32058012e-01 1.40558884e-01
1.24509051e-01 9.08581913e-01 -1.00536302e-01 7.65500546e-01
9.97242093e-01 9.17106092e-01 -6.15515709e-01 8.27548683e-01
-2.99086332e-01 -1.62594672e-02 -4.83271807e-01 -1.04759991e-01
5.16968191e-01 2.27438267e-02 6.27576292e-01 7.67582476e-01
2.93892503e-01 7.28165433e-02 -7.91090205e-02 8.89085531e-01
-8.03083554e-02 -2.73846507e-01 -2.29528978e-01 4.22740102e-01
3.86469960e-01 1.31029224e+00 -1.03716493e+00 -4.15499419e-01
-5.01338720e-01 9.92816985e-01 2.88217515e-01 2.48640433e-01
-9.24402893e-01 -6.73193395e-01 7.18658566e-01 3.47613320e-02
5.60710013e-01 1.65637638e-02 -6.45447433e-01 -9.44147646e-01
2.42448196e-01 -4.38647062e-01 2.22536713e-01 -9.95955586e-01
-8.87124479e-01 6.12442434e-01 -1.06155045e-01 -1.52803195e+00
3.90752286e-01 -5.91674387e-01 -3.69265169e-01 1.08056176e+00
-2.02002883e+00 -1.36358953e+00 -8.09024155e-01 8.04030716e-01
2.64654547e-01 9.07023251e-02 1.26783058e-01 2.45474771e-01
-6.72434032e-01 5.73140144e-01 -1.71954557e-01 -1.79057166e-01
4.41284627e-01 -1.11943972e+00 1.29965752e-01 9.70173657e-01
-5.09461403e-01 2.42231786e-01 2.22150952e-01 -6.88640893e-01
-1.60672593e+00 -1.24163115e+00 2.39551574e-01 7.91573972e-02
3.27285379e-01 -3.46990019e-01 -8.28781009e-01 1.15333989e-01
-3.32678497e-01 1.49641007e-01 -1.55833205e-02 -4.69329238e-01
-3.68535250e-01 -6.54476821e-01 -1.21362245e+00 5.92082202e-01
1.28503299e+00 -4.75150406e-01 -5.15648007e-01 -2.01640204e-01
1.14242220e+00 -5.63865781e-01 -6.44831002e-01 6.36528015e-01
5.78236401e-01 -1.38877261e+00 1.03429377e+00 8.10528994e-02
7.16893256e-01 -6.74091935e-01 -1.74503043e-01 -8.89971733e-01
-3.23439538e-02 -4.17570978e-01 -3.84172827e-01 1.09494793e+00
5.02497330e-03 -5.25168300e-01 9.18140948e-01 2.35436961e-01
-5.60472906e-01 -1.03008533e+00 -1.22879732e+00 -3.34875971e-01
-3.00144374e-01 -4.56842035e-01 1.00373256e+00 5.03444910e-01
-4.01063412e-01 1.23180702e-01 -2.29334384e-02 2.41151154e-02
5.82561374e-01 4.58844632e-01 5.31302691e-01 -1.13455105e+00
2.09185015e-02 -9.24364388e-01 -4.27721709e-01 -1.45732558e+00
-3.48614782e-01 -7.65564442e-01 4.19225931e-01 -1.52693486e+00
-6.41212165e-02 -8.04401934e-01 -5.12409151e-01 5.49422920e-01
-3.94065350e-01 3.58080536e-01 1.94483206e-01 1.81457683e-01
-5.13876975e-01 8.66919756e-01 1.64094448e+00 -2.17673704e-01
-4.08530772e-01 -2.99517483e-01 -9.43435073e-01 5.36862671e-01
6.34125352e-01 -1.91662177e-01 -4.84087974e-01 -5.15951991e-01
3.05661649e-01 4.43467051e-02 6.19267285e-01 -1.01047587e+00
4.11913306e-01 -3.50710265e-02 6.83293641e-01 -1.08783031e+00
5.79935968e-01 -6.75888538e-01 -2.92934656e-01 2.37728730e-01
4.79024239e-02 -2.32921213e-01 2.58779049e-01 4.75854516e-01
-3.25494945e-01 2.03770861e-01 8.10449243e-01 -2.28438172e-02
-8.99530947e-01 4.98188078e-01 9.07109156e-02 -2.44806558e-01
1.10676038e+00 -6.60634577e-01 -5.57070792e-01 4.85248864e-02
-2.04755977e-01 3.03870857e-01 6.41600728e-01 4.45632130e-01
8.53919923e-01 -1.38537335e+00 -4.29971904e-01 4.80812669e-01
1.81316435e-01 8.25585365e-01 4.75643635e-01 9.65749979e-01
-3.82969081e-01 2.54383087e-01 -4.40571934e-01 -9.13174450e-01
-7.99193978e-01 6.26033604e-01 1.98341027e-01 -5.63583849e-03
-5.94456732e-01 1.25525844e+00 4.42188799e-01 -1.38326129e-02
2.66421497e-01 -5.27747393e-01 -7.85271600e-02 2.60870188e-01
3.51651698e-01 3.02514821e-01 1.20399281e-01 -7.20165014e-01
-5.39597750e-01 7.22660422e-01 9.93364155e-02 6.71905652e-02
1.20433831e+00 -3.22437584e-01 -2.72601575e-01 4.43237305e-01
1.28021157e+00 -3.45700413e-01 -1.61184502e+00 -3.43234032e-01
-4.14312273e-01 -5.81471682e-01 4.31879669e-01 -5.02591789e-01
-1.60103619e+00 9.59099472e-01 6.61990702e-01 8.92637148e-02
1.77685130e+00 2.41293699e-01 1.04458988e+00 -1.25264794e-01
3.90560597e-01 -9.23426449e-01 2.25829870e-01 3.11343670e-01
8.01282227e-01 -1.10810292e+00 1.47702754e-01 -4.49947774e-01
-4.86083686e-01 1.06455946e+00 7.99068272e-01 3.55050340e-03
4.83000487e-01 8.51235352e-03 -2.05036551e-01 -3.90370280e-01
-2.22484335e-01 -6.11166835e-01 1.98592901e-01 5.43089092e-01
-4.47741672e-02 -1.02619186e-01 7.70155638e-02 7.51502752e-01
5.99586312e-03 -9.26293358e-02 1.15326904e-01 1.07115829e+00
-4.18136716e-01 -6.63638175e-01 -3.93938661e-01 5.50563812e-01
-3.25896777e-02 -1.94622919e-01 -2.50613660e-01 7.26137400e-01
5.55719852e-01 7.32805669e-01 2.15829790e-01 -6.23934090e-01
4.62970197e-01 -3.66864324e-01 4.46311951e-01 -4.11152661e-01
-5.94606102e-01 2.04364657e-01 -3.57641727e-01 -8.15218985e-01
-6.04199946e-01 -5.61418176e-01 -1.51438069e+00 -1.46570623e-01
-3.33138674e-01 -3.37512523e-01 6.62978530e-01 8.32206368e-01
3.16419244e-01 7.35692322e-01 7.96496332e-01 -1.23246765e+00
-1.79044474e-02 -6.96039081e-01 -5.77077627e-01 6.36131316e-02
7.10959077e-01 -9.47784185e-01 -2.75864184e-01 -4.90603924e-01] | [9.635773658752441, -0.8099986910820007] |
89219f7b-bbc1-4ff2-bda1-2467837abe7c | pixel-sampling-for-style-preserving-face-pose | 2106.07310 | null | https://arxiv.org/abs/2106.07310v1 | https://arxiv.org/pdf/2106.07310v1.pdf | Pixel Sampling for Style Preserving Face Pose Editing | The existing auto-encoder based face pose editing methods primarily focus on modeling the identity preserving ability during pose synthesis, but are less able to preserve the image style properly, which refers to the color, brightness, saturation, etc. In this paper, we take advantage of the well-known frontal/profile optical illusion and present a novel two-stage approach to solve the aforementioned dilemma, where the task of face pose manipulation is cast into face inpainting. By selectively sampling pixels from the input face and slightly adjust their relative locations with the proposed ``Pixel Attention Sampling" module, the face editing result faithfully keeps the identity information as well as the image style unchanged. By leveraging high-dimensional embedding at the inpainting stage, finer details are generated. Further, with the 3D facial landmarks as guidance, our method is able to manipulate face pose in three degrees of freedom, i.e., yaw, pitch, and roll, resulting in more flexible face pose editing than merely controlling the yaw angle as usually achieved by the current state-of-the-art. Both the qualitative and quantitative evaluations validate the superiority of the proposed approach. | ['Liming Chen', 'Yunhong Wang', 'Zehua Fu', 'Hongyu Yang', 'Di Huang', 'Xiangnan Yin'] | 2021-06-14 | null | null | null | null | ['facial-inpainting'] | ['computer-vision'] | [ 2.76942909e-01 2.76944488e-01 -2.24819966e-03 -3.83528471e-01
-1.50489807e-01 -6.11248314e-01 5.11256158e-01 -4.12781894e-01
-2.52519101e-01 6.12321377e-01 2.36300886e-01 3.31176460e-01
1.01913914e-01 -7.51762211e-01 -6.78484857e-01 -8.39513302e-01
6.46673501e-01 -7.82020390e-02 -2.43693098e-01 -1.93466216e-01
3.60090315e-01 8.61992836e-01 -1.60352576e+00 -1.20849900e-01
8.33932579e-01 1.01823854e+00 -1.60206124e-01 2.17910498e-01
-8.78127739e-02 3.27716202e-01 -4.73820180e-01 -6.28772080e-01
3.81666780e-01 -4.98338133e-01 -1.97981939e-01 5.00095606e-01
6.44194841e-01 -6.36862934e-01 -3.88478786e-01 1.20272279e+00
4.92529839e-01 -5.77549413e-02 3.82525206e-01 -1.03945327e+00
-7.93194354e-01 -3.93415149e-03 -9.21582997e-01 -3.51710886e-01
5.10466099e-01 1.16368383e-01 7.28854835e-01 -9.04175580e-01
7.56063163e-01 1.33061385e+00 4.56686497e-01 7.83657849e-01
-1.42411852e+00 -7.63131022e-01 2.08062634e-01 5.47547229e-02
-1.53013527e+00 -6.92294955e-01 1.33027363e+00 -3.80380869e-01
1.20809488e-01 4.22573060e-01 6.84547603e-01 8.53915334e-01
1.99757174e-01 3.85019273e-01 1.19405258e+00 -4.49492514e-01
1.72010157e-02 2.56180584e-01 -6.49749339e-01 8.80077422e-01
4.13433462e-02 1.42449111e-01 -4.26585436e-01 7.61195123e-02
1.33188617e+00 2.48392120e-01 -6.30051374e-01 -6.29800916e-01
-1.09318340e+00 5.09378016e-01 3.66329819e-01 6.73614629e-03
-3.91888291e-01 -4.55908813e-02 1.29691347e-01 1.44013271e-01
5.77371478e-01 4.64462399e-01 -1.84861496e-01 1.95764408e-01
-9.89058256e-01 1.91106558e-01 3.71460676e-01 9.69157577e-01
9.29824293e-01 1.52435437e-01 -3.26497585e-01 6.31418943e-01
4.58862513e-01 4.17864293e-01 2.11667016e-01 -1.08726943e+00
2.37556159e-01 5.82411528e-01 1.89720973e-01 -1.24733889e+00
-8.32046941e-02 -2.47192383e-01 -8.06011021e-01 4.66508538e-01
2.37428978e-01 -1.49526000e-01 -9.12250996e-01 2.01770687e+00
5.99564373e-01 -4.07238677e-02 -2.28248149e-01 9.63071704e-01
5.20457089e-01 3.84081632e-01 -1.41729057e-01 -4.49670911e-01
1.48117852e+00 -7.82621861e-01 -1.23450005e+00 -1.55210540e-01
-7.35272020e-02 -1.00292039e+00 1.11668611e+00 5.80182038e-02
-1.24027407e+00 -5.07614255e-01 -1.06334317e+00 -1.74746618e-01
9.27665271e-03 4.23728049e-01 5.08998096e-01 5.67795575e-01
-9.48598504e-01 5.08853316e-01 -7.37201214e-01 -9.08742398e-02
3.41433406e-01 2.33727515e-01 -6.49798274e-01 1.09262653e-01
-9.39952731e-01 7.03754485e-01 -1.18982628e-01 3.18470448e-01
-5.15075207e-01 -8.09875190e-01 -8.32880259e-01 7.49098882e-02
4.02267158e-01 -6.12793922e-01 8.87359083e-01 -1.07437062e+00
-2.07796097e+00 9.42663074e-01 -2.85290718e-01 3.39325607e-01
8.17015946e-01 -2.89331198e-01 -2.15125099e-01 1.81208909e-01
-4.43209708e-02 7.79789150e-01 1.20618904e+00 -1.42592096e+00
-1.29215032e-01 -6.21699512e-01 1.23672105e-01 2.99432099e-01
-6.04259789e-01 -1.28499657e-01 -7.13791788e-01 -1.00232732e+00
2.43901476e-01 -9.23691988e-01 6.81985915e-02 8.59855175e-01
-3.39085966e-01 3.17246020e-01 9.99774039e-01 -7.90239751e-01
1.18553197e+00 -2.40486193e+00 3.84862065e-01 1.76899359e-01
8.62032250e-02 2.32084155e-01 -5.17097861e-03 3.24837059e-01
-1.94573864e-01 -8.12451616e-02 -1.52393833e-01 -5.74282348e-01
-2.09417492e-01 1.17958069e-01 -1.72899634e-01 5.96047640e-01
3.76086622e-01 6.84750080e-01 -6.90281808e-01 -4.75869417e-01
3.34755599e-01 9.16803777e-01 -8.29163790e-01 2.27717891e-01
-1.25570953e-01 7.65891194e-01 -5.15650272e-01 6.76443636e-01
1.00318646e+00 2.35085458e-01 1.94852874e-01 -5.79158902e-01
-2.24409744e-01 -1.03829607e-01 -1.14733052e+00 1.80611300e+00
-4.78019029e-01 4.07012910e-01 3.98367077e-01 -3.05502862e-01
1.00064051e+00 3.32373798e-01 4.39608186e-01 -5.78426123e-01
2.38791570e-01 3.67813231e-03 -2.54858404e-01 -2.80225843e-01
2.96808869e-01 -1.42978773e-01 3.57024819e-01 2.78252929e-01
-2.93365747e-01 -2.60533780e-01 -2.64288515e-01 -2.41582885e-01
3.72093469e-01 4.69666213e-01 2.97238827e-02 -1.26915872e-01
7.81390309e-01 -5.52345097e-01 6.35533810e-01 -1.12003170e-01
-7.85435587e-02 8.62986684e-01 6.11543894e-01 -2.54903913e-01
-1.05080855e+00 -8.33676338e-01 -7.24754035e-02 6.23116910e-01
1.17391989e-01 -2.47146830e-01 -1.17310274e+00 -4.53085393e-01
-7.88031742e-02 4.38232481e-01 -8.60932708e-01 -3.15704405e-01
-6.69031858e-01 -2.38293543e-01 2.91772693e-01 3.56175750e-01
7.30874717e-01 -9.77439046e-01 -5.54925144e-01 7.76175596e-03
7.92010315e-03 -8.74051571e-01 -1.07726109e+00 -5.32933652e-01
-7.46354997e-01 -7.99357712e-01 -8.77726018e-01 -7.85360932e-01
1.19439220e+00 1.05304226e-01 4.46182042e-01 8.33830610e-02
-2.61573881e-01 -9.72304568e-02 -1.68985248e-01 -4.95374948e-02
-1.28349975e-01 -2.47018576e-01 1.04367752e-02 5.80708623e-01
-2.70398200e-01 -8.00926924e-01 -9.77003038e-01 3.62583041e-01
-1.01833773e+00 3.30869377e-01 5.89520574e-01 7.67020524e-01
5.94738066e-01 -2.34975088e-02 3.38472247e-01 -8.64640117e-01
4.17223096e-01 1.74277514e-01 -4.50239301e-01 1.21589921e-01
-6.50217474e-01 5.86899780e-02 6.57198191e-01 -5.59197068e-01
-1.30155122e+00 2.25533292e-01 -2.62072325e-01 -7.41552651e-01
2.32097823e-02 -5.35111949e-02 -6.06616318e-01 -3.66166264e-01
1.93031728e-01 3.49283367e-01 5.02259433e-01 -4.70234036e-01
4.50595021e-01 4.76089478e-01 4.38483119e-01 -4.65534240e-01
9.69626427e-01 7.30133176e-01 7.25690797e-02 -5.36716819e-01
-5.35584569e-01 2.04949215e-01 -6.81995511e-01 -2.79743582e-01
7.17045844e-01 -7.14121282e-01 -8.42516065e-01 5.90182424e-01
-1.17246878e+00 2.37447783e-01 -2.37493798e-01 1.95157528e-01
-5.31963229e-01 4.11633968e-01 -5.17855942e-01 -5.86167693e-01
-2.76891947e-01 -1.26013172e+00 1.29957259e+00 2.71515667e-01
-2.02417988e-02 -6.87989056e-01 -2.37482876e-01 1.86631352e-01
5.04388392e-01 5.18060029e-01 1.02213919e+00 2.57080555e-01
-5.16417325e-01 -2.63375759e-01 -1.91674024e-01 3.32591593e-01
6.11322582e-01 1.18396319e-01 -8.53570640e-01 -3.80115449e-01
6.15051948e-02 1.07021734e-01 4.41325188e-01 4.42675874e-02
1.13644493e+00 -5.80393314e-01 -4.74056676e-02 8.42538536e-01
1.15856409e+00 1.55803680e-01 7.99302220e-01 1.17512897e-01
7.93253064e-01 8.14935327e-01 6.35859191e-01 5.36307037e-01
1.62727058e-01 8.64589214e-01 4.87891108e-01 -2.46269256e-01
-2.23914221e-01 -5.50553024e-01 2.46915638e-01 4.04341906e-01
-2.57436991e-01 1.56009376e-01 -2.00950786e-01 9.87908468e-02
-1.53603208e+00 -8.00365984e-01 5.16339600e-01 2.31254530e+00
1.18757463e+00 -2.44057208e-01 -2.77461082e-01 1.35925561e-01
7.96561897e-01 3.61361712e-01 -4.74781752e-01 -2.51342058e-01
1.45399854e-01 1.29834384e-01 2.64290661e-01 5.45090556e-01
-8.18691611e-01 9.47683930e-01 5.79109573e+00 7.33074486e-01
-1.61737061e+00 -2.23686710e-01 5.22955477e-01 -1.99933991e-01
-5.64753830e-01 9.66606569e-03 -5.34446836e-01 4.93136942e-01
9.33497474e-02 9.28736851e-02 6.31180346e-01 6.03216708e-01
3.86312038e-01 2.41552025e-01 -9.18103874e-01 9.45482969e-01
2.82587022e-01 -1.22953916e+00 2.73104966e-01 1.13274902e-01
6.06354594e-01 -1.09180331e+00 3.34362596e-01 -1.93687767e-01
-6.58424973e-01 -8.56773794e-01 9.89112675e-01 7.81859279e-01
1.25285327e+00 -9.62322891e-01 2.62245864e-01 7.52189681e-02
-9.68566477e-01 1.22247897e-01 -4.20991555e-02 1.71309084e-01
2.40362570e-01 4.39163685e-01 -3.92560810e-01 3.88400555e-01
4.26973313e-01 4.93836910e-01 -4.21694249e-01 4.92207885e-01
-4.56867456e-01 6.42499253e-02 -1.42691717e-01 3.93236965e-01
-1.86672106e-01 -4.73208934e-01 5.88080108e-01 6.78678393e-01
2.97541320e-01 1.69063434e-01 -1.41762465e-01 1.12426245e+00
-1.64298654e-01 5.57834767e-02 -3.99335891e-01 -5.41484766e-02
5.49637556e-01 1.33901644e+00 -3.83898020e-01 7.25407377e-02
-1.75621435e-01 1.25312352e+00 5.58257438e-02 3.61548007e-01
-8.50704253e-01 -4.37230051e-01 8.89759839e-01 5.28587103e-01
3.09889168e-01 -1.56876162e-01 -2.88815171e-01 -1.05228460e+00
3.03903729e-01 -9.08782840e-01 -3.36493462e-01 -6.27762079e-01
-8.10433149e-01 6.79988623e-01 -1.75969586e-01 -1.20386755e+00
-2.51674186e-02 -4.20065343e-01 -6.94854975e-01 9.37170029e-01
-1.54075348e+00 -1.40442026e+00 -4.22352970e-01 5.83962083e-01
3.66441429e-01 9.73169506e-02 6.89044714e-01 4.69917357e-01
-7.13267207e-01 1.03764236e+00 -1.39646366e-01 5.47615066e-02
9.70856249e-01 -7.14008927e-01 1.79216176e-01 6.82004631e-01
-3.42585236e-01 8.85398805e-01 7.62114048e-01 -4.03849214e-01
-1.60825944e+00 -1.03563404e+00 6.01422966e-01 -3.70194688e-02
1.74164534e-01 -5.33705771e-01 -7.32478201e-01 5.45675278e-01
5.76990172e-02 1.00236796e-01 3.31547678e-01 -4.53600049e-01
-3.44341844e-01 -3.79682750e-01 -1.53841472e+00 8.58930528e-01
9.63007987e-01 -5.37601411e-01 -2.32475430e-01 -2.22489219e-02
6.23681366e-01 -5.05483866e-01 -7.53910363e-01 4.61395562e-01
8.22059453e-01 -9.83856618e-01 9.46085453e-01 -1.78952798e-01
4.09933388e-01 -5.60895503e-01 6.37079477e-02 -1.20955551e+00
-3.61081183e-01 -8.26224685e-01 3.11754327e-02 1.60501659e+00
1.35461509e-01 -7.11147964e-01 8.45103264e-01 4.92432207e-01
6.70410022e-02 -8.68572891e-01 -8.87876213e-01 -2.15045035e-01
-2.37093091e-01 1.90850377e-01 8.45012426e-01 8.60395014e-01
-3.27802002e-01 7.96267122e-04 -5.14549732e-01 2.14013413e-01
4.78188097e-01 1.81119710e-01 6.97823286e-01 -9.55342591e-01
-1.71366706e-01 -2.46788710e-01 -3.15863132e-01 -1.02516842e+00
8.80968124e-02 -3.27925980e-01 -3.66500877e-02 -9.28003967e-01
4.97981273e-02 -3.05635959e-01 2.73955218e-03 4.78309035e-01
-1.14621513e-01 4.82422441e-01 2.72796154e-01 1.69919103e-01
2.69743174e-01 8.19363058e-01 1.91640556e+00 1.00761481e-01
-1.67682022e-01 -1.68054342e-01 -8.87884676e-01 6.74928367e-01
5.63625515e-01 -1.24913909e-01 -5.14910758e-01 -5.47010958e-01
-7.14518055e-02 1.58344164e-01 2.56835133e-01 -6.96024239e-01
1.46023966e-02 -3.45030427e-01 6.40790999e-01 -1.53356433e-01
6.51066959e-01 -9.45973039e-01 3.06862265e-01 3.37142348e-01
-2.10061789e-01 1.55671254e-01 1.54580265e-01 4.30979669e-01
-1.87754735e-01 1.02999195e-01 1.16739917e+00 1.52798742e-02
-2.90188819e-01 4.79360372e-01 1.02187835e-01 -3.60819727e-01
1.24289942e+00 -4.04676557e-01 -1.14274304e-02 -4.07393247e-01
-5.64579070e-01 -1.12213843e-01 9.61696804e-01 4.94632959e-01
6.36749685e-01 -1.45619583e+00 -5.23811221e-01 9.13192332e-01
-4.35703844e-02 -1.13060521e-02 3.30716580e-01 9.25393641e-01
-6.49482489e-01 9.56322998e-02 -4.37680423e-01 -4.19969708e-01
-1.23611856e+00 5.07663965e-01 2.95113206e-01 1.55048057e-01
-5.45773208e-01 6.18073821e-01 4.07519251e-01 -2.02379048e-01
1.72120005e-01 6.96608722e-02 -1.96909621e-01 6.17925934e-02
5.14035881e-01 1.36206597e-01 -1.05283007e-01 -6.95003271e-01
-3.30404103e-01 1.08000517e+00 -1.60906047e-01 -1.54833153e-01
1.12669933e+00 -2.81382948e-01 -3.03602189e-01 -6.24539144e-02
1.08610976e+00 4.62918460e-01 -1.74070978e+00 9.56839323e-02
-6.47669911e-01 -1.01148927e+00 2.09133253e-02 -5.54370999e-01
-1.40102208e+00 8.49351227e-01 5.88400185e-01 -2.25794747e-01
1.31164646e+00 -3.56691837e-01 6.69055700e-01 -3.01745415e-01
3.41390759e-01 -9.51737642e-01 8.02989081e-02 1.13664210e-01
1.22644675e+00 -7.78800309e-01 -8.57486576e-02 -7.28459001e-01
-6.07562840e-01 1.02599037e+00 7.48117685e-01 -1.65381640e-01
4.51474637e-01 2.27338001e-01 7.25411698e-02 6.64650798e-02
-2.80928850e-01 2.94126540e-01 2.68049568e-01 4.11078572e-01
5.60799956e-01 -1.56088158e-01 -3.23805183e-01 2.54418224e-01
-1.79557562e-01 -1.72503680e-01 3.00312996e-01 8.14554691e-01
-2.28118017e-01 -1.08429241e+00 -4.78256702e-01 -1.40835261e-02
-3.70190084e-01 3.19378078e-02 -3.81565064e-01 7.43807554e-01
3.29132676e-01 5.78592300e-01 6.91837892e-02 -2.57231772e-01
6.02174401e-01 -1.35454640e-01 6.94261074e-01 -6.14975631e-01
-2.22490087e-01 2.51971960e-01 -3.12539130e-01 -6.80948317e-01
-2.18070164e-01 -5.36036968e-01 -1.13299584e+00 -5.04721880e-01
-3.44801694e-01 -1.44899130e-01 6.45182133e-01 6.69278502e-01
5.14699459e-01 4.75891739e-01 8.98217201e-01 -1.12535334e+00
-5.30484438e-01 -8.12816024e-01 -6.09945059e-01 4.89085466e-01
5.44174969e-01 -8.14994276e-01 -4.07552272e-01 7.71840522e-03] | [12.722853660583496, -0.16818831861019135] |
d3b5758e-f720-4ada-b6fd-04d5f0193709 | long-short-temporal-co-teaching-for-weakly | 2303.18044 | null | https://arxiv.org/abs/2303.18044v2 | https://arxiv.org/pdf/2303.18044v2.pdf | Long-Short Temporal Co-Teaching for Weakly Supervised Video Anomaly Detection | Weakly supervised video anomaly detection (WS-VAD) is a challenging problem that aims to learn VAD models only with video-level annotations. In this work, we propose a Long-Short Temporal Co-teaching (LSTC) method to address the WS-VAD problem. It constructs two tubelet-based spatio-temporal transformer networks to learn from short- and long-term video clips respectively. Each network is trained with respect to a multiple instance learning (MIL)-based ranking loss, together with a cross-entropy loss when clip-level pseudo labels are available. A co-teaching strategy is adopted to train the two networks. That is, clip-level pseudo labels generated from each network are used to supervise the other one at the next training round, and the two networks are learned alternatively and iteratively. Our proposed method is able to better deal with the anomalies with varying durations as well as subtle anomalies. Extensive experiments on three public datasets demonstrate that our method outperforms state-of-the-art WS-VAD methods. | ['Xiaojin Gong', 'Shengyang Sun'] | 2023-03-31 | null | null | null | null | ['video-anomaly-detection', 'multiple-instance-learning'] | ['computer-vision', 'methodology'] | [ 1.94367409e-01 -1.75909296e-01 -1.41018257e-01 -4.19155985e-01
-8.98316264e-01 -3.11691761e-01 5.67275584e-01 1.40854731e-01
-3.90781373e-01 2.67890453e-01 -4.06590961e-02 -3.41656446e-01
7.41145983e-02 -4.08889711e-01 -1.00600088e+00 -6.53456807e-01
-5.20421088e-01 3.70038509e-01 4.50919449e-01 1.04282811e-01
-1.35936648e-01 2.08147511e-01 -1.46652138e+00 4.67610300e-01
8.69622111e-01 1.52064252e+00 -1.80771321e-01 5.50953925e-01
-2.97638755e-02 1.49726355e+00 -2.55886078e-01 -3.81550193e-01
4.21495914e-01 -4.92719769e-01 -5.55055082e-01 8.07710066e-02
7.28652835e-01 -5.41532993e-01 -7.25802004e-01 8.89086902e-01
2.35574469e-01 2.60027975e-01 7.08711088e-01 -1.63816512e+00
-2.81510293e-01 2.34428033e-01 -6.06575489e-01 7.43496001e-01
1.62769109e-01 3.31964910e-01 1.10001910e+00 -9.43665147e-01
3.77683401e-01 1.08594513e+00 6.78523183e-01 6.60071194e-01
-9.21203315e-01 -6.41859949e-01 6.49175406e-01 6.57024801e-01
-1.07446373e+00 -2.47741014e-01 1.12099719e+00 -5.72429001e-01
7.50052631e-01 -9.58527252e-02 5.98310113e-01 1.42436695e+00
-1.23351634e-01 1.12218440e+00 5.73130727e-01 -1.41774446e-01
1.52345777e-01 -2.11250946e-01 -7.83512294e-02 8.67803991e-01
-2.56081790e-01 -1.83102161e-01 -5.70330262e-01 -2.22601518e-01
4.13748711e-01 1.24359362e-01 5.09115448e-03 -5.61723769e-01
-1.00109673e+00 5.93433440e-01 1.06255762e-01 2.30890810e-01
-5.60972095e-01 4.80262376e-02 1.01460826e+00 6.69119716e-01
1.01919532e+00 1.18723638e-01 -5.95488846e-01 -3.60426962e-01
-9.52736020e-01 1.71618283e-01 3.21562707e-01 7.57204592e-01
2.55987555e-01 2.85504311e-01 -6.70221210e-01 8.14039707e-01
1.80378571e-01 8.00858587e-02 4.66021270e-01 -7.93993831e-01
6.68751419e-01 3.24650824e-01 -4.82394733e-02 -9.23606396e-01
-7.49009326e-02 -4.61713165e-01 -7.45819807e-01 2.02118933e-01
4.12122041e-01 -1.81452744e-02 -1.15060115e+00 1.91603088e+00
2.95445710e-01 1.37760210e+00 -1.49528444e-01 7.65295506e-01
5.04891455e-01 8.50173235e-01 2.34965324e-01 -4.87916470e-01
9.14734721e-01 -1.22716856e+00 -7.49248207e-01 -1.10719718e-01
6.63662016e-01 -3.85971725e-01 1.01971698e+00 5.33636510e-01
-1.25093484e+00 -5.06613553e-01 -9.01065528e-01 2.30506122e-01
-1.17749907e-01 4.55862097e-02 -1.15570806e-01 3.05866394e-02
-9.04632270e-01 5.99840283e-01 -1.07392406e+00 -3.48186903e-02
8.65337789e-01 -3.25876810e-02 -5.97759187e-02 -8.46495330e-02
-1.20839226e+00 6.59362912e-01 1.14199512e-01 1.28569230e-01
-1.60522544e+00 -8.80468071e-01 -9.30470526e-01 -9.97868404e-02
6.80552483e-01 -2.65596330e-01 1.28824008e+00 -1.22492218e+00
-1.34752512e+00 9.83928561e-01 -1.16014093e-01 -8.21516156e-01
8.36889386e-01 -5.27025640e-01 -5.67798555e-01 3.30803335e-01
1.38861939e-01 2.91264415e-01 1.16867888e+00 -1.14264047e+00
-8.01124215e-01 -2.06708416e-01 -6.58374727e-02 2.31680438e-01
-4.74125415e-01 -7.26218745e-02 -6.37502849e-01 -1.05356169e+00
-9.59226489e-02 -6.77550912e-01 8.20838287e-02 1.96687549e-01
-4.16043341e-01 -5.22264779e-01 1.26405120e+00 -9.24300075e-01
1.46481717e+00 -2.25388527e+00 3.94125164e-01 2.93632865e-01
3.88720810e-01 6.22824788e-01 -4.05778259e-01 -1.41384769e-02
-2.81499475e-01 -1.86956659e-01 -2.48931676e-01 -6.03516281e-01
-1.77224502e-01 1.23549104e-01 -3.50982875e-01 4.72739160e-01
3.45009238e-01 4.18726504e-01 -1.16257215e+00 -5.85608482e-01
2.12204844e-01 2.62249768e-01 -4.16534394e-01 7.71716118e-01
-4.87905711e-01 4.11741912e-01 -3.39909256e-01 7.29980230e-01
4.23205316e-01 7.73859918e-02 -3.17812383e-01 -8.64023194e-02
1.17666200e-02 5.17661348e-02 -7.79687464e-01 1.80381203e+00
-3.17964882e-01 7.06785858e-01 -2.49402359e-01 -1.40043962e+00
5.94851732e-01 5.07447183e-01 6.80519164e-01 -7.63415933e-01
-9.40015074e-03 1.91074386e-01 -3.12280357e-01 -9.65820372e-01
-1.88706666e-01 1.21365257e-01 2.00518072e-01 2.30746269e-01
4.27359879e-01 5.34451783e-01 1.43618315e-01 3.07842553e-01
1.34077895e+00 3.19541931e-01 -1.98789582e-01 2.18405411e-01
8.64887893e-01 -4.71148729e-01 8.91181946e-01 7.20068812e-01
-5.08767307e-01 3.81553352e-01 7.74283648e-01 -7.48677731e-01
-1.19812775e+00 -1.14472675e+00 2.54075080e-01 1.13777423e+00
-9.20545682e-03 -4.31755722e-01 -7.16121972e-01 -1.36856771e+00
-1.17607117e-01 7.86856592e-01 -8.64093006e-01 -4.24520463e-01
-5.20049989e-01 -9.18847173e-02 5.13333499e-01 4.36209261e-01
3.64138126e-01 -1.01040125e+00 -3.41629893e-01 2.30529696e-01
-3.39386195e-01 -1.17619014e+00 -6.27496839e-01 3.52123417e-02
-7.46980369e-01 -9.89954531e-01 -7.51887143e-01 -8.35166991e-01
6.10972106e-01 1.16548046e-01 1.03531075e+00 8.66035149e-02
-1.34981051e-01 4.17295694e-01 -5.33422649e-01 -3.01279843e-01
-3.59080613e-01 -7.43159801e-02 2.99670786e-01 6.63480937e-01
5.07215083e-01 -8.19399834e-01 -4.65701848e-01 2.70088345e-01
-9.18748677e-01 -1.86292931e-01 4.29255724e-01 9.47879195e-01
9.01185811e-01 -9.44905728e-02 4.68341768e-01 -8.00785363e-01
3.82367581e-01 -7.72616088e-01 -5.80310702e-01 2.38251746e-01
-4.29068327e-01 -1.37266740e-02 8.22054029e-01 -6.15821958e-01
-9.76422966e-01 -6.55617714e-02 -9.84188691e-02 -1.51778567e+00
-3.24735105e-01 4.78725582e-01 -5.54868504e-02 2.49299571e-01
4.73975480e-01 4.10335034e-01 -1.29021615e-01 -3.86904955e-01
-3.04543041e-02 2.70671666e-01 7.84318864e-01 -4.58852589e-01
1.15927863e+00 2.90686995e-01 -1.78276464e-01 -6.09748602e-01
-1.28473961e+00 -4.18917745e-01 -6.52853191e-01 -5.09615362e-01
1.03717065e+00 -1.03998840e+00 -1.88252047e-01 8.90636027e-01
-8.28869522e-01 -5.62833369e-01 -4.64832008e-01 3.29670697e-01
-5.19106150e-01 2.11937204e-01 -5.76062679e-01 -9.12095070e-01
-1.27125382e-01 -8.86884272e-01 8.44713151e-01 -3.82225998e-02
2.63786465e-01 -9.50929344e-01 3.32916468e-01 1.37450501e-01
1.80445984e-01 5.09263217e-01 8.42920065e-01 -1.02211428e+00
-3.71804804e-01 -2.19848037e-01 5.81558188e-03 7.59091854e-01
6.49985671e-02 -8.86331592e-03 -1.01356161e+00 -3.72802049e-01
-1.38431534e-01 -5.73001444e-01 1.09621906e+00 5.24097800e-01
1.93647766e+00 -5.05459905e-01 -3.57002877e-02 6.36453331e-01
9.63530183e-01 3.18802685e-01 5.91941953e-01 3.64881843e-01
1.00055015e+00 3.21281493e-01 9.49722350e-01 5.56208491e-01
3.82455140e-01 6.37340784e-01 7.22467542e-01 7.18166074e-03
1.40313327e-01 -4.06756341e-01 7.26782799e-01 7.12652743e-01
-9.54990685e-02 -3.43486786e-01 -7.18633711e-01 6.37497246e-01
-1.94163370e+00 -1.28755248e+00 1.72121733e-01 2.16629672e+00
4.49521631e-01 5.76314986e-01 5.04849136e-01 2.38942057e-01
5.52967668e-01 6.14973187e-01 -8.66528630e-01 -1.25449806e-01
4.30959575e-02 -1.78538412e-01 -6.50616512e-02 1.02099836e-01
-1.60096693e+00 7.71671534e-01 4.85299635e+00 9.37084794e-01
-9.64736879e-01 2.28813782e-01 8.97574544e-01 -3.60660166e-01
-1.08250584e-02 -3.29547584e-01 -4.66641970e-02 8.37353110e-01
9.70861852e-01 1.65479258e-01 2.84917176e-01 8.36286008e-01
1.84154347e-01 2.93637812e-01 -1.19676101e+00 9.91260052e-01
2.62319565e-01 -1.13367403e+00 1.88118249e-01 -3.48007768e-01
6.36528134e-01 7.53591657e-02 2.49644473e-01 5.20282149e-01
-1.49268866e-01 -7.26060092e-01 8.88093889e-01 5.63691974e-01
8.66130769e-01 -1.03219509e+00 8.07412326e-01 2.66354322e-01
-1.47313488e+00 -3.05245489e-01 -1.35913014e-01 2.07927763e-01
1.25514776e-01 4.34718430e-01 -4.33342218e-01 3.84851277e-01
9.24077451e-01 1.23039520e+00 -4.02796537e-01 1.14631402e+00
-1.53291732e-01 1.16270077e+00 -1.04571447e-01 4.27510828e-01
5.39552033e-01 -2.06946358e-01 9.40062821e-01 1.07892549e+00
4.16779608e-01 1.30132318e-01 3.47217053e-01 4.25732285e-01
-2.78857797e-01 -7.41417930e-02 -6.11729383e-01 -3.10432073e-03
3.96123499e-01 9.81842756e-01 -2.43649393e-01 -3.66225809e-01
-6.62354946e-01 1.24809909e+00 4.14750397e-01 3.71829301e-01
-9.91544843e-01 -2.16237038e-01 6.74155772e-01 7.54501298e-02
3.92622888e-01 1.62051186e-01 3.47452492e-01 -1.35007286e+00
2.50344545e-01 -7.75131106e-01 9.97210741e-01 -6.64374530e-01
-1.58784115e+00 6.10912085e-01 -7.68430531e-02 -1.79821980e+00
-1.73321739e-01 -2.57932425e-01 -1.10392046e+00 3.63001883e-01
-1.70605612e+00 -1.26418555e+00 -3.50094378e-01 9.23021853e-01
9.19226289e-01 -5.09207547e-01 4.95628595e-01 6.85567617e-01
-8.51325393e-01 9.08811867e-01 -1.30442753e-02 4.19020414e-01
7.01058149e-01 -1.39337289e+00 2.03498647e-01 1.10805523e+00
1.82278827e-01 -3.29475552e-01 7.09953785e-01 -4.59032595e-01
-8.19944084e-01 -1.47815645e+00 5.39045036e-01 -3.57163161e-01
5.99489212e-01 -2.71709502e-01 -1.26177418e+00 9.25601721e-01
-7.14133168e-03 5.73982835e-01 5.28692186e-01 -9.53987390e-02
-5.21817565e-01 -4.00356591e-01 -1.09327900e+00 2.87348211e-01
1.20336437e+00 -6.45979822e-01 -5.39577186e-01 3.75110626e-01
8.62444341e-01 -6.46834731e-01 -4.37810659e-01 4.01639313e-01
3.01337719e-01 -8.21598470e-01 8.75136375e-01 -9.48355317e-01
5.84034801e-01 -2.18796119e-01 1.14218771e-01 -1.46190560e+00
-1.32140666e-01 -5.56850076e-01 -8.10734808e-01 1.30511558e+00
3.66186768e-01 -3.02542984e-01 6.27371907e-01 1.98463753e-01
-5.83377302e-01 -1.08543134e+00 -1.10296726e+00 -9.23307896e-01
-3.00618321e-01 -6.62134826e-01 3.01880568e-01 1.14517140e+00
-2.96359628e-01 -3.07950210e-02 -8.95750940e-01 4.48227584e-01
7.73239970e-01 -3.51915687e-01 4.72743154e-01 -1.31603539e+00
-1.91271722e-01 -3.02695513e-01 -5.99732816e-01 -8.42700005e-01
2.93999463e-01 -6.89643562e-01 3.74955870e-02 -9.62821662e-01
-3.28523219e-02 -2.26303086e-01 -8.26575220e-01 5.03158152e-01
-2.28815198e-01 7.49143884e-02 -8.48673731e-02 2.19382048e-02
-1.08579123e+00 8.23995888e-01 8.16809118e-01 -1.82752460e-01
-2.65003175e-01 2.84995109e-01 7.68856518e-03 9.33573604e-01
5.99797130e-01 -5.97278655e-01 -6.55197084e-01 -3.47108811e-01
-4.26313877e-02 9.95893180e-02 4.12172139e-01 -1.22164500e+00
1.71662062e-01 -3.76739018e-02 2.96168715e-01 -7.58059382e-01
2.28654876e-01 -9.00679946e-01 -3.58724564e-01 2.83886015e-01
-7.24694371e-01 1.06321216e-01 6.64572567e-02 9.83841360e-01
-3.37493569e-01 6.82602674e-02 8.18339705e-01 1.44024923e-01
-7.23286033e-01 1.00362182e+00 -3.35465342e-01 4.06498671e-01
1.34248960e+00 5.58995828e-02 2.60438398e-02 -6.17469490e-01
-8.94726217e-01 7.03286827e-01 -2.72846334e-02 5.66844642e-01
8.86142254e-01 -1.68809724e+00 -8.39855433e-01 1.92934975e-01
5.03882408e-01 2.01203898e-01 6.51818156e-01 8.52976799e-01
-2.86830723e-01 1.61359441e-02 -1.53500155e-01 -7.67209470e-01
-1.33059204e+00 6.68627441e-01 4.98498231e-01 -4.83690590e-01
-7.02320039e-01 1.11812115e+00 1.95347006e-03 -1.04395077e-01
8.03557217e-01 -1.09289892e-01 -3.08607340e-01 6.14554211e-02
5.98603070e-01 4.01061505e-01 -8.23319107e-02 -6.21137381e-01
-2.07880527e-01 3.56991112e-01 -3.77776176e-01 1.31603658e-01
1.52864134e+00 -6.93024248e-02 2.91771978e-01 7.82129407e-01
1.18899977e+00 -4.21649933e-01 -1.77909708e+00 -7.08146453e-01
-1.03677459e-01 -5.34655452e-01 5.06053455e-02 -6.44705236e-01
-1.39408386e+00 9.46749985e-01 9.12352204e-01 1.56240255e-01
1.42343116e+00 -1.35321422e-02 9.91254747e-01 3.16027313e-01
-1.58288032e-02 -1.21950030e+00 7.01254964e-01 4.16145682e-01
6.18638217e-01 -1.17512739e+00 -4.80026603e-01 8.47485289e-02
-7.93794811e-01 8.53009582e-01 1.22732842e+00 -2.10533544e-01
6.44942939e-01 -2.78567940e-01 -7.95552973e-03 -2.11516172e-01
-8.76686454e-01 1.71889029e-02 5.40778816e-01 4.04632092e-01
-1.50156468e-01 -3.77859503e-01 2.41625726e-01 3.99241686e-01
5.53684592e-01 -9.79685187e-02 2.09627330e-01 8.71337295e-01
-2.16837436e-01 -7.72593319e-01 3.68100479e-02 8.71210515e-01
-4.18017536e-01 1.89947233e-01 -2.88708713e-02 4.02710438e-01
3.51955563e-01 5.75733423e-01 3.13389987e-01 -6.79960549e-01
4.38056618e-01 3.38486135e-01 8.65461305e-02 -2.95669973e-01
-2.27348477e-01 1.11792553e-02 -1.25634419e-02 -9.53069627e-01
-4.16947633e-01 -9.02138591e-01 -9.28863525e-01 -7.48718577e-03
-1.04259523e-02 1.69079229e-01 3.26226577e-02 1.13836896e+00
3.10211062e-01 7.49514699e-01 1.09189904e+00 -8.17914307e-01
-4.09285158e-01 -7.83712268e-01 -4.94803399e-01 7.29679763e-01
9.28481042e-01 -7.95923650e-01 -7.06258118e-01 6.83227479e-02] | [7.853674411773682, 1.5900557041168213] |
ac2206af-da7c-4f0d-9752-8af068f0a0d8 | capnet-continuous-approximation-projection | 1811.11731 | null | http://arxiv.org/abs/1811.11731v1 | http://arxiv.org/pdf/1811.11731v1.pdf | CAPNet: Continuous Approximation Projection For 3D Point Cloud Reconstruction Using 2D Supervision | Knowledge of 3D properties of objects is a necessity in order to build
effective computer vision systems. However, lack of large scale 3D datasets can
be a major constraint for data-driven approaches in learning such properties.
We consider the task of single image 3D point cloud reconstruction, and aim to
utilize multiple foreground masks as our supervisory data to alleviate the need
for large scale 3D datasets. A novel differentiable projection module, called
'CAPNet', is introduced to obtain such 2D masks from a predicted 3D point
cloud. The key idea is to model the projections as a continuous approximation
of the points in the point cloud. To overcome the challenges of sparse
projection maps, we propose a loss formulation termed 'affinity loss' to
generate outlier-free reconstructions. We significantly outperform the existing
projection based approaches on a large-scale synthetic dataset. We show the
utility and generalizability of such a 2D supervised approach through
experiments on a real-world dataset, where lack of 3D data can be a serious
concern. To further enhance the reconstructions, we also propose a test stage
optimization procedure to obtain reconstructions that display high
correspondence with the observed input image. | ['Priyanka Mandikal', 'R. Venkatesh Babu', 'Mayank Agarwal', 'Navaneet K L'] | 2018-11-28 | null | null | null | null | ['3d-point-cloud-reconstruction', 'point-cloud-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 3.05824697e-01 3.02274805e-02 1.67427853e-01 -3.14272434e-01
-8.24858665e-01 -3.71415585e-01 5.80407679e-01 -4.07009214e-01
-6.66060075e-02 3.86213690e-01 4.80328463e-02 -2.12329820e-01
-4.85728420e-02 -6.26850545e-01 -1.14581847e+00 -6.24463260e-01
4.02951866e-01 7.86091268e-01 5.25447249e-01 1.09088518e-01
4.82431501e-01 9.48146462e-01 -1.46753073e+00 8.96878913e-02
8.67821276e-01 8.78369391e-01 5.51542461e-01 1.28882751e-01
-3.63693357e-01 1.71299815e-01 -2.22397879e-01 -1.27936840e-01
7.75466740e-01 -1.58262730e-01 -3.47193927e-01 8.26227009e-01
5.17629087e-01 -1.65869951e-01 -9.84827951e-02 1.23411536e+00
2.86499143e-01 -1.44746974e-01 6.64890885e-01 -1.23253596e+00
-3.33014578e-01 -1.84553325e-01 -8.40528131e-01 -2.75690615e-01
3.48714203e-01 1.88121915e-01 6.96928501e-01 -1.30414212e+00
7.23367393e-01 1.21648252e+00 7.32280493e-01 2.94917256e-01
-1.40963185e+00 -4.05914992e-01 1.27741873e-01 -4.06098925e-02
-1.40840769e+00 -3.58699679e-01 1.33326292e+00 -5.30789435e-01
5.86103737e-01 2.92828143e-01 6.68932915e-01 9.15609181e-01
-1.38486311e-01 6.25917554e-01 1.28687727e+00 -4.05738950e-01
2.40538567e-01 3.61758381e-01 -1.40929908e-01 6.32464230e-01
2.57745624e-01 -1.18369218e-02 -5.95567048e-01 -2.43804276e-01
1.16847706e+00 4.14845705e-01 -4.19546157e-01 -1.10980368e+00
-1.32629931e+00 5.29049635e-01 5.21210194e-01 -3.20443749e-01
-5.98506153e-01 -1.77346244e-01 -1.59133360e-01 7.02222884e-02
6.04743183e-01 2.52539307e-01 -2.64304399e-01 2.07036898e-01
-8.80989790e-01 1.76206470e-01 6.10732853e-01 1.06487095e+00
9.98719454e-01 -1.99274458e-02 2.47615680e-01 7.22569525e-01
5.72293103e-01 7.05727160e-01 1.91278964e-01 -1.08019722e+00
6.14163101e-01 9.30557430e-01 2.49987096e-01 -1.04594743e+00
4.44159210e-02 -2.79689997e-01 -7.69374609e-01 5.39642632e-01
1.12828821e-01 5.56887209e-01 -9.63506222e-01 1.32990706e+00
6.95297897e-01 4.01186526e-01 -7.18721971e-02 1.03231776e+00
3.50568384e-01 5.51275790e-01 -7.30318427e-01 -3.20035219e-01
7.58405566e-01 -7.03224063e-01 -1.99239194e-01 -1.11496165e-01
1.82449847e-01 -7.73938894e-01 1.35143781e+00 2.28758261e-01
-1.00408697e+00 -3.94748658e-01 -9.55888331e-01 -3.55542055e-03
3.02472748e-02 -2.10365653e-02 2.74408728e-01 3.12596083e-01
-6.32659972e-01 5.82979500e-01 -1.05140865e+00 -3.41291755e-01
6.74287498e-01 1.83026478e-01 -5.51313519e-01 -2.82797217e-01
-1.90702006e-01 8.31252754e-01 2.47664884e-01 1.43179893e-01
-8.91866922e-01 -8.81264627e-01 -6.46912456e-01 -2.26482332e-01
4.00672704e-01 -6.10753894e-01 6.39831305e-01 -5.68606257e-01
-1.20378458e+00 1.05526972e+00 -1.25505835e-01 -2.99754199e-02
6.99456275e-01 -3.04766297e-01 1.50322765e-02 1.43043652e-01
1.38214499e-01 6.04866743e-01 1.01429045e+00 -1.81869638e+00
-3.54050398e-01 -5.16837120e-01 -2.84918129e-01 4.39422220e-01
-1.47163514e-02 -2.73775786e-01 -7.46631384e-01 -3.02575886e-01
7.98274040e-01 -1.04470932e+00 -4.77191627e-01 4.96792197e-01
-4.87919748e-01 3.84857744e-01 1.00980306e+00 -5.28353631e-01
4.20609444e-01 -2.15376902e+00 1.45960167e-01 3.91294122e-01
1.49773926e-01 1.35616258e-01 -1.90533586e-02 1.91692725e-01
1.04212977e-01 -2.18679726e-01 -7.15100646e-01 -5.99831879e-01
-9.90703404e-02 4.94716853e-01 -4.61973399e-01 7.07336903e-01
4.76360679e-01 6.31062269e-01 -6.06921136e-01 -3.85240227e-01
4.36676234e-01 5.84858537e-01 -6.82314396e-01 3.91738117e-01
-3.77200335e-01 7.39516616e-01 -5.96750677e-01 6.70593262e-01
1.01114380e+00 -3.82382989e-01 -3.03866267e-01 -3.67724359e-01
-8.17171782e-02 9.49384198e-02 -1.35709643e+00 1.98731697e+00
-1.88775510e-01 1.70453519e-01 1.33016557e-01 -1.11664975e+00
1.16446269e+00 6.94449395e-02 6.20299757e-01 -3.22302639e-01
-1.62018269e-01 2.63717949e-01 -1.56844109e-01 -3.10984671e-01
2.58234203e-01 -2.71702766e-01 3.23424429e-01 3.76663327e-01
-2.69299269e-01 -6.02919757e-01 -5.05916715e-01 1.77706346e-01
1.00029159e+00 5.39655685e-01 6.85693547e-02 -1.82374418e-01
4.49579239e-01 2.53618091e-01 7.80763268e-01 4.12873715e-01
3.10268011e-02 1.22133791e+00 2.97917455e-01 -4.33376729e-01
-1.53173614e+00 -1.21466351e+00 -1.98006347e-01 6.41199797e-02
4.20954287e-01 -6.21447749e-02 -3.46724033e-01 -7.32092440e-01
1.33446813e-01 4.39896405e-01 -2.93509781e-01 7.90529028e-02
-6.41974688e-01 -6.92893147e-01 -1.27878919e-01 3.78345340e-01
4.31438535e-01 -8.05632949e-01 -6.63110197e-01 -1.09176501e-03
-1.39385432e-01 -1.27986372e+00 -3.76980096e-01 1.54903680e-01
-1.16054094e+00 -9.68605816e-01 -5.58726430e-01 -5.60481310e-01
1.23088026e+00 5.55653036e-01 9.59243119e-01 -1.74338222e-02
-9.52461809e-02 4.03736055e-01 -2.53702760e-01 -3.06209773e-01
-3.86358410e-01 -3.54466289e-01 3.47228050e-01 2.18662113e-01
2.29088485e-01 -1.14899051e+00 -5.47006607e-01 5.27761757e-01
-8.23921442e-01 3.43183011e-01 8.07926834e-01 7.92787373e-01
1.17260468e+00 -6.31466061e-02 2.29245737e-01 -9.28315103e-01
1.26190007e-01 -2.67332852e-01 -8.85370731e-01 6.26155511e-02
-6.01455808e-01 1.39965013e-01 4.68706787e-01 -4.46309507e-01
-9.68657494e-01 6.85719132e-01 -4.49627265e-03 -1.17203557e+00
-1.12095647e-01 1.88083917e-01 -4.81600285e-01 -2.79508859e-01
5.25341928e-01 5.16856253e-01 3.43111932e-01 -8.43591809e-01
2.00473189e-01 4.57794249e-01 6.30594671e-01 -6.25348449e-01
1.33466005e+00 9.15581822e-01 3.91376346e-01 -7.13854074e-01
-8.48710597e-01 -6.85240507e-01 -1.05813563e+00 -6.35461360e-02
4.58357364e-01 -9.36990917e-01 -2.35960349e-01 3.03401854e-02
-1.14484596e+00 1.57287568e-02 -4.12011623e-01 5.00367343e-01
-7.93660164e-01 5.57674229e-01 -1.56622827e-01 -6.15727782e-01
-1.22529320e-01 -1.21451950e+00 1.21211398e+00 -1.17862523e-01
1.85054153e-01 -5.81621289e-01 3.79663780e-02 4.41286951e-01
-4.65211757e-02 2.72169888e-01 6.80511236e-01 -3.51344615e-01
-1.18439531e+00 -1.97510257e-01 -3.14532161e-01 5.76429248e-01
1.87858552e-01 -1.73824370e-01 -9.56046045e-01 -2.99581438e-01
5.33435702e-01 -2.45152667e-01 5.64132512e-01 1.84259608e-01
1.10176325e+00 -1.14202969e-01 -3.75085324e-01 8.92572284e-01
1.58866549e+00 -2.65001297e-01 5.29632807e-01 1.23147503e-01
1.03097320e+00 5.88886559e-01 8.42730939e-01 4.22425717e-01
1.48512498e-01 6.48994386e-01 6.92933321e-01 6.10315800e-02
-1.16691656e-01 -6.90636635e-01 1.89229906e-01 1.16191757e+00
-2.46859975e-02 2.46470109e-01 -8.95113051e-01 4.84701991e-01
-1.86900795e+00 -5.69834948e-01 -1.53380066e-01 2.48621440e+00
5.24981260e-01 2.70696461e-01 -1.13556333e-01 3.73216867e-02
6.16714180e-01 3.36553194e-02 -7.32246101e-01 3.39581907e-01
-9.31705758e-02 -1.75896019e-01 3.70179087e-01 4.98285830e-01
-7.30667293e-01 8.05453837e-01 5.62504101e+00 5.92145383e-01
-1.08024275e+00 -2.65015047e-02 2.27579966e-01 -1.55800665e-02
-5.52121937e-01 3.46664548e-01 -8.16565573e-01 4.72010851e-01
3.48389536e-01 1.23712011e-01 2.41937399e-01 9.26330328e-01
3.17312837e-01 4.14186493e-02 -1.29255390e+00 1.29850519e+00
2.01248303e-01 -1.43258131e+00 3.54550987e-01 3.70860577e-01
7.95273960e-01 1.73886403e-01 -1.03050999e-01 -2.82062590e-01
7.55903050e-02 -7.80911803e-01 6.78051651e-01 6.59046948e-01
6.88012660e-01 -4.48667526e-01 2.51147687e-01 8.46522212e-01
-8.39802802e-01 3.24720114e-01 -7.66504228e-01 5.25357500e-02
4.53894347e-01 8.23909879e-01 -9.49896932e-01 5.95371366e-01
7.14611113e-01 7.40999103e-01 -4.63946164e-01 1.22625589e+00
-9.43618268e-02 4.06461865e-01 -7.86457121e-01 3.86460543e-01
5.44796186e-03 -7.07986534e-01 9.25992012e-01 5.86116254e-01
5.50811887e-01 1.18151225e-01 3.22976530e-01 1.14104640e+00
-4.80929613e-02 2.54812036e-02 -9.16207254e-01 3.27949673e-01
4.82270926e-01 1.20686233e+00 -8.03266287e-01 -7.07789212e-02
-5.34629643e-01 1.09398079e+00 3.74518692e-01 2.10329637e-01
-5.55875003e-01 2.59668112e-01 4.39829618e-01 3.53015870e-01
4.22558278e-01 -3.99999529e-01 -3.50901783e-01 -1.31897759e+00
5.34646928e-01 -7.43761122e-01 -1.18157879e-01 -9.08488393e-01
-1.33501983e+00 2.94353366e-01 -3.27815413e-02 -1.79001331e+00
9.81670618e-02 -4.90082830e-01 -6.34287179e-01 9.49072778e-01
-1.41871536e+00 -1.25545490e+00 -4.91251439e-01 6.60706401e-01
4.97467726e-01 -5.99269718e-02 4.61492240e-01 2.20367894e-01
-1.90287456e-01 -5.39588109e-02 1.15029737e-01 -3.13522071e-01
4.06559110e-01 -9.95457709e-01 3.44713300e-01 9.74130988e-01
5.20718396e-01 3.89377862e-01 5.73819816e-01 -7.59329140e-01
-1.72724986e+00 -1.16871560e+00 3.78608584e-01 -8.42903554e-01
2.11674020e-01 -5.09811342e-01 -1.25281405e+00 8.04773569e-01
-4.63074654e-01 3.31578195e-01 2.27824897e-01 -1.89936996e-01
-1.75958782e-01 -1.18278779e-01 -1.27104938e+00 4.84463274e-01
1.25269043e+00 -4.31314558e-01 -7.30688632e-01 4.05119002e-01
7.97662675e-01 -4.84896362e-01 -7.31384516e-01 5.45258939e-01
2.60924667e-01 -9.49250221e-01 1.24824572e+00 -2.80594468e-01
4.51928109e-01 -6.93491280e-01 -3.96252036e-01 -1.24544418e+00
-7.84763768e-02 -5.14626205e-01 -1.63664654e-01 1.13263702e+00
2.18864009e-01 -4.77761596e-01 1.28776586e+00 6.45922244e-01
-3.55117023e-01 -6.43215597e-01 -9.51635361e-01 -9.19514954e-01
-1.61017805e-01 -5.25778592e-01 5.08902669e-01 9.57589388e-01
-5.84999144e-01 2.34198734e-01 -4.84136164e-01 5.88563383e-01
1.04324186e+00 4.89827484e-01 1.17495978e+00 -1.25549984e+00
-2.43293136e-01 -4.57881242e-02 -5.10622621e-01 -1.41184175e+00
-3.31008099e-02 -8.76605392e-01 1.75266832e-01 -1.28487945e+00
2.54969329e-01 -7.44685352e-01 -1.17683552e-01 7.89071620e-02
-4.23293784e-02 2.68223077e-01 2.14573205e-01 6.89947307e-01
-3.69099885e-01 9.59858835e-01 1.43083298e+00 1.60239980e-01
-1.85417667e-01 2.18767777e-01 -4.55881178e-01 9.54660594e-01
4.17599887e-01 -5.80141544e-01 -5.08109272e-01 -5.09929597e-01
-4.98467721e-02 -2.90652551e-02 5.09025037e-01 -1.01096690e+00
2.46114612e-01 -3.33874017e-01 4.57823962e-01 -1.21924627e+00
8.21160734e-01 -1.29729438e+00 4.86641228e-01 7.80213326e-02
-2.98636761e-02 -2.52017640e-02 -1.57324135e-01 7.53663063e-01
-9.16568786e-02 -9.13034007e-02 8.12406242e-01 -2.66988099e-01
-5.03759801e-01 6.70456648e-01 4.29537326e-01 -2.43937254e-01
1.02282798e+00 -5.21876574e-01 1.81965664e-01 -6.31074309e-02
-3.62101734e-01 1.71423569e-01 1.02592766e+00 2.32113853e-01
1.14004159e+00 -1.38513076e+00 -6.34477913e-01 6.06832087e-01
2.24459216e-01 7.18577325e-01 -7.61252120e-02 7.66516387e-01
-7.75366068e-01 1.58472344e-01 -2.39674225e-01 -1.10862446e+00
-1.17701793e+00 6.80583715e-01 1.46845639e-01 1.77244261e-01
-1.40853500e+00 5.55777490e-01 3.60012144e-01 -8.36612582e-01
2.04225436e-01 -2.80216426e-01 3.11601698e-01 -5.61273813e-01
2.06782326e-01 1.08481228e-01 3.18625718e-02 -6.33917272e-01
-2.11313292e-01 7.87202179e-01 7.54083972e-03 -1.80539832e-01
1.63261867e+00 -2.54326165e-01 -2.42377315e-02 4.98170286e-01
1.08739924e+00 7.33442977e-02 -1.82342124e+00 -5.08266687e-01
-9.25538503e-03 -1.06300461e+00 -4.22466584e-02 -3.47416669e-01
-9.87871885e-01 9.27484453e-01 5.24816930e-01 -1.35355070e-01
8.78852546e-01 2.26088077e-01 6.78504825e-01 3.95848006e-01
4.67714012e-01 -7.53766119e-01 1.00522153e-01 1.22196704e-01
1.08906114e+00 -1.51462853e+00 3.04038852e-01 -6.66833341e-01
-5.47615230e-01 7.70686865e-01 5.98100781e-01 -3.21288764e-01
7.08949506e-01 -6.17897213e-02 -7.11596105e-03 -3.75955880e-01
-4.78638083e-01 7.58548155e-02 2.79074371e-01 7.43954360e-01
-3.17555606e-01 -1.27005935e-01 1.40429676e-01 1.07735693e-01
-7.72311166e-02 -1.04690418e-01 4.42709088e-01 9.23814178e-01
-4.25464332e-01 -1.10162640e+00 -6.72449231e-01 4.84915942e-01
-1.81522984e-02 1.03029065e-01 -2.79869109e-01 5.15167415e-01
-2.02510685e-01 2.00260267e-01 5.15230522e-02 -2.84040421e-01
3.89957279e-01 -2.77431495e-02 5.31074643e-01 -7.06522226e-01
9.61591825e-02 2.20120504e-01 -3.17994624e-01 -6.06073737e-01
-5.79456091e-01 -7.58739531e-01 -9.52282906e-01 -1.79670289e-01
-2.75476605e-01 -1.29148826e-01 8.01644266e-01 7.87780583e-01
5.19437790e-01 -3.74143459e-02 9.39260066e-01 -1.11488450e+00
-7.06331730e-01 -6.76775098e-01 -7.11948097e-01 8.22276950e-01
2.02925518e-01 -8.51971030e-01 -5.40011406e-01 2.22786814e-01] | [8.377989768981934, -3.1927459239959717] |
0068244c-e16d-4222-9ab5-f5d0b33c23a5 | an-comparative-analysis-of-different-pitch | 2301.13383 | null | https://arxiv.org/abs/2301.13383v1 | https://arxiv.org/pdf/2301.13383v1.pdf | An Comparative Analysis of Different Pitch and Metrical Grid Encoding Methods in the Task of Sequential Music Generation | Pitch and meter are two fundamental music features for symbolic music generation tasks, where researchers usually choose different encoding methods depending on specific goals. However, the advantages and drawbacks of different encoding methods have not been frequently discussed. This paper presents a integrated analysis of the influence of two low-level feature, pitch and meter, on the performance of a token-based sequential music generation model. First, the commonly used MIDI number encoding and a less used class-octave encoding are compared. Second, an dense intra-bar metric grid is imposed to the encoded sequence as auxiliary features. Different complexity and resolutions of the metric grid are compared. For complexity, the single token approach and the multiple token approach are compared; for grid resolution, 0 (ablation), 1 (bar-level), 4 (downbeat-level) 12, (8th-triplet-level) up to 64 (64th-note-grid-level) are compared; for duration resolution, 4, 8, 12 and 16 subdivisions per beat are compared. All different encodings are tested on separately trained Transformer-XL models for a melody generation task. Regarding distribution similarity of several objective evaluation metrics to the test dataset, results suggest that the class-octave encoding significantly outperforms the taken-for-granted MIDI encoding on pitch-related metrics; finer grids and multiple-token grids improve the rhythmic quality, but also suffer from over-fitting at early training stage. Results display a general phenomenon of over-fitting from two aspects, the pitch embedding space and the test loss of the single-token grid encoding. From a practical perspective, we both demonstrate the feasibility and raise the concern of easy over-fitting problem of using smaller networks and lower embedding dimensions on the generation task. The findings can also contribute to futural models in terms of feature engineering. | ['George Fazekas', 'Shengchen Li', 'Yuqiang Li'] | 2023-01-31 | null | null | null | null | ['music-generation', 'feature-engineering', 'music-generation'] | ['audio', 'methodology', 'music'] | [ 2.66958088e-01 -7.29047805e-02 -2.65307836e-02 1.40951037e-01
-6.65025473e-01 -5.78979075e-01 5.41484296e-01 2.85126597e-01
-5.00486195e-01 7.58687794e-01 2.74987221e-01 -6.78891018e-02
-6.42300069e-01 -6.21793151e-01 -1.63572565e-01 -8.02390873e-01
-2.83195257e-01 2.04412282e-01 7.19440207e-02 -3.67637843e-01
3.32765192e-01 3.37103754e-01 -1.93321931e+00 3.08833897e-01
6.20469153e-01 9.73214149e-01 1.08523950e-01 6.64575458e-01
1.05404109e-01 3.43262136e-01 -1.00262427e+00 -3.60434592e-01
3.17892015e-01 -7.78197408e-01 -5.39527774e-01 -3.29966605e-01
2.86136061e-01 3.99373733e-02 1.57109067e-01 8.93799722e-01
8.81829858e-01 2.05021456e-01 7.51998007e-01 -1.11126494e+00
-1.87746346e-01 1.10084915e+00 -3.15935731e-01 9.37407836e-02
4.44730192e-01 3.26990448e-02 1.23297369e+00 -6.57267749e-01
5.12521625e-01 8.86433840e-01 9.92896259e-01 8.57456774e-02
-1.31977010e+00 -6.74579024e-01 -3.99603546e-01 2.36814752e-01
-1.56883597e+00 -1.95852995e-01 1.00862217e+00 -5.65732419e-01
8.86802733e-01 5.15331864e-01 8.64314795e-01 1.01102543e+00
2.86953598e-01 8.80273655e-02 1.32149124e+00 -6.21200085e-01
6.94188401e-02 1.52697086e-01 -1.11314513e-01 2.15635315e-01
2.11238489e-01 4.22535837e-01 -7.14646459e-01 -1.23039573e-01
7.32765973e-01 -6.78060234e-01 -2.50360817e-01 1.71603467e-02
-1.15965378e+00 8.91381979e-01 4.89218384e-02 9.87614632e-01
-4.06850666e-01 8.39066878e-02 6.20966792e-01 4.67866302e-01
6.19720854e-02 8.63750160e-01 -3.24381739e-01 -6.49909616e-01
-1.38859487e+00 5.72655201e-01 7.49593616e-01 4.68551427e-01
5.84473491e-01 6.29194975e-01 -6.39032722e-01 9.71558750e-01
-7.88321942e-02 8.79184902e-02 9.43773687e-01 -7.42674530e-01
3.37764919e-01 2.10796610e-01 -2.13167772e-01 -1.14297473e+00
-5.81798732e-01 -8.04038465e-01 -8.46356690e-01 3.23865533e-01
7.28434443e-01 -5.59730008e-02 -4.20258731e-01 1.83218205e+00
4.64940034e-02 4.89296280e-02 -7.61777833e-02 9.84074295e-01
6.13982081e-01 6.61381721e-01 -1.36688843e-01 -4.80517179e-01
1.58936012e+00 -6.31245136e-01 -8.28899503e-01 2.95273662e-01
3.73949528e-01 -1.13672888e+00 1.41269004e+00 7.50738084e-01
-1.29826415e+00 -1.10149944e+00 -1.28278446e+00 2.07904756e-01
-4.28814918e-01 3.18409443e-01 3.59182954e-01 9.01863277e-01
-8.41361403e-01 1.18339026e+00 -2.28372857e-01 -7.66903684e-02
-2.81424940e-01 2.76405066e-01 -1.56449586e-01 6.22942686e-01
-1.41600502e+00 7.59047210e-01 6.74467862e-01 -9.58716795e-02
-2.76098192e-01 -8.39121878e-01 -5.36230683e-01 2.72292584e-01
-2.96870351e-01 -4.06426668e-01 1.09722757e+00 -7.74885297e-01
-1.75931704e+00 5.86727858e-01 3.94418985e-01 -4.81253177e-01
3.55814546e-01 -1.69950724e-02 -5.56383908e-01 3.57466117e-02
-7.67290369e-02 4.94530767e-01 9.61043656e-01 -9.27335024e-01
-3.87767971e-01 -5.99538088e-02 -9.09722373e-02 2.48603731e-01
-3.09985429e-01 -1.33014396e-01 3.55283171e-01 -1.27646005e+00
3.46232913e-02 -7.97654867e-01 2.17296913e-01 -5.74827313e-01
-2.55913347e-01 -5.37303500e-02 5.23812115e-01 -6.77313924e-01
1.80396402e+00 -2.33305168e+00 1.28579170e-01 2.22684771e-01
-4.04939801e-01 2.06946045e-01 -1.27933025e-01 6.98731720e-01
-4.25057739e-01 8.32728222e-02 -1.04124658e-01 2.12925430e-02
1.99823126e-01 -2.17976458e-02 -3.25535446e-01 1.37901545e-01
6.87382892e-02 4.74497199e-01 -6.72429860e-01 -2.50396281e-01
8.32827389e-02 7.24098027e-01 -6.39397919e-01 -8.98635015e-02
1.93677783e-01 3.73792499e-01 6.08435720e-02 2.41811305e-01
3.02146733e-01 3.17337424e-01 -4.60091569e-02 -6.61064088e-01
-3.82172704e-01 8.17852974e-01 -1.47731698e+00 1.70055366e+00
-3.34238380e-01 4.71314847e-01 -4.35971737e-01 -7.77334452e-01
1.19682300e+00 6.33027792e-01 3.75081092e-01 -7.08427310e-01
8.40638578e-02 5.24966240e-01 7.48729944e-01 -7.84532875e-02
7.85670519e-01 -4.80460703e-01 -1.55712426e-01 2.64726162e-01
1.49015039e-01 -3.50498468e-01 4.09774065e-01 -4.64571983e-01
7.39179790e-01 3.17989945e-01 4.46894258e-01 -3.15058529e-01
5.55130541e-01 -4.79459196e-01 3.97980303e-01 5.04419565e-01
3.54771651e-02 7.57642686e-01 6.86652720e-01 -1.48671523e-01
-9.82098937e-01 -8.80971193e-01 -3.50138992e-01 9.86127019e-01
-3.26576084e-01 -8.47566128e-01 -7.77952135e-01 -8.90416950e-02
6.07064031e-02 8.87444735e-01 -5.08621275e-01 -2.20409945e-01
-6.12507164e-01 -6.21232331e-01 1.14743960e+00 2.67459512e-01
2.45749101e-01 -1.38804805e+00 -1.05520296e+00 4.25132632e-01
-1.70115799e-01 -7.11585462e-01 -3.61589909e-01 4.11538273e-01
-9.31322515e-01 -5.71934998e-01 -7.87200570e-01 -4.68387306e-01
-8.31638202e-02 -4.62841958e-01 1.03848910e+00 -6.67565688e-02
-2.76622683e-01 1.54792875e-01 -6.78574026e-01 -3.60153854e-01
-4.22397465e-01 2.00916886e-01 2.07074597e-01 -5.50188012e-02
-1.04814477e-01 -1.07040477e+00 -5.43028295e-01 1.58014029e-01
-7.46920526e-01 -1.75188959e-01 5.93747795e-01 9.50308800e-01
4.74688321e-01 6.01724312e-02 7.99385250e-01 -2.84444541e-01
9.60805714e-01 3.62132825e-02 -1.36905715e-01 -2.16473684e-01
-7.61862516e-01 8.12879354e-02 6.75638735e-01 -6.92573428e-01
-4.77192193e-01 -4.36183244e-01 -2.41583630e-01 -4.02144760e-01
-4.76393066e-02 6.29128814e-01 1.09775923e-01 1.60605252e-01
7.57541001e-01 2.38033846e-01 -1.61403015e-01 -4.74388003e-01
1.96665272e-01 4.61859554e-01 3.71753752e-01 -8.36893797e-01
7.65958905e-01 -2.49602124e-01 9.18769762e-02 -9.75393116e-01
-4.40540403e-01 -3.62696350e-02 -4.55747038e-01 -1.26814470e-01
6.77743435e-01 -5.22665083e-01 -3.22408408e-01 3.87184441e-01
-9.51291978e-01 -2.48070121e-01 -9.33081031e-01 6.73497736e-01
-8.92431200e-01 2.77190328e-01 -6.23518348e-01 -8.33006322e-01
-5.29884398e-01 -9.85828876e-01 8.22199047e-01 1.38879642e-01
-7.54241467e-01 -6.95534706e-01 1.41868070e-01 -2.03709565e-02
4.65352148e-01 4.76482213e-01 1.17746603e+00 -5.56509435e-01
4.15233672e-02 6.07148670e-02 2.61119038e-01 4.45432991e-01
1.84561461e-01 1.52031025e-02 -1.07687986e+00 -4.05723125e-01
3.75356041e-02 -1.59067377e-01 4.98242855e-01 2.87881434e-01
8.23943079e-01 -4.97742593e-01 3.17236423e-01 5.55749476e-01
1.21444881e+00 3.70631903e-01 8.55049074e-01 4.18665588e-01
2.43123502e-01 5.24463356e-01 7.53496468e-01 6.37299776e-01
-1.97403237e-01 1.04450321e+00 7.48277232e-02 1.28173068e-01
-4.26201284e-01 -2.44055018e-01 5.98076940e-01 1.21825778e+00
-5.78263819e-01 4.33328971e-02 -4.97933298e-01 4.68114704e-01
-1.29812646e+00 -1.29679477e+00 -3.50123532e-02 2.52876854e+00
1.11366200e+00 3.02348584e-01 5.23510873e-01 1.08487499e+00
4.31592405e-01 1.96915478e-01 3.94626148e-02 -8.52056086e-01
-2.35159501e-01 7.66245067e-01 3.22354622e-02 3.11367840e-01
-7.28112459e-01 5.79625070e-01 5.51553345e+00 1.28695810e+00
-1.37546277e+00 1.97072923e-02 1.67514920e-01 -3.38996691e-03
-1.65693626e-01 4.78837714e-02 -7.41233945e-01 5.92931151e-01
1.04248595e+00 -2.81828225e-01 4.73324269e-01 5.49995124e-01
1.72860324e-01 3.22827324e-02 -1.18469787e+00 1.12713313e+00
-1.46651193e-02 -1.02039850e+00 2.93935299e-01 1.32785678e-01
5.34087479e-01 -4.86141652e-01 3.29225630e-01 5.23428857e-01
-7.18805075e-01 -1.11018336e+00 1.21033525e+00 4.28589612e-01
1.05743659e+00 -9.28119779e-01 7.14830697e-01 8.27043578e-02
-1.61415219e+00 -1.29808888e-01 -2.50722766e-01 -3.08858424e-01
1.88257873e-01 4.27790791e-01 -7.73138344e-01 8.14338148e-01
4.87173617e-01 2.70229608e-01 -5.09549856e-01 1.03098738e+00
2.23095194e-02 7.16813445e-01 -2.14542776e-01 1.16697386e-01
2.13700160e-01 -3.92637193e-01 7.31543779e-01 1.51993144e+00
8.52205157e-01 -1.58200875e-01 -1.47990599e-01 9.29780066e-01
5.89708567e-01 3.03131670e-01 -4.34795290e-01 -2.41875544e-01
5.42108953e-01 1.18370020e+00 -7.39729047e-01 -7.04303244e-03
-3.47641185e-02 7.24061847e-01 -9.01541635e-02 1.31074995e-01
-8.97154450e-01 -8.27349365e-01 5.61208129e-01 3.92351061e-01
4.40005243e-01 -6.48761094e-02 -3.59222084e-01 -6.72780097e-01
-8.07111859e-02 -1.17353058e+00 2.80175567e-01 -6.72357321e-01
-1.01281762e+00 6.65947735e-01 6.79208785e-02 -1.52521253e+00
-7.07508087e-01 -4.72898245e-01 -6.03945196e-01 1.08103311e+00
-9.80847359e-01 -8.98492932e-01 -4.66127740e-03 4.48740125e-01
4.41137522e-01 -2.30674759e-01 1.21646321e+00 4.32246864e-01
-3.33764628e-02 9.57382143e-01 -2.75515378e-01 -3.68728004e-02
5.19333184e-01 -1.13541877e+00 -1.12155542e-01 3.43971968e-01
5.48277795e-01 5.72818279e-01 8.00575316e-01 -2.13960797e-01
-7.55070210e-01 -6.61394596e-01 1.10486794e+00 -4.79309037e-02
4.81546193e-01 -1.97849631e-01 -8.29283834e-01 6.07719123e-02
2.69125491e-01 -5.44847548e-01 8.81647050e-01 1.49293616e-01
-2.71740526e-01 -2.93836266e-01 -1.00641072e+00 5.61235309e-01
6.36942744e-01 -6.10427141e-01 -8.35412920e-01 -2.77150452e-01
4.41391081e-01 -2.53325135e-01 -1.35473537e+00 5.39899170e-01
9.45392847e-01 -1.43337798e+00 8.42252433e-01 -1.14732251e-01
2.74509937e-01 -4.30065095e-01 -2.21894875e-01 -1.36671782e+00
-4.97471899e-01 -7.42499709e-01 -2.93133203e-02 1.33064497e+00
4.58293438e-01 -4.96018857e-01 4.07693863e-01 -3.89429957e-01
-2.59542286e-01 -8.21104527e-01 -1.26364481e+00 -9.15862381e-01
1.32496759e-01 -3.17412496e-01 4.28952694e-01 9.37007904e-01
2.29752541e-01 5.03904283e-01 -2.80862868e-01 -3.32396388e-01
2.09931716e-01 -3.40523534e-02 5.80955863e-01 -1.43426204e+00
-7.84738719e-01 -9.05758262e-01 -6.18959188e-01 -3.55483353e-01
-3.33278835e-01 -9.24180984e-01 -3.64785403e-01 -1.03642905e+00
-3.65363359e-01 -4.09811199e-01 -5.93085587e-01 3.33835334e-01
1.61966279e-01 5.03400564e-01 6.60657167e-01 1.22360103e-01
3.97189647e-01 4.85011756e-01 1.15967619e+00 1.27128642e-02
-4.21805233e-01 -3.69993560e-02 -3.21772963e-01 6.62602901e-01
8.35716784e-01 -4.78364468e-01 -5.30241430e-01 1.33173048e-01
3.68536860e-01 1.84823856e-01 2.38643706e-01 -1.56089711e+00
-1.83283910e-01 2.08050460e-01 3.01077306e-01 -5.21871924e-01
6.82909966e-01 -4.79017764e-01 5.90043902e-01 6.69500232e-01
-3.29893976e-01 3.17690939e-01 4.32463437e-01 3.37278023e-02
-4.27626133e-01 -5.44470549e-01 7.69094825e-01 5.68537265e-02
-2.43463278e-01 -3.14688742e-01 -1.81571901e-01 1.39577582e-03
5.97672403e-01 -7.23435342e-01 1.43202558e-01 -2.79697478e-01
-9.38989341e-01 -6.91656411e-01 1.05788842e-01 4.27919269e-01
2.88646519e-01 -1.50698602e+00 -7.86882222e-01 3.31477016e-01
-9.23915431e-02 -5.06826639e-01 1.64523557e-01 1.14817357e+00
-2.78412223e-01 3.67909789e-01 -5.21927774e-01 -4.64449823e-01
-1.13138294e+00 2.47033387e-01 2.05570325e-01 -5.69015324e-01
-5.71337163e-01 6.40544891e-01 -8.54230544e-04 -1.57848999e-01
2.95275092e-01 -5.84655344e-01 -4.08925325e-01 5.96104085e-01
2.85466492e-01 6.18355274e-01 1.76230416e-01 -7.40994096e-01
-1.30190685e-01 8.76796007e-01 3.39220524e-01 -4.83369559e-01
9.51553762e-01 2.38683581e-01 4.89555439e-03 1.03657722e+00
8.19017410e-01 2.55758256e-01 -6.90682471e-01 3.55721563e-01
9.32508931e-02 -2.01347604e-01 -2.90761203e-01 -8.61727715e-01
-5.89861393e-01 8.79551888e-01 7.42886543e-01 4.63386089e-01
1.22454655e+00 -5.98183572e-01 3.76923651e-01 -8.26161727e-03
3.90407294e-01 -1.14437115e+00 8.83201696e-03 5.69502652e-01
1.26931047e+00 -2.12596565e-01 -6.33934140e-02 -1.42161474e-01
-5.57780862e-01 1.33063531e+00 2.21152678e-01 -2.10363671e-01
4.50551599e-01 2.50477523e-01 -1.19479515e-01 1.69416681e-01
-4.54417914e-01 -1.71695545e-01 5.31089902e-01 4.18091923e-01
1.00630736e+00 1.00762751e-02 -9.94093120e-01 9.12812412e-01
-1.17156029e+00 -2.03414097e-01 3.87253225e-01 3.30071270e-01
-2.42198631e-01 -1.30105102e+00 -4.94314075e-01 2.49422193e-01
-4.97581303e-01 -2.72779763e-01 -1.62420392e-01 1.14316356e+00
7.04899371e-01 7.46630669e-01 1.17953345e-01 -6.01826131e-01
5.12210429e-01 4.48027909e-01 7.38885641e-01 -4.39838052e-01
-1.18190372e+00 3.69483978e-01 2.37502933e-01 -2.41176456e-01
-3.67320865e-01 -8.14271808e-01 -9.92482305e-01 1.72129236e-02
-3.88516784e-01 4.77595985e-01 5.86304426e-01 5.20974457e-01
1.38051018e-01 7.93204784e-01 4.81144756e-01 -1.02386928e+00
-9.30022776e-01 -1.45617557e+00 -9.79419589e-01 4.06628668e-01
1.26844108e-01 -6.23713374e-01 -5.63447177e-01 -1.23199411e-02] | [15.79648208618164, 5.337507247924805] |
cf3e836e-c7c3-42fc-b548-7ef145102261 | word-segmentation-as-unsupervised | null | null | https://aclanthology.org/2022.acl-long.283 | https://aclanthology.org/2022.acl-long.283.pdf | Word Segmentation as Unsupervised Constituency Parsing | Word identification from continuous input is typically viewed as a segmentation task. Experiments with human adults suggest that familiarity with syntactic structures in their native language also influences word identification in artificial languages; however, the relation between syntactic processing and word identification is yet unclear. This work takes one step forward by exploring a radically different approach of word identification, in which segmentation of a continuous input is viewed as a process isomorphic to unsupervised constituency parsing. Besides formalizing the approach, this study reports simulations of human experiments with DIORA (Drozdov et al., 2020), a neural unsupervised constituency parser. Results show that this model can reproduce human behavior in word identification experiments, suggesting that this is a viable approach to study word identification and its relation to syntactic processing. | ['Raquel Alhama'] | null | null | null | null | acl-2022-5 | ['constituency-parsing'] | ['natural-language-processing'] | [ 2.53571033e-01 4.28760886e-01 -6.13841936e-02 -5.51815450e-01
3.51987518e-02 -8.31193209e-01 4.67404753e-01 4.98412997e-01
-1.02112925e+00 9.43921655e-02 2.64250308e-01 -5.47707617e-01
2.76414454e-01 -7.43081748e-01 -2.73074061e-01 -3.88457119e-01
2.18314677e-01 6.39028370e-01 3.96051705e-02 -1.22037999e-01
4.72621977e-01 4.20444489e-01 -1.28772295e+00 6.49481937e-02
6.16303504e-01 1.19581521e-01 4.14920181e-01 6.45448148e-01
-5.55275440e-01 2.73624510e-01 -5.39113998e-01 -3.11239034e-01
1.49891376e-01 -4.87249881e-01 -1.36198783e+00 -1.48555279e-01
4.50978458e-01 -1.20243274e-01 3.84766370e-01 1.45688367e+00
3.37098449e-01 -2.81698983e-02 6.24055147e-01 -5.25496840e-01
-7.09822059e-01 1.31417680e+00 9.08353925e-02 4.58628476e-01
6.25974953e-01 1.18940115e-01 1.18775022e+00 -8.33334982e-01
4.83338445e-01 1.55840468e+00 5.78367472e-01 7.44723380e-01
-1.69519758e+00 -5.15669167e-01 3.36137295e-01 -5.48889756e-01
-1.27667081e+00 -2.09060848e-01 5.63153028e-01 -5.85180879e-01
1.17588663e+00 3.77179384e-02 8.68587613e-01 9.90674913e-01
9.84916538e-02 7.00991035e-01 1.51058912e+00 -9.94623780e-01
3.34588021e-01 1.69309899e-01 1.33539867e+00 4.88422453e-01
6.76162541e-01 3.29732925e-01 -4.94993240e-01 7.20381588e-02
7.37726748e-01 -7.83741593e-01 1.82319388e-01 1.86092600e-01
-1.07650042e+00 9.73510563e-01 -4.57907431e-02 9.15690422e-01
-3.04293483e-01 -9.85093415e-03 3.61226618e-01 3.60409439e-01
1.64765790e-01 8.83329690e-01 -2.92633414e-01 -6.86250478e-02
-4.66787994e-01 1.81098953e-01 9.01241720e-01 6.90695882e-01
4.54811394e-01 1.74079046e-01 1.43206060e-01 7.44866610e-01
5.83534360e-01 5.51196039e-01 1.02398109e+00 -6.16745114e-01
1.61378562e-01 5.49802065e-01 -2.95239270e-01 -7.34037459e-01
-6.91401064e-01 1.01965412e-01 -3.90435830e-02 2.33266234e-01
9.86691058e-01 -1.54046640e-01 -9.57919061e-01 2.15282559e+00
2.53829267e-02 -6.48850322e-01 1.56635240e-01 7.90321946e-01
7.39992917e-01 5.28570414e-01 1.03678143e+00 -3.18095505e-01
1.83708096e+00 -2.61618167e-01 -6.74665749e-01 -5.56751192e-01
8.37190092e-01 -5.63615561e-01 1.26662397e+00 4.77739900e-01
-1.23698926e+00 -7.10520148e-01 -7.86001921e-01 -1.29934698e-01
-6.68628573e-01 -3.23668450e-01 1.11358392e+00 1.41590893e+00
-1.18766630e+00 2.95058876e-01 -7.25661933e-01 -9.01694357e-01
-1.12428464e-01 6.83786392e-01 -4.22362000e-01 3.39890122e-01
-1.26378429e+00 9.14370537e-01 9.23278213e-01 -8.30866173e-02
-1.79486826e-01 -5.05557135e-02 -1.05145597e+00 -1.26233533e-01
-1.72495935e-02 -4.31534410e-01 1.50369465e+00 -1.54241502e+00
-1.31991446e+00 1.48806965e+00 -3.09741169e-01 -2.44400471e-01
-2.32905552e-01 -1.38286367e-01 -1.81891844e-01 -7.38898590e-02
1.76318571e-01 8.02388251e-01 6.49536014e-01 -1.08158946e+00
-4.05402392e-01 -5.67373991e-01 2.21353229e-02 1.47699192e-01
-2.23642945e-01 6.37128592e-01 3.26313019e-01 -8.38451445e-01
4.70839262e-01 -7.38400638e-01 -1.28313959e-01 -7.53565907e-01
-1.58407792e-01 -6.64036334e-01 -1.59584552e-01 -5.04086673e-01
1.22211778e+00 -1.97411478e+00 -1.19311854e-01 3.17464888e-01
2.53367126e-01 2.55037159e-01 -3.19601983e-01 4.43787307e-01
-4.72009927e-01 6.61095858e-01 -3.58659953e-01 -3.56074452e-01
1.88595489e-01 2.64975458e-01 -1.12237543e-01 3.25783134e-01
1.72223493e-01 1.03462350e+00 -9.22399461e-01 -4.20259506e-01
-1.83187410e-01 3.49858068e-02 -4.36031193e-01 1.15323946e-01
3.55114825e-02 2.17845872e-01 -4.01879013e-01 5.25384903e-01
3.78448665e-01 2.46265113e-01 5.07928073e-01 1.48340210e-01
-5.22780716e-01 5.19456804e-01 -1.20167851e+00 1.11802471e+00
-2.73114473e-01 6.97064936e-01 -1.04354762e-01 -1.01422071e+00
7.09871054e-01 3.37728173e-01 -2.64131516e-01 -7.35462070e-01
7.63090432e-01 1.96433946e-01 8.25920403e-01 -3.58286113e-01
5.07836699e-01 -2.69827574e-01 -4.39806640e-01 7.80151904e-01
2.42062628e-01 -7.33145028e-02 2.57906318e-01 -1.22960191e-02
8.16386938e-01 -1.33289710e-01 6.40633821e-01 -1.06007624e+00
4.19513851e-01 1.61210164e-01 4.18751001e-01 9.64153171e-01
-2.33616471e-01 2.16283470e-01 3.33417028e-01 -5.39797127e-01
-8.59224200e-01 -1.38039303e+00 -3.94038647e-01 1.38413441e+00
7.34587759e-02 -2.56450117e-01 -1.49054015e+00 -2.62994021e-01
-3.56283158e-01 1.14328969e+00 -5.72235882e-01 -9.22885761e-02
-9.29566443e-01 -6.32652521e-01 6.87336087e-01 6.66696012e-01
3.44100483e-02 -1.60463405e+00 -1.03676617e+00 3.40729207e-01
2.92446941e-01 -1.05460322e+00 -2.96862602e-01 4.81002420e-01
-1.12042058e+00 -8.22007120e-01 -3.02199215e-01 -1.33568776e+00
9.94621277e-01 -1.36736169e-01 1.23853946e+00 3.93222660e-01
-3.98654133e-01 7.69501746e-01 -3.49034935e-01 -4.20628637e-01
-7.90643394e-01 1.22133680e-01 4.41553980e-01 -3.70251656e-01
9.71228838e-01 -3.95572811e-01 -1.96839154e-01 -8.93859342e-02
-1.07176030e+00 -9.62682441e-02 3.18112820e-01 6.34828627e-01
1.45388305e-01 -3.93798590e-01 2.88096726e-01 -1.27183735e+00
1.05398715e+00 -3.03032815e-01 -7.70889878e-01 7.72406682e-02
-7.32750356e-01 2.51107782e-01 5.64312637e-01 -7.30043411e-01
-1.09885049e+00 4.42467988e-01 -2.29207218e-01 5.35811186e-01
-7.95275092e-01 2.76756078e-01 -2.50199765e-01 9.44444910e-03
5.38080871e-01 2.37672180e-01 -1.25323132e-01 -4.46974516e-01
3.98343921e-01 4.69482720e-01 4.90605831e-01 -9.00565803e-01
9.22274411e-01 -1.19416259e-01 -4.32435870e-01 -1.18999088e+00
-4.54331636e-01 -3.47236335e-01 -1.10159540e+00 -6.65286407e-02
1.22695649e+00 -6.18281901e-01 -8.44112456e-01 5.90098619e-01
-1.53520620e+00 -4.46036488e-01 -1.61984384e-01 7.47589529e-01
-3.97414774e-01 4.64395255e-01 -6.49442673e-01 -1.02167559e+00
-2.47385368e-01 -9.71688390e-01 6.74412191e-01 3.09058130e-01
-1.02366900e+00 -1.21951687e+00 3.03884327e-01 -2.38394942e-02
1.54514879e-01 -3.82459998e-01 1.25880313e+00 -1.44594479e+00
-8.24370533e-02 1.34810964e-02 1.72318444e-02 3.06876183e-01
-6.38474748e-02 -9.15120021e-02 -1.11293936e+00 9.30185542e-02
4.44820195e-01 -3.40060085e-01 5.63127816e-01 2.07429603e-01
3.36971074e-01 -9.11922082e-02 -4.40319777e-02 3.46873283e-01
1.31039035e+00 4.99731928e-01 2.78794885e-01 1.79634437e-01
5.94420671e-01 1.11433792e+00 3.62819493e-01 -2.12160259e-01
3.84774148e-01 1.49145111e-01 -1.78866416e-01 1.07246719e-01
9.56806093e-02 -1.33623451e-01 4.25546229e-01 9.75018799e-01
9.99160185e-02 -2.42252350e-01 -1.33170891e+00 7.16109335e-01
-1.28594315e+00 -7.16882706e-01 -3.78886670e-01 2.13814449e+00
9.08883452e-01 3.73301506e-01 1.38066798e-01 1.74544960e-01
1.06697631e+00 -5.87705933e-02 3.76749374e-02 -1.07867193e+00
-1.73860654e-01 6.07801855e-01 4.31266189e-01 8.80129516e-01
-9.83559072e-01 1.54903042e+00 7.40673494e+00 2.92660117e-01
-9.57873344e-01 -1.00188017e-01 7.37515092e-01 8.43115866e-01
-3.33856702e-01 1.83357950e-02 -9.22241390e-01 2.89325207e-01
1.06648636e+00 -1.99996635e-01 3.27985972e-01 6.79854870e-01
-1.22446176e-02 -6.11537516e-01 -1.48298383e+00 6.28052115e-01
2.05743074e-01 -4.64467347e-01 1.32433280e-01 -1.45807013e-01
4.04755473e-02 -4.47129041e-01 -9.46231484e-02 1.35950431e-01
3.31627846e-01 -1.17381012e+00 8.08663130e-01 3.24251615e-02
2.75595516e-01 -2.73061872e-01 5.39140701e-01 4.03904855e-01
-1.03657293e+00 1.29473701e-01 -2.00321689e-01 -5.97617209e-01
2.94065833e-01 7.24722818e-03 -5.80254018e-01 -4.54442620e-01
3.39169711e-01 -2.02883169e-01 -8.02812278e-01 6.56940341e-01
-4.75170493e-01 1.02016735e+00 -4.75931823e-01 -4.41396236e-01
1.09418407e-01 -2.28807479e-01 5.16786397e-01 1.55355215e+00
-1.81318939e-01 6.14408195e-01 1.09003007e-01 9.72001910e-01
3.99854034e-01 6.63412094e-01 -8.22454751e-01 -2.76627898e-01
4.11214322e-01 1.09879279e+00 -1.42680383e+00 -2.05728710e-01
-3.77360106e-01 8.65382731e-01 4.07436073e-01 5.84124662e-02
-2.93811888e-01 -5.08646369e-02 4.75101262e-01 1.03248857e-01
-1.48940518e-01 -4.16383415e-01 -7.04743624e-01 -7.63728380e-01
-2.21829742e-01 -7.35045671e-01 1.59333676e-01 -2.76518971e-01
-1.21181226e+00 6.76894188e-01 4.35327590e-01 -5.12619913e-01
-3.12708527e-01 -1.19943714e+00 -6.06886566e-01 1.05838048e+00
-8.50744247e-01 -8.26024115e-01 8.55274424e-02 1.45089656e-01
7.49143243e-01 -6.84105754e-02 1.19279003e+00 -3.67707103e-01
-4.83435988e-01 6.48459971e-01 -6.07402861e-01 5.17746508e-01
1.97185978e-01 -1.49531043e+00 9.70323324e-01 9.10389364e-01
5.14646411e-01 1.21143627e+00 6.57102227e-01 -8.12696517e-01
-1.24116600e+00 -3.50058913e-01 1.32009649e+00 -5.10444760e-01
7.33235896e-01 -6.18820071e-01 -1.00695741e+00 5.16333163e-01
5.37811160e-01 -4.29653019e-01 9.18619931e-01 -2.16313694e-02
-2.40829065e-01 5.21273494e-01 -9.46504056e-01 7.88779914e-01
1.31742561e+00 -5.27553797e-01 -1.49876070e+00 2.61586279e-01
7.84338713e-01 9.57746208e-02 -6.34231806e-01 9.98919085e-02
4.43231910e-01 -8.62163603e-01 7.84249842e-01 -7.02199399e-01
-1.37406150e-02 2.03678608e-01 1.02634557e-01 -1.14812422e+00
-5.10210276e-01 -4.24657255e-01 6.34901702e-01 1.45407712e+00
4.82194096e-01 -9.76820111e-01 4.20881778e-01 9.98073995e-01
1.54502347e-01 1.29206911e-01 -7.10889995e-01 -7.46527970e-01
6.25013471e-01 -6.20332479e-01 3.70446816e-02 8.29466701e-01
3.45442981e-01 7.15841115e-01 5.55037975e-01 -1.97416940e-03
5.69615722e-01 -2.07729444e-01 2.31617421e-01 -1.60061669e+00
-1.69875890e-01 -6.76912546e-01 -2.69256085e-01 -9.26697969e-01
7.22237527e-01 -9.21079934e-01 4.87614721e-01 -9.67513621e-01
-2.09446505e-01 -2.42947787e-01 -2.40016915e-02 1.76920459e-01
-2.17755139e-01 1.66227613e-02 2.26802126e-01 -3.14632878e-02
-7.37062022e-02 -7.53409117e-02 5.43410957e-01 2.35446781e-01
-2.42450893e-01 -1.91839695e-01 -6.11899555e-01 1.31211674e+00
1.01507115e+00 -7.47888863e-01 -2.79894531e-01 -6.30064785e-01
2.22320259e-01 -2.29075298e-01 -4.46826126e-03 -8.35619628e-01
3.33749563e-01 -9.71055701e-02 1.28410891e-01 -1.56245202e-01
-9.49681699e-02 -7.29552269e-01 -2.88336515e-01 6.71201527e-01
-4.43692654e-01 7.46727586e-01 3.33883971e-01 -5.55712953e-02
-4.04883362e-02 -1.03368998e+00 7.29237854e-01 -5.62258363e-01
-9.81562257e-01 -3.67211193e-01 -1.19895422e+00 3.92827421e-01
7.79353082e-01 -5.50665796e-01 5.37851639e-02 1.95297793e-01
-9.06150937e-01 -1.48208156e-01 4.14577186e-01 2.05868632e-01
4.18944925e-01 -8.10211480e-01 -5.06145537e-01 3.73009413e-01
-3.52401473e-02 -2.87172109e-01 -4.48620498e-01 3.49190265e-01
-7.96372771e-01 4.32995677e-01 -1.82657406e-01 -2.55906492e-01
-1.27734578e+00 6.98142648e-01 7.97138736e-02 1.91677064e-01
-4.25069869e-01 1.05548453e+00 4.68333095e-01 -3.87143284e-01
2.70547152e-01 -4.99772400e-01 -5.58130622e-01 3.06222498e-01
4.96632457e-01 -1.33636728e-01 -4.94765103e-01 -1.04469836e+00
-2.98793346e-01 8.36715400e-01 -7.62136877e-02 -3.78439903e-01
7.94108629e-01 -4.09671627e-02 -5.01164794e-01 8.82894516e-01
9.14330542e-01 2.64480025e-01 -2.60086209e-01 -1.82651371e-01
7.14317441e-01 -3.95267121e-02 -3.18860292e-01 -4.95359778e-01
-5.49595833e-01 6.59910917e-01 5.90762854e-01 4.93973315e-01
8.45921695e-01 9.97390971e-02 4.32974368e-01 6.84505880e-01
3.68503481e-01 -1.34687161e+00 -2.97020704e-01 8.30593348e-01
4.35283273e-01 -1.30266917e+00 -4.10443515e-01 -8.32951546e-01
-2.90198207e-01 1.29116821e+00 7.61773050e-01 -3.67206305e-01
5.51423252e-01 4.54302251e-01 2.67695427e-01 -1.55496702e-01
-3.57622206e-01 -4.70100909e-01 -4.93995799e-03 6.73928022e-01
9.17175293e-01 1.59692988e-01 -1.10849750e+00 9.53124404e-01
-5.10736048e-01 -5.55017352e-01 5.22250116e-01 8.75477970e-01
-6.88775778e-01 -1.40575492e+00 -7.00304091e-01 3.57746556e-02
-5.78915954e-01 -6.97626591e-01 -6.31226897e-01 9.04226303e-01
4.08567190e-02 8.15950096e-01 3.34145397e-01 -8.27345997e-02
3.32498044e-01 6.12348080e-01 3.83804321e-01 -1.30222392e+00
-1.18980742e+00 -2.05780029e-01 1.80505663e-01 -3.01658273e-01
-3.84951413e-01 -9.56237316e-01 -1.59914923e+00 2.02481523e-01
-1.39498278e-01 3.11837912e-01 4.70413864e-01 1.05436039e+00
-5.72418630e-01 1.19156964e-01 1.35073140e-01 -6.15901530e-01
-4.98273075e-01 -8.78906071e-01 -5.32264531e-01 4.64267194e-01
-1.19438484e-01 -3.40389043e-01 -4.68679368e-01 2.62789488e-01] | [10.451141357421875, 9.54215145111084] |
55a027a9-3fcb-4aa5-8f6d-1524edeb79e2 | bbw-matching-csv-to-wikidata-via-meta-lookup | null | null | https://drive.google.com/file/d/1b5sdsJfhXGGXP-xNj3xVu3ULH5tV4Nap/view | https://drive.google.com/file/d/1b5sdsJfhXGGXP-xNj3xVu3ULH5tV4Nap/view | bbw: Matching CSV to Wikidata via Meta-lookup | We present our publicly available semantic annotator bbw (boosted by wiki) tested at the second Semantic Web Challenge on Tabular Data to Knowledge Graph Matching (SemTab2020). It annotates a raw CSV-table using the entities, types and properties in Wikidata. Our key ideas are meta-lookup over the SearX metasearch API and contextual matching with at least two features. Avoiding the use of dump files, we kept the storage requirements low, used only up-to-date values in Wikidata and ranked third in the challenge. | ['Irene Schumm', 'Jörg Mechnich', 'Lars Oberländer', 'Jan Kamlah', 'Philipp Zumstein', 'Renat Shigapov'] | 2021-03-01 | null | null | null | null | ['table-annotation', 'table-annotation'] | ['knowledge-base', 'natural-language-processing'] | [-4.89224374e-01 8.55371773e-01 -2.87462294e-01 -1.78865999e-01
-5.92085600e-01 -9.80829954e-01 7.58741617e-01 8.25802207e-01
-4.78357226e-01 9.20814157e-01 6.59589231e-01 -1.27375662e-01
-7.77622521e-01 -1.08256674e+00 -7.83331335e-01 1.83574095e-01
-2.06069313e-02 9.85021710e-01 7.92155325e-01 -5.73507905e-01
-2.14083157e-02 -2.39246503e-01 -1.62350488e+00 6.11377180e-01
7.33821452e-01 1.22826195e+00 -4.97352108e-02 1.18840434e-01
-7.85825074e-01 9.60148096e-01 -1.41598448e-01 -9.52129543e-01
1.94574848e-01 -2.87542604e-02 -1.44326913e+00 -9.44316566e-01
8.69438648e-01 7.47611105e-01 -5.21993876e-01 1.35550284e+00
2.76568145e-01 -5.33859767e-02 8.03922787e-02 -1.55749738e+00
-5.70384800e-01 1.08850431e+00 3.98793697e-01 1.99562553e-02
6.76953077e-01 -4.32247281e-01 1.40825891e+00 -6.28794074e-01
1.66690731e+00 1.15963948e+00 1.09288728e+00 5.92344820e-01
-8.32594633e-01 -2.46755391e-01 -2.45939627e-01 7.29882956e-01
-1.36308241e+00 -3.69293302e-01 -1.43156230e-01 -3.12234223e-01
1.57460749e+00 6.52637541e-01 5.34680963e-01 5.16770780e-01
-3.83412838e-01 1.68017857e-02 9.44890320e-01 -3.53472024e-01
1.98818728e-01 2.03337371e-01 3.12205464e-01 9.89093721e-01
7.03945696e-01 -4.29853618e-01 -8.74388218e-01 -4.57581401e-01
-1.57737538e-01 -6.92995191e-01 -1.28227368e-01 -7.20539868e-01
-1.21112192e+00 1.94336057e-01 4.46319133e-01 9.42906588e-02
-1.45047575e-01 1.97862074e-01 8.86552453e-01 3.63414079e-01
3.40093732e-01 8.24265063e-01 -9.61127043e-01 -4.71320525e-02
-3.01497102e-01 3.79508406e-01 1.11233902e+00 1.34064102e+00
7.45577753e-01 -6.41063511e-01 2.57390797e-01 7.87979662e-01
1.50063574e-01 2.97625571e-01 3.04752380e-01 -9.96875286e-01
8.38972867e-01 1.13028741e+00 2.20561177e-01 -7.85582602e-01
-6.71995997e-01 2.27168836e-02 1.47292614e-01 -5.41892089e-02
5.05839229e-01 2.53212333e-01 -5.62147141e-01 1.39037561e+00
5.62809646e-01 -1.73216999e-01 3.59876692e-01 7.25127518e-01
1.64733493e+00 9.24565792e-02 4.17869836e-01 2.04510987e-01
1.61243832e+00 -8.36959958e-01 -8.88029337e-01 -7.15289414e-02
1.05917013e+00 -4.61800486e-01 6.75548971e-01 2.49436609e-02
-9.49294806e-01 -8.12102780e-02 -1.14794993e+00 -2.36835822e-01
-1.53507423e+00 -7.82294750e-01 9.87653434e-01 3.32373023e-01
-1.07156813e+00 9.15860295e-01 -4.07066107e-01 -9.87270057e-01
1.64036170e-01 -6.91549480e-02 -1.05382061e+00 6.31466210e-02
-1.76940370e+00 1.40716219e+00 1.04537892e+00 -3.97287577e-01
-2.01379627e-01 -1.33472121e+00 -1.02614498e+00 -5.46780899e-02
1.00804794e+00 -6.94420159e-01 8.73489976e-01 -2.13887081e-01
-6.30613446e-01 1.12681949e+00 1.77162394e-01 -7.15201735e-01
2.58067936e-01 6.52197450e-02 -1.14864337e+00 -1.55063691e-02
2.94899315e-01 4.20068949e-01 -3.68306458e-01 -7.42850959e-01
-6.80247307e-01 -5.38740039e-01 2.67191023e-01 4.68569621e-02
-1.79126799e-01 2.01067552e-01 -7.32253194e-01 -2.01331586e-01
2.07307771e-01 -5.22798359e-01 5.98164238e-02 2.90196799e-02
-2.49224648e-01 -3.03831458e-01 5.17632842e-01 -1.03146815e+00
1.39297473e+00 -1.62279320e+00 1.65220544e-01 4.08536255e-01
3.15687239e-01 -4.31943759e-02 1.01699844e-01 9.82567906e-01
-1.30583376e-01 2.98289418e-01 8.90650898e-02 2.42446914e-01
4.26136702e-01 3.86065841e-01 -1.21647812e-01 1.02321014e-01
-4.40043092e-01 1.19321156e+00 -1.28744662e+00 -7.98229158e-01
-7.76852965e-02 1.76589899e-02 -4.23462957e-01 -2.95500338e-01
-6.97778106e-01 -5.85177004e-01 -2.65604526e-01 7.33165383e-01
4.60828692e-01 -2.53887057e-01 8.44265401e-01 -7.17801750e-01
-3.25963169e-01 1.07228720e+00 -1.24489176e+00 2.20259643e+00
-2.77237684e-01 2.96852052e-01 -4.70972024e-02 -4.80511963e-01
6.61599994e-01 2.24027485e-01 4.26679343e-01 -1.07953990e+00
-3.66162449e-01 6.73468053e-01 -5.87848306e-01 -4.00678366e-01
6.51781082e-01 4.11279112e-01 -4.61392164e-01 -1.09260254e-01
6.67502582e-01 1.05044246e-01 5.24773359e-01 6.74569547e-01
1.35783041e+00 5.12069106e-01 5.24748743e-01 -8.82630408e-01
2.86800474e-01 7.44532228e-01 5.67718863e-01 3.40098292e-01
5.22391498e-01 1.82754826e-02 5.85621536e-01 -7.61314034e-01
-1.01219273e+00 -9.08755660e-01 -1.98369518e-01 9.45529819e-01
2.40959644e-01 -1.48339069e+00 -6.02448404e-01 -9.43422318e-01
3.94538492e-01 6.79718614e-01 -6.89953804e-01 6.42456040e-02
-3.11832458e-01 -2.35290438e-01 8.89840543e-01 4.39700663e-01
3.50359887e-01 -9.49286938e-01 -4.17876065e-01 4.89850007e-02
-1.99111298e-01 -1.19332623e+00 -3.64767313e-02 2.61181831e-01
-3.45703512e-01 -1.57608783e+00 2.78769493e-01 -5.50346136e-01
3.65601242e-01 -4.87361580e-01 1.56666803e+00 1.71161264e-01
-4.38156933e-01 3.21141094e-01 -3.53311926e-01 -2.83394873e-01
-3.05189013e-01 1.60429209e-01 -1.04703970e-01 -8.94471228e-01
3.56126636e-01 -2.09359452e-01 -2.23204151e-01 2.10056260e-01
-6.44701779e-01 9.93760377e-02 -3.84373188e-01 5.76683104e-01
4.49980229e-01 -1.48470953e-01 5.36813855e-01 -1.23492658e+00
4.73305099e-02 -6.50578737e-01 -1.18287611e+00 7.78137565e-01
-1.13490391e+00 3.41002703e-01 4.10524577e-01 5.41935265e-01
-8.59828770e-01 2.59266645e-02 -1.05257042e-01 1.86280295e-01
2.34721750e-01 9.65548754e-01 -5.50888002e-01 -1.48758128e-01
5.44055164e-01 -3.36513758e-01 -1.43490836e-01 -9.13259149e-01
8.26213717e-01 3.14438760e-01 1.02754867e+00 -7.58369744e-01
6.68350935e-01 2.99762279e-01 2.41291299e-01 7.26578012e-02
-9.26477730e-01 -4.19307649e-01 -5.93816578e-01 -5.14003150e-02
7.58903742e-01 -1.00985885e+00 -1.12407815e+00 6.12613671e-02
-7.85113394e-01 -2.47651428e-01 -4.46285933e-01 -1.80943578e-01
-3.38987052e-01 3.41670394e-01 -1.42845184e-01 -1.47735879e-01
-6.64131105e-01 -3.50462854e-01 5.20094216e-01 -1.44344419e-01
-3.05759341e-01 -1.27492046e+00 2.70718306e-01 3.32472414e-01
4.52799886e-01 3.02279830e-01 1.17050171e+00 -1.17677355e+00
-3.86610925e-01 -4.66762818e-02 -4.48150814e-01 -3.44238728e-01
-1.14283122e-01 -2.32710928e-01 -6.58093750e-01 1.33163691e-01
-1.17384148e+00 -3.06840330e-01 6.44981980e-01 -6.30739927e-01
9.94427800e-01 -5.63076079e-01 -6.09748244e-01 7.14095473e-01
1.93397915e+00 -1.64183855e-01 5.87334514e-01 1.10694432e+00
6.16015255e-01 8.09244156e-01 6.23414993e-01 1.59046099e-01
8.64769697e-01 1.02758133e+00 7.40448236e-01 3.91238481e-01
-5.44601500e-01 -5.84838748e-01 -4.24340397e-01 6.05200887e-01
8.52179974e-02 -2.29385123e-01 -1.20683670e+00 9.94049668e-01
-2.19216824e+00 -9.98664021e-01 -4.62943882e-01 2.27484751e+00
1.10255575e+00 4.17077616e-02 -4.82609048e-02 -3.82457763e-01
3.98518860e-01 -1.54783428e-01 -1.80535913e-02 -5.19654620e-03
-4.55922842e-01 4.56652790e-01 1.09644198e+00 7.55483270e-01
-7.65204430e-01 1.12096214e+00 6.36282063e+00 7.25963295e-01
-3.97041589e-01 3.40634704e-01 -4.45579588e-01 -1.23823471e-02
-6.98770046e-01 6.50239706e-01 -1.04819524e+00 5.88775754e-01
1.13864911e+00 -5.32002211e-01 6.07908726e-01 8.43879402e-01
-8.72367144e-01 9.81338918e-02 -7.62969136e-01 5.13061404e-01
-2.06832021e-01 -2.24424100e+00 -2.86147371e-02 -1.10093094e-01
2.27553651e-01 5.67321479e-01 -7.65925586e-01 3.65942627e-01
7.68974066e-01 -9.19160247e-01 1.00841200e+00 6.90467358e-01
9.59846437e-01 -4.64418352e-01 6.71016395e-01 -3.54794204e-01
-1.35757911e+00 2.41396129e-01 -3.49538058e-01 4.01004493e-01
5.38906269e-03 2.15925038e-01 -5.21100402e-01 1.28508055e+00
1.04570711e+00 7.13259399e-01 -8.58443916e-01 9.51226771e-01
2.90751103e-02 1.20994292e-01 -4.64703321e-01 1.19687304e-01
-2.70247702e-02 1.00758292e-01 6.00337029e-01 1.23681128e+00
1.91923827e-01 -1.96400195e-01 -8.45891386e-02 3.76334697e-01
-3.36066425e-01 2.37457782e-01 -3.63333374e-01 -4.78899740e-02
1.00667214e+00 1.34533060e+00 -2.85274982e-01 -6.39872551e-01
-3.50617021e-01 6.16470695e-01 5.74329257e-01 -1.56486601e-01
-5.19233465e-01 -7.48625696e-01 1.00277698e+00 3.79152149e-01
1.76821604e-01 3.30901086e-01 -8.63478556e-02 -1.07938516e+00
3.49291563e-02 -6.44861996e-01 1.30384147e+00 -1.08489037e+00
-1.14976144e+00 5.97811162e-01 4.26197082e-01 -7.18466759e-01
-2.33741730e-01 -7.69834399e-01 5.23338169e-02 6.57556117e-01
-1.35078990e+00 -1.24565744e+00 -6.13961101e-01 4.48764682e-01
-1.26798019e-01 -1.20809987e-01 1.14238477e+00 7.17270672e-01
-9.56528112e-02 5.62207282e-01 -8.67597982e-02 7.34043941e-02
8.05162966e-01 -1.46890032e+00 9.85793710e-01 3.93614709e-01
1.14051886e-01 8.03463578e-01 9.27475154e-01 -1.01531875e+00
-1.52641106e+00 -1.01773095e+00 1.37380791e+00 -8.90587389e-01
1.34031284e+00 -4.82042015e-01 -1.05077779e+00 8.44882071e-01
4.08997476e-01 3.72171521e-01 2.61528701e-01 2.70309955e-01
-9.76833820e-01 -1.17423937e-01 -1.22455728e+00 2.33968586e-01
1.66385579e+00 -5.21309793e-01 -6.80412352e-01 4.31814641e-01
8.69772136e-01 -6.92060173e-01 -1.58603978e+00 4.88484681e-01
6.11130536e-01 -2.95237243e-01 9.83373284e-01 -1.12201667e+00
-5.76262996e-02 -6.25259459e-01 -3.62318486e-01 -1.36930501e+00
-7.63994604e-02 -3.92492115e-01 -2.27458671e-01 1.46895885e+00
9.10104871e-01 -8.76727343e-01 5.26271105e-01 7.36014366e-01
-4.61570621e-01 -1.45060763e-01 -9.82525468e-01 -1.09922016e+00
-4.64898020e-01 -1.94102541e-01 1.02050149e+00 1.42783844e+00
8.98525953e-01 -8.40266719e-02 -4.09806445e-02 1.50696188e-01
7.04048216e-01 1.29773673e-02 4.70550865e-01 -1.63063776e+00
-1.54884681e-01 2.34231930e-02 -6.01714551e-01 1.13928020e-01
9.82951820e-02 -1.57461309e+00 -4.36712325e-01 -1.95500588e+00
1.56122476e-01 -5.61129928e-01 -4.73515540e-01 1.23869252e+00
1.94909915e-01 2.40179211e-01 -4.11829576e-02 4.36112890e-03
-1.10586417e+00 -7.16376975e-02 2.70144194e-01 -4.49312925e-02
6.23536110e-01 -1.02640402e+00 -6.80768669e-01 4.10544187e-01
4.01926577e-01 -8.20671022e-01 -1.29350424e-01 -4.16727990e-01
1.26154006e+00 -1.51100755e-01 3.53940785e-01 -9.51796293e-01
5.59515595e-01 -1.55283734e-01 -2.52649039e-01 -6.07745230e-01
1.75544128e-01 -9.12267864e-01 1.05253649e+00 2.66247869e-01
-3.29565406e-01 1.17340170e-01 4.34651464e-01 1.35728151e-01
-7.25400448e-02 -5.05879402e-01 3.67299885e-01 -3.15441221e-01
-1.34374499e+00 -6.17241263e-02 2.18398541e-01 5.53331912e-01
6.59958780e-01 1.82178438e-01 -1.23983538e+00 3.00924271e-01
-7.09600806e-01 5.54720581e-01 8.38019490e-01 6.70048952e-01
-9.06824395e-02 -1.36564827e+00 -3.47621053e-01 -4.37858582e-01
9.70033824e-01 -4.36248332e-01 3.11468076e-02 4.88857001e-01
-7.89113522e-01 8.04310262e-01 -5.15028656e-01 3.04217458e-01
-1.33610082e+00 6.34869874e-01 2.72196740e-01 -2.49649316e-01
-7.11701572e-01 4.85985518e-01 -6.75993204e-01 -9.05269563e-01
7.27020577e-02 9.60337594e-02 -6.40494153e-02 2.33353212e-01
4.25340176e-01 5.23302794e-01 7.67474473e-01 -2.74189532e-01
-9.37886059e-01 2.41997048e-01 3.37201983e-01 1.36728749e-01
1.46074462e+00 8.05653781e-02 -6.31431520e-01 1.48269325e-01
9.02132988e-01 1.76394209e-01 -3.40948284e-01 -3.89120728e-01
8.12115312e-01 -2.01717883e-01 -6.47460818e-02 -1.46398270e+00
-8.31640065e-01 -3.57962400e-02 1.01921342e-01 1.67478651e-01
4.91073132e-01 4.10794944e-01 7.21455872e-01 6.59310341e-01
6.81141496e-01 -1.31499577e+00 -7.43169010e-01 5.16225755e-01
7.60268450e-01 -7.70062625e-01 2.79745162e-01 -7.37077713e-01
-4.04977828e-01 9.02355790e-01 7.43628263e-01 3.32708657e-01
4.19460624e-01 3.06802213e-01 4.49452251e-02 -9.38061476e-01
-1.25035870e+00 -5.53510725e-01 6.05673671e-01 6.64010584e-01
2.97706157e-01 -2.51280874e-01 -5.31955481e-01 6.57946646e-01
-2.24824309e-01 -8.16462263e-02 8.77510607e-02 8.03586841e-01
-2.83381879e-01 -1.26455975e+00 2.76727051e-01 3.71904016e-01
-4.38714415e-01 -4.71501380e-01 -4.67648953e-01 1.12124670e+00
3.91073637e-02 3.36637944e-01 1.28137216e-01 -3.83588314e-01
6.75180793e-01 4.45282042e-01 3.17134857e-01 -5.56656837e-01
-8.70417476e-01 -8.90311003e-01 1.21376932e+00 -9.77592111e-01
-8.61962065e-02 -4.50791627e-01 -1.58699596e+00 -3.35179567e-01
-9.40316916e-02 5.63305378e-01 1.03093660e+00 7.07412899e-01
6.55078351e-01 3.34907085e-01 -3.84710431e-01 3.18510771e-01
-5.60185872e-02 -7.42055178e-01 -2.85746962e-01 4.98102248e-01
-5.25896549e-01 -6.55018985e-01 -1.82336971e-01 -3.24844941e-02] | [9.268011093139648, 8.009982109069824] |
b23b0216-97a8-4816-a6b2-5e8d07e02d35 | modelling-technical-and-biological-effects-in | 2209.06716 | null | https://arxiv.org/abs/2209.06716v2 | https://arxiv.org/pdf/2209.06716v2.pdf | Modelling Technical and Biological Effects in scRNA-seq data with Scalable GPLVMs | Single-cell RNA-seq datasets are growing in size and complexity, enabling the study of cellular composition changes in various biological/clinical contexts. Scalable dimensionality reduction techniques are in need to disentangle biological variation in them, while accounting for technical and biological confounders. In this work, we extend a popular approach for probabilistic non-linear dimensionality reduction, the Gaussian process latent variable model, to scale to massive single-cell datasets while explicitly accounting for technical and biological confounders. The key idea is to use an augmented kernel which preserves the factorisability of the lower bound allowing for fast stochastic variational inference. We demonstrate its ability to reconstruct latent signatures of innate immunity recovered in Kumasaka et al. (2021) with 9x lower training time. We further analyze a COVID dataset and demonstrate across a cohort of 130 individuals, that this framework enables data integration while capturing interpretable signatures of infection. Specifically, we explore COVID severity as a latent dimension to refine patient stratification and capture disease-specific gene expression. | ['Neil D. Lawrence', 'Sarah A. Teichmann', 'Shaista Madad', 'Rik G. H. Lindeboom', 'Dinithi Sumanaweera', 'Natsuhiko Kumasaka', 'Emma Dann', 'Aditya Ravuri', 'Vidhi Lalchand'] | 2022-09-14 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [ 3.96023780e-01 -3.70032728e-01 -8.81800354e-02 -2.21455559e-01
-9.10043955e-01 -8.16483736e-01 5.99431753e-01 1.88975096e-01
-1.64590001e-01 8.97391498e-01 6.13352060e-01 -3.72796599e-03
-4.64704096e-01 -6.07617736e-01 -2.35407382e-01 -1.16838408e+00
-5.12268692e-02 1.05528247e+00 -4.94995415e-01 3.84028673e-01
-1.24288127e-01 5.58005154e-01 -1.22641075e+00 1.16388261e-01
6.72848880e-01 3.27534795e-01 -2.69391835e-01 1.00453675e+00
1.67999983e-01 -8.64631981e-02 -1.54182076e-01 -9.39446241e-02
4.30364273e-02 -1.55753508e-01 -1.76677749e-01 -1.56991094e-01
2.02999175e-01 -1.24168225e-01 -9.75911170e-02 6.86090350e-01
7.74059534e-01 -2.98434973e-01 1.13819420e+00 -1.09093750e+00
-3.08986068e-01 1.34362265e-01 -4.17665869e-01 -5.30662835e-02
-1.95657521e-01 4.30563986e-01 8.14294994e-01 -6.57642603e-01
9.25558150e-01 1.36524665e+00 8.44172895e-01 7.88224459e-01
-2.15287995e+00 -3.01277637e-01 -3.38243395e-01 -3.56520265e-01
-1.45586109e+00 -6.29799366e-01 2.72997797e-01 -1.02697814e+00
8.54896903e-01 5.32259941e-01 4.19864893e-01 1.39638746e+00
3.17611098e-01 3.24577600e-01 1.18498504e+00 -5.12734540e-02
4.18542564e-01 -2.43156210e-01 3.32597733e-01 5.79816341e-01
4.14846718e-01 1.39630243e-01 -4.14501160e-01 -9.89763141e-01
5.78652203e-01 6.80087984e-01 -6.57724887e-02 -6.38612926e-01
-1.37055326e+00 7.57206917e-01 -5.93398392e-01 -1.64879590e-01
-5.40934086e-01 3.47145677e-01 5.22377253e-01 -2.98280239e-01
6.76170528e-01 2.00791270e-01 -8.68688107e-01 -3.09727132e-01
-9.18327868e-01 3.10042053e-01 6.62108243e-01 3.65919232e-01
4.48604494e-01 -2.55154520e-01 -5.99626362e-01 7.95212626e-01
4.14943814e-01 8.87452543e-01 5.40622361e-02 -1.21335471e+00
-7.43955448e-02 5.76008320e-01 1.54641688e-01 -6.42869115e-01
-6.93131685e-01 -1.02136709e-01 -1.27471042e+00 -7.95956478e-02
4.49070573e-01 -3.05998415e-01 -8.45607698e-01 1.89653480e+00
5.93183458e-01 3.86376023e-01 -9.83132720e-02 5.02438545e-01
1.52113780e-01 2.76387036e-01 4.21051562e-01 -5.10441124e-01
1.79305816e+00 -5.70119582e-02 -5.28140843e-01 3.65758717e-01
5.29936790e-01 -2.65367359e-01 7.64178753e-01 5.12940288e-01
-6.73485696e-01 2.89491117e-02 -4.37135726e-01 -2.04055369e-01
-1.70459375e-01 3.77299264e-02 7.79549420e-01 7.25128770e-01
-9.04572666e-01 5.60127735e-01 -1.24801052e+00 -4.54756200e-01
5.39258897e-01 3.98014665e-01 -3.78975302e-01 -1.82348460e-01
-9.16603267e-01 3.32635880e-01 -4.45801318e-02 7.51542822e-02
-1.04306185e+00 -1.38516390e+00 -5.26372433e-01 1.65925041e-01
-8.55899137e-03 -1.29235148e+00 5.33499479e-01 2.38352969e-01
-1.42597198e+00 8.78492773e-01 -5.45172751e-01 -1.50150239e-01
3.32889557e-01 3.01644560e-02 -1.58379644e-01 1.27477460e-02
-6.92252964e-02 2.70663142e-01 7.31601655e-01 -8.59955490e-01
4.72596474e-02 -8.40517759e-01 -6.33368254e-01 -2.22755849e-01
-1.64576098e-01 -1.25824124e-01 -2.89125174e-01 -4.94332105e-01
6.09291382e-02 -1.19382250e+00 -3.93625945e-01 2.10743949e-01
-4.25887316e-01 9.99286845e-02 3.43108356e-01 -6.99133158e-01
9.30320561e-01 -2.10363674e+00 6.11288130e-01 8.95865038e-02
7.05244243e-01 3.74230705e-02 -1.94410890e-01 4.73627567e-01
4.93166409e-02 4.17006105e-01 -4.33725834e-01 -7.03105688e-01
2.33428210e-01 3.32536608e-01 -2.36244157e-01 8.99051070e-01
3.32191736e-01 1.15637541e+00 -7.49516785e-01 -2.36622274e-01
1.16719566e-01 1.05210280e+00 -8.75718117e-01 -3.92890088e-02
-2.99473286e-01 6.07029200e-01 -4.25838083e-01 6.01255357e-01
7.77990758e-01 -4.44390684e-01 6.26752019e-01 -2.17335314e-01
1.54013967e-03 -1.47572041e-01 -1.07365143e+00 1.49372065e+00
-1.34831324e-01 3.25841188e-01 5.73840201e-01 -6.41529739e-01
5.31280875e-01 2.39230990e-01 7.87195265e-01 9.65752527e-02
1.87343836e-01 -3.25342298e-01 -3.13824981e-01 -4.13408965e-01
-9.19160098e-02 -7.57025778e-01 -1.86786607e-01 3.55963916e-01
-8.27142596e-02 1.06350377e-01 -7.89926127e-02 4.66061719e-02
1.29866087e+00 -3.97239290e-02 2.79080898e-01 -6.08651340e-01
3.69008839e-01 -1.58960506e-01 8.22234809e-01 5.64111769e-01
-3.59927148e-01 6.90575123e-01 9.03032064e-01 -1.82169184e-01
-1.09718335e+00 -1.62402809e+00 -5.27905524e-01 8.11483562e-01
-6.40269399e-01 -2.79321194e-01 -3.63232523e-01 -3.15645814e-01
4.16875213e-01 3.22217643e-01 -9.01785672e-01 4.25749552e-03
-7.23213851e-02 -1.70611000e+00 8.63239586e-01 1.50307819e-01
-4.99694586e-01 -2.87430644e-01 -7.57798702e-02 4.38181758e-02
-6.82924613e-02 -6.51298702e-01 -3.02491874e-01 1.55853182e-01
-1.07570839e+00 -1.29579496e+00 -4.56761420e-01 4.50369976e-02
6.43873453e-01 -2.47565359e-01 8.09730351e-01 -2.29662731e-01
-9.24077213e-01 4.70241070e-01 2.73501962e-01 -3.36881548e-01
-3.69212419e-01 -1.69238627e-01 6.27103150e-01 -1.57461166e-01
4.50674325e-01 -6.04071736e-01 -8.01797032e-01 -8.81840289e-02
-9.39458370e-01 -1.43354684e-01 3.49536657e-01 9.58743513e-01
9.22318876e-01 -2.62187541e-01 4.26958650e-01 -1.07905293e+00
4.14787143e-01 -7.94979990e-01 -6.29577160e-01 1.30496681e-01
-7.75807202e-01 3.39048266e-01 3.16328049e-01 -2.34973669e-01
-1.01225507e+00 -2.95249950e-02 7.89846852e-03 -2.66141772e-01
-3.27543139e-01 2.12529972e-01 -1.91294074e-01 5.08774459e-01
4.09907550e-01 2.07999036e-01 4.39064622e-01 -6.58208013e-01
5.04808843e-01 5.94381392e-01 2.29777098e-01 -5.97194672e-01
5.01552403e-01 1.07204342e+00 8.01197410e-01 -8.21858883e-01
-3.62127602e-01 -6.14218295e-01 -8.47580612e-01 4.63503778e-01
9.40397620e-01 -1.13831377e+00 -1.23904490e+00 5.37339687e-01
-7.67003953e-01 -4.09752876e-01 -2.16082066e-01 6.14412844e-01
-6.48810446e-01 3.87719780e-01 -1.07677674e+00 -7.56684005e-01
-3.90091568e-01 -1.03123081e+00 1.52527988e+00 -3.47943097e-01
-3.93204480e-01 -1.24886906e+00 9.41931307e-01 3.15324157e-01
1.79406151e-01 5.67686200e-01 1.20917463e+00 -6.06461048e-01
-4.19165373e-01 9.49330907e-03 -2.64156878e-01 1.71562284e-02
3.14889431e-01 2.64239848e-01 -9.42362487e-01 -2.77630240e-01
-9.07747075e-02 1.20931558e-01 7.89746106e-01 7.01951623e-01
7.91927278e-01 -1.69906110e-01 -5.60631275e-01 8.48263025e-01
1.48076403e+00 -5.44041038e-01 4.93760288e-01 -5.18006682e-01
7.65636444e-01 6.59269571e-01 3.07719857e-02 7.01699376e-01
2.07374617e-01 5.96349597e-01 -1.87025353e-01 7.04184175e-02
2.63767503e-02 -3.62476408e-02 1.58755153e-01 7.91283548e-01
2.33073995e-01 -2.05311939e-01 -1.05944395e+00 4.98778224e-01
-1.55224133e+00 -8.34914684e-01 -6.21558249e-01 2.05576706e+00
1.02108848e+00 -4.38861012e-01 1.19158514e-01 -3.96689028e-01
4.99398023e-01 -2.18694001e-01 -6.87036693e-01 -2.01428503e-01
-2.59742618e-01 1.60543740e-01 4.50026900e-01 6.84358776e-01
-6.34709537e-01 3.40062827e-01 7.35216141e+00 5.93735278e-01
-7.18863368e-01 2.01836646e-01 6.89913869e-01 -4.64859098e-01
-4.27484840e-01 -1.66999176e-01 -9.23127890e-01 4.43011105e-01
1.25328374e+00 9.48213320e-03 4.35996324e-01 3.74635369e-01
6.45668805e-01 1.18066976e-02 -1.15159249e+00 7.68492699e-01
-1.24828622e-01 -1.31485856e+00 -6.85953647e-02 5.07547617e-01
8.45136523e-01 1.70616597e-01 1.98425099e-01 1.08840719e-01
4.35231626e-01 -1.07266772e+00 -2.89812744e-01 1.09740281e+00
1.06537342e+00 -4.85403180e-01 5.90404391e-01 2.91053355e-01
-5.90664089e-01 2.71745920e-01 -2.49948055e-01 -1.46488883e-02
3.42693925e-01 1.10862982e+00 -1.12371254e+00 2.20098719e-01
3.03833395e-01 3.98463875e-01 -2.74732500e-01 4.97372210e-01
1.62907660e-01 8.52286518e-01 -4.47071463e-01 2.51042008e-01
-5.89572728e-01 -5.30906737e-01 6.53197944e-01 1.22620165e+00
3.31491649e-01 2.27806821e-01 -2.95621544e-01 9.89678323e-01
8.16925019e-02 -1.03917718e-01 -2.88661540e-01 -3.37825865e-01
2.82637030e-01 1.35684443e+00 -6.25526965e-01 -3.21798205e-01
-1.30420178e-01 9.87269402e-01 2.24639773e-01 3.75257820e-01
-5.17313063e-01 3.14865649e-01 1.64340734e+00 1.29307881e-01
2.90502816e-01 -4.04558957e-01 -4.73271906e-01 -1.31620061e+00
-5.38518250e-01 -6.71014845e-01 4.03373629e-01 -3.47961605e-01
-1.74330449e+00 -2.64121920e-01 5.65675274e-03 -5.42438149e-01
-2.91887164e-01 -8.04021955e-01 -9.30854380e-02 1.28734398e+00
-1.21503663e+00 -1.10356677e+00 2.21133679e-02 9.56484601e-02
6.43159845e-04 2.41110891e-01 1.15681982e+00 2.38074660e-01
-8.79907191e-01 1.63476825e-01 8.29793811e-01 -3.15389752e-01
7.91909754e-01 -1.15177989e+00 2.84879565e-01 5.60456276e-01
-3.21896344e-01 1.24024653e+00 9.13256466e-01 -9.88888621e-01
-1.77846003e+00 -1.22240973e+00 7.23681867e-01 -1.13585377e+00
6.76674187e-01 -7.27736175e-01 -8.47461045e-01 5.62209308e-01
-5.14694393e-01 -5.77268749e-02 1.38333082e+00 3.80461365e-01
-5.27740598e-01 2.50265338e-02 -1.41259217e+00 6.41105413e-01
9.93408680e-01 -7.84118950e-01 -2.48812139e-01 2.38649130e-01
5.50246954e-01 1.31952256e-01 -1.23636878e+00 3.13060790e-01
6.99342012e-01 -5.96594512e-01 1.13969088e+00 -9.93618548e-01
9.05276090e-02 -4.65895116e-01 -2.24344373e-01 -1.06691706e+00
-5.74249744e-01 -6.48237169e-01 -3.86381924e-01 1.29763687e+00
-1.79123767e-02 -5.25367081e-01 9.91906881e-01 9.50444341e-01
4.29562300e-01 -5.33929646e-01 -1.02823591e+00 -4.31683153e-01
2.18201250e-01 -3.14383775e-01 4.99973923e-01 9.15513158e-01
-5.15848510e-02 8.65276158e-02 -4.07996833e-01 2.68684089e-01
9.88207042e-01 -8.46175924e-02 7.31059849e-01 -1.65166652e+00
-4.48660374e-01 -2.01202095e-01 -2.84953743e-01 -5.21553993e-01
1.26556084e-01 -8.79574895e-01 -2.50286125e-02 -1.30906200e+00
8.71748090e-01 -3.75605524e-01 -3.78354847e-01 1.83672786e-01
-4.07643437e-01 3.70938927e-01 -2.55068868e-01 2.48371344e-02
-3.99999022e-01 5.10409236e-01 7.30911732e-01 4.03969809e-02
-2.04548523e-01 -4.09291595e-01 -7.45762825e-01 3.03243518e-01
6.95980906e-01 -5.85880339e-01 -2.69214958e-01 -3.83258350e-02
2.63209730e-01 -7.82417785e-03 5.91779530e-01 -4.40504789e-01
-1.04863033e-01 -4.23732907e-01 5.75623333e-01 -5.02177477e-01
3.88749093e-01 -4.86038089e-01 8.89129639e-01 6.94962919e-01
-1.60382479e-01 -2.35270724e-01 3.40636522e-01 1.01182306e+00
5.44037044e-01 2.72921294e-01 5.49911857e-01 2.17955455e-01
2.22071037e-02 4.61642385e-01 -6.42300189e-01 2.16682315e-01
7.97105551e-01 2.60981530e-01 -4.35462952e-01 1.68622151e-01
-8.28108191e-01 1.57223329e-01 6.43581152e-01 5.95015213e-02
3.31122309e-01 -1.03078008e+00 -9.97045159e-01 3.10345262e-01
1.62680924e-01 -3.01467180e-01 8.78651023e-01 1.07787693e+00
-4.00334239e-01 5.96050620e-01 1.19544711e-04 -8.29216659e-01
-1.32446802e+00 7.11483955e-01 8.59184861e-02 -4.08380270e-01
-1.96739644e-01 6.86641634e-01 3.28020006e-01 -6.84546530e-01
-4.21969295e-01 -2.44733870e-01 2.31259704e-01 3.64116400e-01
5.59760451e-01 6.72132909e-01 -1.94337323e-01 -3.48343879e-01
-5.25788844e-01 4.64500666e-01 2.05401555e-01 -3.63068916e-02
1.54048932e+00 -4.64632213e-01 -5.81643224e-01 6.94398999e-01
1.41470230e+00 3.63355935e-01 -1.31567144e+00 5.14965169e-02
-1.91183146e-02 -1.98179826e-01 -1.02297634e-01 -7.02102721e-01
-4.37861353e-01 8.48745883e-01 8.07843804e-01 -3.12413573e-01
8.23415399e-01 2.79648099e-02 4.13955420e-01 -8.11042148e-04
1.33383214e-01 -4.93492603e-01 -5.57179928e-01 2.24978596e-01
4.19524431e-01 -8.87450278e-01 2.01941878e-01 -3.59756082e-01
-1.08252011e-01 7.51649201e-01 -9.77012813e-02 7.39810318e-02
6.84046745e-01 5.35606742e-01 -6.58074245e-02 -3.85191649e-01
-1.16448343e+00 1.91361219e-01 2.12945148e-01 8.92196715e-01
3.01137626e-01 3.34360689e-01 -2.30544016e-01 7.62463868e-01
2.28494868e-01 2.19042718e-01 4.12765175e-01 3.37692440e-01
7.45887961e-03 -1.24770880e+00 -2.44614244e-01 7.57357597e-01
-7.04772472e-01 -3.02234530e-01 -5.19754700e-02 2.32058361e-01
1.23102084e-01 5.86894929e-01 1.82608172e-01 1.00556545e-01
1.04136057e-01 5.31542957e-01 3.34252715e-01 -7.06649482e-01
-1.47866562e-01 1.91760898e-01 -1.11815974e-01 -6.18811905e-01
-9.40029174e-02 -1.26273072e+00 -1.01240480e+00 -3.57701212e-01
1.27155378e-01 -1.19867779e-01 6.15570545e-01 6.92681253e-01
1.00137115e+00 4.29516315e-01 1.94899365e-01 -4.53652412e-01
-6.35344028e-01 -7.62187779e-01 -8.83453965e-01 3.85914385e-01
6.67631209e-01 -5.37361324e-01 -4.64732140e-01 2.85974920e-01] | [6.720701694488525, 5.210686206817627] |
c7253f9b-cb71-49fd-a4b7-8cb27726a471 | maximum-bayes-smatch-ensemble-distillation | 2112.07790 | null | https://arxiv.org/abs/2112.07790v2 | https://arxiv.org/pdf/2112.07790v2.pdf | Maximum Bayes Smatch Ensemble Distillation for AMR Parsing | AMR parsing has experienced an unprecendented increase in performance in the last three years, due to a mixture of effects including architecture improvements and transfer learning. Self-learning techniques have also played a role in pushing performance forward. However, for most recent high performant parsers, the effect of self-learning and silver data augmentation seems to be fading. In this paper we propose to overcome this diminishing returns of silver data by combining Smatch-based ensembling techniques with ensemble distillation. In an extensive experimental setup, we push single model English parser performance to a new state-of-the-art, 85.9 (AMR2.0) and 84.3 (AMR3.0), and return to substantial gains from silver data augmentation. We also attain a new state-of-the-art for cross-lingual AMR parsing for Chinese, German, Italian and Spanish. Finally we explore the impact of the proposed technique on domain adaptation, and show that it can produce gains rivaling those of human annotated data for QALD-9 and achieve a new state-of-the-art for BioAMR. | ['Salim Roukos', 'Radu Florian', 'Tahira Naseem', 'Thanh Lam Hoang', 'Ramon Fernandez Astudillo', 'Young-suk Lee'] | 2021-12-14 | null | https://aclanthology.org/2022.naacl-main.393 | https://aclanthology.org/2022.naacl-main.393.pdf | naacl-2022-7 | ['self-learning'] | ['natural-language-processing'] | [ 1.35095537e-01 5.20705044e-01 5.94736896e-02 -5.39343059e-01
-1.38769853e+00 -6.12332582e-01 5.77203989e-01 1.85292691e-01
-7.37950623e-01 7.66685307e-01 2.08377406e-01 -5.47586083e-01
2.93532401e-01 -5.16167164e-01 -7.92317390e-01 -4.64032888e-01
-1.81280170e-02 7.74942458e-01 1.83913305e-01 -5.79400480e-01
-2.03637853e-01 1.29947424e-01 -1.05082476e+00 5.00252068e-01
9.35161054e-01 6.03461564e-01 1.81651324e-01 8.46819818e-01
-5.22821963e-01 8.60416472e-01 -6.40622139e-01 -8.31146300e-01
4.16251756e-02 -2.72063583e-01 -1.05531061e+00 -4.63698536e-01
4.01924521e-01 4.73200418e-02 -8.20515305e-02 8.35776985e-01
7.67849684e-01 -2.13191837e-01 3.04504901e-01 -5.99735856e-01
-8.37020218e-01 1.25146878e+00 -7.27658272e-01 7.80737698e-02
2.15699121e-01 8.09557661e-02 1.18034875e+00 -6.27229333e-01
6.06538177e-01 1.30589700e+00 6.73940241e-01 1.00388765e+00
-1.17584562e+00 -6.83630228e-01 2.46309310e-01 -6.66451156e-02
-8.91348124e-01 -6.38132155e-01 5.92733204e-01 -1.18073680e-01
1.48088443e+00 -9.21077561e-03 1.35985821e-01 1.15633655e+00
-6.61328584e-02 9.69044924e-01 1.20979190e+00 -9.09044683e-01
2.69868642e-01 3.24949920e-02 1.38994247e-01 5.71356952e-01
1.07836686e-01 -5.24549298e-02 -4.75932926e-01 7.74161518e-02
3.65187466e-01 -7.34543443e-01 1.50771409e-01 8.07129219e-02
-9.78235245e-01 8.83101106e-01 4.28389341e-01 3.75501722e-01
-3.00802350e-01 -1.24094216e-02 6.76816165e-01 4.32991624e-01
7.88918793e-01 6.92986369e-01 -1.11452150e+00 -4.34645176e-01
-7.76693642e-01 5.08119613e-02 7.70657778e-01 8.03087413e-01
4.37282085e-01 2.17059001e-01 1.08859755e-01 1.07418823e+00
1.13896199e-01 4.52798009e-01 7.37081587e-01 -7.66407073e-01
9.45683241e-01 3.78092200e-01 -1.30424902e-01 -6.07852787e-02
-6.03236318e-01 -3.93397897e-01 -7.07671881e-01 1.66839004e-01
7.69783437e-01 -3.88350040e-01 -1.23334265e+00 2.00504136e+00
1.47119910e-01 -4.40837830e-01 5.38329661e-01 4.92847204e-01
8.55931997e-01 7.20046878e-01 5.71971595e-01 -2.08663866e-02
1.50032520e+00 -9.76364911e-01 -5.81667662e-01 -6.69495523e-01
1.01304924e+00 -7.45196462e-01 1.16895211e+00 5.49206674e-01
-1.28412318e+00 -7.79226720e-01 -9.74903882e-01 -1.37094632e-01
-3.94856364e-01 1.19526416e-01 7.05869734e-01 8.15010309e-01
-9.87500131e-01 5.38052797e-01 -1.10506225e+00 -4.60723817e-01
3.66789669e-01 4.03054297e-01 -5.67328930e-01 -9.43680331e-02
-1.18523681e+00 1.14221501e+00 6.32641971e-01 -8.79214928e-02
-4.91850257e-01 -8.18920910e-01 -7.83712327e-01 -2.30699673e-01
1.68959141e-01 -6.34524703e-01 1.56899571e+00 -8.96588683e-01
-1.82296348e+00 1.19017148e+00 1.43946335e-01 -8.45535874e-01
5.09155452e-01 -7.47414827e-01 -6.86621726e-01 -2.55399495e-01
-6.96822181e-02 9.78251755e-01 2.07836255e-01 -9.55553949e-01
-7.62698293e-01 -6.30379677e-01 -4.00897339e-02 1.28345847e-01
-2.51460224e-01 2.64009506e-01 -1.70759454e-01 -4.61369634e-01
2.04613507e-02 -9.54920352e-01 -4.15921539e-01 -7.86600471e-01
-8.14658776e-02 -3.75743330e-01 3.23549449e-01 -9.72300947e-01
1.03898346e+00 -2.00489902e+00 1.85583696e-01 -5.45737147e-01
-7.27777854e-02 6.81440234e-01 -4.07147259e-01 4.15899724e-01
-1.68518230e-01 1.35006264e-01 -4.49743390e-01 -6.44367516e-01
-2.55054712e-01 3.80615115e-01 1.27080071e-03 1.38463095e-01
8.02869081e-01 9.41231072e-01 -7.98512876e-01 -2.93861955e-01
7.29978383e-02 4.11157876e-01 -5.75023830e-01 4.02162582e-01
-3.91060650e-01 6.28554344e-01 -2.70090997e-01 4.23704296e-01
6.59553051e-01 2.32093692e-01 2.52996027e-01 1.81499615e-01
-2.90439487e-01 6.38018548e-01 -8.09128761e-01 2.25293589e+00
-7.07825661e-01 2.31083795e-01 1.33173510e-01 -9.86724615e-01
1.12676799e+00 2.72987843e-01 1.62369683e-01 -9.45378780e-01
9.68421809e-03 4.98915076e-01 4.90358740e-01 -2.75625169e-01
6.22510135e-01 -1.47556826e-01 -3.45752418e-01 -6.25512563e-03
4.54467177e-01 -1.49978593e-01 9.25103500e-02 1.16050780e-01
1.22251475e+00 5.40287316e-01 3.96979183e-01 -3.81682009e-01
5.25596559e-01 1.53676733e-01 4.68779534e-01 7.02650428e-01
-1.02321044e-01 4.73993063e-01 3.37654084e-01 -4.23355043e-01
-1.21783876e+00 -1.00076246e+00 -1.47698492e-01 1.49523330e+00
-6.10006273e-01 -2.62725741e-01 -1.04718924e+00 -8.29123616e-01
-2.97256380e-01 8.29479456e-01 -5.27884066e-01 -1.46709740e-01
-1.20650065e+00 -1.18075311e+00 9.28052902e-01 8.05735528e-01
6.09429777e-01 -1.26220834e+00 -4.77161556e-01 5.55743814e-01
-1.39422059e-01 -1.51930845e+00 2.36587748e-01 7.36638844e-01
-1.06529403e+00 -4.51612204e-01 -7.89882600e-01 -8.72285366e-01
1.11978315e-01 -3.53065550e-01 1.62508726e+00 -3.23811889e-01
-2.40575913e-02 -1.47946691e-02 -7.30549812e-01 -5.19726396e-01
-1.02887917e+00 6.99607372e-01 -1.11629084e-01 -5.31395733e-01
2.93428272e-01 -4.22569424e-01 -2.72177398e-01 -8.00270811e-02
-5.41039884e-01 -4.19550808e-03 9.30068672e-01 8.60257924e-01
3.82383943e-01 -3.12691748e-01 8.90933037e-01 -1.33191288e+00
3.67437243e-01 -3.55641276e-01 -4.56153899e-01 5.65951131e-03
-5.54492056e-01 3.84663880e-01 7.77210295e-01 -3.45845312e-01
-1.43647075e+00 3.10265452e-01 -7.21573353e-01 3.11637431e-01
-3.96994680e-01 2.38242820e-01 -4.50444072e-01 4.32447016e-01
8.25637758e-01 -2.19481185e-01 -1.68413043e-01 -8.35296333e-01
9.96326208e-01 8.74437571e-01 8.53897929e-01 -8.19498777e-01
4.30829108e-01 2.93767806e-02 -3.48984212e-01 -7.71587074e-01
-9.82113242e-01 -3.43284816e-01 -6.81371629e-01 4.91679639e-01
1.17070341e+00 -1.15247548e+00 -2.64007419e-01 5.57925105e-01
-1.16551459e+00 -5.43662846e-01 -4.38602149e-01 2.01743469e-01
-4.44129318e-01 2.44471774e-01 -9.51258600e-01 -8.17005157e-01
-7.34849274e-01 -9.65016723e-01 1.03401577e+00 1.16810881e-01
-1.09450772e-01 -8.12897325e-01 3.10398251e-01 4.52612728e-01
4.75991398e-01 -6.87306002e-02 8.95202219e-01 -1.07500970e+00
-1.43022276e-02 -1.49970934e-01 -1.81691855e-01 5.81134856e-01
-9.01462808e-02 -4.13622797e-01 -1.28121400e+00 -2.49762714e-01
-2.11815059e-01 -3.60514432e-01 9.37101305e-01 1.80495411e-01
6.43587232e-01 4.17502224e-02 -4.93747294e-02 4.31880534e-01
1.22687447e+00 2.41934150e-01 7.18849063e-01 5.95487654e-01
5.79466105e-01 5.05523622e-01 5.26604414e-01 -3.35278884e-02
5.69458544e-01 5.85823298e-01 4.41645414e-01 -2.79012173e-01
-4.14262056e-01 -1.94982469e-01 4.18253630e-01 1.14313507e+00
-1.74660608e-01 -2.64975429e-01 -1.17115223e+00 5.23452759e-01
-1.65346479e+00 -4.49243367e-01 -2.38850862e-01 2.12824821e+00
9.72721040e-01 5.32314897e-01 2.81340510e-01 4.48672324e-02
4.90462869e-01 -4.82028611e-02 -2.70442128e-01 -9.28650379e-01
-3.25476199e-01 7.02621341e-01 5.81725061e-01 3.41440916e-01
-1.26668465e+00 1.63280785e+00 6.32851505e+00 5.78718245e-01
-8.71343195e-01 3.22286069e-01 7.70717978e-01 4.74307716e-01
4.55283597e-02 -1.16937794e-01 -1.04270685e+00 2.20613614e-01
1.59242570e+00 4.00543004e-01 2.03220874e-01 9.56115246e-01
-3.45851839e-01 6.69919094e-03 -1.02546942e+00 5.87550640e-01
-1.02449991e-01 -9.42033768e-01 -2.22710758e-01 -5.37657589e-02
5.98048627e-01 5.98440945e-01 -6.66174889e-02 8.90614569e-01
7.38118052e-01 -9.70958769e-01 5.07675350e-01 -1.58792019e-01
7.27041185e-01 -7.92182028e-01 1.01806748e+00 3.43308836e-01
-9.60513473e-01 -8.08086991e-02 -5.82447708e-01 -2.14866772e-01
2.23254859e-01 5.01390100e-01 -1.01036906e+00 7.51344979e-01
7.78874695e-01 2.74782091e-01 -7.04923868e-01 5.20845413e-01
-4.33418006e-01 9.01696801e-01 -2.30680913e-01 -4.38332483e-02
2.66268641e-01 1.10672072e-01 2.79373497e-01 1.69862735e+00
6.21923879e-02 -1.36157349e-01 -4.09194380e-02 4.64802921e-01
-1.67117283e-01 3.86142462e-01 -3.16380948e-01 -1.41990278e-02
3.05441827e-01 1.34904397e+00 -4.45951879e-01 -4.44392592e-01
-4.62854534e-01 1.11990893e+00 8.00693870e-01 -7.55983889e-02
-7.29360759e-01 -1.65479079e-01 6.08655453e-01 -1.41165808e-01
4.31993902e-01 -2.22952500e-01 -2.66212136e-01 -9.30913568e-01
-4.31851484e-02 -1.12081206e+00 6.23564124e-01 -3.89674604e-01
-1.35279429e+00 1.07229292e+00 -1.87119231e-01 -7.45139837e-01
-4.19603854e-01 -8.79879236e-01 -3.63026112e-01 8.67002189e-01
-1.48478973e+00 -1.38018179e+00 4.23300900e-02 9.20697525e-02
7.61013627e-01 -4.07246083e-01 1.28898323e+00 3.62293392e-01
-4.47819173e-01 8.29706609e-01 -2.47846581e-02 2.79234409e-01
7.84236670e-01 -1.68025768e+00 1.27127159e+00 9.25992548e-01
4.12734002e-01 2.94234246e-01 5.94395041e-01 -3.60651165e-01
-1.30950594e+00 -9.01310086e-01 7.09967673e-01 -7.09470809e-01
8.39174509e-01 -5.50500035e-01 -9.36996698e-01 7.27951348e-01
3.95960450e-01 -5.84671274e-02 4.03451204e-01 6.06019258e-01
-7.19056726e-01 -4.61611785e-02 -1.12738240e+00 4.28136349e-01
1.05914938e+00 -1.96801737e-01 -8.78327310e-01 7.44926976e-03
1.02611160e+00 -4.71641630e-01 -9.79992926e-01 6.00913584e-01
4.34013039e-01 -7.47136116e-01 8.95007193e-01 -8.46168756e-01
4.70984578e-01 9.96465981e-02 -3.03915918e-01 -1.67691267e+00
-3.75009120e-01 -6.73652768e-01 5.79820275e-02 1.74595404e+00
8.07901263e-01 -6.05425775e-01 8.28828514e-01 2.65747666e-01
-5.40489793e-01 -3.36135656e-01 -1.14241910e+00 -7.40822971e-01
8.80431235e-01 -4.47312444e-01 4.15542185e-01 8.75546575e-01
1.62952822e-02 9.27396178e-01 -2.54211068e-01 6.72692526e-03
3.88916522e-01 -3.82690161e-01 7.86179543e-01 -1.13075912e+00
-5.94506860e-01 -2.66510218e-01 -3.20616633e-01 -8.75882924e-01
1.16719708e-01 -9.68091130e-01 1.66242033e-01 -1.52147627e+00
7.54723996e-02 -4.09989834e-01 -4.84157950e-01 5.62868297e-01
-4.35298026e-01 1.27171636e-01 4.08305585e-01 -1.84373647e-01
-5.45610428e-01 1.38677537e-01 8.50771904e-01 2.28517860e-01
-3.85271043e-01 -2.80301183e-01 -8.44747186e-01 7.16258228e-01
1.08702111e+00 -4.51492459e-01 2.43327077e-02 -8.98233891e-01
1.60778582e-01 -3.15725245e-02 -3.07438225e-01 -1.15040481e+00
-1.99314550e-01 5.96772611e-01 2.91160256e-01 -2.50472784e-01
3.03697467e-01 -5.08770466e-01 -2.40402400e-01 3.92546356e-01
-2.81650484e-01 4.11305338e-01 6.54472589e-01 2.55361706e-01
-9.85995382e-02 -1.48464859e-01 8.90762269e-01 -4.02937889e-01
-6.95443928e-01 -2.11739987e-01 -2.12493420e-01 4.89067554e-01
3.61148059e-01 1.24470815e-01 -4.80011702e-01 4.82248403e-02
-6.87095881e-01 2.37415340e-02 2.40214989e-02 6.16139770e-01
7.38373306e-03 -8.14998150e-01 -1.26249015e+00 1.91403568e-01
1.42367944e-01 8.99123475e-02 1.16145067e-01 4.91850376e-01
-3.90675634e-01 2.95484304e-01 -1.91835776e-01 -5.42455733e-01
-1.07627404e+00 3.49559009e-01 9.88283008e-02 -8.30161810e-01
-6.99570954e-01 1.04966867e+00 6.84102997e-02 -8.36530924e-01
1.22818232e-01 -2.52548844e-01 -7.00588301e-02 -9.34143066e-02
4.00449842e-01 1.33893281e-01 5.39632857e-01 -4.23902124e-01
-4.69284147e-01 3.23601395e-01 -4.70469832e-01 -1.48069605e-01
1.59696317e+00 1.68103337e-01 2.68951297e-01 3.95317793e-01
9.13185179e-01 1.56232804e-01 -9.18688953e-01 -1.14737883e-01
3.81543189e-01 2.53719956e-01 -1.03147559e-01 -1.37625122e+00
-8.04467022e-01 1.04697788e+00 7.04519331e-01 6.29740283e-02
1.09484482e+00 3.79149556e-01 8.83371413e-01 3.39053780e-01
3.32732856e-01 -1.02808833e+00 -6.62705526e-02 6.60214901e-01
4.52187538e-01 -1.46761584e+00 -2.57795006e-01 -3.02391529e-01
-8.29462171e-01 8.83210719e-01 5.75549901e-01 -1.60718083e-01
2.25342259e-01 6.08399212e-01 3.02579254e-01 1.91272348e-01
-6.96648180e-01 -3.09735477e-01 9.61234141e-03 7.17589319e-01
1.10426211e+00 3.56972218e-01 -3.18925261e-01 9.12136555e-01
-5.04640460e-01 -3.27307999e-01 2.59302109e-01 7.03879058e-01
-4.25232887e-01 -1.60568357e+00 -9.64480937e-02 -2.95706000e-02
-9.86525178e-01 -4.23665345e-01 -2.50400811e-01 1.21585822e+00
-3.72394659e-02 1.00059342e+00 1.07681960e-01 -1.80082083e-01
6.39242768e-01 4.46125835e-01 6.30098522e-01 -5.87960124e-01
-8.10853899e-01 6.36263639e-02 6.42271638e-01 -2.79369026e-01
-2.57252991e-01 -5.36615670e-01 -1.29013824e+00 4.84356396e-02
-3.88921499e-01 2.57712871e-01 8.97082806e-01 7.46830106e-01
2.51752973e-01 7.88566232e-01 1.58572733e-01 -4.62698162e-01
-8.34406435e-01 -1.43192410e+00 -5.24603650e-02 3.37484479e-01
-7.00576529e-02 -1.89730734e-01 -2.04099864e-01 -8.87155086e-02] | [10.420746803283691, 9.716361999511719] |
509881a4-d6ef-41c3-bf92-3e601ec23aaf | translating-sumo-k-to-higher-order-set-theory | 2305.07903 | null | https://arxiv.org/abs/2305.07903v1 | https://arxiv.org/pdf/2305.07903v1.pdf | Translating SUMO-K to Higher-Order Set Theory | We describe a translation from a fragment of SUMO (SUMO-K) into higher-order set theory. The translation provides a formal semantics for portions of SUMO which are beyond first-order and which have previously only had an informal interpretation. It also for the first time embeds a large common-sense ontology into a very secure interactive theorem proving system. We further extend our previous work in finding contradictions in SUMO from first order constructs to include a portion of SUMO's higher order constructs. Finally, using the translation, we can create problems that can be proven using higher-order interactive and automated theorem provers. This is tested in several systems and can be used to form a corpus of higher-order common-sense reasoning problems. | ['Josef Urban', 'Adam Pease', 'Chad Brown'] | 2023-05-13 | null | null | null | null | ['automated-theorem-proving', 'automated-theorem-proving', 'common-sense-reasoning'] | ['miscellaneous', 'reasoning', 'reasoning'] | [ 4.01051313e-01 9.45941508e-01 1.12103611e-01 -1.94296107e-01
-4.81960744e-01 -1.07777786e+00 6.30464792e-01 2.60128289e-01
7.06399605e-02 9.78206336e-01 -1.05826944e-01 -1.35765553e+00
-5.18853903e-01 -1.15249360e+00 -7.37651348e-01 3.06178451e-01
-3.36361289e-01 7.62757242e-01 1.17191374e+00 -8.51166725e-01
8.58579129e-02 1.95865124e-01 -1.51887798e+00 5.16528189e-01
7.87623823e-01 1.05494581e-01 -2.39947230e-01 7.19875336e-01
7.79907480e-02 9.88028705e-01 -3.18135649e-01 -6.16909683e-01
1.73316464e-01 -3.36775333e-01 -1.69986653e+00 -4.56200629e-01
3.24476093e-01 -4.06603754e-01 -5.14204539e-02 1.15583694e+00
-2.36227661e-01 -1.36859030e-01 1.62673205e-01 -2.18567681e+00
-5.04108965e-02 8.76941144e-01 2.89875776e-01 -1.19405910e-01
1.22373879e+00 -2.93855052e-02 1.36305761e+00 1.70341209e-01
1.05901146e+00 1.74136245e+00 5.38353503e-01 6.82151973e-01
-1.48929524e+00 -7.10451007e-01 -4.14422065e-01 4.06277120e-01
-1.50810933e+00 -2.75456160e-01 4.67448503e-01 -1.65630668e-01
1.39222860e+00 9.97092724e-01 7.30863214e-01 5.65587342e-01
4.57350612e-01 5.31281650e-01 1.07546115e+00 -1.01449203e+00
4.87333834e-02 5.22895813e-01 2.79808015e-01 9.37214911e-01
6.57955170e-01 6.80601150e-02 -1.74489602e-01 -8.70560408e-01
7.55651176e-01 -6.67205930e-01 -3.56408209e-02 -3.65975618e-01
-1.15707207e+00 8.76437426e-01 -2.66707778e-01 5.43156743e-01
6.66081131e-01 2.01269433e-01 2.41651595e-01 6.62303209e-01
-2.88985014e-01 8.16871405e-01 -7.69960523e-01 -2.16641109e-02
-5.43439269e-01 8.89486611e-01 1.74013829e+00 1.07542372e+00
4.35363621e-01 -6.12520576e-01 5.85338414e-01 -2.38471955e-01
6.25739992e-01 2.15005562e-01 -1.96001738e-01 -1.64234948e+00
2.19693124e-01 8.96811783e-01 5.24820507e-01 -9.24502909e-01
-3.01307231e-01 9.39249527e-03 2.80877024e-01 5.14573157e-01
3.51063222e-01 1.81524336e-01 6.17728457e-02 1.62198424e+00
4.00231481e-01 -1.62969697e-02 5.86783171e-01 4.93913800e-01
4.93377328e-01 4.86857772e-01 -3.05777848e-01 -4.14637357e-01
1.50693035e+00 -1.87087610e-01 -5.02937913e-01 1.97093934e-01
1.11311793e+00 -2.42430717e-01 8.71394813e-01 6.88625038e-01
-1.35229003e+00 4.78971541e-01 -1.45711136e+00 -3.18227023e-01
-4.44576144e-01 -6.69975936e-01 1.17819309e+00 5.39005637e-01
-1.17840314e+00 5.49172997e-01 -4.25449729e-01 -2.86464602e-01
2.20397096e-02 6.08528793e-01 -3.65298659e-01 1.61253158e-02
-1.69391751e+00 9.48524833e-01 1.01222837e+00 -7.06020713e-01
-4.65058595e-01 -6.83013201e-01 -9.89036024e-01 -8.36995393e-02
1.07584572e+00 -7.42883563e-01 1.62739086e+00 -5.57506263e-01
-1.29259074e+00 9.48730886e-01 -1.73795104e-01 -6.56039834e-01
3.92133534e-01 3.79539222e-01 -8.60075355e-01 2.72087455e-01
4.30942059e-01 4.24190998e-01 -1.42179832e-01 -1.33021164e+00
-7.83929825e-01 -2.77211696e-01 1.34688640e+00 -1.21657379e-01
3.06134939e-01 5.28812706e-01 3.26020479e-01 -4.18353938e-02
2.57100195e-01 -1.00496376e+00 -1.22569911e-01 -5.87641373e-02
-4.31107581e-01 -4.45437849e-01 7.78947651e-01 -2.80611575e-01
1.42871034e+00 -1.79908907e+00 -6.02630638e-02 5.43755770e-01
2.64645666e-01 -1.42790318e-01 3.54510456e-01 7.08725393e-01
-3.47473830e-01 8.40862155e-01 -6.74695075e-02 6.20052278e-01
8.42792690e-01 5.87030768e-01 -5.43699861e-01 2.48523690e-02
2.88229972e-01 9.56898510e-01 -9.60288644e-01 -8.94124389e-01
7.61016831e-02 -5.27438283e-01 -1.08405912e+00 -2.35623255e-01
-8.13069463e-01 -3.02339494e-01 -3.77331227e-01 3.49481016e-01
3.84587497e-01 -2.61679739e-01 6.46552563e-01 2.04230100e-01
-1.72362238e-01 5.13961673e-01 -1.63769019e+00 1.51902521e+00
-1.21164747e-01 3.19360018e-01 -1.85626417e-01 -3.13139319e-01
4.49246019e-02 5.68629146e-01 -2.46620122e-02 -2.10428387e-01
-1.70262996e-02 2.06884861e-01 3.06461692e-01 -4.66909945e-01
2.72443146e-01 -8.42050552e-01 -4.54840273e-01 7.70982385e-01
-8.27993229e-02 -6.62457764e-01 6.61023617e-01 7.15397596e-01
1.13627732e+00 1.17354371e-01 7.53435254e-01 -5.16017973e-01
7.79108107e-01 6.06940866e-01 6.35801375e-01 5.52402377e-01
1.10519871e-01 -1.34120241e-01 1.04792428e+00 -5.55675507e-01
-8.85494947e-01 -9.32356894e-01 -1.83102429e-01 7.41767645e-01
2.68526644e-01 -1.26106644e+00 -6.17305040e-01 -6.51834249e-01
-6.93512261e-02 1.40139961e+00 -1.59672216e-01 -3.56433392e-02
-2.27037936e-01 1.97825357e-01 1.30589426e+00 1.24150038e-01
3.10388535e-01 -7.65198648e-01 -6.62642419e-01 8.48031491e-02
-4.22693133e-01 -1.37548292e+00 3.51101875e-01 5.65067008e-02
-4.66615200e-01 -1.77295887e+00 1.03292966e+00 -4.41695452e-01
5.58784068e-01 -2.53485262e-01 1.00256455e+00 7.15915442e-01
-7.98049346e-02 2.72546083e-01 -2.84749001e-01 -9.67763186e-01
-9.00722861e-01 -3.52469683e-01 3.64618376e-02 -9.06302035e-01
3.93930435e-01 -5.46751976e-01 4.26715076e-01 2.20598385e-01
-1.30276346e+00 1.36386514e-01 -2.96653599e-01 5.10261595e-01
-1.44066423e-01 8.34875703e-01 -1.97484374e-01 -1.00279927e+00
6.20640278e-01 3.82013619e-02 -1.00938308e+00 4.36702400e-01
-2.59867102e-01 3.27030867e-01 7.57468402e-01 9.59070772e-03
-9.10073936e-01 -3.96094352e-01 2.63516486e-01 2.96862185e-01
-2.45672911e-01 5.24091363e-01 -4.15651321e-01 -1.53542086e-01
6.54901981e-01 -2.67516732e-01 -1.62138790e-01 3.69031310e-01
2.59689152e-01 3.00623268e-01 4.77208197e-01 -1.34184873e+00
1.36613286e+00 4.03607041e-01 6.40487552e-01 -2.61719018e-01
-4.44150567e-01 3.92282493e-02 -3.44434589e-01 3.07470649e-01
3.69624674e-01 -3.42873454e-01 -1.45821273e+00 -3.08474660e-01
-1.30712688e+00 -5.90898633e-01 -3.69988650e-01 5.20194918e-02
-9.02485490e-01 6.80619717e-01 -3.28756094e-01 -1.07994187e+00
2.12649435e-01 -9.22359824e-01 7.97891915e-01 -1.79561868e-01
-9.42408323e-01 -1.01510751e+00 3.21527794e-02 1.71809554e-01
-3.86511475e-01 2.28360355e-01 1.47552490e+00 -8.90860736e-01
-3.35522413e-01 -7.11896271e-02 -2.26830076e-02 -5.44688702e-02
-1.57726437e-01 1.36243835e-01 -4.84513462e-01 -8.86396840e-02
-2.06949189e-01 -2.89568633e-01 -4.67482388e-01 -3.22257787e-01
5.59905589e-01 -6.75663531e-01 -5.52693188e-01 -2.15481654e-01
1.50405169e+00 1.72683731e-01 1.01477706e+00 7.80754864e-01
-3.22303981e-01 2.99862742e-01 6.90833688e-01 -2.76286621e-02
6.99942052e-01 7.14539111e-01 1.28293857e-01 2.88838446e-01
5.10698140e-01 -4.03520733e-01 1.07836574e-01 4.88785841e-02
-2.58523405e-01 3.27376306e-01 -1.15246725e+00 4.34058875e-01
-1.89633441e+00 -1.61242485e+00 -4.31479633e-01 1.69756591e+00
1.23004377e+00 4.59594935e-01 8.77688751e-02 6.66382492e-01
2.98197180e-01 -5.69509387e-01 2.73891568e-01 -9.26588953e-01
9.87110361e-02 5.22411644e-01 2.12728247e-01 1.19308293e+00
-8.25197875e-01 1.11312044e+00 7.33808231e+00 4.98905987e-01
-3.43399256e-01 -6.46119714e-02 -4.90277290e-01 1.86358660e-01
-8.91130686e-01 9.78428364e-01 -3.66770387e-01 -1.00543581e-01
8.46854091e-01 -8.04991901e-01 7.69206107e-01 5.35484612e-01
2.02646889e-02 -2.74108171e-01 -1.50977993e+00 9.55590978e-02
-1.71330139e-01 -1.34746373e+00 5.77692874e-02 8.65939632e-02
3.97529870e-01 -6.43025637e-01 -6.88321233e-01 3.25940996e-01
8.84858429e-01 -9.54991698e-01 9.56247807e-01 2.21859381e-01
4.85842556e-01 -9.33627546e-01 7.46852875e-01 6.92483425e-01
-9.01728392e-01 -4.32708472e-01 -1.01421075e-02 -8.09033155e-01
-2.06084937e-01 6.76497966e-02 -9.90710139e-01 1.22573388e+00
3.56569827e-01 -3.62878926e-02 -1.61561370e-01 7.61265516e-01
-4.26728517e-01 2.37079442e-01 -8.34191322e-01 -4.60549556e-02
8.64050537e-02 -1.40402257e-01 7.80533135e-01 9.59699571e-01
9.69876349e-02 8.86791885e-01 2.44097382e-01 1.07739484e+00
4.82179672e-01 -5.78817368e-01 -7.35832632e-01 1.19680673e-01
3.50415677e-01 7.66387701e-01 -5.14301479e-01 -5.94399750e-01
-2.68267781e-01 3.79711241e-01 -2.18542114e-01 1.87908947e-01
-9.12261546e-01 -5.37432671e-01 5.33945262e-01 2.97603369e-01
-1.21977426e-01 -4.13302809e-01 6.64615855e-02 -1.28017592e+00
2.86438257e-01 -1.21257627e+00 5.66486120e-01 -1.16219115e+00
-8.37557077e-01 1.31159453e-02 8.24367106e-01 -7.94155478e-01
-4.40158039e-01 -8.21689248e-01 -5.53944826e-01 5.69533050e-01
-1.28894877e+00 -1.29469848e+00 4.63991821e-01 6.61639631e-01
-1.66198865e-01 2.48313621e-01 1.50356734e+00 -1.05275646e-01
1.36452228e-01 3.25722337e-01 -9.60388601e-01 -2.05731377e-01
1.59407347e-01 -1.32753575e+00 1.10149190e-01 1.01450169e+00
-1.42862514e-01 1.28402078e+00 1.21371889e+00 -5.36400616e-01
-1.58694053e+00 -5.48602343e-01 1.40394258e+00 -6.29305899e-01
1.20275879e+00 -2.16749385e-01 -4.67048705e-01 1.42905676e+00
1.36474580e-01 -3.23788881e-01 6.15406513e-01 1.80565789e-01
-5.85014522e-01 3.49310576e-03 -1.47894037e+00 1.01013434e+00
1.43370771e+00 -9.26607966e-01 -1.43856978e+00 3.78379852e-01
8.20912957e-01 -5.25741100e-01 -8.86211514e-01 3.30907494e-01
6.50844395e-01 -6.92722321e-01 6.91887975e-01 -1.07070398e+00
2.15693861e-01 -1.01574743e+00 -3.40932459e-01 -7.05533504e-01
-2.15322614e-01 -1.10256922e+00 8.68909061e-02 1.06894314e+00
6.37012362e-01 -1.05696404e+00 2.76273489e-01 1.09352672e+00
1.87887534e-01 -2.26800933e-01 -1.10227525e+00 -7.32055843e-01
2.50014454e-01 -1.13956833e+00 7.89336324e-01 1.20380414e+00
1.38471448e+00 2.79585421e-01 4.85881269e-01 5.70928037e-01
5.20531476e-01 1.95130900e-01 7.29662776e-01 -1.76443398e+00
-3.25908571e-01 -1.40248880e-01 -7.37076223e-01 -2.66586751e-01
5.94241738e-01 -1.29779160e+00 -1.77075073e-01 -1.55475485e+00
2.64950186e-01 -2.02781126e-01 2.05770507e-01 1.10396039e+00
5.61705887e-01 3.29014915e-03 2.09126055e-01 -2.58812696e-01
-7.18189836e-01 -2.42185682e-01 1.04168379e+00 -2.05017313e-01
3.21434081e-01 -5.02005816e-01 -1.06628680e+00 1.03059804e+00
4.84715611e-01 -4.47313547e-01 -5.92439055e-01 1.73969403e-01
9.64331925e-01 2.07882479e-01 8.40262711e-01 -9.96690571e-01
3.45795900e-01 -9.70211744e-01 -5.01582384e-01 1.02894239e-01
-2.02971339e-01 -1.13974261e+00 7.18259454e-01 5.29838026e-01
-1.55830830e-01 -1.41288891e-01 5.75791955e-01 -2.63731360e-01
2.61630956e-02 -4.81425554e-01 1.93432078e-01 -2.70193070e-01
-8.61287951e-01 -5.40206015e-01 -6.38878107e-01 -4.23495546e-02
1.24384308e+00 -9.30839032e-02 -3.59347224e-01 -2.19695330e-01
-8.69532466e-01 3.69710624e-01 9.59435403e-01 3.65014262e-02
4.35282767e-01 -1.11043823e+00 -3.47398788e-01 -9.10925120e-02
1.08913317e-01 -2.77588457e-01 -4.04023558e-01 7.45616913e-01
-8.20927143e-01 6.18706703e-01 -2.19854936e-01 -1.06092378e-01
-1.67025721e+00 5.59181690e-01 5.13691306e-01 -2.98696280e-01
-4.64273602e-01 1.13422267e-01 -4.28359136e-02 -7.44191647e-01
-3.47548276e-01 -7.29919136e-01 2.65254945e-01 -8.71636093e-01
6.80249214e-01 -3.64056081e-02 -2.23925143e-01 -2.86518544e-01
-9.34847832e-01 2.85051495e-01 4.11374539e-01 -4.44296718e-01
1.30189192e+00 6.99930117e-02 -9.26496983e-01 3.91250491e-01
5.56407332e-01 1.97023377e-01 1.51201740e-01 3.45618784e-01
1.27645940e-01 -7.89706185e-02 -3.58759284e-01 -8.64802539e-01
2.59554595e-01 1.82340920e-01 -2.49536380e-01 7.24713027e-01
7.96185374e-01 2.44945094e-01 4.03539717e-01 9.62437749e-01
1.08761060e+00 -8.40606689e-01 -6.40751064e-01 6.29911780e-01
8.39600742e-01 -5.67830861e-01 2.75052786e-01 -1.05148661e+00
-2.39668027e-01 1.23192406e+00 2.32047960e-01 7.01868236e-02
4.20060940e-02 9.20568883e-01 -2.29784518e-01 -3.14507157e-01
-9.94638741e-01 -8.44477043e-02 -3.12180966e-01 5.33090889e-01
8.76556486e-02 1.20344877e-01 -6.20182753e-01 8.39089632e-01
-7.34236777e-01 7.62394667e-01 9.86435831e-01 1.42383695e+00
-3.00696224e-01 -1.61550319e+00 -7.01210976e-01 -1.04417749e-01
-5.64317763e-01 2.04577036e-02 -7.62111187e-01 1.35495842e+00
4.29216176e-01 1.33687592e+00 -5.20329893e-01 -2.28009656e-01
3.27036917e-01 2.38172829e-01 1.16993415e+00 -6.58678830e-01
-2.65183121e-01 -5.89172542e-01 1.05334389e+00 -6.55759215e-01
-3.89172912e-01 -5.19945204e-01 -1.84446645e+00 -8.94023895e-01
-5.37912369e-01 6.36833727e-01 -5.96339665e-02 1.46839499e+00
-2.16141611e-01 2.63847888e-01 1.04838304e-01 -2.02521130e-01
-3.14477563e-01 -1.35260567e-01 -5.34976125e-01 1.83365092e-01
-4.65238793e-03 -4.27492738e-01 -2.85954088e-01 2.18237653e-01] | [8.786149024963379, 6.848349571228027] |
7cbe1e6a-8755-4fe8-bdaf-81f919e08ce8 | dual-alignment-unsupervised-domain-adaptation | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Hao_Dual_Alignment_Unsupervised_Domain_Adaptation_for_Video-Text_Retrieval_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Hao_Dual_Alignment_Unsupervised_Domain_Adaptation_for_Video-Text_Retrieval_CVPR_2023_paper.pdf | Dual Alignment Unsupervised Domain Adaptation for Video-Text Retrieval | Video-text retrieval is an emerging stream in both computer vision and natural language processing communities, which aims to find relevant videos given text queries. In this paper, we study the notoriously challenging task, i.e., Unsupervised Domain Adaptation Video-text Retrieval (UDAVR), wherein training and testing data come from different distributions. Previous works merely alleviate the domain shift, which however overlook the pairwise misalignment issue in target domain, i.e., there exist no semantic relationships between target videos and texts. To tackle this, we propose a novel method named Dual Alignment Domain Adaptation (DADA). Specifically, we first introduce the cross-modal semantic embedding to generate discriminative source features in a joint embedding space. Besides, we utilize the video and text domain adaptations to smoothly balance the minimization of the domain shifts. To tackle the pairwise misalignment in target domain, we introduce the Dual Alignment Consistency (DAC) to fully exploit the semantic information of both modalities in target domain. The proposed DAC adaptively aligns the video-text pairs which are more likely to be relevant in target domain, enabling that positive pairs are increasing progressively and the noisy ones will potentially be aligned in the later stages. To that end, our method can generate more truly aligned target pairs and ensure the discriminality of target features.Compared with the state-of-the-art methods, DADA achieves 20.18% and 18.61% relative improvements on R@1 under the setting of TGIF->MSRVTT and TGIF->MSVD respectively, demonstrating the superiority of our method. | ['Bo Li', 'Fei Zhu', 'Dayan Wu', 'Wanqian Zhang', 'Xiaoshuai Hao'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['video-text-retrieval', 'unsupervised-domain-adaptation'] | ['computer-vision', 'methodology'] | [ 2.88656533e-01 -6.66761816e-01 -2.14027733e-01 -2.87801147e-01
-9.61365581e-01 -6.57487273e-01 7.31969059e-01 -1.27731696e-01
-5.49004316e-01 4.23521727e-01 4.95145231e-01 2.44369656e-01
-2.97774434e-01 -3.41435432e-01 -4.86149192e-01 -8.45999777e-01
3.25901777e-01 2.70538568e-01 1.88242257e-01 -1.68680564e-01
1.79852366e-01 7.51025463e-03 -1.41555762e+00 1.30554348e-01
8.42703283e-01 9.97346878e-01 3.66066903e-01 3.47980261e-01
4.78025730e-04 3.71955633e-01 -3.64486039e-01 -1.97758421e-01
3.42207044e-01 -4.53938067e-01 -5.02521753e-01 2.19364658e-01
5.37235320e-01 -5.17189622e-01 -8.61940920e-01 1.13140154e+00
7.17872918e-01 4.70111161e-01 7.64591694e-01 -1.47922611e+00
-7.16790080e-01 1.11322105e-01 -8.95337582e-01 4.08828169e-01
3.44899237e-01 1.35104030e-01 1.15415490e+00 -1.15573704e+00
7.69637227e-01 1.33726025e+00 1.27682254e-01 5.05438626e-01
-9.88316298e-01 -8.51526856e-01 2.89229989e-01 4.93383944e-01
-1.51807606e+00 -4.16068792e-01 1.01313019e+00 -4.27207887e-01
4.37419772e-01 1.81568578e-01 2.53931522e-01 1.24449062e+00
-2.06128180e-01 1.03664911e+00 7.23325431e-01 -2.41483986e-01
4.57788864e-03 9.59980190e-02 -7.27802515e-02 1.26912892e-01
1.54730976e-01 -2.26425499e-01 -7.48322248e-01 3.14386771e-03
6.24556363e-01 2.39890203e-01 -3.88478428e-01 -4.88873214e-01
-1.60885942e+00 8.45956147e-01 8.54790509e-02 3.50466907e-01
-3.49310875e-01 -2.41013587e-01 6.76562548e-01 2.97367781e-01
3.25872719e-01 1.69421837e-01 -2.40306601e-01 -1.21735603e-01
-9.41879630e-01 2.42672607e-01 2.65523851e-01 1.25353622e+00
7.40327716e-01 -1.78565010e-01 -4.72478390e-01 1.00105047e+00
5.14821053e-01 6.99517012e-01 7.55905986e-01 -7.36697793e-01
8.27374399e-01 5.79552650e-01 1.49586320e-01 -1.33084786e+00
1.93175092e-01 -1.83222592e-01 -6.65982246e-01 -5.77207506e-01
1.21675365e-01 2.48571066e-03 -8.71462405e-01 1.72617674e+00
4.43827659e-01 3.89164835e-01 3.59140515e-01 1.18872070e+00
7.20484912e-01 8.83580923e-01 -3.32377888e-02 -4.31342185e-01
1.20913935e+00 -1.04416215e+00 -8.04045618e-01 -3.22209716e-01
4.25280422e-01 -1.11309516e+00 1.13298333e+00 1.43098861e-01
-7.29354501e-01 -4.60780472e-01 -8.99814010e-01 -5.54697141e-02
-4.64353338e-02 2.97505736e-01 -4.67670746e-02 1.63205430e-01
-6.20481908e-01 4.74942178e-02 -4.46767956e-01 -4.83275145e-01
2.13073522e-01 1.68682009e-01 -4.89017218e-01 -5.80406249e-01
-1.39623821e+00 3.79092246e-01 4.78960037e-01 -1.17393978e-01
-9.02171850e-01 -5.34815371e-01 -7.43694305e-01 -1.14479631e-01
6.83448851e-01 -4.97405380e-01 1.05254459e+00 -1.22902584e+00
-1.15484512e+00 7.17977524e-01 -2.93985367e-01 -2.28533402e-01
5.36448300e-01 -4.08607781e-01 -5.28819144e-01 3.78159046e-01
2.54411966e-01 6.14140511e-01 1.13698387e+00 -1.09906495e+00
-8.23683858e-01 -3.69924992e-01 -1.38823003e-01 7.08799362e-01
-8.89505148e-01 1.55479154e-02 -1.06939995e+00 -1.00731456e+00
4.69733886e-02 -9.05587494e-01 1.20534621e-01 1.42852124e-02
-1.48532674e-01 -4.04719681e-01 1.20187724e+00 -7.65687525e-01
1.41569173e+00 -2.43227410e+00 4.25197810e-01 1.62048414e-01
1.68654397e-01 3.71510178e-01 -3.46948653e-01 3.91504973e-01
7.13023543e-02 -1.25400379e-01 -5.16196946e-04 -1.81042656e-01
6.62741363e-02 1.83208615e-01 -2.92676985e-01 3.93909603e-01
1.74670085e-01 6.15371346e-01 -1.06793499e+00 -7.55179822e-01
2.45550931e-01 4.33353215e-01 -4.69240367e-01 2.97143966e-01
-2.12102681e-01 4.63280439e-01 -8.56916368e-01 4.91408169e-01
7.39273548e-01 -1.36678666e-01 1.13376439e-01 -5.09235799e-01
1.40500531e-01 -4.01164629e-02 -1.22684455e+00 1.84416544e+00
-2.04538137e-01 6.77102029e-01 -1.00758925e-01 -1.11886811e+00
9.54496205e-01 3.32583874e-01 6.85783267e-01 -8.66510212e-01
7.06262439e-02 2.47452736e-01 -2.17375472e-01 -6.80487096e-01
5.67882001e-01 -3.48756928e-03 -1.16644047e-01 2.28887990e-01
2.48038359e-02 2.70602167e-01 1.66903317e-01 3.84327680e-01
8.44546497e-01 1.25817835e-01 2.47372165e-01 -4.34621461e-02
7.27590203e-01 -6.94651753e-02 6.82066798e-01 4.59069282e-01
-4.17679399e-01 7.46837378e-01 2.20104530e-01 -7.85131603e-02
-1.11971366e+00 -1.00018024e+00 1.40384212e-01 1.16553068e+00
5.75108707e-01 -3.24675649e-01 -4.23042119e-01 -8.41900647e-01
-9.13478434e-02 3.72576594e-01 -3.76604676e-01 -3.26023579e-01
-5.77498794e-01 -4.44831848e-01 2.98099130e-01 3.26274753e-01
6.84119523e-01 -6.61071897e-01 -1.81218177e-01 -3.51680852e-02
-6.01777136e-01 -1.34478509e+00 -1.08855140e+00 -4.77747798e-01
-6.26940131e-01 -9.29780185e-01 -1.07383800e+00 -9.71760452e-01
6.85500264e-01 7.54406452e-01 6.80506825e-01 -2.37267673e-01
-1.23564675e-01 6.08724773e-01 -7.11734295e-01 1.18992753e-01
-8.56889486e-02 -5.85061908e-02 6.77387491e-02 3.65407676e-01
7.58819580e-01 -3.32764804e-01 -8.86293769e-01 6.84034348e-01
-1.30525064e+00 -1.24217784e-02 5.63998699e-01 1.07019663e+00
7.86022484e-01 3.11684348e-02 5.13416290e-01 -3.64011645e-01
5.79557061e-01 -7.97286510e-01 -2.06061244e-01 3.88692021e-01
-4.89406705e-01 -3.55540216e-02 6.84185505e-01 -7.37817049e-01
-8.66697490e-01 6.24535903e-02 2.45599598e-01 -9.39913094e-01
-2.32657436e-02 4.65066284e-01 -4.62641925e-01 2.23498613e-01
3.76234978e-01 5.98651409e-01 -1.44527838e-01 -3.05105746e-01
1.78986281e-01 9.11050916e-01 3.83054316e-01 -5.14113009e-01
9.44226921e-01 4.47437495e-01 -2.92299360e-01 -7.44149148e-01
-4.49794799e-01 -9.98414040e-01 -3.08821142e-01 -2.71297395e-01
6.69035912e-01 -1.16302478e+00 -1.66393846e-01 2.99109191e-01
-8.35720658e-01 1.61287531e-01 1.30267173e-01 7.06233799e-01
-3.81164372e-01 7.63347089e-01 -1.06597699e-01 -4.74958628e-01
-3.92699569e-01 -1.15969598e+00 1.22977817e+00 2.79882729e-01
1.23146556e-01 -7.65198410e-01 2.01247148e-02 3.65277350e-01
1.72429249e-01 4.65434194e-02 6.65815055e-01 -9.80262518e-01
-4.96533185e-01 -9.11704004e-02 -4.33665633e-01 4.65213329e-01
3.92165750e-01 -7.89650157e-02 -5.73021531e-01 -6.18888974e-01
-1.75726086e-01 -1.77249804e-01 7.10141122e-01 1.32017285e-01
8.05664420e-01 -3.79652143e-01 -3.56969714e-01 2.57744372e-01
1.30803502e+00 3.59985560e-01 4.95625228e-01 4.86079186e-01
7.08778739e-01 6.57014430e-01 1.27976418e+00 5.87277055e-01
2.67372310e-01 9.67917144e-01 9.19712931e-02 1.68413714e-01
-7.57220015e-02 -4.11437511e-01 5.05497932e-01 8.90502930e-01
4.25668597e-01 -3.72354060e-01 -8.47192347e-01 7.91346490e-01
-2.08035874e+00 -7.10850716e-01 2.27434173e-01 2.35230613e+00
9.04826939e-01 -1.32536560e-01 4.44151461e-02 -1.04129933e-01
1.10820198e+00 2.73851246e-01 -6.57501519e-01 2.10984811e-01
-1.83332630e-03 -2.46814638e-01 2.92273581e-01 1.86959684e-01
-1.17896438e+00 8.50827038e-01 4.73088217e+00 1.27538204e+00
-1.15965986e+00 1.27826765e-01 3.68635476e-01 -2.98777372e-01
-3.94488901e-01 -9.14157256e-02 -7.32666492e-01 7.16201901e-01
5.60915291e-01 -4.85856891e-01 3.48943025e-01 7.72305906e-01
3.78518373e-01 9.82740521e-02 -1.00047898e+00 1.18639565e+00
3.89758348e-01 -8.90517235e-01 4.35393095e-01 -2.68334419e-01
8.57285559e-01 -2.18461007e-01 2.42670089e-01 3.35660607e-01
-9.09936354e-02 -5.75440705e-01 5.03247023e-01 2.61617482e-01
8.26534986e-01 -6.84229016e-01 6.19932532e-01 1.83104008e-01
-1.24819922e+00 -6.81212358e-03 -3.55935723e-01 6.07454121e-01
2.14054044e-02 3.68210942e-01 -6.64617479e-01 8.85762870e-01
6.87858462e-01 1.11060154e+00 -5.05227745e-01 9.62663531e-01
1.39275983e-01 3.84241581e-01 -3.23280036e-01 1.13693424e-01
3.49137634e-01 -1.79189742e-01 8.58129263e-01 1.14417088e+00
4.49463129e-01 -1.21841788e-01 3.02017957e-01 3.98513228e-01
-2.33517095e-01 4.29870218e-01 -5.97720683e-01 -1.11056738e-01
7.35223114e-01 9.96226251e-01 -2.69993335e-01 -3.13987523e-01
-5.38772285e-01 1.17560375e+00 -8.70084763e-02 6.61577940e-01
-8.95917237e-01 -3.01867247e-01 6.35380805e-01 -1.60223559e-01
4.86364484e-01 -1.38755709e-01 1.78276226e-01 -1.51242173e+00
4.26446944e-01 -1.19697475e+00 5.74966848e-01 -6.77854240e-01
-1.37320137e+00 3.34736764e-01 1.40730515e-01 -1.78379261e+00
-1.57796890e-01 -2.30446935e-01 -2.74503261e-01 6.26304746e-01
-1.71189487e+00 -1.20537567e+00 -4.51982021e-01 9.40377355e-01
9.75681543e-01 -2.03706712e-01 3.29182774e-01 6.91343248e-01
-5.11428952e-01 8.34609330e-01 6.61428630e-01 1.81453511e-01
1.31341386e+00 -7.87103832e-01 1.89012066e-02 9.74498749e-01
6.04733527e-02 5.81533372e-01 6.77914441e-01 -5.63718081e-01
-1.51220822e+00 -1.34831727e+00 6.40350938e-01 -8.52582008e-02
6.52333260e-01 -2.12588653e-01 -9.46290553e-01 5.40543437e-01
2.42192913e-02 -2.79875454e-02 4.48190868e-01 -4.72511619e-01
-6.19584322e-01 -3.59929085e-01 -1.01503837e+00 7.66817629e-01
1.00452781e+00 -5.85493863e-01 -6.18345857e-01 1.82917997e-01
7.87889779e-01 -2.37242565e-01 -7.81590462e-01 4.75685298e-01
5.26181698e-01 -5.80241919e-01 9.73025024e-01 -4.08435881e-01
4.20214266e-01 -5.00322938e-01 -4.32020724e-01 -1.01160669e+00
-5.03114723e-02 -5.19449294e-01 -1.22296596e-02 1.48626208e+00
1.81329977e-02 -4.49470341e-01 4.98502791e-01 3.77806872e-01
6.60800040e-02 -4.05546099e-01 -1.07397294e+00 -7.83371568e-01
-3.82749029e-02 -7.28045180e-02 3.33572209e-01 1.13304710e+00
-2.11680248e-01 3.39612961e-01 -5.68411231e-01 1.60907060e-01
4.81882036e-01 1.62191883e-01 7.58391500e-01 -8.54172111e-01
-1.40631050e-01 -2.35395923e-01 -4.17373478e-01 -1.44079149e+00
8.13739821e-02 -6.42842174e-01 1.71171784e-01 -1.30492163e+00
4.52297747e-01 -2.66585320e-01 -5.45000374e-01 2.21849784e-01
-4.35716391e-01 1.07000291e-01 2.17679873e-01 5.94520688e-01
-9.41008389e-01 8.10454965e-01 1.28525674e+00 -2.28952825e-01
-1.20026775e-01 -3.57338458e-01 -5.92281938e-01 3.74533564e-01
6.16047859e-01 -4.40026522e-01 -6.07102394e-01 -5.91022313e-01
-6.79397136e-02 -6.24800436e-02 3.21506411e-01 -8.07639658e-01
2.27746427e-01 -2.23180786e-01 1.97247282e-01 -5.33814609e-01
2.38295093e-01 -1.00679123e+00 -9.21593159e-02 -1.35719646e-02
-4.36195880e-01 6.03089854e-03 1.05229847e-01 9.73488390e-01
-6.06938124e-01 -1.26457334e-01 5.80275059e-01 2.53222615e-01
-1.00553405e+00 4.71866548e-01 -1.47891223e-01 2.74561673e-01
1.19611430e+00 -1.90796509e-01 -2.69671291e-01 -4.58647877e-01
-3.70567620e-01 6.25648737e-01 4.76629615e-01 8.37572753e-01
7.93840587e-01 -1.60281146e+00 -8.44450653e-01 6.13181666e-02
4.94063139e-01 2.91774794e-02 5.60154319e-01 8.49747658e-01
-5.68437800e-02 4.14446920e-01 -6.87015848e-03 -7.14941680e-01
-1.45537364e+00 5.83301425e-01 1.07199699e-02 -2.28772506e-01
-3.35915178e-01 5.74555457e-01 5.41473687e-01 -6.27980083e-02
2.14255974e-01 1.29182845e-01 -2.75903404e-01 1.21714436e-01
4.47638273e-01 1.85118631e-01 -1.81766868e-01 -8.37616622e-01
-3.90921801e-01 6.75399899e-01 -4.79592085e-01 1.95859987e-02
1.04865015e+00 -5.28239787e-01 1.21946208e-01 2.40631834e-01
1.55534077e+00 -1.03267767e-01 -1.33076704e+00 -7.29827106e-01
-9.87873077e-02 -7.99835503e-01 4.17780317e-02 -5.57215035e-01
-1.01541197e+00 8.52981865e-01 7.75765061e-01 -1.66037157e-01
1.46916914e+00 1.93644986e-02 1.03161585e+00 2.70602345e-01
1.74839776e-02 -1.23291755e+00 5.21944821e-01 2.87455767e-01
7.57289708e-01 -1.29325306e+00 -4.42545898e-02 -2.45155618e-01
-9.07651365e-01 9.48566198e-01 6.97615325e-01 -1.55764427e-02
2.20478743e-01 -4.65701461e-01 -1.80928916e-01 1.07933022e-01
-6.79011106e-01 -1.83189183e-01 4.85257268e-01 4.31871265e-01
1.12963103e-01 -2.22441331e-01 -5.21660447e-01 4.90346879e-01
4.53881264e-01 -2.50468459e-02 1.48760453e-01 9.41403031e-01
-3.13685149e-01 -1.09845400e+00 -4.90242511e-01 2.53863871e-01
-4.32522029e-01 -1.39176264e-01 -2.10329324e-01 7.14032888e-01
-5.67493923e-02 9.41655040e-01 -4.47967611e-02 -5.32240510e-01
3.24374050e-01 -6.13194890e-02 1.29469305e-01 -3.95974666e-01
-6.14879727e-02 5.11607289e-01 -1.59482062e-01 -3.17025781e-01
-5.97595334e-01 -7.36996412e-01 -1.02714038e+00 -9.31178704e-02
-4.16816711e-01 1.90781653e-01 2.41552889e-01 9.14209425e-01
5.88587224e-01 2.90199816e-01 1.01398993e+00 -5.77489793e-01
-6.69099212e-01 -8.52324188e-01 -3.82268906e-01 7.78956413e-01
3.59763414e-01 -7.51343191e-01 -4.80816990e-01 3.44085783e-01] | [10.344731330871582, 0.9001376628875732] |
22324de4-9203-4081-a4ce-d3b8c0108ad4 | neural-network-acceptability-judgments | 1805.12471 | null | https://arxiv.org/abs/1805.12471v3 | https://arxiv.org/pdf/1805.12471v3.pdf | Neural Network Acceptability Judgments | This paper investigates the ability of artificial neural networks to judge the grammatical acceptability of a sentence, with the goal of testing their linguistic competence. We introduce the Corpus of Linguistic Acceptability (CoLA), a set of 10,657 English sentences labeled as grammatical or ungrammatical from published linguistics literature. As baselines, we train several recurrent neural network models on acceptability classification, and find that our models outperform unsupervised models by Lau et al (2016) on CoLA. Error-analysis on specific grammatical phenomena reveals that both Lau et al.'s models and ours learn systematic generalizations like subject-verb-object order. However, all models we test perform far below human level on a wide range of grammatical constructions. | ['Samuel R. Bowman', 'Alex Warstadt', 'Amanpreet Singh'] | 2018-05-31 | neural-network-acceptability-judgments-1 | https://aclanthology.org/Q19-1040 | https://aclanthology.org/Q19-1040.pdf | tacl-2019-3 | ['linguistic-acceptability'] | ['natural-language-processing'] | [ 5.48201986e-02 6.30315721e-01 1.55929387e-01 -9.69226718e-01
-7.63793826e-01 -8.13434124e-01 4.15726155e-01 4.95362073e-01
-5.60040593e-01 3.92399758e-01 4.30105150e-01 -9.99631166e-01
6.28598407e-02 -8.50969791e-01 -7.31998920e-01 -3.00856587e-02
-1.91273857e-02 4.25545633e-01 -3.93862963e-01 -5.64609826e-01
3.37394863e-01 -8.76611546e-02 -1.32765245e+00 4.45808291e-01
1.25449550e+00 8.73439968e-01 -1.09999940e-01 5.32718301e-01
1.45952404e-01 8.21624219e-01 -6.21928155e-01 -9.59943056e-01
-1.73763782e-01 -1.38726696e-01 -1.18447268e+00 -5.77752054e-01
8.44011664e-01 -1.59271166e-01 1.03410706e-01 1.09328282e+00
3.71847689e-01 -1.22465640e-02 7.84917355e-01 -6.33610725e-01
-1.32840896e+00 1.36971080e+00 3.44435692e-01 3.16079617e-01
6.29008770e-01 3.82765293e-01 1.67124474e+00 -1.10586524e+00
6.50001168e-01 1.70254993e+00 8.45757186e-01 5.91826499e-01
-1.14268899e+00 -4.35154200e-01 3.48997116e-01 -8.17866158e-03
-6.86701119e-01 -5.32532036e-01 6.41477048e-01 -3.23031366e-01
1.67673147e+00 -6.25380126e-05 5.84484994e-01 1.44127023e+00
5.23105145e-01 5.46522141e-01 1.40684903e+00 -5.79073489e-01
1.02183707e-01 -3.32721591e-01 7.52127409e-01 5.12660384e-01
2.65550882e-01 2.11320072e-01 -4.39957440e-01 -1.19087659e-02
-2.21448764e-02 -9.67383742e-01 1.61248446e-01 5.54288447e-01
-8.11783075e-01 9.09135938e-01 3.76135916e-01 1.85629964e-01
-2.13405818e-01 5.99014796e-02 6.59850776e-01 7.76617348e-01
6.69668257e-01 8.60898376e-01 -6.28651142e-01 -8.33971426e-02
-5.98172545e-02 2.02420935e-01 7.28967309e-01 8.43198299e-01
3.53727162e-01 3.78264636e-01 2.00591832e-01 1.00492930e+00
6.00997210e-01 6.00039065e-01 5.53045809e-01 -1.00030851e+00
6.13615036e-01 5.67431986e-01 -2.70918906e-01 -9.65771079e-01
-6.78063393e-01 -6.82379976e-02 -3.04364949e-01 -3.90633754e-02
4.78669584e-01 -4.60140780e-02 -6.56466722e-01 2.24662209e+00
-2.28703078e-02 -6.91123366e-01 4.63339865e-01 4.03282940e-01
1.37033129e+00 7.34033167e-01 9.10716176e-01 -1.13675192e-01
1.14491260e+00 -5.27615964e-01 -4.91611451e-01 -7.93215930e-01
1.26717603e+00 -6.10953033e-01 1.69264078e+00 4.30842489e-01
-1.57183969e+00 -6.77846789e-01 -1.01168609e+00 -4.32033658e-01
-5.54841340e-01 -5.12603298e-02 9.28813219e-01 7.90732622e-01
-1.35810399e+00 3.68811578e-01 -3.80160093e-01 -3.04047644e-01
2.08198935e-01 3.34562719e-01 -4.59213048e-01 1.90113306e-01
-1.54843462e+00 1.36924756e+00 5.76485336e-01 3.18251103e-01
-4.84231263e-01 -1.47345364e-01 -1.31749773e+00 -1.53191894e-01
-6.91848993e-02 -3.65381092e-01 1.42047584e+00 -1.23617637e+00
-1.44788313e+00 1.56813061e+00 -1.67678818e-01 -4.83779162e-01
-4.36646491e-01 -4.66982394e-01 -6.14303172e-01 -3.45619977e-01
9.62035060e-02 6.41758800e-01 3.77905875e-01 -6.82236254e-01
-4.63502228e-01 -3.00638199e-01 4.34735984e-01 1.63767487e-01
-3.45616072e-01 4.34526801e-01 6.16947651e-01 -6.81323111e-01
2.61026084e-01 -9.42625701e-01 1.73083916e-01 -8.43625188e-01
-4.28867996e-01 -1.30850279e+00 1.22938223e-01 -7.29653180e-01
1.29643703e+00 -2.09681988e+00 1.44533008e-01 2.06237510e-01
-1.83949471e-02 -1.94465164e-02 -4.65494603e-01 2.55337507e-01
-3.66246343e-01 7.04177618e-01 -3.33767444e-01 -2.90437341e-01
3.13531607e-01 2.00659871e-01 -3.80648077e-01 2.52300262e-01
5.93038619e-01 1.24972355e+00 -8.37020218e-01 -2.95745701e-01
-1.34531036e-01 4.22450788e-02 -7.98113525e-01 1.26469851e-01
-3.36427927e-01 1.45305634e-01 7.32221827e-02 8.51722956e-01
3.11810941e-01 1.97303399e-01 3.74943435e-01 1.42022491e-01
6.76505491e-02 1.40403199e+00 -4.82039630e-01 1.32946336e+00
-8.58204603e-01 6.42573416e-01 -2.42960021e-01 -7.92910755e-01
9.11042094e-01 2.75589615e-01 -6.38235986e-01 -8.76098752e-01
2.96525329e-01 6.58711672e-01 7.79111683e-01 -4.56212908e-01
6.24289334e-01 -1.05031826e-01 -8.25835407e-01 6.59298182e-01
7.81424642e-02 -4.81848001e-01 3.08159500e-01 1.69201970e-01
8.06927085e-01 1.82076320e-01 2.14096904e-01 -8.36671650e-01
3.70772094e-01 -4.04012918e-01 6.64133906e-01 9.27118361e-01
-2.57931262e-01 2.85838515e-01 6.30930066e-01 -5.77979386e-01
-9.22477543e-01 -1.25869799e+00 -5.82020938e-01 1.67639565e+00
-3.93897384e-01 -4.63271827e-01 -9.04123664e-01 -6.42796040e-01
-3.15564990e-01 1.23330641e+00 -7.68151343e-01 -2.95118570e-01
-9.03617501e-01 -6.83003247e-01 6.09760702e-01 6.10152900e-01
5.91514856e-02 -2.04924130e+00 -6.22276783e-01 3.15645427e-01
-1.95261732e-01 -9.96802151e-01 1.00353032e-01 2.11809710e-01
-6.00389123e-01 -7.98492432e-01 4.87573028e-01 -1.06529224e+00
4.28358465e-01 -7.80651271e-01 1.87440038e+00 5.11777461e-01
3.00735116e-01 1.36970937e-01 -3.53999496e-01 -5.97214758e-01
-9.32544172e-01 3.01484287e-01 2.13826492e-01 -6.16646051e-01
6.96412027e-01 -4.74208325e-01 -1.36587098e-01 -2.25689277e-01
-5.66875160e-01 -4.63015229e-01 2.25824550e-01 8.53179693e-01
2.47933939e-01 -5.76089621e-01 1.06260908e+00 -1.27971399e+00
1.27786994e+00 -4.64553833e-01 -4.92856562e-01 1.06910586e-01
-6.12930238e-01 -1.89175501e-01 6.26223147e-01 -6.26303703e-02
-1.01330280e+00 -4.01260376e-01 -6.56629264e-01 5.80696166e-01
-3.40062588e-01 8.57641578e-01 -1.65063217e-01 3.10609698e-01
7.94032514e-01 -3.15970629e-01 -3.88031483e-01 -6.73172325e-02
4.03167307e-01 6.97581410e-01 7.34690368e-01 -9.88550127e-01
2.83111632e-01 -3.54398698e-01 -3.34126234e-01 -4.96557772e-01
-1.29925716e+00 2.27146760e-01 -6.24722004e-01 -1.55375972e-01
7.48198152e-01 -8.27147961e-01 -8.15693855e-01 4.70154077e-01
-1.43705797e+00 -5.02427638e-01 -4.35525253e-02 2.84802765e-01
-5.61600149e-01 5.21271341e-02 -1.16420984e+00 -5.59251130e-01
-5.65645158e-01 -1.10166860e+00 9.45653021e-01 -8.24992210e-02
-1.13808656e+00 -1.03972268e+00 -6.35386929e-02 5.49209118e-02
4.23754632e-01 9.80120450e-02 1.60032499e+00 -7.89964139e-01
-1.34432912e-01 -8.38939026e-02 1.70223653e-01 4.30048227e-01
-1.28885657e-01 3.74499828e-01 -1.03124011e+00 -1.07220657e-01
2.26024941e-01 -1.01956761e+00 8.79619062e-01 4.22245532e-01
1.18266165e+00 -5.54219067e-01 3.82715285e-01 5.12170196e-01
9.78202641e-01 1.65635079e-01 6.89592242e-01 4.79065686e-01
4.45807517e-01 9.47167575e-01 4.91592079e-01 -3.53087038e-01
6.64685667e-01 8.54323804e-02 3.76077265e-01 5.50149202e-01
2.18942285e-01 -1.98414803e-01 8.13493609e-01 1.27504277e+00
1.12051889e-02 -5.91426253e-01 -1.19605196e+00 8.06086540e-01
-1.44021642e+00 -6.68010831e-01 -2.27679759e-01 1.91000700e+00
1.12192786e+00 5.89673877e-01 -6.06973246e-02 1.02968059e-01
4.19184893e-01 2.25393996e-01 -1.92785934e-01 -1.69061959e+00
-5.56388080e-01 5.23682177e-01 -3.16183895e-01 9.32797015e-01
-1.15520191e+00 1.57620347e+00 7.40335083e+00 2.62102425e-01
-7.60058701e-01 -6.62810355e-02 1.02069068e+00 2.35185131e-01
-8.26092005e-01 -8.79036114e-02 -6.06866956e-01 3.03700000e-01
1.58590996e+00 1.17754690e-01 2.33445972e-01 7.71149695e-01
-2.13080704e-01 -8.31715390e-02 -1.48261189e+00 2.40350083e-01
2.15465128e-01 -7.81718493e-01 1.25462532e-01 -4.56717551e-01
4.93193477e-01 1.33750260e-01 3.81975234e-01 7.03290462e-01
6.30223155e-01 -1.54283452e+00 8.80000353e-01 2.27149174e-01
6.35611355e-01 -8.16048861e-01 7.98572421e-01 3.61245833e-02
-4.68743682e-01 -1.33894026e-01 -4.49955314e-01 -7.01946795e-01
1.39028355e-01 3.66539270e-01 -3.54295492e-01 -2.40797743e-01
1.00537121e+00 6.11690998e-01 -1.09089744e+00 9.89996940e-02
-8.70308280e-01 1.21576941e+00 -4.53085840e-01 -3.91268462e-01
3.38495284e-01 -1.61432430e-01 3.78808230e-01 1.22628868e+00
-1.55269802e-01 1.92258775e-01 -2.18666226e-01 8.82147729e-01
-3.79017264e-01 4.04552758e-01 -9.81792152e-01 -1.47938073e-01
6.25818193e-01 8.86556566e-01 -5.19395232e-01 -3.05092484e-01
-3.45543891e-01 3.59752029e-01 9.79537845e-01 1.46534801e-01
-2.95558214e-01 -4.03889030e-01 2.42390752e-01 -3.09885442e-01
-2.33521581e-01 -1.69051826e-01 -6.23369515e-01 -7.84478366e-01
3.11225533e-01 -9.79265332e-01 3.71094525e-01 -7.94178009e-01
-1.59788287e+00 7.94755757e-01 -7.95215294e-02 -4.62439388e-01
-7.31545985e-01 -9.88268077e-01 -8.36990595e-01 8.30672264e-01
-9.78303492e-01 -1.21150708e+00 3.48707229e-01 4.31042984e-02
5.29854953e-01 -1.95241809e-01 1.11301899e+00 -1.73203826e-01
-5.57409167e-01 7.94712126e-01 -3.94602656e-01 2.85513461e-01
5.25026441e-01 -1.41109109e+00 1.11324692e+00 7.34174073e-01
1.88439727e-01 1.10638487e+00 5.57282269e-01 -6.33061349e-01
-8.41811836e-01 -7.34885037e-01 1.68826711e+00 -1.09191263e+00
6.77822173e-01 -5.24883389e-01 -1.18057966e+00 1.25614667e+00
7.98417270e-01 -2.17704028e-01 5.48493981e-01 8.40837061e-01
-5.34288764e-01 2.97859102e-01 -9.93892074e-01 6.45356059e-01
1.24727619e+00 -1.01443565e+00 -1.29978323e+00 4.00908440e-01
8.75729382e-01 -3.32445592e-01 -8.44089985e-01 6.45109594e-01
3.79824311e-01 -1.05893731e+00 5.94001710e-01 -1.01807630e+00
1.01504982e+00 2.72352874e-01 -1.12671316e-01 -1.63864541e+00
-1.95040375e-01 -3.37169319e-01 2.65114576e-01 1.18573356e+00
9.41572726e-01 -9.12016749e-01 2.73401067e-02 7.92664111e-01
-5.80925167e-01 -9.52036917e-01 -1.08470941e+00 -4.73565608e-01
9.79907095e-01 -7.53600836e-01 5.46738744e-01 9.33809817e-01
3.68066460e-01 7.67321050e-01 4.40056205e-01 -1.03231244e-01
3.40363294e-01 4.93143946e-02 1.32398888e-01 -1.26426470e+00
-5.05623184e-02 -6.69998467e-01 -1.81089953e-01 -4.07690644e-01
1.11442852e+00 -1.31914711e+00 1.18816614e-01 -1.29745662e+00
-1.14330083e-01 -3.56948704e-01 -1.31571651e-01 5.48584461e-01
-2.94899493e-01 8.09600428e-02 2.55504511e-02 -2.63273150e-01
-4.49204057e-01 4.50544208e-01 6.10326231e-01 -1.08027779e-01
1.65642258e-02 -4.00210679e-01 -9.21929300e-01 1.27155471e+00
1.14064729e+00 -3.14209014e-01 -2.59883285e-01 -1.00458860e+00
1.00554276e+00 -4.68887806e-01 1.54146358e-01 -6.31904721e-01
-2.72214592e-01 1.37555853e-01 1.83602408e-01 -3.17157418e-01
-8.36356208e-02 -1.85308263e-01 -6.83959007e-01 1.35834709e-01
-8.49637449e-01 8.35061133e-01 4.41533953e-01 4.42522541e-02
-2.86195785e-01 -4.26437914e-01 5.46606421e-01 -1.26408383e-01
-6.14172697e-01 -1.53276294e-01 -6.22875690e-01 6.98495567e-01
4.46734041e-01 4.54051383e-02 -5.07240117e-01 -3.89035970e-01
-8.79636943e-01 8.38221833e-02 2.77057320e-01 7.46389866e-01
7.01670706e-01 -1.23182964e+00 -9.46043253e-01 1.79176986e-01
2.06942961e-01 6.85831830e-02 -1.86979890e-01 3.95621896e-01
-4.43921328e-01 3.12566787e-01 4.11646292e-02 -3.95263106e-01
-7.69578576e-01 1.07480064e-01 3.43846351e-01 1.79845065e-01
-2.79181480e-01 1.10803068e+00 1.91484466e-01 -1.05467474e+00
-1.01938799e-01 -5.91677368e-01 -3.11969638e-01 -1.22483619e-01
7.41862357e-02 -1.34184688e-01 -3.01468950e-02 -1.02215719e+00
-1.96107581e-01 3.90944451e-01 -1.91416070e-01 2.44450346e-02
1.24614298e+00 -7.76655450e-02 -5.57670951e-01 9.04692769e-01
1.08359540e+00 1.28050894e-01 -3.65071356e-01 1.79165974e-02
3.61199796e-01 1.64421201e-01 -3.58873308e-01 -9.92831111e-01
-4.88980830e-01 9.52233911e-01 1.68177292e-01 2.19232365e-01
7.36710727e-01 2.27150828e-01 5.87613881e-01 8.94622803e-01
1.20956488e-01 -1.55064487e+00 -3.74806136e-01 1.36150849e+00
1.16855896e+00 -1.41015649e+00 -4.38288569e-01 -3.10961336e-01
-5.23356497e-01 9.34775949e-01 9.94516075e-01 -3.60057861e-01
4.14995849e-01 -1.23103909e-01 2.67122149e-01 -3.73128116e-01
-9.68312442e-01 1.46678030e-01 3.29022072e-02 3.88503879e-01
1.13178313e+00 3.81560594e-01 -7.26080000e-01 7.23144829e-01
-1.14616227e+00 -6.35152876e-01 6.78184807e-01 6.38070881e-01
-3.28599483e-01 -8.53879452e-01 4.54685874e-02 6.34792566e-01
-6.55168951e-01 -7.62020469e-01 -6.47275627e-01 7.85328329e-01
-1.98639348e-01 1.07658792e+00 2.98932165e-01 -6.49197996e-02
4.48651493e-01 4.97058004e-01 2.94485211e-01 -8.53956342e-01
-9.58899260e-01 -4.53539521e-01 7.29673743e-01 -5.71086049e-01
-4.26949114e-01 -8.73065591e-01 -1.30036068e+00 -4.80921306e-02
-1.58567727e-01 1.81197431e-02 3.65232766e-01 1.12806261e+00
-1.52363302e-02 1.46291226e-01 4.45550740e-01 -3.04706961e-01
-5.43469012e-01 -1.34981775e+00 -2.51615167e-01 7.87080586e-01
2.26606280e-01 -2.19226494e-01 -6.89193070e-01 -2.70179540e-01] | [10.683989524841309, 9.429229736328125] |
547f71d2-a0ca-4fa9-84f1-92745f396f5c | improving-few-shot-text-classification-via | 1908.08788 | null | https://arxiv.org/abs/1908.08788v2 | https://arxiv.org/pdf/1908.08788v2.pdf | When Low Resource NLP Meets Unsupervised Language Model: Meta-pretraining Then Meta-learning for Few-shot Text Classification | Text classification tends to be difficult when data are deficient or when it is required to adapt to unseen classes. In such challenging scenarios, recent studies have often used meta-learning to simulate the few-shot task, thus negating implicit common linguistic features across tasks. This paper addresses such problems using meta-learning and unsupervised language models. Our approach is based on the insight that having a good generalization from a few examples relies on both a generic model initialization and an effective strategy for adapting this model to newly arising tasks. We show that our approach is not only simple but also produces a state-of-the-art performance on a well-studied sentiment classification dataset. It can thus be further suggested that pretraining could be a promising solution for few-shot learning of many other NLP tasks. The code and the dataset to replicate the experiments are made available at https://github.com/zxlzr/FewShotNLP. | ['Ningyu Zhang', 'Shumin Deng', 'Jiaoyan Chen', 'Huajun Chen', 'Zhanlin Sun'] | 2019-08-22 | null | null | null | null | ['few-shot-text-classification'] | ['natural-language-processing'] | [ 1.99835747e-01 -2.35559884e-02 -1.58661664e-01 -5.75558066e-01
-7.38827705e-01 -2.56409377e-01 8.81341934e-01 4.68126595e-01
-7.41224349e-01 7.35622048e-01 4.37944829e-02 -1.74922809e-01
-8.30973387e-02 -6.74339950e-01 -4.41913664e-01 -5.86744964e-01
2.58589387e-01 6.38134181e-01 2.08518803e-01 -4.61394668e-01
3.93072814e-01 1.91343874e-01 -1.85079622e+00 5.51043153e-01
6.88827753e-01 5.79832494e-01 4.12020892e-01 6.15707397e-01
-5.30481517e-01 6.03718281e-01 -5.24431229e-01 -6.14691854e-01
5.15635386e-02 -3.46378833e-01 -9.06083822e-01 -2.08225823e-03
2.63570130e-01 1.00559033e-01 1.50865242e-01 1.00808346e+00
6.53265774e-01 7.07297683e-01 7.11434186e-01 -1.05962789e+00
-4.19310778e-01 6.03331923e-01 -3.05581808e-01 3.06591779e-01
1.00063801e-01 9.33121219e-02 8.06830227e-01 -1.04086542e+00
5.13399065e-01 9.47289348e-01 5.47438204e-01 9.07539666e-01
-1.22588611e+00 -5.78928709e-01 1.38428301e-01 2.34911978e-01
-1.05071306e+00 -7.43618369e-01 7.22842515e-01 -3.32747728e-01
1.13270700e+00 -6.34138510e-02 3.48162681e-01 1.44207335e+00
1.56729266e-01 6.92778885e-01 1.09132385e+00 -9.36576247e-01
5.63677907e-01 5.72900355e-01 4.82647449e-01 5.02880454e-01
2.27105588e-01 -6.67854697e-02 -5.14028192e-01 -1.98726997e-01
7.45192543e-02 1.07430786e-01 -1.00815475e-01 -3.94782096e-01
-8.83946240e-01 9.59553003e-01 1.54157370e-01 7.69976199e-01
-2.08343193e-01 -1.53271496e-01 6.35674417e-01 5.11717141e-01
7.61091769e-01 7.42169261e-01 -8.51907670e-01 -2.63413042e-01
-9.65977490e-01 1.31637037e-01 9.76560771e-01 6.36912465e-01
8.88443708e-01 -5.70048094e-02 8.43964219e-02 1.17912912e+00
2.28493344e-02 -1.84655085e-01 1.03078246e+00 -6.80337071e-01
2.88770139e-01 3.85015547e-01 8.64638686e-02 -5.03885210e-01
-4.95371342e-01 -2.74047524e-01 -6.40670955e-01 1.49025887e-01
4.49984431e-01 -4.16827112e-01 -1.02693069e+00 1.78873646e+00
2.12215990e-01 1.58045068e-01 1.73894480e-01 3.26055408e-01
6.96603715e-01 7.07954347e-01 3.83367598e-01 -2.96056390e-01
1.20330262e+00 -1.02093160e+00 -6.64705813e-01 -6.25753582e-01
1.18587518e+00 -6.50692225e-01 1.35289645e+00 3.25428903e-01
-9.15380001e-01 -6.82362080e-01 -9.28017795e-01 -5.59090041e-02
-9.69802380e-01 -2.42262527e-01 6.26796007e-01 7.11578131e-01
-8.30879867e-01 7.46341825e-01 -6.72365785e-01 -8.85433853e-01
3.36740673e-01 2.00705647e-01 -2.74019182e-01 -2.08443671e-01
-1.20673752e+00 1.09147608e+00 7.51331568e-01 -1.16482988e-01
-6.07050776e-01 -6.21123016e-01 -9.36576068e-01 2.96414383e-02
5.44792712e-01 -6.15076005e-01 1.47802627e+00 -1.13229358e+00
-1.53039455e+00 9.84427750e-01 -2.63554454e-01 -3.04733515e-01
3.60133052e-01 -1.74117550e-01 -3.84159058e-01 -1.76504508e-01
4.17526513e-02 5.67470968e-01 9.18667197e-01 -1.01182616e+00
-5.36420822e-01 -4.09767985e-01 -1.41963825e-01 1.76356807e-01
-6.68798566e-01 3.24330255e-02 -2.72459164e-02 -6.83940351e-01
-3.61117750e-01 -7.42989361e-01 -3.24148238e-01 -3.53376716e-01
1.39226750e-01 -4.85523045e-01 4.79575723e-01 -1.26138568e-01
1.11590862e+00 -1.99756396e+00 1.16113298e-01 -3.09378982e-01
-2.81926751e-01 6.85986698e-01 -3.76905680e-01 7.10028350e-01
-2.16378391e-01 1.73698530e-01 -3.63673091e-01 -5.65084815e-01
-1.47171125e-01 2.06639424e-01 -3.12890828e-01 1.78624719e-01
4.65876758e-02 8.87092888e-01 -1.02287686e+00 -2.08384380e-01
3.04694533e-01 2.50451177e-01 -2.46477321e-01 7.34292865e-02
-3.59609187e-01 2.44130045e-01 -4.00537491e-01 3.35173488e-01
3.65851134e-01 -2.99156517e-01 1.85906887e-01 1.25841305e-01
-7.15656951e-02 1.71517521e-01 -1.08282030e+00 2.01359224e+00
-6.28216207e-01 4.32326257e-01 -3.58256727e-01 -1.45873427e+00
6.81191385e-01 4.64952260e-01 2.61139780e-01 -5.27471960e-01
5.50894260e-01 2.11278141e-01 7.45868459e-02 -6.35861933e-01
4.48626816e-01 -5.28357625e-01 -1.56239092e-01 6.53030813e-01
6.04436100e-01 -2.87419677e-01 5.13876379e-01 4.67221849e-02
8.72698307e-01 8.35893080e-02 5.32610774e-01 -2.78467298e-01
4.89336163e-01 1.27572745e-01 3.05978537e-01 1.04049134e+00
-2.15552449e-01 3.64157408e-01 1.00802071e-01 -4.53058600e-01
-9.31834161e-01 -6.51435077e-01 -3.19737375e-01 1.47448921e+00
-3.55904102e-01 -5.24513721e-01 -5.40681124e-01 -7.10385978e-01
-2.21228212e-01 1.27684486e+00 -7.34322250e-01 -3.53716284e-01
-1.89244762e-01 -9.16083872e-01 2.07062960e-01 3.51699382e-01
2.88904548e-01 -1.27396798e+00 -5.30354798e-01 3.19758832e-01
4.91177142e-02 -8.40556800e-01 8.28258395e-02 6.57262862e-01
-1.07395566e+00 -8.73917997e-01 -7.85342693e-01 -9.49839830e-01
5.90571761e-01 4.55092311e-01 1.06180310e+00 2.20076799e-01
-2.06837431e-01 4.61096138e-01 -6.82384133e-01 -7.21690059e-01
-5.79319596e-01 4.03118581e-01 1.70244090e-02 -1.09937973e-01
8.88663888e-01 -4.08793718e-01 -1.35719731e-01 4.95992377e-02
-1.02788281e+00 -4.82532308e-02 2.84066439e-01 1.03401053e+00
3.32238734e-01 1.56245977e-01 8.73883724e-01 -1.31784272e+00
8.35985541e-01 -6.65926576e-01 -2.99471796e-01 3.09332371e-01
-6.62377238e-01 6.25820458e-02 8.62271249e-01 -5.57221234e-01
-1.20836902e+00 -4.05485667e-02 -3.08128566e-01 -1.64469361e-01
-5.94200909e-01 5.50900340e-01 8.25202540e-02 1.26360461e-01
9.15687025e-01 1.50939152e-01 -1.42607510e-01 -6.02369189e-01
4.87528622e-01 8.68246198e-01 -8.29310715e-02 -6.35519922e-01
5.24730861e-01 4.61418122e-01 -3.70736122e-01 -1.10374582e+00
-1.26034594e+00 -6.04497015e-01 -8.76576900e-01 1.23117035e-02
6.27933323e-01 -7.88567424e-01 -2.18310882e-03 4.47770894e-01
-9.48959649e-01 -6.40414417e-01 -4.54772592e-01 4.03449088e-01
-5.94879508e-01 1.82557195e-01 -4.71368343e-01 -7.29232728e-01
-2.63906956e-01 -8.09799016e-01 7.10269928e-01 2.88785696e-01
-4.86062676e-01 -1.36074793e+00 3.53683442e-01 2.36158580e-01
3.70716780e-01 -2.69180208e-01 8.95860374e-01 -1.22504711e+00
1.50957659e-01 -3.03743482e-01 3.11775208e-01 4.16407377e-01
2.42130160e-01 -1.94485247e-01 -1.31259406e+00 -3.87403935e-01
4.33945805e-01 -8.77997220e-01 1.06186581e+00 2.73004234e-01
9.90878224e-01 -1.55146653e-02 -1.98944196e-01 3.93268764e-01
1.43364835e+00 5.59286997e-02 4.20245081e-01 5.41423082e-01
3.57092500e-01 1.01908326e+00 7.10574031e-01 2.52628118e-01
1.54876873e-01 3.73301029e-01 -1.59931481e-01 2.55254924e-01
1.39455810e-01 3.26759950e-03 1.45052269e-01 7.94511080e-01
2.21929699e-01 -2.50037402e-01 -1.14242995e+00 6.64217472e-01
-2.01490164e+00 -1.13157856e+00 1.61820322e-01 2.13267994e+00
9.63615954e-01 2.76105672e-01 1.14355907e-02 5.80467109e-04
5.86766601e-01 2.79384077e-01 -3.34248751e-01 -6.55617297e-01
1.00022823e-01 4.48897511e-01 -1.61949396e-02 5.19489169e-01
-1.06256747e+00 1.17074871e+00 5.87186146e+00 7.57395864e-01
-1.06865728e+00 4.21089292e-01 6.85329437e-01 -2.20473737e-01
3.56298089e-02 -3.28661874e-03 -9.68561172e-01 4.76035088e-01
1.35579610e+00 -3.91884863e-01 2.17560157e-01 7.25907087e-01
-1.14450557e-03 -2.09603935e-01 -9.99892414e-01 7.77217805e-01
3.09370697e-01 -1.18422115e+00 1.36264339e-01 -3.77978861e-01
7.52243102e-01 2.38205016e-01 -6.81371689e-02 7.21007109e-01
3.41829926e-01 -7.56485701e-01 3.26594889e-01 3.33513081e-01
3.98235887e-01 -6.71110332e-01 6.02683723e-01 8.33459556e-01
-7.50627875e-01 -2.16545165e-01 -6.80864334e-01 -2.54082114e-01
-6.11941591e-02 3.44892591e-01 -6.54489458e-01 3.85339439e-01
6.14413083e-01 7.59953141e-01 -8.18557978e-01 1.02354860e+00
-3.08505774e-01 4.56079066e-01 -1.20404117e-01 -2.18170837e-01
2.81818032e-01 2.51406506e-02 3.68591398e-01 1.30486226e+00
2.35355318e-01 -1.12745529e-02 1.65364265e-01 4.80699450e-01
4.72001918e-02 4.80886549e-01 -8.86470795e-01 -1.83808029e-01
1.92971498e-01 1.35852551e+00 -8.45609546e-01 -5.74710011e-01
-7.05610514e-01 8.74134898e-01 6.73703372e-01 4.15416092e-01
-3.80554676e-01 -4.82904255e-01 3.54884952e-01 -1.49181157e-01
3.64634782e-01 -1.09465979e-02 -2.04720795e-01 -1.52108991e+00
-1.82390228e-01 -8.68850768e-01 5.65206885e-01 -7.15424716e-01
-1.43544126e+00 6.70737922e-01 1.11688435e-01 -1.03397489e+00
-4.53909963e-01 -7.25650787e-01 -7.49370813e-01 7.15628982e-01
-1.59238291e+00 -9.27883029e-01 -1.17692374e-01 5.94950378e-01
1.05182374e+00 -2.79093891e-01 1.17578137e+00 4.96206135e-02
-6.21303797e-01 3.34321022e-01 2.91125298e-01 -1.21799313e-01
1.03929782e+00 -1.08989453e+00 2.98345178e-01 7.36450374e-01
4.44240779e-01 7.00213313e-01 8.68427813e-01 -5.25558472e-01
-9.81218934e-01 -9.37866688e-01 1.11820877e+00 -6.71892762e-01
8.89389277e-01 -5.58949769e-01 -1.33068740e+00 8.23810935e-01
3.88271987e-01 -9.53228027e-02 1.09667516e+00 3.01601887e-01
-2.15875670e-01 -2.59568151e-02 -9.03548479e-01 6.18407786e-01
7.41968870e-01 -4.59857762e-01 -1.13869905e+00 5.85775137e-01
4.05845314e-01 4.44286540e-02 -5.29244125e-01 1.47818923e-01
2.42802128e-01 -8.42470109e-01 5.42832851e-01 -1.03575349e+00
3.78143519e-01 1.40120551e-01 -2.77364124e-02 -1.64372468e+00
-2.75652707e-01 -4.00909036e-01 -1.55637592e-01 1.27176714e+00
5.81931293e-01 -7.08302140e-01 6.24015212e-01 6.66023672e-01
3.01592108e-02 -6.71147704e-01 -7.08703399e-01 -8.74228895e-01
4.76557434e-01 -5.88544607e-01 1.86515525e-01 1.29027724e+00
2.07657456e-01 6.98216558e-01 -1.92032993e-01 -1.97300062e-01
4.90306944e-01 -5.76099046e-02 7.29909062e-01 -1.48557258e+00
-2.70199418e-01 -3.80977482e-01 -2.52595115e-02 -5.63190103e-01
6.18363380e-01 -1.08202243e+00 4.15411629e-02 -1.41322672e+00
1.92671880e-01 -3.31932187e-01 -6.24784648e-01 6.21718585e-01
-1.81263074e-01 1.72916893e-02 1.06462605e-01 1.66075990e-01
-6.20885909e-01 4.73125011e-01 9.35412765e-01 -6.29914328e-02
-4.41471189e-01 1.47348434e-01 -9.16839182e-01 8.48218858e-01
1.18721712e+00 -7.71548748e-01 -5.92004120e-01 -4.06628966e-01
2.37729564e-01 -2.78005302e-01 6.11664802e-02 -8.88280869e-01
4.86783803e-01 -8.21077749e-02 3.33311051e-01 -2.10361674e-01
4.84840542e-01 -7.38122582e-01 -3.10386449e-01 3.75050992e-01
-4.94822562e-01 -1.08864807e-01 5.30420959e-01 4.93019313e-01
-9.41066742e-02 -9.28193152e-01 9.36087906e-01 -6.21322215e-01
-1.05926275e+00 1.79977983e-01 -4.07848805e-01 1.81686908e-01
9.70667958e-01 -1.48088112e-01 -2.86411017e-01 -1.68122128e-01
-7.93900549e-01 2.32361138e-01 5.65986037e-01 6.46410644e-01
3.30970019e-01 -8.54937315e-01 -6.67385280e-01 2.46274576e-01
3.99466634e-01 -3.25687259e-01 2.49832034e-01 6.34360969e-01
-1.28109604e-01 3.88521820e-01 -1.61175564e-01 -2.92782694e-01
-1.03520465e+00 7.89328873e-01 3.03728044e-01 -2.09658831e-01
-6.56699479e-01 8.44881892e-01 -2.17788643e-03 -5.46705723e-01
3.06833118e-01 5.26569895e-02 -3.54352176e-01 5.26749372e-01
7.71314859e-01 1.17703572e-01 2.18018308e-01 -3.87099415e-01
-8.96030068e-02 3.77155989e-01 -5.90014160e-01 -1.82742447e-01
1.51548314e+00 -2.29220614e-01 1.61868185e-01 1.07553089e+00
1.14090192e+00 -3.54580253e-01 -9.68618691e-01 -4.12504852e-01
3.21221143e-01 -3.39496464e-01 8.31992775e-02 -8.16350043e-01
-5.74756503e-01 1.03667331e+00 4.08737451e-01 2.19744042e-01
9.40060735e-01 -3.55154797e-02 4.51891094e-01 8.05616140e-01
3.76809120e-01 -1.33908486e+00 1.10113770e-01 6.95894361e-01
6.09848440e-01 -1.64882553e+00 6.53274283e-02 1.03782713e-01
-7.00198948e-01 1.21864522e+00 5.89426756e-01 -1.25805035e-01
9.21168685e-01 5.94767481e-02 1.04846403e-01 -1.22188747e-01
-1.12975478e+00 -2.28873536e-01 8.88293162e-02 4.69895750e-01
6.91421270e-01 -2.95330614e-01 -3.08642626e-01 4.99065787e-01
2.04808768e-02 1.47087321e-01 5.50527751e-01 1.29213607e+00
-5.90150118e-01 -1.44665873e+00 1.62968040e-02 5.23390710e-01
-4.77384180e-01 -2.47315183e-01 -3.69023234e-01 8.03088069e-01
-3.85734811e-02 9.32687581e-01 3.69651914e-02 1.00938147e-02
2.67776519e-01 8.21412742e-01 4.32534069e-01 -1.33731520e+00
-6.83166504e-01 -1.03017434e-01 -1.07571622e-02 -1.70408562e-01
-6.07722640e-01 -5.48724413e-01 -9.41959858e-01 -5.47373742e-02
-3.27506006e-01 2.82625318e-01 5.80111146e-01 1.05749309e+00
2.89368302e-01 3.98933798e-01 3.70504856e-01 -8.72492790e-01
-7.12575078e-01 -1.15032852e+00 -6.66271448e-01 4.67469603e-01
1.91283941e-01 -6.55831754e-01 -5.75345397e-01 -3.04299258e-02] | [10.706286430358887, 7.73544454574585] |
78f9bcb0-4e17-486b-b4f4-6e529cea7c1c | opt-one-shot-pose-controllable-talking-head | 2302.08197 | null | https://arxiv.org/abs/2302.08197v1 | https://arxiv.org/pdf/2302.08197v1.pdf | OPT: One-shot Pose-Controllable Talking Head Generation | One-shot talking head generation produces lip-sync talking heads based on arbitrary audio and one source face. To guarantee the naturalness and realness, recent methods propose to achieve free pose control instead of simply editing mouth areas. However, existing methods do not preserve accurate identity of source face when generating head motions. To solve the identity mismatch problem and achieve high-quality free pose control, we present One-shot Pose-controllable Talking head generation network (OPT). Specifically, the Audio Feature Disentanglement Module separates content features from audios, eliminating the influence of speaker-specific information contained in arbitrary driving audios. Later, the mouth expression feature is extracted from the content feature and source face, during which the landmark loss is designed to enhance the accuracy of facial structure and identity preserving quality. Finally, to achieve free pose control, controllable head pose features from reference videos are fed into the Video Generator along with the expression feature and source face to generate new talking heads. Extensive quantitative and qualitative experimental results verify that OPT generates high-quality pose-controllable talking heads with no identity mismatch problem, outperforming previous SOTA methods. | ['Jizhong Han', 'Jiao Dai', 'Cai Yu', 'Yesheng Chai', 'Xiaomeng Fu', 'Xi Wang', 'Jin Liu'] | 2023-02-16 | null | null | null | null | ['talking-head-generation'] | ['computer-vision'] | [ 3.68286371e-02 2.92317808e-01 -4.47214842e-02 -4.33001190e-01
-9.62044358e-01 -3.38843733e-01 3.28766257e-01 -8.13345432e-01
1.62371174e-01 5.64294934e-01 6.68274939e-01 6.29722476e-01
1.68249503e-01 -3.13029796e-01 -5.41930139e-01 -9.98924971e-01
3.63027513e-01 -1.37334302e-01 -2.71782547e-01 -1.84167176e-01
8.51192921e-02 3.74954104e-01 -2.02636266e+00 1.51827922e-02
8.06672096e-01 1.19165504e+00 6.05313107e-03 3.62280220e-01
5.74284270e-02 3.47704828e-01 -7.42847025e-01 -3.34104121e-01
2.15662092e-01 -5.29082477e-01 4.16633859e-03 4.62910175e-01
6.08554482e-01 -5.66981912e-01 -2.27877453e-01 1.08579862e+00
1.14688742e+00 -2.08227243e-02 4.70132351e-01 -1.65267265e+00
-3.91252905e-01 2.80621946e-01 -5.11990368e-01 -5.83035529e-01
8.93093169e-01 5.02940655e-01 7.52602458e-01 -1.08238447e+00
6.08957767e-01 1.60864413e+00 5.32305241e-01 9.20641840e-01
-1.15192544e+00 -1.28820050e+00 7.61726946e-02 8.08262154e-02
-1.77455342e+00 -1.44740045e+00 1.23506260e+00 -1.76949486e-01
1.45845667e-01 2.96732754e-01 7.28304744e-01 1.06283343e+00
-9.94548574e-02 6.90562665e-01 5.42011857e-01 -1.12753719e-01
-8.37427676e-02 1.68997809e-01 -5.54415047e-01 6.47130966e-01
-2.57017583e-01 1.81012645e-01 -9.58572865e-01 -2.04903278e-02
5.43275595e-01 -2.33717173e-01 -8.90089691e-01 -3.73841166e-01
-1.22387791e+00 6.03449702e-01 1.30967975e-01 1.04732208e-01
-1.79738119e-01 -1.90009370e-01 3.40152949e-01 1.11243449e-01
2.68902570e-01 7.20452443e-02 -5.13192080e-03 -8.72036964e-02
-1.02047801e+00 1.76667050e-01 6.01500571e-01 1.25716758e+00
7.12992847e-01 5.17009914e-01 -4.79867399e-01 8.99381518e-01
4.17742074e-01 8.50741744e-01 6.20143592e-01 -1.20595300e+00
2.93074399e-01 3.86213422e-01 -1.66066289e-02 -1.15850675e+00
-1.54215917e-01 -1.24702536e-01 -8.49437356e-01 -4.26164130e-03
-7.69849047e-02 -2.99325526e-01 -5.55936992e-01 2.11846304e+00
5.98402858e-01 9.69994888e-02 -8.12914670e-02 1.01622534e+00
9.15516615e-01 6.44490600e-01 -4.01688427e-01 -8.42192829e-01
1.40801334e+00 -6.69143796e-01 -1.36180520e+00 4.72772941e-02
8.33791643e-02 -8.40616286e-01 1.18771923e+00 3.16252410e-01
-1.00080097e+00 -6.58965290e-01 -8.88478100e-01 -9.79972188e-04
3.20093423e-01 3.35754871e-01 -1.74842943e-02 7.42701948e-01
-1.00564957e+00 6.26791641e-02 -1.84417382e-01 6.68300167e-02
1.41026482e-01 5.75682402e-01 -6.97224855e-01 1.87234417e-01
-1.19093871e+00 3.79246086e-01 -1.57311365e-01 4.45803516e-02
-6.61843002e-01 -7.69517303e-01 -1.09853303e+00 -1.66263670e-01
2.92902797e-01 -6.33248627e-01 1.28822851e+00 -1.09891081e+00
-2.32179642e+00 6.63688838e-01 -6.14076018e-01 1.87945127e-01
5.65765917e-01 -1.19314954e-01 -5.84759772e-01 3.02282155e-01
2.29859799e-01 9.89964664e-01 1.34703028e+00 -1.38150251e+00
-4.37294245e-01 -3.58856946e-01 -4.75317478e-01 3.80249888e-01
-5.51584303e-01 -5.44365458e-02 -5.19908071e-01 -5.84842861e-01
2.01269224e-01 -7.89952993e-01 5.24168134e-01 4.33958024e-01
-6.76171720e-01 2.50559002e-02 1.21780348e+00 -6.64212167e-01
1.11844695e+00 -2.50888157e+00 1.99664652e-01 -1.00254886e-01
1.62762776e-01 1.03668985e-03 -2.70246923e-01 6.45591412e-03
-2.56695822e-02 -2.26354793e-01 3.13411765e-02 -6.45212352e-01
-3.21114622e-02 -7.70461559e-02 -2.58235067e-01 6.95559561e-01
2.06138849e-01 6.09387279e-01 -6.39296830e-01 -8.29080999e-01
2.27687970e-01 7.71148145e-01 -7.98905313e-01 3.91400546e-01
-4.66478959e-04 5.85613668e-01 -2.32029974e-01 6.84241652e-01
7.94497132e-01 6.05187535e-01 -1.16555981e-01 -7.16762245e-01
-1.30681455e-01 4.58112769e-02 -1.33666778e+00 1.69726193e+00
-4.87860292e-01 4.47473913e-01 8.08880389e-01 -3.75868648e-01
1.20139420e+00 7.89451838e-01 5.35866559e-01 -6.08024359e-01
2.69338489e-01 1.46279335e-01 -2.50752181e-01 -6.40426159e-01
2.40467623e-01 -3.34402680e-01 -5.68994246e-02 7.69571960e-02
1.53804556e-01 -6.16984546e-01 -3.45968723e-01 -2.08599642e-01
3.43372256e-01 -1.12799160e-01 -7.25343004e-02 -1.42651081e-01
8.82702351e-01 -8.92061234e-01 9.05170977e-01 -2.63721496e-01
-3.79268587e-01 9.78831172e-01 2.85375208e-01 2.66509265e-01
-7.84409583e-01 -1.00010848e+00 1.05548561e-01 9.89729047e-01
2.50090927e-01 -3.69384885e-01 -1.13155401e+00 -1.23669118e-01
-2.12418675e-01 5.13235271e-01 -4.46600646e-01 -4.14162844e-01
-5.56976557e-01 2.61016544e-02 6.81675136e-01 1.26895875e-01
6.42715335e-01 -8.16157818e-01 -1.26698345e-01 1.24672286e-01
-6.29239559e-01 -9.95570779e-01 -1.36848474e+00 -7.67386377e-01
-2.10070878e-01 -7.32247949e-01 -1.02883267e+00 -8.84649038e-01
6.49367094e-01 1.38844758e-01 3.52315366e-01 -3.06526810e-01
-2.49947533e-01 2.13625073e-01 8.48677903e-02 -3.47241998e-01
-3.54532152e-01 -1.69997096e-01 6.52185559e-01 6.94119096e-01
-4.80210334e-02 -8.26293290e-01 -6.77079022e-01 5.61076581e-01
-5.37845075e-01 8.34349319e-02 2.65624195e-01 8.54886413e-01
3.10638249e-01 -1.53614834e-01 8.31012845e-01 1.64006174e-01
6.66369975e-01 3.19321677e-02 -3.68310213e-01 -1.56555712e-01
-2.14672849e-01 -2.04917818e-01 6.90179527e-01 -7.89694011e-01
-1.25663877e+00 2.82188863e-01 -2.63326049e-01 -8.80235016e-01
3.07791661e-02 -3.59753728e-01 -1.20003593e+00 1.20487414e-01
3.52776885e-01 4.94054526e-01 6.15765214e-01 -1.82900548e-01
5.24047017e-01 1.09696031e+00 9.29159164e-01 -2.66162068e-01
8.48804235e-01 5.07937968e-01 -2.98925072e-01 -1.08117604e+00
-4.00211036e-01 -2.12305859e-01 -3.51701498e-01 -5.96455514e-01
7.78879821e-01 -1.03687501e+00 -1.14578366e+00 7.75176883e-01
-1.13827097e+00 1.54000655e-01 -7.28833377e-02 4.00617570e-01
-8.37484062e-01 1.87765166e-01 -4.83697146e-01 -9.76114869e-01
-5.39268672e-01 -1.32776654e+00 1.60605288e+00 2.80255675e-01
-3.11583549e-01 -2.73811013e-01 -1.85395762e-01 4.05964315e-01
3.58704805e-01 1.31978899e-01 4.22488540e-01 1.06677666e-01
-3.24596822e-01 -2.53327578e-01 2.12278754e-01 1.70358703e-01
7.17545271e-01 2.27608234e-01 -1.32328641e+00 -3.32298785e-01
2.52415448e-01 -1.43072203e-01 2.55497783e-01 3.80229563e-01
6.04725778e-01 -9.21422303e-01 -2.35008255e-01 7.23988533e-01
7.35119402e-01 1.70118958e-01 5.43881655e-01 -4.62974727e-01
6.44552708e-01 1.14179754e+00 5.88863015e-01 5.63912868e-01
2.49603987e-01 8.87421846e-01 2.49293298e-01 9.15019512e-02
-3.81824464e-01 -6.49002612e-01 7.54007161e-01 9.49367404e-01
4.22745019e-01 -7.93258697e-02 -1.96540609e-01 5.30679524e-01
-1.29179251e+00 -9.86578822e-01 4.73508805e-01 2.10849524e+00
1.10841882e+00 -1.74625069e-01 2.50465482e-01 4.21215564e-01
1.15053320e+00 1.35049984e-01 -6.51759624e-01 1.14810549e-01
-1.44478977e-01 -3.11348796e-01 -1.53903410e-01 6.51286900e-01
-7.07671583e-01 7.97333837e-01 4.96558475e+00 9.79297876e-01
-1.55976820e+00 -1.40717447e-01 3.36974174e-01 -5.82176924e-01
-4.31640893e-01 -4.15323913e-01 -8.36490870e-01 5.31536579e-01
4.96897995e-01 -5.99616587e-01 2.07296848e-01 8.43169630e-01
7.18912601e-01 3.17026138e-01 -9.99606192e-01 1.33972704e+00
3.58907789e-01 -9.88126159e-01 1.57758489e-01 -4.07586694e-02
3.44661653e-01 -7.75768518e-01 1.85458913e-01 -2.28173062e-02
-3.87490630e-01 -8.45822930e-01 1.04968989e+00 6.66343331e-01
1.48022509e+00 -9.21721637e-01 1.21182397e-01 2.40116060e-01
-1.45654643e+00 -1.30685419e-01 7.12743402e-02 3.85760963e-01
4.83822763e-01 3.37960929e-01 -6.50515616e-01 3.19670498e-01
4.06210154e-01 4.98436153e-01 -2.21305452e-02 7.09078431e-01
-1.86902210e-01 1.76169783e-01 -3.18778366e-01 9.75648835e-02
-4.02374148e-01 1.04372792e-01 1.09071302e+00 7.89193630e-01
3.67401361e-01 1.45729352e-03 -1.17978046e-03 9.63448584e-01
-1.49477690e-01 1.79685429e-01 -7.73696542e-01 2.77557045e-01
8.07751715e-01 1.10816085e+00 1.87534466e-02 3.66376378e-02
7.05840290e-02 1.05554473e+00 -4.65079218e-01 2.25918055e-01
-6.99236691e-01 -7.29289174e-01 1.22211611e+00 3.90896946e-01
-1.44094033e-02 1.09224148e-01 1.16280295e-01 -1.08759439e+00
2.46988535e-01 -1.01391482e+00 -3.37333143e-01 -9.61238921e-01
-9.49097455e-01 5.68354845e-01 -2.29870230e-01 -1.51149929e+00
-5.16414106e-01 -8.39200392e-02 -8.06690454e-01 7.69510090e-01
-1.11927211e+00 -1.25706863e+00 -3.90415579e-01 9.26376581e-01
7.71051526e-01 -1.09197631e-01 7.15806961e-01 4.95592505e-01
-6.32378221e-01 1.23113823e+00 -1.86555386e-01 1.13775313e-01
1.08156288e+00 -4.96654600e-01 -3.64989750e-02 3.99120480e-01
-5.22110403e-01 5.92693150e-01 7.60331273e-01 -3.29720974e-01
-1.54552972e+00 -1.08189511e+00 6.71980143e-01 9.09354389e-02
1.10688522e-01 -6.58211112e-01 -7.44214833e-01 1.17373623e-01
-1.76290330e-02 -8.27699080e-02 3.63675237e-01 -7.15647280e-01
-2.32742682e-01 -6.50486588e-01 -1.36022258e+00 7.50169873e-01
1.07187843e+00 -7.42120087e-01 -4.62190777e-01 -1.36514500e-01
1.05524886e+00 -1.96371287e-01 -6.59697115e-01 5.26960254e-01
7.85950363e-01 -8.81536484e-01 7.03520477e-01 1.73395455e-01
5.79081886e-02 -4.65121001e-01 -4.81225327e-02 -1.17118192e+00
1.20738231e-01 -1.38837492e+00 1.77900612e-01 1.99446237e+00
5.56660034e-02 -5.62495351e-01 6.46742165e-01 3.56972307e-01
-1.78827360e-01 -3.22255701e-01 -1.11827481e+00 -7.48416007e-01
-1.67578548e-01 -1.97952092e-01 9.77845013e-01 7.46579170e-01
1.55994564e-01 5.39246857e-01 -6.17305517e-01 2.93149315e-02
7.17636049e-01 1.72209721e-02 1.04459858e+00 -1.04960847e+00
2.32275441e-01 -3.03134859e-01 -3.15277994e-01 -1.00826585e+00
4.56881136e-01 -2.80115694e-01 5.22367060e-01 -1.03193402e+00
-2.34877303e-01 3.63080268e-04 4.46699798e-01 1.38224512e-01
2.87207160e-02 1.54374406e-01 2.00843692e-01 1.43518567e-01
-4.84871008e-02 1.18202472e+00 1.61132050e+00 -1.33209035e-01
-3.97219181e-01 8.89670998e-02 -5.25505662e-01 7.68238068e-01
4.84197795e-01 -5.53512983e-02 -6.24114573e-01 -4.00287360e-02
-4.66631353e-01 5.69286764e-01 2.41940171e-01 -1.17531455e+00
3.28315109e-01 -8.65350291e-03 1.39585704e-01 -3.66762996e-01
8.29118907e-01 -6.50687754e-01 1.24687843e-01 2.91296691e-01
-2.28847414e-01 -3.44063967e-01 1.09467588e-01 3.51426989e-01
-3.48788321e-01 3.12728435e-01 1.14030468e+00 2.31670335e-01
-8.98085162e-02 5.62285721e-01 -2.51868308e-01 -8.22218414e-03
9.24490809e-01 -4.98163491e-01 1.09350994e-01 -1.00721359e+00
-5.68218887e-01 1.33817986e-01 5.54484427e-01 6.53041363e-01
7.02163219e-01 -1.60795653e+00 -7.03579426e-01 7.73414493e-01
8.56594071e-02 1.01900071e-01 3.94160807e-01 6.88683867e-01
2.24713534e-02 1.70138136e-01 -1.39305577e-01 -7.27674842e-01
-1.41722035e+00 4.86491293e-01 5.57380915e-01 6.70494795e-01
-5.24520099e-01 8.33256900e-01 5.10847509e-01 -3.31566691e-01
5.09077728e-01 1.36954403e-02 -1.61515281e-01 3.18629861e-01
6.95343256e-01 2.58697301e-01 -1.84053317e-01 -1.09602153e+00
-2.89361477e-01 8.49686921e-01 2.78913528e-01 -4.20617104e-01
9.42384362e-01 -5.78316748e-01 1.77109852e-01 3.05076599e-01
1.55754578e+00 5.66899180e-01 -1.48502612e+00 -1.34291556e-02
-6.74837530e-01 -5.28565049e-01 -1.24854326e-01 -2.75009751e-01
-1.35248303e+00 1.02993739e+00 6.39710248e-01 -3.84350836e-01
1.36019087e+00 -2.33849853e-01 9.82823670e-01 -2.80176736e-02
4.18068051e-01 -1.01013601e+00 3.53414685e-01 3.40568200e-02
1.29449821e+00 -8.30591202e-01 -4.43700671e-01 -6.01486027e-01
-7.76055753e-01 8.34058225e-01 7.99127817e-01 3.58506501e-01
6.40851080e-01 4.09534514e-01 1.95680663e-01 2.35464975e-01
-5.44179499e-01 9.64051262e-02 2.50682503e-01 9.73304689e-01
1.90741464e-01 -1.98969141e-01 9.55307484e-02 8.44478905e-01
-7.83412099e-01 -3.69289257e-02 1.35438845e-01 4.53704625e-01
-4.40406740e-01 -7.51932323e-01 -6.48479104e-01 -5.18691950e-02
-3.01096112e-01 8.15895200e-03 -5.39049387e-01 3.77898306e-01
2.05281243e-01 1.19392550e+00 8.03450346e-02 -6.60523653e-01
5.56504369e-01 1.79928541e-01 2.85022199e-01 -3.09754759e-01
-2.11634442e-01 5.23715556e-01 -6.04776964e-02 -5.96503496e-01
-1.31170675e-01 -4.22775656e-01 -1.18107533e+00 -4.35182005e-01
-6.59267128e-01 1.60774812e-01 3.64304960e-01 5.39711058e-01
6.70918763e-01 1.95675120e-01 1.18963087e+00 -1.34119368e+00
-4.38465834e-01 -1.11218250e+00 -6.90250397e-01 4.60097671e-01
7.75521874e-01 -7.52280712e-01 -7.50801563e-01 2.61880040e-01] | [13.229835510253906, -0.4219840168952942] |
36bad896-db02-4af1-9991-0b741e15b7bf | on-the-automatic-generation-of-medical | 1711.08195 | null | http://arxiv.org/abs/1711.08195v3 | http://arxiv.org/pdf/1711.08195v3.pdf | On the Automatic Generation of Medical Imaging Reports | Medical imaging is widely used in clinical practice for diagnosis and
treatment. Report-writing can be error-prone for unexperienced physicians, and
time- consuming and tedious for experienced physicians. To address these
issues, we study the automatic generation of medical imaging reports. This task
presents several challenges. First, a complete report contains multiple
heterogeneous forms of information, including findings and tags. Second,
abnormal regions in medical images are difficult to identify. Third, the re-
ports are typically long, containing multiple sentences. To cope with these
challenges, we (1) build a multi-task learning framework which jointly performs
the pre- diction of tags and the generation of para- graphs, (2) propose a
co-attention mechanism to localize regions containing abnormalities and
generate narrations for them, (3) develop a hierarchical LSTM model to generate
long paragraphs. We demonstrate the effectiveness of the proposed methods on
two publicly available datasets. | ['Baoyu Jing', 'Eric Xing', 'Pengtao Xie'] | 2017-11-22 | on-the-automatic-generation-of-medical-1 | https://aclanthology.org/P18-1240 | https://aclanthology.org/P18-1240.pdf | acl-2018-7 | ['medical-report-generation'] | ['medical'] | [ 4.16167915e-01 1.17010176e-01 -1.79550517e-02 -3.60071599e-01
-1.42245877e+00 -3.63544613e-01 2.50068665e-01 4.83368576e-01
-2.69622579e-02 9.27508116e-01 6.19133294e-01 -3.30254972e-01
1.10008776e-01 -7.45390058e-01 -5.51891744e-01 -6.69001460e-01
-1.52931675e-01 3.16227853e-01 6.81092292e-02 5.60855031e-01
3.81837279e-01 2.40320250e-01 -9.67782676e-01 6.29631162e-01
8.16141188e-01 8.23287785e-01 4.74223197e-01 7.10702121e-01
-1.11778013e-01 1.14068854e+00 -7.91016281e-01 -3.84986192e-01
-5.45636117e-01 -6.91566646e-01 -7.83526957e-01 6.44003689e-01
1.21475391e-01 -3.90561581e-01 -3.24656755e-01 1.13570380e+00
7.48325586e-01 -1.03166126e-01 8.13865364e-01 -6.89836860e-01
-9.70818639e-01 8.78500581e-01 -7.98683882e-01 5.21731734e-01
1.94901481e-01 -2.05213930e-02 6.48271501e-01 -7.00285554e-01
7.39997864e-01 9.13291991e-01 3.95165056e-01 5.77086926e-01
-7.11714983e-01 -5.27314603e-01 8.77659991e-02 6.13947921e-02
-1.37567627e+00 -3.87465298e-01 6.89848542e-01 -6.46397710e-01
4.18723136e-01 1.35078713e-01 3.86207342e-01 1.38943827e+00
6.34000659e-01 5.33201516e-01 1.01165688e+00 1.36588365e-02
-2.23356690e-02 5.60969040e-02 -5.45674376e-02 1.03413916e+00
1.90823659e-01 -3.16851854e-01 -2.73707569e-01 -2.10022479e-01
8.62939060e-01 3.05409133e-01 -4.38964754e-01 4.76986498e-01
-1.53505039e+00 7.77850389e-01 2.51451969e-01 6.23714626e-01
-7.29998648e-01 5.76854916e-03 5.81522405e-01 -2.29953259e-01
6.93640590e-01 1.31504998e-01 1.04049416e-02 7.08944723e-02
-1.00533664e+00 2.62664091e-02 3.50592822e-01 9.97544646e-01
1.55001745e-01 -1.68658540e-01 -7.63260067e-01 8.95102620e-01
2.22223043e-01 4.36066091e-01 8.39101374e-01 -3.61619264e-01
7.96848774e-01 2.95174122e-01 1.28841400e-02 -1.36935031e+00
-4.70791340e-01 -7.22123146e-01 -1.33285463e+00 -4.57356066e-01
-2.62065142e-01 -4.97708112e-01 -9.05757606e-01 1.36055315e+00
1.04121797e-01 3.08005422e-01 5.95362149e-02 7.98607528e-01
1.43294990e+00 9.19410050e-01 4.48472559e-01 -5.26102245e-01
1.63800240e+00 -1.05184937e+00 -1.10555708e+00 -2.21389100e-01
6.36333823e-01 -8.63260150e-01 8.26066017e-01 1.37380049e-01
-1.34657216e+00 -5.31838775e-01 -9.64622736e-01 -5.31296469e-02
-1.15958743e-01 6.66455925e-01 2.37784326e-01 -1.01418629e-01
-8.01680446e-01 3.68235528e-01 -8.03850114e-01 -1.70172676e-01
6.84314013e-01 -3.31074260e-02 -2.83356518e-01 -1.30501240e-01
-9.54473138e-01 7.80483246e-01 7.14152455e-01 2.53121912e-01
-8.86251569e-01 -6.37249231e-01 -1.00915945e+00 1.78250484e-02
4.24136311e-01 -9.15213585e-01 1.29774380e+00 -4.44949448e-01
-1.17319882e+00 1.09960532e+00 -2.70248502e-01 -2.12640211e-01
4.78233546e-01 5.90169281e-02 -6.53715551e-01 2.10659966e-01
4.87522453e-01 3.55937302e-01 6.58233583e-01 -1.14488411e+00
-6.53979421e-01 -3.76136214e-01 -2.68346250e-01 2.00077623e-01
-1.14364810e-01 8.12336504e-02 -5.27045727e-01 -1.29042697e+00
1.24502763e-01 -6.02034390e-01 -5.21342516e-01 -3.18637937e-01
-1.01320028e+00 -1.54437944e-01 4.87819910e-01 -9.42171156e-01
1.49042392e+00 -2.14116478e+00 -1.25697826e-03 -1.79113626e-01
5.47489047e-01 5.91749139e-02 1.12689041e-01 1.56851038e-01
-6.33794218e-02 4.28840160e-01 -2.96620727e-01 -4.07068700e-01
-3.59478861e-01 8.35220590e-02 -3.22197706e-01 1.74594253e-01
2.83690900e-01 8.81067634e-01 -8.45497429e-01 -1.13475227e+00
-1.15766950e-01 2.82777190e-01 -4.00481783e-02 6.09753788e-01
-9.60953012e-02 6.90657496e-01 -9.75721180e-01 6.83726549e-01
3.82777750e-01 -9.56849098e-01 -1.22270532e-01 -4.72000003e-01
9.12134349e-03 3.18026304e-01 -7.18068898e-01 1.82153583e+00
-5.31853795e-01 1.16103224e-01 -2.08648026e-01 -6.97714806e-01
6.40281975e-01 6.84480369e-01 4.51706797e-01 -3.95274013e-01
1.57824770e-01 1.43147916e-01 -3.17578852e-01 -1.20607185e+00
1.84057906e-01 -3.66098434e-01 -1.25623241e-01 6.96660638e-01
-3.68532836e-02 3.71611491e-02 1.96902379e-01 1.62075564e-01
1.33575177e+00 -2.38459930e-01 4.00777608e-01 2.24975064e-01
7.13358104e-01 -7.72867724e-02 6.82018518e-01 5.67799985e-01
6.25332445e-02 9.14833367e-01 4.30775821e-01 -3.06840926e-01
-7.87656724e-01 -9.71101463e-01 1.59750924e-01 4.86182839e-01
-1.21026956e-01 -3.15576762e-01 -5.69122970e-01 -1.00413847e+00
-4.46382225e-01 6.69062495e-01 -6.54941559e-01 -2.09407657e-02
-5.12556136e-01 -8.30088854e-01 3.71064991e-01 6.29622102e-01
5.63736200e-01 -1.35829151e+00 -4.62771714e-01 3.83563459e-01
-5.61260760e-01 -1.33872569e+00 -8.03630114e-01 -2.05899581e-01
-7.70733953e-01 -9.20120060e-01 -1.03753960e+00 -1.01079118e+00
1.16434479e+00 -5.29547594e-03 1.20833135e+00 2.97915041e-01
-4.66607809e-01 7.90991709e-02 -4.83959347e-01 -4.34971988e-01
-5.92410624e-01 1.74143896e-01 -6.07399404e-01 -5.02774678e-03
-1.09611981e-01 -3.92873675e-01 -5.48397005e-01 -2.26916090e-01
-1.10879874e+00 5.30618072e-01 1.04629064e+00 9.35684800e-01
9.23636198e-01 1.46158203e-01 6.22653961e-01 -1.39911962e+00
9.47894156e-01 -7.42391884e-01 -2.08122090e-01 5.19020259e-01
-3.62232298e-01 -6.98814318e-02 5.52287877e-01 -1.84526548e-01
-1.24335492e+00 -7.66615272e-02 -1.87753662e-01 -1.67296112e-01
-2.84372002e-01 8.77119362e-01 2.50522345e-01 2.29257643e-01
2.72211909e-01 5.24154961e-01 -3.42000365e-01 -2.79218763e-01
4.02887762e-02 7.16415644e-01 6.49970710e-01 -2.94149548e-01
6.55378282e-01 2.00059682e-01 -6.57081679e-02 -2.76286483e-01
-1.49465001e+00 -6.28511980e-02 -5.31362176e-01 -2.94571728e-01
1.14639878e+00 -6.49173260e-01 -3.00104618e-01 1.37242526e-01
-1.59225821e+00 1.31938204e-01 2.61503786e-01 5.21846235e-01
-2.40208074e-01 4.34507489e-01 -8.91481936e-01 -4.01214272e-01
-8.03899348e-01 -1.29125404e+00 1.11626160e+00 3.88274819e-01
-5.90150096e-02 -1.16078603e+00 -1.44268200e-01 4.19108242e-01
-1.88367034e-03 6.04673207e-01 1.08484793e+00 -4.29643661e-01
-5.29532552e-01 -2.99424119e-02 -5.48276842e-01 -1.34667726e-02
4.39997822e-01 -5.85416928e-02 -5.99425852e-01 -8.53937864e-02
9.83055010e-02 -3.48350435e-01 8.20955634e-01 4.38370675e-01
1.80749381e+00 -5.20026684e-01 -4.89440441e-01 4.24687415e-01
1.11641371e+00 3.29520553e-01 2.29835019e-01 -1.71133026e-01
7.03519285e-01 5.89053035e-01 6.61462605e-01 6.10216320e-01
7.63527334e-01 1.40702486e-01 1.27852801e-02 -4.02952313e-01
-1.98951229e-01 -3.52826059e-01 -4.48025204e-02 1.43727458e+00
3.09101287e-02 -4.52206105e-01 -9.24613893e-01 7.17844844e-01
-1.83504295e+00 -7.14317143e-01 -1.39925897e-01 1.65324533e+00
1.14652109e+00 3.03088166e-02 -3.48373324e-01 -3.95961910e-01
8.09816658e-01 3.18581909e-01 -3.83469880e-01 8.43549669e-02
1.75694525e-01 4.32875119e-02 3.15942079e-01 1.48810372e-01
-1.32810879e+00 6.43057585e-01 6.24759769e+00 6.55374587e-01
-1.26647735e+00 5.05322933e-01 1.02373135e+00 1.55333383e-02
-3.09535533e-01 -5.86782575e-01 -5.98523259e-01 8.60292971e-01
6.64103568e-01 -2.96024323e-01 -2.11126223e-01 5.41727126e-01
3.54970157e-01 1.18136466e-01 -1.01023841e+00 1.11968493e+00
5.60930490e-01 -1.53379941e+00 2.21919522e-01 -1.63491443e-01
6.74679220e-01 -2.19625264e-01 1.13471262e-01 1.50510356e-01
1.91727638e-01 -1.00631618e+00 4.74348754e-01 6.92584932e-01
9.50475514e-01 -5.74648261e-01 9.86832738e-01 2.99201906e-01
-1.02339721e+00 2.18560264e-01 -3.68550330e-01 5.79311669e-01
4.97035295e-01 7.82272995e-01 -9.30079997e-01 7.58043110e-01
4.05281752e-01 7.71849692e-01 -5.92732847e-01 1.11475897e+00
-4.79354501e-01 4.83660996e-01 2.56780326e-01 -4.26806286e-02
3.08389962e-01 1.41304478e-01 2.30284110e-01 1.37677002e+00
7.53872156e-01 4.62970972e-01 6.01444900e-01 1.08103931e+00
-2.47415051e-01 3.43194276e-01 -5.03647268e-01 -4.45560455e-01
2.36332670e-01 1.52700353e+00 -1.01947057e+00 -6.12648726e-01
-3.51780474e-01 9.52249885e-01 2.08641440e-01 2.95227259e-01
-1.02570641e+00 -4.12730366e-01 -2.59125501e-01 -2.82493420e-03
2.76084859e-02 -4.32804488e-02 -2.97651291e-01 -1.30859017e+00
1.67425811e-01 -6.74380541e-01 6.00743771e-01 -9.86652136e-01
-1.45793343e+00 9.51924741e-01 -3.76098394e-01 -1.45310795e+00
-3.45467508e-01 -1.21057771e-01 -6.44205749e-01 6.31603062e-01
-1.55803680e+00 -1.30767679e+00 -5.24163425e-01 4.43992287e-01
7.25711882e-01 -1.26185596e-01 9.13702190e-01 6.83461607e-01
-7.32244790e-01 3.91632318e-01 -4.05465782e-01 2.95906603e-01
7.59660840e-01 -1.12847090e+00 6.02423698e-02 8.36274683e-01
6.88376427e-02 5.14354765e-01 4.59140092e-01 -8.51867378e-01
-8.98636341e-01 -1.41637540e+00 1.04688895e+00 2.62771011e-03
5.91114223e-01 2.61994228e-02 -9.66495574e-01 6.98786378e-01
2.16123015e-01 2.56471485e-02 9.06036139e-01 -2.29193687e-01
8.41326341e-02 1.91719010e-02 -9.83222663e-01 4.87737060e-01
7.25849748e-01 -4.48140353e-01 -5.24592280e-01 9.31286156e-01
8.84677470e-01 -7.37524807e-01 -1.10485435e+00 3.33393365e-01
7.17845783e-02 -5.95798969e-01 5.70658505e-01 -5.29582798e-01
8.80744100e-01 -1.20254479e-01 3.98739755e-01 -1.34003448e+00
-3.53413194e-01 -2.55672008e-01 1.06938116e-01 1.40866411e+00
6.77291691e-01 -3.15468878e-01 4.79156733e-01 3.01079184e-01
-4.26331103e-01 -1.06985414e+00 -5.80155253e-01 -3.80697474e-02
-2.24661529e-01 -9.30455700e-02 3.02131921e-01 1.04290950e+00
-1.90686241e-01 4.07417715e-01 -3.93438876e-01 3.31186324e-01
4.23955798e-01 3.23801279e-01 1.38241306e-01 -9.89174068e-01
-2.68791437e-01 -2.71954209e-01 1.67792905e-02 -8.09288740e-01
9.34426561e-02 -8.34520876e-01 2.65216410e-01 -2.02099395e+00
5.57561338e-01 -3.85211855e-01 -2.77303785e-01 6.06043458e-01
-4.33931470e-01 6.45196065e-02 -2.59105504e-01 1.64805114e-01
-7.71155119e-01 3.09755415e-01 1.51554453e+00 -3.64073247e-01
1.61373317e-01 3.61884609e-02 -7.79287457e-01 7.52610147e-01
6.34211600e-01 -6.72825456e-01 -4.03886884e-01 -6.05623662e-01
1.79642960e-01 6.28879964e-01 2.48519629e-01 -8.89030457e-01
2.56508976e-01 -2.24567622e-01 5.20487905e-01 -9.39777374e-01
-7.13831112e-02 -4.67503756e-01 -1.58120900e-01 5.14773607e-01
-5.13932705e-01 3.77230227e-01 -3.33840325e-02 4.45999235e-01
-5.63775837e-01 -3.78628910e-01 5.69426537e-01 -6.69828773e-01
-2.46250659e-01 6.18046403e-01 -3.85327697e-01 1.16528146e-01
1.06233299e+00 4.34693873e-01 -3.58490795e-01 -3.12661707e-01
-7.91556001e-01 4.51804012e-01 -1.25697210e-01 3.81973565e-01
9.21466231e-01 -1.33405697e+00 -9.83412743e-01 -1.85040370e-01
1.13607481e-01 3.06177497e-01 8.02724302e-01 8.82111788e-01
-6.86946929e-01 3.55492413e-01 1.63083866e-01 -3.81484896e-01
-1.40727854e+00 5.91331363e-01 -5.77227119e-03 -6.81719005e-01
-7.78231859e-01 8.71743739e-01 3.03474009e-01 -1.85834803e-02
1.62467420e-01 -3.91497970e-01 -4.38981354e-01 1.27533957e-01
7.30426073e-01 -2.23742098e-01 -1.58037953e-02 -4.52104360e-01
-1.11702837e-01 5.35612166e-01 -4.38393652e-01 -3.78672779e-02
1.41653883e+00 -1.09403647e-01 -3.94084275e-01 4.71596688e-01
1.09235299e+00 -6.88576698e-02 -6.03180766e-01 -2.56012499e-01
-1.22511081e-01 -7.54081830e-02 1.75371110e-01 -8.87507737e-01
-1.28342867e+00 1.09258890e+00 2.54207224e-01 1.89331934e-01
1.06505799e+00 1.22060552e-01 1.12613046e+00 1.06511332e-01
1.06944121e-01 -8.22566271e-01 2.38039419e-01 4.34511378e-02
1.08159578e+00 -1.14278078e+00 5.67001887e-02 -4.51811194e-01
-8.78061712e-01 1.16080379e+00 6.54850423e-01 2.24303260e-01
5.51680923e-01 2.50006318e-01 2.20065862e-01 -4.78159964e-01
-7.80500293e-01 4.95777093e-02 4.38644409e-01 3.39048952e-01
7.42552161e-01 4.95183580e-02 -5.23722053e-01 9.08542395e-01
4.19597812e-02 1.18194908e-01 5.91122389e-01 9.29188967e-01
-1.70136273e-01 -8.71461749e-01 -3.73486876e-01 6.97376311e-01
-9.95201290e-01 -1.47439405e-01 -7.99914002e-02 1.65360436e-01
2.98862666e-01 8.05071235e-01 -1.11489713e-01 -3.94948155e-01
6.16322942e-02 -1.43930644e-01 1.28938377e-01 -1.06983376e+00
-3.86638790e-01 3.48204017e-01 6.19796850e-02 -1.88991711e-01
-3.22818756e-01 -4.63616878e-01 -1.35371268e+00 2.75602549e-01
-3.04126646e-02 1.33232623e-01 5.15642524e-01 1.09347367e+00
6.39559329e-01 1.28788149e+00 5.08083284e-01 -2.61772186e-01
-3.63410860e-01 -1.16689157e+00 -4.44329441e-01 4.51183170e-01
2.76857048e-01 -5.59186518e-01 1.05425268e-01 4.83878940e-01] | [15.044610023498535, -1.4010933637619019] |
64d90caf-6803-457d-9201-5fb09bddc675 | cluster-guided-contrastive-graph-clustering | 2301.01098 | null | https://arxiv.org/abs/2301.01098v1 | https://arxiv.org/pdf/2301.01098v1.pdf | Cluster-guided Contrastive Graph Clustering Network | Benefiting from the intrinsic supervision information exploitation capability, contrastive learning has achieved promising performance in the field of deep graph clustering recently. However, we observe that two drawbacks of the positive and negative sample construction mechanisms limit the performance of existing algorithms from further improvement. 1) The quality of positive samples heavily depends on the carefully designed data augmentations, while inappropriate data augmentations would easily lead to the semantic drift and indiscriminative positive samples. 2) The constructed negative samples are not reliable for ignoring important clustering information. To solve these problems, we propose a Cluster-guided Contrastive deep Graph Clustering network (CCGC) by mining the intrinsic supervision information in the high-confidence clustering results. Specifically, instead of conducting complex node or edge perturbation, we construct two views of the graph by designing special Siamese encoders whose weights are not shared between the sibling sub-networks. Then, guided by the high-confidence clustering information, we carefully select and construct the positive samples from the same high-confidence cluster in two views. Moreover, to construct semantic meaningful negative sample pairs, we regard the centers of different high-confidence clusters as negative samples, thus improving the discriminative capability and reliability of the constructed sample pairs. Lastly, we design an objective function to pull close the samples from the same cluster while pushing away those from other clusters by maximizing and minimizing the cross-view cosine similarity between positive and negative samples. Extensive experimental results on six datasets demonstrate the effectiveness of CCGC compared with the existing state-of-the-art algorithms. | ['En Zhu', 'Liming Fang', 'Xinwang Liu', 'Qun Zheng', 'Wenxuan Tu', 'Siwei Wang', 'Sihang Zhou', 'Yue Liu', 'Xihong Yang'] | 2023-01-03 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [-1.25671715e-01 1.48573726e-01 -3.58745068e-01 -4.58260089e-01
-5.32804787e-01 -2.98861474e-01 3.90684992e-01 6.61275387e-02
-1.54914185e-01 4.92605478e-01 1.16604648e-03 3.24818105e-01
-4.34986353e-01 -7.87553728e-01 -6.01493955e-01 -1.13473415e+00
-1.95228800e-01 6.20691299e-01 1.26690879e-01 2.00961664e-01
6.33309111e-02 1.80316359e-01 -1.39438879e+00 1.54938146e-01
1.34890068e+00 8.18086028e-01 1.98056713e-01 -5.09518348e-02
-3.37768942e-02 4.17310357e-01 -6.05025470e-01 -4.25781459e-01
3.03979158e-01 -5.73052108e-01 -3.98034096e-01 4.90575820e-01
2.91488945e-01 1.69933170e-01 -8.09385106e-02 1.50595605e+00
2.98338503e-01 -5.97661994e-02 4.62448925e-01 -1.78622401e+00
-8.46989810e-01 6.77865684e-01 -1.18230999e+00 -4.58635017e-02
-9.09605026e-02 2.30107769e-01 1.11316919e+00 -1.04017305e+00
6.36919022e-01 1.17477012e+00 4.75604951e-01 5.22711575e-01
-1.30734861e+00 -9.03426528e-01 5.25628388e-01 1.55623883e-01
-1.53365552e+00 -2.81353205e-01 1.31046879e+00 -2.99622566e-01
1.60599127e-01 1.96540639e-01 9.53288913e-01 1.16108119e+00
-6.08490407e-01 9.71131563e-01 7.59809613e-01 4.68343124e-02
3.47585797e-01 3.38431418e-01 4.18306030e-02 8.86321008e-01
4.23363477e-01 -1.76681459e-01 -4.61167514e-01 -9.65052173e-02
6.04401588e-01 2.90204883e-01 -4.83332455e-01 -1.10811830e+00
-1.33224392e+00 8.78526568e-01 7.71237850e-01 3.73243511e-01
-2.34822452e-01 -1.18726991e-01 4.16835427e-01 7.16539845e-02
4.69110996e-01 2.15710476e-01 -2.97451675e-01 3.37914795e-01
-8.48290801e-01 -1.76351696e-01 3.82526517e-01 1.17366767e+00
1.02143204e+00 -1.35733128e-01 1.23889670e-02 7.93646455e-01
4.00689572e-01 3.06086570e-01 3.33433628e-01 -8.21757436e-01
7.13047981e-01 1.18899906e+00 -2.84179360e-01 -1.52582407e+00
-1.46981090e-01 -8.10751677e-01 -1.39683032e+00 -1.10829301e-01
3.55645180e-01 -1.03576981e-01 -9.68633890e-01 1.86166382e+00
4.57399040e-01 2.38125414e-01 -3.45435217e-02 1.03923726e+00
6.15618706e-01 3.53968829e-01 -2.03184187e-01 -3.95123750e-01
8.76554787e-01 -1.03249383e+00 -5.75002909e-01 -2.23068640e-01
6.10997140e-01 -2.53769338e-01 1.23983884e+00 3.64966393e-01
-8.32937837e-01 -4.62063581e-01 -1.16877151e+00 3.43485355e-01
-1.46617621e-01 2.49053061e-01 7.27403939e-01 3.15095395e-01
-7.43053019e-01 6.35866523e-01 -8.35750937e-01 -9.58415791e-02
8.34034503e-01 4.17617887e-01 -4.30613935e-01 -2.29528382e-01
-8.44166517e-01 6.23928308e-02 3.41053635e-01 1.96563527e-01
-7.74172604e-01 -6.60179257e-01 -7.53565609e-01 1.49095118e-01
7.04442382e-01 -5.45233488e-01 3.23708683e-01 -1.39910972e+00
-9.02143776e-01 8.21155608e-01 -1.48307666e-01 -8.29044133e-02
6.13199532e-01 1.16326392e-01 -5.55901945e-01 3.34053844e-01
5.37150323e-01 7.40564525e-01 8.76527488e-01 -1.81092572e+00
-6.34848893e-01 -6.90533400e-01 -4.07129765e-01 2.98442900e-01
-6.32735610e-01 -5.89484394e-01 -8.51035893e-01 -6.42770588e-01
5.27108133e-01 -7.19726980e-01 -3.44636530e-01 1.02660783e-01
-7.79263318e-01 -1.96423993e-01 1.13171661e+00 -2.02387705e-01
1.16007495e+00 -2.37115884e+00 3.03385854e-01 6.55312896e-01
6.24119282e-01 6.89114407e-02 -1.76366434e-01 1.77212991e-02
-2.73193210e-01 3.33718151e-01 -4.91484106e-01 -3.29018444e-01
-1.81318983e-01 2.35201925e-01 1.17473774e-01 4.62428510e-01
3.86199147e-01 6.76888824e-01 -1.37881804e+00 -5.36311388e-01
2.28217859e-02 2.32135594e-01 -4.35716271e-01 1.23214155e-01
9.15373713e-02 3.77836049e-01 -4.99081463e-01 5.38250923e-01
7.99214125e-01 -7.15741813e-01 4.65701610e-01 -3.94073546e-01
5.81788778e-01 4.66375314e-02 -1.23381853e+00 1.56599176e+00
3.52414735e-02 4.68672514e-02 3.33614320e-01 -1.42551947e+00
9.24843729e-01 2.79837679e-02 5.55521011e-01 -5.27929008e-01
-2.85745114e-01 1.77986920e-01 8.40983465e-02 -3.99497837e-01
1.19507998e-01 -4.38805446e-02 1.95922866e-01 3.29376429e-01
-4.01210487e-02 3.17614049e-01 2.59083956e-01 6.48164451e-01
6.76440775e-01 -1.98120415e-01 -2.11480126e-01 -2.62062073e-01
6.01058543e-01 -2.56880522e-01 1.07227433e+00 2.65096605e-01
-4.18373197e-01 6.99243009e-01 8.39315236e-01 -1.13079816e-01
-8.31212103e-01 -1.25979304e+00 1.93196148e-01 6.49511635e-01
4.87152785e-01 -4.24972296e-01 -8.05108190e-01 -1.37043905e+00
-3.62755731e-02 3.00181985e-01 -8.05001318e-01 -5.27600348e-01
-5.19909441e-01 -9.73036528e-01 1.65018767e-01 6.42262518e-01
4.20723498e-01 -8.30288827e-01 2.23216906e-01 -8.88716429e-02
-2.78247446e-01 -7.29381442e-01 -5.97412765e-01 2.31116474e-01
-8.85059118e-01 -1.34879625e+00 -5.02095342e-01 -1.03545964e+00
1.25390863e+00 7.12578416e-01 1.16563034e+00 3.30329835e-01
1.93320900e-01 -8.27917177e-03 -2.77081788e-01 1.64697826e-01
-1.71439201e-01 9.33338255e-02 1.02978937e-01 3.41319203e-01
5.42238235e-01 -7.73405254e-01 -6.70559168e-01 3.79264981e-01
-7.90818989e-01 6.53767353e-03 5.21717429e-01 9.31532800e-01
7.67423093e-01 2.95754164e-01 1.07235181e+00 -1.00230360e+00
4.27310735e-01 -6.95263326e-01 -3.25599164e-01 2.45059520e-01
-8.05285156e-01 -8.26042071e-02 9.81923878e-01 -3.87764037e-01
-7.22735107e-01 4.18412499e-02 3.55788380e-01 -8.10275555e-01
-4.18162309e-02 2.98825741e-01 -5.79607666e-01 4.21330154e-01
3.44283938e-01 2.13173166e-01 2.49562725e-01 -2.48451084e-01
3.70055765e-01 3.32956672e-01 4.50121731e-01 -4.15692329e-01
9.51647997e-01 8.25007141e-01 -1.58067897e-01 -5.18271565e-01
-8.50604951e-01 -5.18982828e-01 -6.84715569e-01 -1.28216773e-01
7.66292512e-01 -8.68923068e-01 -5.29503822e-01 2.64911354e-01
-5.71995795e-01 8.81880596e-02 -3.54247272e-01 3.33558291e-01
-1.57803357e-01 7.08236217e-01 -4.58720624e-01 -6.75795138e-01
-1.98204786e-01 -1.13192451e+00 1.01265979e+00 2.54007369e-01
7.64294015e-03 -1.00783622e+00 -3.60175729e-01 3.72825235e-01
-2.75688916e-01 2.20804244e-01 9.19585586e-01 -7.38857388e-01
-7.04627931e-01 -2.49946401e-01 -2.97694713e-01 4.36346173e-01
3.33247602e-01 2.78243273e-01 -7.58759618e-01 -5.15852511e-01
-9.26480815e-02 -2.97352642e-01 8.80227923e-01 2.45419443e-01
1.32679248e+00 -3.54047239e-01 -7.18280852e-01 6.59436882e-01
1.21658111e+00 -2.64423853e-03 3.38988066e-01 -3.83652467e-03
1.15645814e+00 8.48390877e-01 5.40859699e-01 2.53559619e-01
3.42998475e-01 1.94555953e-01 7.08988905e-01 -1.92600742e-01
2.71941036e-01 -6.45575106e-01 1.72270074e-01 1.01607645e+00
1.40709847e-01 -1.10070638e-01 -5.68016410e-01 8.21414769e-01
-1.93089116e+00 -9.70030546e-01 -3.01441610e-01 2.44490814e+00
7.24821568e-01 3.07666361e-01 2.33881965e-01 1.91281363e-01
1.16714013e+00 3.09984207e-01 -7.34748244e-01 5.15718639e-01
-1.40979692e-01 -2.92203248e-01 6.46849498e-02 1.37562618e-01
-9.85612869e-01 6.76548362e-01 4.65234852e+00 9.96548653e-01
-9.18822765e-01 -1.94024250e-01 1.05494487e+00 -3.62120360e-01
-7.44317353e-01 4.35055159e-02 -5.93288839e-01 7.68085837e-01
3.13032210e-01 9.77449268e-02 3.02809417e-01 8.67587090e-01
1.86979517e-01 1.16655573e-01 -1.20252490e+00 8.94931138e-01
9.58486926e-03 -1.25540936e+00 6.73110560e-02 1.16809167e-01
8.92875373e-01 -1.01859376e-01 8.48715678e-02 1.76982209e-01
3.56967866e-01 -6.94049597e-01 4.72404540e-01 2.30109245e-01
5.49658120e-01 -1.12524402e+00 4.42358226e-01 5.28361261e-01
-1.08676767e+00 2.99889296e-02 -4.58064586e-01 4.08423126e-01
-1.18688136e-01 1.16317618e+00 -5.33233762e-01 6.66518569e-01
9.46874022e-01 1.29279649e+00 -7.35657036e-01 6.82110310e-01
-2.80054152e-01 5.63820004e-01 -2.88675219e-01 1.47144318e-01
2.78866559e-01 -8.48672628e-01 5.14929712e-01 6.94800675e-01
1.59424677e-01 -3.59675705e-01 3.52667212e-01 1.20808113e+00
-3.10363740e-01 -8.40112101e-03 -6.22094214e-01 5.69674373e-02
7.07348526e-01 1.49657762e+00 -9.04390931e-01 -5.72246969e-01
-3.98365378e-01 1.00922799e+00 6.64400995e-01 5.11845529e-01
-7.49367535e-01 -2.52404988e-01 5.21119058e-01 1.01884112e-01
3.25041682e-01 1.49415731e-01 -4.72230881e-01 -1.33862460e+00
5.14109433e-01 -8.39979887e-01 4.86554295e-01 -4.37382430e-01
-1.74244368e+00 3.20811987e-01 -4.07395601e-01 -1.50208080e+00
2.15661332e-01 -9.39587578e-02 -7.48490572e-01 6.94777906e-01
-1.06794822e+00 -9.58406150e-01 -3.53237510e-01 5.46289861e-01
1.02032505e-01 -9.28530470e-02 3.80373508e-01 3.97587061e-01
-8.24913800e-01 7.41451144e-01 1.13271751e-01 3.53944719e-01
7.06199646e-01 -1.35481000e+00 1.17475785e-01 9.12959814e-01
6.15084507e-02 7.70096421e-01 1.05300903e-01 -8.50345969e-01
-9.74799395e-01 -1.25228548e+00 3.88639241e-01 -1.90832868e-01
5.77435970e-01 -6.28782332e-01 -1.27239084e+00 5.41621745e-01
5.33805527e-02 2.68213809e-01 7.71487176e-01 1.03851460e-01
-5.18949747e-01 -3.26507241e-01 -1.21419668e+00 7.54635096e-01
1.19447827e+00 -3.02540779e-01 -2.13426441e-01 4.23621029e-01
8.00800622e-01 2.83565372e-01 -7.86381185e-01 5.62508523e-01
9.84395742e-02 -1.27917695e+00 9.23219979e-01 -5.86740673e-01
6.24145210e-01 -5.58836102e-01 2.57411569e-01 -1.47570682e+00
-4.02588248e-01 -3.01638514e-01 -2.02632919e-01 1.44617057e+00
4.10629928e-01 -4.84444052e-01 1.34145820e+00 3.20663154e-01
-2.44476765e-01 -9.33076560e-01 -6.75937295e-01 -6.79593503e-01
-3.69665064e-02 -2.91126370e-02 5.53087115e-01 1.58112657e+00
1.01228528e-01 6.99993372e-01 -3.15874159e-01 2.52023935e-01
8.95381510e-01 3.13644737e-01 6.97122276e-01 -1.31501055e+00
-3.05879682e-01 -4.20473725e-01 -2.23904714e-01 -9.18317616e-01
1.10800546e-02 -1.04575372e+00 -3.55898328e-02 -1.34436560e+00
4.95006949e-01 -7.67332315e-01 -2.04717591e-01 2.98496664e-01
-7.00517356e-01 5.42514361e-02 -5.46608418e-02 2.98117638e-01
-7.61668086e-01 9.06962395e-01 1.53412080e+00 -1.42770544e-01
-1.57094926e-01 -8.62568840e-02 -8.94051790e-01 8.56600463e-01
5.92316151e-01 -5.33959091e-01 -8.27951014e-01 -2.58401006e-01
2.74013162e-01 2.04823259e-02 2.25491956e-01 -7.99998939e-01
1.02281526e-01 -6.91199824e-02 6.88871324e-01 -7.85778105e-01
1.09699756e-01 -1.07138336e+00 -1.57643631e-01 4.76781964e-01
-2.72432178e-01 7.18481513e-03 -4.61465806e-01 1.13572991e+00
-3.00896138e-01 -5.74533306e-02 8.96077514e-01 -2.15008274e-01
-3.58616471e-01 6.11407220e-01 -4.14792709e-02 3.94376457e-01
1.19695914e+00 -3.52686137e-01 -3.07486534e-01 -2.21470445e-01
-7.45098054e-01 6.86919689e-01 7.37689435e-01 4.19701010e-01
6.36174202e-01 -1.60330284e+00 -4.30726826e-01 5.47147870e-01
2.61472434e-01 6.52321279e-01 2.35518664e-01 9.78685379e-01
-1.90383736e-02 -2.51599282e-01 1.52062625e-01 -8.75756025e-01
-1.06082296e+00 1.03031015e+00 2.84989297e-01 -2.17083529e-01
-5.70238590e-01 8.34156275e-01 5.14827907e-01 -6.37036383e-01
2.46735588e-01 -1.02074211e-02 -2.34394625e-01 1.56013876e-01
2.70063579e-01 3.49929750e-01 -1.63735881e-01 -3.60414147e-01
-4.24620539e-01 1.90079734e-01 -2.85349667e-01 2.36505240e-01
1.39684427e+00 -1.08731948e-01 -2.56915659e-01 3.42484295e-01
1.48342633e+00 6.15544617e-02 -1.31129122e+00 -2.29206711e-01
1.69288576e-01 -5.61386943e-01 -3.58217567e-01 -4.56484616e-01
-1.86827803e+00 9.24872637e-01 3.00047129e-01 3.74352723e-01
1.12823606e+00 5.25833927e-02 6.57730937e-01 8.88179466e-02
7.53757283e-02 -1.26034820e+00 4.48805481e-01 -6.43768087e-02
3.99478078e-01 -1.52093291e+00 -7.83846006e-02 -8.13872397e-01
-7.88271308e-01 7.08278298e-01 1.04281306e+00 -2.45752335e-01
5.75995684e-01 8.19781050e-02 -6.74754456e-02 -6.07383966e-01
-6.37037694e-01 9.29376408e-02 2.13008538e-01 7.99654245e-01
2.74139404e-01 1.17264226e-01 -9.66499224e-02 6.00866735e-01
1.58886537e-01 -3.88564348e-01 4.06658113e-01 5.32058299e-01
-2.11737201e-01 -8.58969331e-01 -1.23141386e-01 7.78996825e-01
-1.50519282e-01 -2.00424921e-02 -7.72328258e-01 7.83210278e-01
3.30750719e-02 8.38309348e-01 1.87595487e-01 -5.48772812e-01
7.96215162e-02 -1.37302920e-01 1.63200706e-01 -5.93711317e-01
-3.71107459e-01 2.81582326e-01 -3.73206623e-02 -4.09326226e-01
-5.49806237e-01 -6.28176570e-01 -1.44587386e+00 -1.18408002e-01
-4.75405455e-01 3.65801573e-01 5.00384010e-02 6.99420273e-01
5.56495607e-01 4.47111934e-01 9.93782520e-01 -5.54047346e-01
-5.25861740e-01 -8.09109926e-01 -7.59814382e-01 8.12456429e-01
1.16834268e-01 -5.71377814e-01 -7.23128498e-01 -1.84140518e-01] | [7.458194255828857, 5.923952102661133] |
3ee271cb-a9cd-42b6-8e17-9da686e90abd | numerical-smoothing-with-hierarchical | 2111.01874 | null | https://arxiv.org/abs/2111.01874v2 | https://arxiv.org/pdf/2111.01874v2.pdf | Numerical Smoothing with Hierarchical Adaptive Sparse Grids and Quasi-Monte Carlo Methods for Efficient Option Pricing | When approximating the expectations of a functional of a solution to a stochastic differential equation, the numerical performance of deterministic quadrature methods, such as sparse grid quadrature and quasi-Monte Carlo (QMC) methods, may critically depend on the regularity of the integrand. To overcome this issue and improve the regularity structure of the problem, we consider cases in which analytic smoothing (bias-free mollification) cannot be performed and introduce a novel numerical smoothing approach by combining a root-finding method with a one-dimensional numerical integration with respect to a single well-chosen variable. We prove that, under appropriate conditions, the resulting function of the remaining variables is highly smooth, potentially affording the improved efficiency of adaptive sparse grid quadrature (ASGQ) and QMC methods, particularly when combined with hierarchical transformations (ie., the Brownian bridge and Richardson extrapolation on the weak error). This approach facilitates the effective treatment of high dimensionality. Our study is motivated by option pricing problems, focusing on dynamics where the discretization of the asset price is necessary. Based on our analysis and numerical experiments, we demonstrate the advantages of combining numerical smoothing with the ASGQ and QMC methods over these methods without smoothing and the Monte Carlo approach. Finally, our approach is generic and can be applied to solve a broad class of problems, particularly approximating distribution functions, computing financial Greeks, and estimating risk quantities. | ['Raúl Tempone', 'Chiheb Ben Hammouda', 'Christian Bayer'] | 2021-11-02 | null | null | null | null | ['numerical-integration'] | ['miscellaneous'] | [-1.94275439e-01 -1.70636430e-01 4.68838036e-01 3.76945108e-01
-9.01982784e-01 -2.64448613e-01 4.82603997e-01 1.74996823e-01
-5.29301763e-01 1.38944018e+00 -1.64503574e-01 -2.62376785e-01
-2.51969039e-01 -1.06346941e+00 -4.43738788e-01 -1.15983689e+00
-3.03841149e-03 5.09505868e-01 9.41208377e-02 -2.88209915e-01
2.64607340e-01 7.70743012e-01 -1.46543109e+00 -5.25508404e-01
1.39835346e+00 1.27680302e+00 -4.48605448e-01 5.19984424e-01
-2.07355097e-01 4.77518380e-01 -1.72487557e-01 -4.10340458e-01
2.16811195e-01 -4.46689695e-01 -6.08657300e-01 6.98499233e-02
-5.81624359e-02 -1.71090290e-01 3.57675523e-01 1.33986020e+00
6.64613068e-01 6.04760587e-01 1.04378772e+00 -7.51091123e-01
-2.50081629e-01 1.22522585e-01 -5.91811240e-01 -2.04467885e-02
8.22937936e-02 -8.87067467e-02 6.28451586e-01 -1.09322214e+00
4.37865347e-01 1.02468014e+00 1.33780491e+00 2.65209019e-01
-1.66748917e+00 -6.93444982e-02 -5.72241604e-01 -3.78366381e-01
-1.55161095e+00 -1.98020577e-01 6.76855206e-01 -8.25779438e-01
4.59066510e-01 2.76604384e-01 5.24914324e-01 3.45390141e-01
4.21406657e-01 7.80709833e-02 1.24775350e+00 -5.24576783e-01
8.52015972e-01 -2.02019443e-03 3.62635292e-02 5.91498375e-01
4.51566428e-01 2.75464892e-01 1.93562075e-01 -8.93559813e-01
9.20756340e-01 -4.98539321e-02 -1.73019066e-01 -4.39024746e-01
-8.55780721e-01 1.29487062e+00 -7.61804134e-02 4.34578925e-01
-9.00509834e-01 1.54045656e-01 3.67995232e-01 2.53221467e-02
1.11432874e+00 3.28162432e-01 -1.72207206e-01 -2.04607978e-01
-1.31605411e+00 6.45726383e-01 9.95780468e-01 4.43482190e-01
6.73976183e-01 2.91632444e-01 -2.00857192e-01 5.11996806e-01
1.54673070e-01 7.99296737e-01 1.44934267e-01 -1.35822701e+00
5.99404685e-02 3.00299767e-02 7.27658272e-01 -7.36818731e-01
-3.52601141e-01 -2.72843450e-01 -1.00139701e+00 7.19842911e-01
8.58572721e-01 -4.60087836e-01 -2.64312208e-01 1.46222270e+00
8.17118227e-01 1.07340150e-01 1.22296074e-02 5.63100159e-01
-2.27127224e-01 5.82161844e-01 6.00060597e-02 -6.90102756e-01
1.05823803e+00 -2.79271424e-01 -6.97328508e-01 7.56257117e-01
1.09416294e+00 -8.68517041e-01 7.27930069e-01 4.63233232e-01
-1.46535301e+00 -3.07363272e-01 -6.96516216e-01 2.29109868e-01
-1.24092050e-01 -1.78082492e-02 3.47870022e-01 6.48032546e-01
-9.45112705e-01 1.31769669e+00 -7.83020139e-01 1.71169385e-01
3.53630692e-01 -1.64596196e-02 7.80616328e-02 4.07773554e-01
-1.19228578e+00 9.54452634e-01 -2.20992491e-01 1.38648078e-01
-1.42562598e-01 -1.02905762e+00 -4.98003453e-01 1.33982658e-01
3.17553505e-02 -8.03837001e-01 9.79649305e-01 -7.68019974e-01
-1.59106159e+00 3.68917286e-01 1.45788059e-01 -5.40458679e-01
9.85269427e-01 -8.51813331e-03 4.61673178e-02 2.67518312e-01
1.82968065e-01 -1.85188949e-01 1.02162373e+00 -7.31401742e-01
-2.29017138e-01 -3.68244469e-01 -2.81651884e-01 -1.19773880e-01
1.92060620e-01 -1.16981998e-01 7.42660403e-01 -9.22212243e-01
-1.63942575e-01 -7.51482189e-01 -6.20499432e-01 1.89699128e-01
1.73031509e-01 -9.15380940e-02 2.05580428e-01 -8.02603722e-01
1.17056215e+00 -2.24944806e+00 1.37964159e-01 3.15759867e-01
-7.61316940e-02 1.61828384e-01 4.28443551e-01 6.25781953e-01
-4.94522937e-02 9.76840872e-03 -8.53678823e-01 -4.12453532e-01
3.22213545e-02 2.39640344e-02 -5.36252439e-01 9.09849942e-01
1.54758155e-01 5.00735760e-01 -7.25332260e-01 -5.02492785e-01
7.41930753e-02 6.56374037e-01 -8.48991752e-01 -1.51791915e-01
-1.29196361e-01 6.70414448e-01 -6.79258585e-01 1.19361989e-01
9.97307420e-01 -1.96818933e-01 -2.61930615e-01 2.16058895e-01
-5.68688095e-01 -3.40716355e-02 -1.67328858e+00 1.15598345e+00
-8.64966989e-01 1.74695849e-02 4.63639617e-01 -1.17342281e+00
7.46368289e-01 3.97583902e-01 4.97735620e-01 -1.69506788e-01
-5.47170127e-03 7.94068336e-01 -4.25839156e-01 -1.91239223e-01
2.46761292e-01 -1.03533697e+00 1.38301939e-01 2.25207433e-01
-3.92055571e-01 -3.09252590e-01 1.50443181e-01 -9.44989994e-02
7.26565242e-01 1.06650539e-01 4.82654989e-01 -9.31343436e-01
8.10218513e-01 -1.04430936e-01 3.03595215e-01 4.98837590e-01
-6.27887920e-02 4.11720723e-01 8.32141995e-01 -8.45515504e-02
-1.22057569e+00 -7.79213071e-01 -9.27888691e-01 3.35773766e-01
-2.41633132e-01 1.39629632e-01 -9.40377891e-01 -1.86870992e-01
5.31747162e-01 9.97803271e-01 -6.88108504e-01 1.30565360e-01
-5.46275377e-01 -9.65205014e-01 1.63971126e-01 4.54150677e-01
4.94580060e-01 -5.85245132e-01 -4.26034123e-01 6.10284686e-01
3.13000590e-01 -5.75519562e-01 -2.86798686e-01 7.38036702e-04
-1.14455318e+00 -8.59855354e-01 -1.20882809e+00 -1.25297904e-01
3.18062574e-01 -2.90136695e-01 8.18903983e-01 5.65516856e-03
4.53629950e-03 4.74850923e-01 8.96995366e-02 2.07308322e-01
-5.37360013e-01 -3.01541656e-01 1.09014057e-01 3.88889402e-01
-2.65317023e-01 -7.21071839e-01 -6.04005992e-01 1.63342759e-01
-8.11395168e-01 -4.95140791e-01 -5.30605875e-02 1.08323526e+00
4.79314566e-01 1.45334736e-01 7.28020549e-01 -7.29716420e-01
7.20623732e-01 -5.76096594e-01 -1.24317312e+00 -1.35769323e-01
-5.30767798e-01 1.56916335e-01 7.95616269e-01 -4.28255588e-01
-1.05846310e+00 -2.34016910e-01 -3.50886941e-01 -2.52786815e-01
3.98976386e-01 4.88755912e-01 2.40622625e-01 -6.52244449e-01
5.93062997e-01 -5.59740774e-02 1.89383835e-01 -8.17023098e-01
2.05243006e-01 3.99048120e-01 2.18238965e-01 -8.39750886e-01
5.15214503e-01 7.84989357e-01 8.05919528e-01 -1.09847701e+00
-5.80078244e-01 -2.59175628e-01 -3.92987370e-01 -9.13265999e-03
9.20839846e-01 -4.24537480e-01 -9.07033920e-01 5.82072914e-01
-1.21191847e+00 -3.32285702e-01 -1.06442189e+00 6.40400946e-01
-9.28745091e-01 6.13106608e-01 -7.82392442e-01 -1.59104276e+00
-1.79230291e-02 -9.69934344e-01 9.55087781e-01 -5.49451299e-02
-1.83596134e-01 -1.52947080e+00 3.41867030e-01 -2.19634309e-01
6.74305975e-01 6.11297190e-01 8.14907432e-01 -3.75405818e-01
-2.22772002e-01 -1.89160004e-01 2.04252545e-02 5.02575994e-01
-2.85414606e-01 4.18033600e-02 -6.16307914e-01 -1.09022975e-01
8.55649769e-01 1.13564812e-01 5.97502470e-01 7.29980826e-01
6.51576281e-01 -3.33557129e-01 -7.90327862e-02 5.54415703e-01
1.51629436e+00 -2.29425311e-01 4.33762819e-01 6.83043823e-02
2.27691993e-01 5.72092414e-01 4.97713596e-01 8.60735655e-01
-2.04282835e-01 7.43957043e-01 4.19807024e-02 1.19355910e-01
2.94908106e-01 8.00118446e-02 1.32936731e-01 5.84681392e-01
-3.55616331e-01 4.73556638e-01 -7.26104200e-01 2.89474070e-01
-1.68757534e+00 -1.02535892e+00 -5.61546803e-01 2.60133100e+00
9.41252887e-01 -3.27196941e-02 3.28959405e-01 3.74463350e-01
7.19818354e-01 -9.89686176e-02 -2.82639056e-01 -5.49182951e-01
-6.81741759e-02 5.73752224e-01 6.59638941e-01 1.00549865e+00
-9.63593841e-01 2.13808313e-01 6.24768400e+00 1.26559532e+00
-7.31962025e-01 4.07040328e-01 4.08729166e-01 4.74591047e-01
-3.82297546e-01 9.69665200e-02 -8.78885090e-01 8.28935146e-01
1.01057720e+00 -2.36631602e-01 3.70823264e-01 8.59732330e-01
3.76932591e-01 -3.49855006e-01 -4.74101663e-01 6.02915406e-01
-5.89816272e-01 -1.38313293e+00 -4.18260723e-01 4.67485607e-01
9.62079406e-01 -4.64103818e-01 -4.10549925e-04 1.43147647e-01
-8.43465421e-03 -8.77998173e-01 4.96655911e-01 9.13383186e-01
6.14712715e-01 -8.93538296e-01 7.54062593e-01 3.83274764e-01
-1.00307870e+00 2.02890441e-01 -2.64039457e-01 -3.31719428e-01
4.99396354e-01 1.20224679e+00 2.14123614e-02 4.98856753e-01
1.31832480e-01 2.99501777e-01 1.03495121e-01 1.01526022e+00
1.84349626e-01 5.52233696e-01 -6.10787451e-01 -6.21845797e-02
2.54248649e-01 -1.09991252e+00 8.14752579e-01 7.69283295e-01
8.02699864e-01 3.72349501e-01 -1.50249064e-01 1.11034369e+00
4.28785801e-01 4.36680853e-01 -3.93897831e-01 2.52691746e-01
5.24456389e-02 9.62202549e-01 -6.36635125e-01 -4.79383290e-01
-4.45719540e-01 4.81272906e-01 8.30608383e-02 5.58027565e-01
-5.82065821e-01 -3.85960430e-01 7.86466658e-01 4.52102005e-01
6.83797002e-01 -5.03045142e-01 -5.49282014e-01 -1.08191156e+00
1.10909164e-01 -4.64620531e-01 3.80963802e-01 -3.75020981e-01
-1.39765203e+00 3.66203308e-01 2.22776402e-02 -1.12781489e+00
-4.43839312e-01 -5.40180802e-01 -6.40843570e-01 1.38965464e+00
-1.47406316e+00 -4.23682064e-01 4.97080475e-01 4.41508830e-01
-1.52052268e-01 1.78044438e-01 6.76888406e-01 2.73543864e-01
-2.01991186e-01 2.39527766e-02 8.72340560e-01 -4.48981494e-01
1.03158236e-01 -1.19132543e+00 6.04073554e-02 6.78726435e-01
-5.99893391e-01 4.84231144e-01 1.03117514e+00 -8.48103285e-01
-1.02349639e+00 -7.20394433e-01 1.00449371e+00 -8.99138451e-02
1.15680718e+00 -5.47294021e-02 -1.35110664e+00 1.64635941e-01
-2.22900942e-01 3.48228782e-01 3.97829026e-01 -2.58739710e-01
3.83357048e-01 6.33564815e-02 -1.55015206e+00 4.01972502e-01
3.97220701e-01 -3.28002572e-01 -4.09477443e-01 3.92015517e-01
2.41592273e-01 -1.02444969e-01 -1.31474602e+00 3.50088477e-01
1.94416463e-01 -9.68298256e-01 9.42993283e-01 -5.55022418e-01
6.72463328e-02 -2.78340340e-01 -1.52387828e-01 -1.09334373e+00
1.61469057e-02 -1.15899396e+00 -1.06661633e-01 1.16743731e+00
-9.25971270e-02 -1.26408398e+00 3.87868971e-01 7.22745478e-01
2.85338491e-01 -8.81299198e-01 -1.41450787e+00 -1.00557721e+00
8.03751528e-01 -2.20648453e-01 3.95504713e-01 7.74456680e-01
-3.15383263e-03 -3.27371180e-01 -2.75507867e-01 -3.57588194e-02
1.00370169e+00 6.31089807e-02 3.49919379e-01 -1.44140840e+00
-5.53042531e-01 -4.82730955e-01 -1.92635432e-01 -8.52936387e-01
1.84240684e-01 -6.64299011e-01 -2.24828377e-01 -9.12335396e-01
-4.81231868e-01 -5.87049723e-01 2.48545617e-01 -3.61808717e-01
-1.49589211e-01 1.22006111e-01 -8.16462189e-02 1.87739685e-01
5.87204099e-02 9.20762599e-01 1.30747902e+00 5.32211542e-01
-8.26127902e-02 3.95768255e-01 -2.52572671e-02 9.46878135e-01
3.73239875e-01 -5.16923785e-01 9.25237536e-02 4.79292959e-01
4.34994251e-01 5.43244421e-01 5.97332954e-01 -8.69604528e-01
6.33169934e-02 -2.73115654e-02 5.46273775e-02 -3.75779599e-01
1.70895845e-01 -5.22432625e-01 3.26806813e-01 6.21356964e-01
-1.30903855e-01 -1.10818215e-01 5.99772371e-02 5.73214591e-01
-1.29629388e-01 -8.77101123e-01 1.19892657e+00 6.41868263e-02
1.03420094e-01 1.26765683e-01 -5.20748198e-01 4.09659058e-01
7.67449021e-01 1.08966179e-01 2.06914291e-01 -3.63850504e-01
-1.00982642e+00 -1.29046410e-01 6.05138242e-01 -9.30004537e-01
2.17906862e-01 -1.67713714e+00 -6.85236037e-01 2.51545310e-01
-6.85583472e-01 -1.94061026e-01 4.21755195e-01 1.39169562e+00
-6.71391785e-01 1.01519056e-01 2.16614202e-01 -2.15918079e-01
-3.85617822e-01 6.60188496e-01 5.88434756e-01 -5.11849582e-01
-6.19409025e-01 3.78876626e-01 4.82194945e-02 -3.62765021e-03
-1.23194799e-01 -3.91251117e-01 8.04366246e-02 2.80490249e-01
4.57483590e-01 1.00496507e+00 6.07967936e-02 -5.44367313e-01
-1.09947756e-01 8.61902475e-01 6.97692215e-01 -4.11712885e-01
1.10354018e+00 -2.27901086e-01 -2.82639414e-01 6.14376843e-01
1.34798360e+00 3.54107052e-01 -1.47448528e+00 -7.59498868e-03
7.76749477e-02 -3.41992378e-01 4.76416610e-02 2.25786828e-02
-7.87639678e-01 9.56944764e-01 7.44354129e-02 8.39164019e-01
8.10189784e-01 -4.99443382e-01 8.35432231e-01 -1.67360216e-01
3.60103399e-01 -1.15906024e+00 -5.64137757e-01 3.99060816e-01
9.26008701e-01 -8.66205633e-01 3.96509767e-02 -4.54457998e-01
-2.59772748e-01 1.21420360e+00 -3.53969604e-01 -7.19394386e-01
1.14889503e+00 3.02631408e-01 -2.97457457e-01 7.66240805e-02
-3.34991992e-01 -2.74503827e-01 1.53399393e-01 2.85601486e-02
4.38357085e-01 -4.55021411e-02 -1.03393304e+00 3.96755397e-01
8.05074647e-02 9.70059261e-02 4.71917689e-01 6.92991316e-01
-2.45602176e-01 -1.03210413e+00 -5.46388865e-01 2.67663479e-01
-6.18319571e-01 -2.43444428e-01 4.68009025e-01 8.36236119e-01
-1.48212850e-01 4.89957362e-01 2.22539641e-02 5.58610439e-01
2.58817166e-01 5.14901459e-01 3.33366513e-01 -2.35121012e-01
-4.66206819e-01 2.96838850e-01 7.75000826e-02 -3.40405405e-01
-4.63608503e-01 -1.21719623e+00 -1.13554132e+00 -4.44853574e-01
-3.42144519e-01 6.97898746e-01 5.18090904e-01 1.00026393e+00
1.09802015e-01 1.32525161e-01 7.71733046e-01 -1.10572040e+00
-1.45257115e+00 -7.62308598e-01 -1.36003649e+00 2.59592414e-01
3.84919763e-01 -1.00360310e+00 -9.92609680e-01 -4.34352577e-01] | [6.371648788452148, 3.7724499702453613] |
36af8e3c-f1da-487d-b9ab-098d62afa69d | gait-based-human-identification-through | 2110.09286 | null | https://arxiv.org/abs/2110.09286v1 | https://arxiv.org/pdf/2110.09286v1.pdf | Gait-based Human Identification through Minimum Gait-phases and Sensors | Human identification is one of the most common and critical tasks for condition monitoring, human-machine interaction, and providing assistive services in smart environments. Recently, human gait has gained new attention as a biometric for identification to achieve contactless identification from a distance robust to physical appearances. However, an important aspect of gait identification through wearables and image-based systems alike is accurate identification when limited information is available, for example, when only a fraction of the whole gait cycle or only a part of the subject body is visible. In this paper, we present a gait identification technique based on temporal and descriptive statistic parameters of different gait phases as the features and we investigate the performance of using only single gait phases for the identification task using a minimum number of sensors. It was shown that it is possible to achieve high accuracy of over 95.5 percent by monitoring a single phase of the whole gait cycle through only a single sensor. It was also shown that the proposed methodology could be used to achieve 100 percent identification accuracy when the whole gait cycle was monitored through pelvis and foot sensors combined. The ANN was found to be more robust to fewer data features compared to SVM and was concluded as the best machine algorithm for the purpose. | ['Jinwook Kim', 'Kyung-Ryoul Mun', 'Mina Park', 'Dawoon Jung', 'Muhammad Zeeshan Arshad'] | 2021-10-15 | null | null | null | null | ['gait-identification'] | ['computer-vision'] | [ 2.72139579e-01 -2.57470727e-01 -2.20690235e-01 1.16853928e-02
-1.11124828e-01 -1.50593206e-01 1.51066720e-01 3.95540923e-01
-6.31236613e-01 6.84223711e-01 -3.85428071e-01 1.02808371e-01
-2.17830002e-01 -6.95285499e-01 -7.03798309e-02 -7.39587069e-01
-1.24866597e-01 4.04830545e-01 2.16712698e-01 -1.88891307e-01
1.16501689e-01 6.58252537e-01 -1.87653422e+00 -4.51609313e-01
6.76249027e-01 1.10897923e+00 -1.66193694e-02 5.04496336e-01
4.79319006e-01 -1.41917437e-01 -7.41399944e-01 -3.35947722e-02
1.76631123e-01 -2.30485871e-01 -3.17611307e-01 6.22902811e-02
5.32720387e-02 -2.60390431e-01 8.94075036e-02 7.23360777e-01
8.41734707e-01 5.79159707e-02 6.38934135e-01 -1.01523185e+00
1.66149922e-02 -1.76316038e-01 -6.22416735e-01 1.75041214e-01
8.27906847e-01 1.26665905e-01 2.90078580e-01 -4.14825827e-01
2.77808368e-01 7.28951514e-01 7.12970197e-01 3.31258565e-01
-1.05104017e+00 -5.06220818e-01 -5.88545620e-01 4.41769332e-01
-1.41726768e+00 -3.04858863e-01 8.10961246e-01 -4.51398343e-01
8.02168787e-01 3.13339740e-01 9.03329432e-01 7.37884283e-01
4.45038289e-01 2.11824983e-01 1.15364933e+00 -6.03540659e-01
1.37902156e-01 1.62728447e-02 3.67746562e-01 8.18702817e-01
9.13555861e-01 3.33485007e-02 -2.76353806e-01 -8.90225396e-02
6.51867509e-01 -1.29349753e-01 -1.30966855e-02 -1.38169348e-01
-1.07764089e+00 4.07827914e-01 9.61818674e-04 6.80022657e-01
-5.35659909e-01 -7.75825307e-02 4.14661884e-01 1.18781276e-01
1.49392486e-01 2.55893290e-01 2.70855688e-02 -6.07578278e-01
-9.02847409e-01 -2.46167630e-02 7.98617303e-01 1.79285154e-01
4.22706038e-01 3.09821427e-01 1.71618000e-01 6.75061285e-01
4.95457202e-02 9.56328571e-01 6.63978100e-01 -6.21286809e-01
2.42205694e-01 8.61703694e-01 2.17934728e-01 -1.21923375e+00
-6.86860740e-01 -2.70884454e-01 -9.05101955e-01 1.85466141e-01
6.58372581e-01 -1.68533713e-01 -6.03166521e-01 1.32242644e+00
4.68993008e-01 -2.42931947e-01 -3.59183252e-01 9.15359020e-01
4.96884197e-01 1.10743992e-01 3.06816120e-02 -3.05174172e-01
1.55397236e+00 -9.30169448e-02 -7.95149267e-01 -1.38130218e-01
3.43063056e-01 -6.40554845e-01 7.48583913e-01 3.32206994e-01
-7.01454759e-01 -7.21916318e-01 -1.55707562e+00 6.35336339e-01
-4.47987616e-01 4.97576177e-01 1.75015211e-01 1.21783495e+00
-5.64818501e-01 7.71640539e-01 -1.11341393e+00 -8.13809514e-01
-1.52760148e-01 7.81635821e-01 -7.69064844e-01 1.36393607e-01
-1.00822508e+00 1.10066319e+00 2.07796246e-01 2.04129621e-01
3.53694767e-01 5.82689941e-02 -8.40706587e-01 -2.61045337e-01
-8.36204812e-02 -5.66948712e-01 4.47198331e-01 -4.51941758e-01
-1.42149901e+00 7.24800587e-01 -3.28749478e-01 -4.05742764e-01
6.46953344e-01 -8.28379542e-02 -6.09121621e-01 4.24318552e-01
7.03988969e-02 -1.23296708e-01 9.14284229e-01 -5.62133610e-01
-1.28097206e-01 -7.71995962e-01 -5.56164861e-01 1.19585916e-02
-3.88285458e-01 -3.09103318e-02 6.48789778e-02 -3.11986029e-01
3.36574554e-01 -1.05007529e+00 3.12530696e-01 -1.73297688e-01
-1.96672186e-01 1.14866666e-01 9.12405133e-01 -1.16130865e+00
1.19840097e+00 -1.93442547e+00 -1.83688417e-01 5.83286762e-01
-6.97582588e-02 5.94121873e-01 4.58695441e-01 3.64545345e-01
5.83336800e-02 -1.85372740e-01 -2.59328187e-01 -1.49813756e-01
-2.17840135e-01 1.96184188e-01 5.35346389e-01 9.03761566e-01
-1.38464034e-01 5.46739936e-01 -5.81184864e-01 -5.20968139e-01
7.52391398e-01 5.26903510e-01 3.02481890e-01 -1.47152238e-03
7.44201720e-01 5.38042724e-01 -2.87192464e-01 1.00271344e+00
3.66640806e-01 1.63033634e-01 1.75353270e-02 -4.78476405e-01
1.56939134e-01 -3.76677573e-01 -1.46953070e+00 9.50882673e-01
-2.14259878e-01 5.15967369e-01 -2.34834701e-01 -1.38438654e+00
1.35228729e+00 6.92831874e-01 9.71785724e-01 -8.80416393e-01
3.30221355e-01 5.51070213e-01 2.38723412e-01 -6.24765575e-01
2.06958339e-01 6.80674091e-02 -1.18462436e-01 2.89624333e-01
-2.68225998e-01 2.96107382e-01 3.23639184e-01 -5.75406075e-01
9.77730691e-01 -1.16492175e-01 9.53546941e-01 -1.22984387e-01
6.62097692e-01 -2.23017797e-01 3.44100505e-01 5.22036731e-01
-5.36205649e-01 4.66930777e-01 -2.71792799e-01 -4.33343351e-01
-8.54145348e-01 -1.01255941e+00 -1.83247343e-01 1.59526944e-01
2.59893924e-01 1.21501312e-01 -7.08612144e-01 -1.61887243e-01
2.92260110e-01 1.79568697e-02 -4.47829306e-01 -3.25897157e-01
-6.97972417e-01 -9.76896167e-01 7.13296235e-01 5.60283065e-01
8.86875987e-01 -7.80908525e-01 -1.22107732e+00 3.33946109e-01
-1.77661613e-01 -1.13977253e+00 3.79981138e-02 -6.02363609e-02
-1.03268409e+00 -1.23541784e+00 -7.89262235e-01 -3.38735551e-01
5.81098437e-01 -1.40597850e-01 4.66167629e-01 2.10444435e-01
-6.98124707e-01 4.29906875e-01 -2.66387373e-01 -1.64875880e-01
-2.26998344e-01 -9.74171981e-02 6.15749180e-01 1.46445602e-01
3.29467744e-01 -6.18017673e-01 -6.56857610e-01 5.56038201e-01
-1.30651027e-01 -3.75564992e-01 2.44676173e-01 7.56316364e-01
2.14800254e-01 -1.13023147e-02 6.33531153e-01 -1.99234381e-01
7.10133135e-01 -1.59065843e-01 -3.05961240e-02 1.67856961e-01
-5.76122761e-01 -2.02712059e-01 4.95648444e-01 -4.11852241e-01
-5.43318927e-01 1.04685739e-01 -9.53178555e-02 -1.89035367e-02
-3.99653196e-01 2.61241406e-01 -1.33665679e-02 -5.42342961e-01
5.39733052e-01 3.83088201e-01 5.12845635e-01 -3.86050493e-01
-3.58873397e-01 7.90003836e-01 6.60713673e-01 -5.46880662e-01
5.17446697e-01 4.39394742e-01 6.07023895e-01 -1.37453854e+00
2.15023682e-01 -9.73729849e-01 -9.04351652e-01 -7.68096924e-01
5.85658312e-01 -3.92473429e-01 -1.20122540e+00 8.08127463e-01
-6.56115294e-01 2.78235942e-01 -1.22959562e-01 6.73843443e-01
-5.56620717e-01 7.42105365e-01 -2.95162827e-01 -1.27215898e+00
-7.58752942e-01 -9.64579999e-01 8.58146548e-01 3.93403143e-01
-7.77442694e-01 -9.10281599e-01 -1.80217803e-01 3.13399196e-01
3.45472932e-01 9.14116859e-01 4.56106901e-01 -4.29721266e-01
3.20032120e-01 -1.03494620e+00 1.26020879e-01 7.72812888e-02
8.19790602e-01 -4.20430675e-02 -6.83619738e-01 -5.07346392e-01
6.62226975e-02 1.31229058e-01 3.04000109e-01 3.63583356e-01
4.40383822e-01 -5.14724217e-02 -3.87448490e-01 2.56855488e-01
1.36011016e+00 6.32444799e-01 6.30220175e-01 5.13957798e-01
5.12493074e-01 4.92485493e-01 7.30511189e-01 2.78951854e-01
3.05396412e-02 9.55610156e-01 1.08224131e-01 -1.13797821e-01
-1.04424007e-01 2.12100238e-01 1.46180421e-01 7.58859336e-01
-7.64727473e-01 1.62436500e-01 -7.94031024e-01 4.73210007e-01
-1.58763003e+00 -9.55400705e-01 -1.27371624e-01 2.66480780e+00
1.93685502e-01 -3.79656106e-02 6.54980123e-01 1.21161795e+00
9.62140262e-01 -3.18722129e-01 -3.99027646e-01 -3.60962898e-01
-5.15933596e-02 4.21632588e-01 7.27293789e-01 1.61104560e-01
-1.05478907e+00 -9.42548085e-03 5.94742155e+00 3.97814095e-01
-1.26639283e+00 -2.18008175e-01 1.75169364e-01 2.44268611e-01
6.46668434e-01 -3.88782114e-01 -5.60874462e-01 9.17130470e-01
7.57223487e-01 -7.58927539e-02 -1.57154456e-01 5.24608314e-01
5.40952027e-01 -7.46590793e-01 -7.42476165e-01 1.09420264e+00
2.27854792e-02 -4.82713223e-01 -2.05723882e-01 1.46221757e-01
1.77203611e-01 -8.11676919e-01 -1.33683816e-01 -5.04097462e-01
-6.97131693e-01 -7.82081664e-01 1.22047111e-01 5.78119636e-01
8.53589535e-01 -6.12274528e-01 1.07038105e+00 3.09381723e-01
-1.61281645e+00 -6.73022717e-02 1.93342775e-01 -3.08822423e-01
2.68039346e-01 4.42864895e-01 -8.62763822e-01 7.40595818e-01
4.38993692e-01 4.81268585e-01 -4.41230983e-01 1.36850476e+00
2.71442831e-01 5.27563393e-01 -7.21237779e-01 -2.98028052e-01
-3.85916650e-01 -2.13577703e-01 7.65543461e-01 9.95827496e-01
6.59840226e-01 -7.49066025e-02 8.24956968e-02 1.89796701e-01
7.62930989e-01 2.37316549e-01 -5.68098307e-01 1.30976647e-01
4.32761699e-01 8.00305247e-01 -9.24248338e-01 -1.01330571e-01
-1.28042966e-01 8.48167360e-01 -2.73352504e-01 2.27318946e-02
-4.26478028e-01 -5.88711977e-01 3.80372405e-01 5.73583245e-01
-1.26022890e-01 -4.93435919e-01 -4.42869663e-01 -7.69107759e-01
2.77358323e-01 -4.33762163e-01 5.16890764e-01 -3.15255433e-01
-7.89643943e-01 1.98260933e-01 1.22378930e-01 -1.52791798e+00
-6.74683988e-01 -5.12865067e-01 -6.11529171e-01 1.01560462e+00
-6.03400648e-01 -9.65939283e-01 -4.72554415e-01 6.35424793e-01
1.07980117e-01 -1.35509551e-01 9.90902185e-01 2.95748234e-01
-5.08293867e-01 7.68412173e-01 -6.53177034e-04 5.79857044e-02
4.98002738e-01 -9.69794631e-01 1.64504543e-01 9.14455354e-01
-2.92987406e-01 5.77766180e-01 8.95666540e-01 -8.55566502e-01
-1.40619218e+00 -2.89116353e-01 7.82994986e-01 8.04626271e-02
2.12711930e-01 1.65848106e-01 -6.71560168e-01 8.64668842e-03
-3.72391999e-01 -1.25319898e-01 4.83205408e-01 -3.26404124e-01
4.13950324e-01 -3.93717349e-01 -1.55020618e+00 1.88432664e-01
5.52802801e-01 -2.77380139e-01 -6.32700980e-01 -3.03111911e-01
-3.05434078e-01 -4.13295805e-01 -1.11984956e+00 7.26386845e-01
1.15504944e+00 -7.50966668e-01 1.10210729e+00 2.60847628e-01
-4.24125046e-01 -4.83068347e-01 -1.13866404e-02 -9.43267703e-01
-6.02607243e-02 -2.12139279e-01 -2.91371197e-01 9.88082945e-01
-1.49073586e-01 -1.00999093e+00 9.18214142e-01 4.65376496e-01
5.07479608e-01 -5.35003185e-01 -1.21179628e+00 -9.69509244e-01
-6.88981473e-01 1.05684409e-02 2.29137883e-01 6.07504487e-01
2.75256813e-01 -1.74724132e-01 -5.18142343e-01 3.22832726e-02
9.37922120e-01 -1.11258000e-01 6.20246768e-01 -1.50421214e+00
-9.06007364e-02 -2.77269119e-03 -1.24532640e+00 -1.90481752e-01
-3.83104146e-01 -2.44059384e-01 -2.65043169e-01 -1.42879796e+00
-1.67835325e-01 -1.69544429e-01 -2.10793227e-01 2.93212920e-01
-2.64847595e-02 6.25408471e-01 -2.62703355e-02 2.04227388e-01
2.11427048e-01 -1.30762696e-01 1.10703504e+00 -6.32807985e-02
-2.63921380e-01 4.73987341e-01 -2.42610447e-04 4.91515309e-01
7.53259361e-01 -1.18268341e-01 -1.78111747e-01 3.26421380e-01
-3.33995074e-01 2.47053385e-01 3.87001604e-01 -1.53453040e+00
2.82844990e-01 1.85052305e-01 8.11837971e-01 -4.26126480e-01
6.60835862e-01 -8.95913720e-01 6.60596609e-01 9.72157836e-01
4.65919316e-01 8.13731551e-02 2.01948851e-01 2.58304179e-01
-1.70818850e-01 -1.25493273e-01 7.31325805e-01 -2.73791328e-02
-8.41552854e-01 -2.24676698e-01 -4.44246173e-01 -4.98044252e-01
1.04708171e+00 -1.18309128e+00 -1.90628860e-02 -3.10104609e-01
-1.08342838e+00 -2.57866502e-01 5.11883318e-01 2.43098676e-01
5.36266387e-01 -1.20790422e+00 -3.86614114e-01 3.41167808e-01
-4.33039060e-03 -7.26284027e-01 2.99808621e-01 9.52554107e-01
-6.37557149e-01 5.32498777e-01 -7.48220265e-01 -8.10869992e-01
-1.90628922e+00 3.55595611e-02 2.69820303e-01 2.87752668e-03
-5.78168094e-01 -2.79888939e-02 -8.14098597e-01 2.13972941e-01
1.47925522e-02 -4.07717079e-02 -4.95937139e-01 1.05729014e-01
3.86310756e-01 9.84212458e-01 3.74579787e-01 -1.01043463e+00
-6.68757379e-01 1.24258304e+00 5.09152293e-01 3.09205130e-02
6.95236504e-01 -2.81009078e-01 1.98530912e-01 5.41982114e-01
9.79131877e-01 -1.86087191e-01 -7.07481742e-01 1.84312135e-01
-1.06539771e-01 -4.22538012e-01 -4.71289098e-01 -6.52518451e-01
-6.93343818e-01 7.13038921e-01 1.13148558e+00 3.13458622e-01
1.24367821e+00 -6.29227161e-01 6.54636443e-01 1.84652358e-01
8.36553216e-01 -1.20088518e+00 -2.55354673e-01 -9.93113443e-02
3.99844676e-01 -1.14730990e+00 3.44432712e-01 -4.66906130e-01
-1.29487425e-01 1.30151737e+00 3.20363939e-01 -1.43960267e-01
6.07142568e-01 1.05492979e-01 -2.34093174e-01 1.91642016e-01
1.24923505e-01 -2.93799847e-01 4.90832716e-01 9.13800001e-01
3.59370440e-01 4.20997441e-01 -8.51507545e-01 3.10538620e-01
-1.34083569e-01 1.37391269e-01 1.38281420e-01 1.18426013e+00
-5.20113707e-01 -1.04451668e+00 -7.48322427e-01 7.98358917e-01
-6.24593854e-01 5.50025165e-01 -1.28438815e-01 1.07082427e+00
1.80169061e-01 1.15961766e+00 -1.08444080e-01 -5.29820263e-01
4.48189437e-01 2.60805666e-01 7.80527353e-01 -1.16331331e-01
-3.57833356e-01 -2.93039173e-01 1.98566973e-01 -6.23215199e-01
-5.91487348e-01 -7.96008110e-01 -9.72078264e-01 -3.98474813e-01
-1.64318264e-01 9.92489904e-02 9.57430720e-01 1.15127540e+00
-2.81984685e-04 1.85422167e-01 4.86566156e-01 -8.42879176e-01
-3.76399547e-01 -9.20366108e-01 -8.58542264e-01 3.39672953e-01
2.09913537e-01 -1.01882148e+00 -2.58917421e-01 1.46506444e-01] | [14.048177719116211, 1.4934331178665161] |
a780c432-5cd6-4baf-8b74-1f04d3644766 | dixit-interactive-visual-storytelling-via | 1903.02230 | null | https://arxiv.org/abs/1903.02230v3 | https://arxiv.org/pdf/1903.02230v3.pdf | Dixit: Interactive Visual Storytelling via Term Manipulation | In this paper, we introduce Dixit, an interactive visual storytelling system that the user interacts with iteratively to compose a short story for a photo sequence. The user initiates the process by uploading a sequence of photos. Dixit first extracts text terms from each photo which describe the objects (e.g., boy, bike) or actions (e.g., sleep) in the photo, and then allows the user to add new terms or remove existing terms. Dixit then generates a short story based on these terms. Behind the scenes, Dixit uses an LSTM-based model trained on image caption data and FrameNet to distill terms from each image and utilizes a transformer decoder to compose a context-coherent story. Users change images or terms iteratively with Dixit to create the most ideal story. Dixit also allows users to manually edit and rate stories. The proposed procedure opens up possibilities for interpretable and controllable visual storytelling, allowing users to understand the story formation rationale and to intervene in the generation process. | ['Lun-Wei Ku', "Ting-Hao 'Kenneth' Huang", 'Hsin-Yu Lin', 'Zi-Yuan Chen', 'Yu-Hua Chen', 'Chao-Chun Hsu'] | 2019-03-06 | null | null | null | null | ['visual-storytelling'] | ['natural-language-processing'] | [ 6.24183714e-01 3.66969109e-01 -4.75633033e-02 -2.83157200e-01
-2.88778454e-01 -6.72225952e-01 1.03057051e+00 1.34870917e-01
8.65406319e-02 6.95135832e-01 5.49057364e-01 -1.20021693e-01
3.36204380e-01 -7.75930464e-01 -7.96425223e-01 -6.19322598e-01
3.11773151e-01 4.56062317e-01 4.75807637e-02 -1.11844771e-01
3.08234453e-01 2.82860100e-01 -1.70530939e+00 7.03402936e-01
3.58310223e-01 5.18265009e-01 9.36945856e-01 8.22273612e-01
-4.92564678e-01 1.35445738e+00 -5.74330330e-01 -2.91878641e-01
-1.02285340e-01 -7.28496969e-01 -9.66481030e-01 5.45237482e-01
2.73539990e-01 -6.23927295e-01 -2.99730092e-01 5.96625507e-01
1.37845218e-01 2.68017769e-01 4.82741743e-01 -1.25597990e+00
-7.64683723e-01 1.07184148e+00 -3.61782670e-01 -2.26411089e-01
9.72847462e-01 3.04787457e-01 8.47715855e-01 -9.18210804e-01
1.18941844e+00 1.06409228e+00 1.45346448e-01 6.22877061e-01
-1.18831265e+00 -3.48894417e-01 6.33347109e-02 3.61661851e-01
-1.09524226e+00 -4.47567493e-01 9.79072452e-01 -5.52119553e-01
6.99274778e-01 6.02059901e-01 1.15838325e+00 1.17752814e+00
-1.15041249e-01 1.00591922e+00 7.32170522e-01 -7.40637422e-01
1.00039132e-01 1.55411333e-01 -2.00551286e-01 6.47940099e-01
-5.97325683e-01 -2.36689225e-01 -7.05134153e-01 1.97904140e-01
1.00711286e+00 4.61923666e-02 -1.58674777e-01 -5.49122334e-01
-1.51774323e+00 4.46186364e-01 4.89171922e-01 5.80451131e-01
-6.79441452e-01 4.49452400e-01 3.53341132e-01 1.23565830e-02
2.34370008e-01 5.18229306e-01 1.31609276e-01 -1.23675086e-01
-1.09292328e+00 4.21725899e-01 4.03206229e-01 1.09830701e+00
6.15642965e-01 8.98110867e-03 -5.67931890e-01 7.22983301e-01
2.21400395e-01 2.41936833e-01 2.46112630e-01 -1.02634120e+00
3.29703569e-01 5.87482035e-01 2.22247913e-01 -9.17899847e-01
-9.23960879e-02 9.70286578e-02 -5.71323633e-01 3.14596087e-01
1.41660154e-01 9.12104845e-02 -1.05140865e+00 1.45007694e+00
3.24581891e-01 -1.76913626e-02 -2.39978671e-01 8.33092749e-01
1.07843971e+00 1.09931684e+00 2.11950287e-01 -2.75225878e-01
1.69363427e+00 -9.73082185e-01 -9.89986241e-01 -2.07298249e-01
3.35444301e-01 -7.81547964e-01 1.26738572e+00 2.22689599e-01
-1.54422200e+00 -5.66807032e-01 -8.32689643e-01 -5.10715902e-01
-1.93808481e-01 2.09452167e-01 1.62049621e-01 -1.53902784e-01
-1.13257420e+00 3.58860224e-01 -6.35872900e-01 -7.49349773e-01
2.14645490e-01 -1.50960132e-01 -1.77849278e-01 9.80327427e-02
-9.07973468e-01 9.34744298e-01 7.78417826e-01 -8.33231658e-02
-1.18680966e+00 -5.97084761e-01 -1.11762178e+00 9.54381824e-02
3.99704367e-01 -1.05706453e+00 1.59791720e+00 -1.38084710e+00
-1.71318102e+00 9.16840613e-01 -4.71086293e-01 -5.14039338e-01
5.64370871e-01 -2.21906409e-01 5.11421300e-02 3.14756960e-01
1.69097960e-01 1.25360715e+00 9.51132655e-01 -1.70784020e+00
-5.90231061e-01 9.31410417e-02 4.61839288e-01 5.52845061e-01
4.80353907e-02 2.78722823e-01 -5.95936298e-01 -6.86384976e-01
-9.16643962e-02 -8.33437502e-01 6.89292550e-02 7.87538886e-02
-8.03000987e-01 6.22408986e-02 1.11955941e+00 -8.11354160e-01
1.24418914e+00 -2.01136732e+00 4.11479831e-01 -7.01890290e-02
3.34495753e-01 8.34329128e-02 -6.27295449e-02 9.23016250e-01
-2.64846563e-01 -6.66701943e-02 -9.50759277e-02 -7.03110158e-01
-2.24098280e-01 1.28373235e-01 -2.33985454e-01 -1.10213988e-01
-1.96081892e-01 8.65704238e-01 -1.02159977e+00 -5.90618074e-01
8.06260169e-01 5.26986599e-01 -2.02183664e-01 4.16088998e-01
-7.73002982e-01 6.91690087e-01 -2.01471269e-01 1.15263157e-01
2.88786769e-01 -4.10503626e-01 1.68701738e-01 -1.35761932e-01
-5.24916112e-01 4.22819614e-01 -7.84701645e-01 1.85963237e+00
-6.66773915e-01 1.27364600e+00 -4.37840849e-01 -6.31162822e-01
6.26483381e-01 5.53629935e-01 1.87160879e-01 -3.28141391e-01
1.31324783e-01 -3.33652318e-01 -6.35113895e-01 -9.79578197e-01
8.03270519e-01 -1.25764519e-01 9.79995579e-02 6.95918202e-01
-2.15130493e-01 -4.50181395e-01 3.53297800e-01 3.95693660e-01
6.86852455e-01 5.39345086e-01 5.89943230e-01 3.29376936e-01
3.67332041e-01 7.39039779e-02 -3.09101880e-01 4.73143101e-01
5.05587459e-01 7.45291591e-01 3.45494777e-01 -7.31610119e-01
-1.38129115e+00 -1.03592765e+00 4.76173609e-01 1.26065266e+00
1.33683354e-01 -4.39056337e-01 -7.64303923e-01 -1.76758796e-01
-6.19611084e-01 1.39936662e+00 -7.41443932e-01 1.30584985e-01
-6.10474408e-01 2.13330805e-01 2.39416659e-02 1.93648323e-01
7.21947730e-01 -1.56557119e+00 -1.05396986e+00 2.42761314e-01
-7.92064071e-01 -7.58773506e-01 -6.35649860e-01 -3.31736237e-01
-5.16879976e-01 -4.89256740e-01 -8.33884537e-01 -1.03674293e+00
9.78595138e-01 3.78751159e-01 9.92848396e-01 -8.85090008e-02
-1.18607022e-01 1.98940918e-01 -4.90628809e-01 -1.36036620e-01
-7.65798748e-01 -7.36668780e-02 -5.83229005e-01 1.41948909e-01
-3.46036136e-01 -6.10114992e-01 -4.68910635e-01 -6.10340275e-02
-1.02128100e+00 1.39588881e+00 8.57010260e-02 5.41085780e-01
4.23661202e-01 -1.37672380e-01 -4.50042039e-02 -6.85757637e-01
4.35461879e-01 -1.98341340e-01 -2.51610160e-01 3.53043318e-01
2.71825064e-02 1.98561251e-01 4.62716281e-01 -5.23098111e-01
-1.43239391e+00 3.22473794e-01 8.74955654e-02 -4.77499485e-01
-2.13036329e-01 4.63937193e-01 -5.97967468e-02 5.39024055e-01
4.62548554e-01 4.49294627e-01 -6.03763402e-01 -1.02372520e-01
1.01818061e+00 3.86188656e-01 8.16170037e-01 -5.12648709e-02
7.53568351e-01 4.80316579e-01 -4.57463592e-01 -8.19335938e-01
-5.94412565e-01 -9.90894139e-02 -7.53688633e-01 -8.72316718e-01
1.21891141e+00 -6.80124581e-01 -6.53108835e-01 3.38731647e-01
-1.63485622e+00 -5.27497709e-01 -5.11421025e-01 1.51842311e-01
-8.82267594e-01 -6.74247891e-02 -4.66928631e-01 -7.81654119e-01
-2.38872677e-01 -9.23642755e-01 1.06788409e+00 3.09563726e-01
-7.51071155e-01 -8.99643064e-01 -1.79299295e-01 4.54187721e-01
-3.90414745e-02 6.86337292e-01 7.97431886e-01 4.76766303e-02
-8.49137187e-01 8.05643722e-02 -1.40775979e-01 -1.93372548e-01
3.06046754e-01 3.50203693e-01 -8.46345842e-01 1.44458741e-01
-4.14201319e-01 -1.13915235e-01 4.45725590e-01 4.86710727e-01
1.06106484e+00 -7.49307871e-01 -4.50565159e-01 2.52964258e-01
1.14062488e+00 7.96069264e-01 9.83716011e-01 3.83044034e-01
8.77077758e-01 5.99922597e-01 4.05360043e-01 5.24276078e-01
6.25559568e-01 8.03527236e-01 3.39408457e-01 -2.40324363e-01
-2.23600566e-01 -8.74398947e-01 4.03827429e-01 3.76101494e-01
-9.37036946e-02 -5.36893964e-01 -7.40992308e-01 6.17732346e-01
-1.83025444e+00 -1.41581070e+00 -2.37684190e-01 1.91589570e+00
6.80651486e-01 -3.29944119e-02 3.40085626e-02 9.29881483e-02
9.51515973e-01 4.48676854e-01 -3.31952840e-01 -5.38431883e-01
1.77505136e-01 -2.17098758e-01 1.73699811e-01 6.56324744e-01
-7.58837044e-01 1.26495349e+00 5.77187872e+00 6.54253781e-01
-1.24826562e+00 3.59656364e-02 6.70612514e-01 -3.37765425e-01
-4.73767221e-01 1.72384709e-01 -2.37024054e-01 2.55431443e-01
3.82782698e-01 -3.82144898e-01 7.39284933e-01 5.95302641e-01
8.79731238e-01 -4.29999173e-01 -1.16668308e+00 1.00617743e+00
4.32973266e-01 -1.65680087e+00 4.08986449e-01 -3.33139151e-01
8.00987482e-01 -6.15544379e-01 -8.68276805e-02 -1.01348788e-01
4.65017200e-01 -8.33582103e-01 1.35170317e+00 7.28881955e-01
9.73657191e-01 -5.45549572e-01 4.17749844e-02 3.95646036e-01
-1.18648374e+00 1.44059852e-01 1.94931880e-01 -4.17060196e-01
7.75494814e-01 1.48767620e-01 -1.22539282e+00 2.98009187e-01
3.15275997e-01 5.39702773e-01 -3.75578046e-01 8.26899350e-01
-6.78700268e-01 2.33862102e-01 -3.18472413e-03 -1.52180389e-01
1.70922682e-01 -2.17815891e-01 7.07571864e-01 1.04601133e+00
6.45341277e-01 3.53695899e-01 4.68786107e-03 9.43287730e-01
-1.55144811e-01 7.19367899e-03 -8.35508049e-01 -1.49704978e-01
3.25344205e-01 1.20171523e+00 -9.42602098e-01 -8.19474101e-01
2.09197197e-02 1.55347848e+00 1.39194448e-02 3.95373583e-01
-9.16048467e-01 -2.50907034e-01 1.35511965e-01 7.33835459e-01
2.14269832e-01 -1.52121052e-01 -1.88272372e-01 -9.30690169e-01
-2.67929602e-02 -7.40383565e-01 6.59457454e-03 -2.08774137e+00
-2.33491138e-01 7.23315954e-01 4.21173573e-01 -1.24706578e+00
-6.10326171e-01 2.65859902e-01 -8.61495495e-01 4.49562192e-01
-7.52031684e-01 -1.46902001e+00 -5.23454428e-01 3.18780929e-01
1.14493906e+00 3.77401829e-01 4.52289820e-01 3.72848772e-02
-1.76036626e-01 -1.60083234e-01 -2.33952537e-01 -7.71103986e-03
4.29247499e-01 -1.16867232e+00 5.97352326e-01 7.30486572e-01
3.69123071e-01 4.06504601e-01 1.21852863e+00 -7.22058654e-01
-7.34586060e-01 -8.61030340e-01 1.33891857e+00 -2.33689204e-01
4.39387172e-01 -5.52146792e-01 -5.37419558e-01 9.21561480e-01
8.29325736e-01 -7.58628905e-01 4.38385725e-01 -5.96782625e-01
6.91651255e-02 1.90845788e-01 -9.24875259e-01 1.20042396e+00
1.13714278e+00 -5.39362192e-01 -8.66573274e-01 4.52548236e-01
8.67932200e-01 -6.06401742e-01 -1.55540928e-01 -3.23412150e-01
6.34784520e-01 -8.85622680e-01 8.22788775e-01 -3.02038968e-01
9.52470839e-01 -5.32187819e-01 4.19761449e-01 -1.41606987e+00
-2.90254653e-01 -8.99228692e-01 3.44296038e-01 1.38167536e+00
3.60092103e-01 4.35359739e-02 4.52933699e-01 6.25800252e-01
-1.72279015e-01 -1.14730604e-01 -5.27997911e-01 -1.79160535e-01
-6.53218925e-01 -5.28173268e-01 7.35468626e-01 8.40448558e-01
5.10112107e-01 3.77734959e-01 -7.99734235e-01 -7.37546682e-02
1.00458480e-01 9.26956013e-02 9.37861800e-01 -8.23171318e-01
-1.22418761e-01 -2.10517004e-01 -4.58683223e-02 -1.24830818e+00
-2.18395054e-01 -7.06847131e-01 2.23434269e-01 -2.17941666e+00
3.08905333e-01 3.53890061e-02 5.14233649e-01 6.37436152e-01
-9.17682797e-03 2.07928091e-01 7.15720356e-01 2.25360036e-01
-6.02541566e-01 4.33676153e-01 1.49081397e+00 -1.90484524e-01
-5.66115856e-01 -3.10149789e-01 -5.01763403e-01 8.27110708e-01
7.05457628e-01 -3.09805989e-01 -6.43623471e-01 -4.80873674e-01
4.50804502e-01 6.29121304e-01 6.22286558e-01 -9.70353901e-01
1.20532773e-01 -2.87045509e-01 4.77109820e-01 -8.65891099e-01
5.27688146e-01 -7.59935975e-01 7.84499466e-01 3.59433055e-01
-8.51216137e-01 1.97886810e-01 4.65031192e-02 6.35533705e-02
-4.87128831e-02 -2.09411696e-01 6.16828442e-01 -2.47522697e-01
-5.00562310e-01 -7.26094544e-02 -8.11947286e-01 -5.35497010e-01
1.35579598e+00 -5.14768660e-01 -3.26248221e-02 -1.04426193e+00
-1.21897185e+00 2.27929309e-01 5.30767441e-01 5.64669609e-01
6.70470595e-01 -1.65912223e+00 -6.16069853e-01 3.39057930e-02
8.72348323e-02 1.21163361e-01 5.76948941e-01 1.68151438e-01
-8.82911444e-01 2.18178183e-01 -2.68444687e-01 -4.00738150e-01
-1.70702040e+00 6.54998183e-01 6.26940578e-02 -1.61498010e-01
-8.07946980e-01 6.93158805e-01 5.74329257e-01 1.91955358e-01
-1.89542212e-02 -6.74867153e-01 -4.80472028e-01 2.71654069e-01
8.17322791e-01 1.30143970e-01 -5.28854430e-01 -8.16945136e-01
1.05048642e-01 3.66267771e-01 1.51583597e-01 -6.32488549e-01
1.23043334e+00 -5.36422431e-01 3.89397480e-02 8.55679214e-01
9.16685820e-01 -3.07411790e-01 -1.44978166e+00 -5.14207482e-02
-2.85799980e-01 -4.36711788e-01 -3.04182321e-01 -8.62635791e-01
-7.10797668e-01 6.97788417e-01 2.72354603e-01 3.44816685e-01
1.23819923e+00 2.26096526e-01 7.93634593e-01 1.01128340e-01
1.69831604e-01 -9.46742952e-01 4.20594186e-01 3.51022124e-01
1.36326480e+00 -5.85312724e-01 -7.46782050e-02 -2.30957419e-01
-9.31201637e-01 1.18807518e+00 2.22979859e-01 2.11187556e-01
8.42019841e-02 -1.41689265e-02 -3.04512791e-02 -3.38809997e-01
-7.43608117e-01 -1.82740167e-01 8.61415341e-02 5.57624519e-01
3.42038721e-01 1.21588849e-01 -2.90949315e-01 -1.52520224e-01
-6.52897358e-01 3.14562708e-01 8.24811697e-01 5.58474243e-01
-4.98363167e-01 -9.16317821e-01 -6.09470069e-01 5.56614585e-02
8.10480192e-02 -1.77937984e-01 -5.66197455e-01 5.86594045e-01
4.91381288e-01 9.20680106e-01 3.46155107e-01 -1.66930154e-01
1.71682343e-01 2.07968757e-01 4.58849639e-01 -6.97166383e-01
-5.31464159e-01 3.24089383e-03 1.98097050e-01 -2.69952923e-01
-4.82106715e-01 -7.82895625e-01 -1.35375559e+00 -2.28403658e-01
8.30378830e-02 -1.25632674e-01 6.45233274e-01 1.03214383e+00
9.53695327e-02 6.45611763e-01 4.78137136e-01 -1.10230303e+00
7.76565969e-01 -9.90310073e-01 -4.79703248e-02 4.71851677e-01
2.73598373e-01 -2.42412895e-01 4.22544032e-02 1.06947744e+00] | [11.188536643981934, 0.7256268858909607] |
4bded800-ee49-427c-8edb-34dd698e9263 | assessing-robustness-of-text-classification | 2010.02004 | null | https://arxiv.org/abs/2010.02004v2 | https://arxiv.org/pdf/2010.02004v2.pdf | Assessing Robustness of Text Classification through Maximal Safe Radius Computation | Neural network NLP models are vulnerable to small modifications of the input that maintain the original meaning but result in a different prediction. In this paper, we focus on robustness of text classification against word substitutions, aiming to provide guarantees that the model prediction does not change if a word is replaced with a plausible alternative, such as a synonym. As a measure of robustness, we adopt the notion of the maximal safe radius for a given input text, which is the minimum distance in the embedding space to the decision boundary. Since computing the exact maximal safe radius is not feasible in practice, we instead approximate it by computing a lower and upper bound. For the upper bound computation, we employ Monte Carlo Tree Search in conjunction with syntactic filtering to analyse the effect of single and multiple word substitutions. The lower bound computation is achieved through an adaptation of the linear bounding techniques implemented in tools CNN-Cert and POPQORN, respectively for convolutional and recurrent network models. We evaluate the methods on sentiment analysis and news classification models for four datasets (IMDB, SST, AG News and NEWS) and a range of embeddings, and provide an analysis of robustness trends. We also apply our framework to interpretability analysis and compare it with LIME. | ['Marta Kwiatkowska', 'Anthony Hartshorn', 'Benjie Wang', 'Luca Laurenti', 'Min Wu', 'Emanuele La Malfa'] | 2020-10-01 | null | https://aclanthology.org/2020.findings-emnlp.266 | https://aclanthology.org/2020.findings-emnlp.266.pdf | findings-of-the-association-for-computational | ['news-classification'] | ['natural-language-processing'] | [ 2.06275403e-01 1.82276070e-01 3.06426808e-02 -5.14320612e-01
-3.45239520e-01 -8.41129184e-01 7.08742082e-01 7.89713502e-01
-9.25495446e-01 5.08555055e-01 2.54605234e-01 -5.88842213e-01
-1.22050464e-01 -7.56212950e-01 -8.90924096e-01 -5.54271817e-01
-4.88618053e-02 1.04613669e-01 2.21137658e-01 -1.65828243e-01
2.51896530e-01 4.97327656e-01 -1.53956246e+00 4.94527191e-01
4.74975646e-01 1.06154406e+00 -4.20366198e-01 6.69802785e-01
-1.82616174e-01 4.20013338e-01 -6.92276955e-01 -8.42247665e-01
2.84204334e-01 -2.74622053e-01 -8.93259943e-01 -4.24649626e-01
2.39867747e-01 8.29312950e-02 1.89887896e-01 1.23850226e+00
4.04991657e-01 2.69716829e-01 8.41507971e-01 -9.64197934e-01
-3.30961287e-01 1.11603665e+00 -1.67333603e-01 3.00462961e-01
2.48306125e-01 -1.91773474e-01 1.17938054e+00 -7.30990767e-01
6.55752063e-01 1.19662476e+00 9.15114760e-01 3.78675431e-01
-1.30081069e+00 -3.36393893e-01 3.13479215e-01 1.28153935e-01
-1.10606074e+00 -4.01048452e-01 3.28406125e-01 -3.68128389e-01
1.16062129e+00 4.06780303e-01 3.24744612e-01 1.13096571e+00
4.41142768e-01 4.47587371e-01 7.76722372e-01 -6.45151734e-01
5.02748489e-01 2.82837152e-01 5.49231052e-01 4.89839584e-01
2.64022648e-01 -6.29565790e-02 -4.28753257e-01 -3.02049220e-01
-1.28812507e-01 -3.29143465e-01 -2.72936672e-01 -5.96216321e-02
-9.30058420e-01 9.85364079e-01 3.91931087e-01 4.42918360e-01
-2.25191459e-01 2.06741050e-01 7.66037226e-01 4.51069653e-01
7.20075369e-01 4.06519502e-01 -7.06975520e-01 -3.42931896e-02
-7.82959104e-01 2.46752501e-01 9.86701965e-01 3.87568206e-01
3.56654882e-01 -2.27629185e-01 1.68554373e-02 7.60279655e-01
1.93713769e-01 3.95891927e-02 8.03686857e-01 -3.13451231e-01
3.35477471e-01 5.78588605e-01 -4.29760851e-02 -1.01977229e+00
-5.99892497e-01 -5.00217855e-01 -7.42456317e-01 1.83630362e-01
6.48299873e-01 -1.82321414e-01 -7.60517716e-01 1.76807296e+00
2.91212440e-01 -1.60331596e-02 1.70979217e-01 5.29511511e-01
5.15280187e-01 7.36379385e-01 1.91133972e-02 -1.58100352e-01
1.35806668e+00 -6.85541928e-01 -5.15735328e-01 -1.94024250e-01
1.15800500e+00 -5.75302243e-01 1.15042746e+00 5.18672168e-01
-7.41832137e-01 -2.70020038e-01 -1.21239007e+00 -6.62438944e-02
-7.28356600e-01 -4.43682149e-02 6.86261952e-02 8.18469644e-01
-6.89068675e-01 9.82150435e-01 -7.29948103e-01 -3.02535027e-01
3.96739244e-02 1.89224213e-01 -4.75567579e-01 3.43104422e-01
-1.58915460e+00 1.08893597e+00 6.85341179e-01 1.62419572e-01
-1.40135780e-01 -6.04617417e-01 -8.01149428e-01 9.07418653e-02
4.08199579e-02 -5.00597239e-01 1.09932375e+00 -1.17745173e+00
-1.39409578e+00 7.76898921e-01 2.22795662e-02 -1.08553803e+00
7.48107731e-01 -2.31883213e-01 -3.73564929e-01 -3.40410262e-01
-3.05211455e-01 3.42903078e-01 8.85112643e-01 -6.43076956e-01
-3.55437607e-01 -3.69369507e-01 1.69440359e-02 -1.39811784e-01
-4.39835101e-01 2.17533156e-01 -2.16038749e-02 -7.18469262e-01
-6.53570667e-02 -9.27871466e-01 -2.44376481e-01 -1.01147041e-01
-5.18136919e-01 -1.80989340e-01 2.21887752e-01 -6.19354367e-01
1.44550717e+00 -2.30823565e+00 1.55055389e-01 4.72770423e-01
-2.58882552e-01 3.23475242e-01 4.43147570e-02 3.36916387e-01
-2.41413549e-01 3.78413230e-01 -3.67800146e-01 -4.74771798e-01
1.37201384e-01 2.48887524e-01 -6.39355183e-01 6.30018651e-01
2.32007831e-01 5.92743039e-01 -5.10633469e-01 -1.32010520e-01
-1.24297710e-02 3.88144702e-01 -6.05275154e-01 -3.55983645e-01
-3.12584698e-01 -4.21826303e-01 -1.26600429e-01 -5.18496670e-02
5.19168496e-01 7.05981813e-03 1.98146120e-01 -2.16395631e-01
-8.28197375e-02 7.19056606e-01 -1.48012018e+00 1.10443234e+00
-8.07942092e-01 4.30044591e-01 -4.28905338e-01 -9.18926120e-01
1.02263308e+00 4.06754501e-02 -3.04908127e-01 -3.40762943e-01
3.38661999e-01 1.95269510e-01 7.65376836e-02 -8.76902044e-02
5.80451906e-01 -1.69983834e-01 -1.01080716e-01 5.61111689e-01
-1.45274594e-01 3.59263629e-01 2.72163332e-01 -6.02144673e-02
1.17609203e+00 -8.95373672e-02 3.76210600e-01 -3.94458562e-01
6.41302824e-01 -3.37047905e-01 3.84578794e-01 6.50070667e-01
9.23036858e-02 4.14871544e-01 1.04356134e+00 -5.82711875e-01
-1.06618536e+00 -7.93113828e-01 -4.03157234e-01 1.25243247e+00
-3.24102223e-01 -6.78905427e-01 -9.37596142e-01 -8.34169388e-01
3.55569199e-02 1.05565453e+00 -1.04124236e+00 -3.95777643e-01
-3.26196820e-01 -1.05365610e+00 7.38616943e-01 4.32678580e-01
1.89619541e-01 -9.24016714e-01 -8.65698278e-01 1.45803660e-01
1.05372347e-01 -9.06721532e-01 -2.52947837e-01 6.63982332e-01
-7.35337436e-01 -8.24921072e-01 -2.55710572e-01 -3.88341248e-01
4.94029671e-01 -4.89319026e-01 8.90788257e-01 3.40430485e-03
7.51154348e-02 -2.50742108e-01 -4.96742040e-01 -3.58997285e-01
-8.04923594e-01 3.56887132e-01 2.66981781e-01 4.54500392e-02
2.64256507e-01 -2.88071990e-01 -1.54398173e-01 2.33858004e-01
-1.22989118e+00 -1.13836899e-01 2.27450266e-01 9.05145407e-01
4.33617502e-01 1.18850313e-01 3.73289764e-01 -9.55495715e-01
8.66120458e-01 -4.43117648e-01 -6.88139498e-01 2.91301012e-01
-6.78237677e-01 5.71699202e-01 1.01035380e+00 -5.24088562e-01
-6.64054573e-01 1.45992056e-01 -3.81551623e-01 -6.21233545e-02
2.02130359e-02 7.86690712e-01 -6.91278279e-02 3.61410737e-01
9.50008452e-01 -1.29702434e-01 -2.66330004e-01 -6.15536809e-01
5.42573750e-01 5.97349465e-01 2.61353016e-01 -1.97097093e-01
4.93507445e-01 2.64058888e-01 -1.57843515e-01 -7.04087198e-01
-8.95097077e-01 -3.97273302e-02 -5.97625196e-01 8.27332661e-02
4.68658566e-01 -3.25898647e-01 -5.90738714e-01 2.58164316e-01
-1.41773605e+00 -6.21133484e-02 -2.99114555e-01 3.02265048e-01
-1.35080278e-01 3.27746928e-01 -3.91445220e-01 -6.86670959e-01
-5.95851123e-01 -1.02415848e+00 5.61591685e-01 -9.18791443e-02
-6.62241399e-01 -1.02679837e+00 1.49707377e-01 -2.17662901e-01
2.82960743e-01 2.89087117e-01 1.31809771e+00 -1.39999747e+00
3.13248664e-01 -5.13124704e-01 6.68671541e-03 7.83178926e-01
-1.89401880e-01 3.39282393e-01 -1.04085529e+00 -1.10001929e-01
-2.66190749e-02 3.44037265e-02 1.15263879e+00 1.58850268e-01
8.80334675e-01 -5.85633576e-01 -1.04470618e-01 5.26004016e-01
1.14033401e+00 -1.05895840e-01 5.96148193e-01 5.71175456e-01
2.41586000e-01 8.16957653e-01 3.98441970e-01 4.31368858e-01
-2.96130180e-01 5.78961730e-01 4.30587441e-01 5.06273508e-01
3.42405736e-01 -1.76625356e-01 5.82217455e-01 5.83474278e-01
3.77792984e-01 -5.19808531e-01 -9.79901075e-01 4.09715474e-01
-1.74631906e+00 -6.85367405e-01 -1.18520983e-01 2.29130292e+00
8.07031631e-01 8.04357469e-01 -4.45622504e-02 5.06999195e-01
6.33705795e-01 1.48681402e-01 -3.53171557e-01 -1.02453959e+00
-1.56308055e-01 2.54117310e-01 5.40859640e-01 7.21700430e-01
-9.54694450e-01 8.05074155e-01 5.93988132e+00 9.10636604e-01
-1.13208330e+00 6.59157187e-02 7.48136044e-01 -2.12663174e-01
-2.77718395e-01 -1.26973778e-01 -8.56456935e-01 5.33541977e-01
1.30991137e+00 -1.71338931e-01 2.00927332e-01 9.32645082e-01
-3.38916630e-02 -6.00658692e-02 -1.22095191e+00 4.51193184e-01
-5.21454290e-02 -1.34505904e+00 1.82620078e-01 -1.85881898e-01
2.47929677e-01 1.67078346e-01 -3.32419910e-02 4.50105555e-02
9.28172916e-02 -9.48487401e-01 9.43960786e-01 3.91860396e-01
3.45044702e-01 -1.12220597e+00 1.04694152e+00 3.55780810e-01
-7.12349296e-01 -3.70331965e-02 -3.66931021e-01 -6.06120005e-02
-1.41740218e-01 7.93846786e-01 -8.30819845e-01 3.68135273e-01
6.05102777e-01 5.06616831e-01 -6.20140254e-01 5.88150859e-01
-2.26520404e-01 5.09715736e-01 -5.89388072e-01 -5.41011453e-01
1.57639608e-01 3.41196358e-02 6.04043841e-01 1.52698028e+00
1.04842596e-01 -4.26800549e-01 -5.37703753e-01 7.76459455e-01
-1.83870867e-01 4.02559370e-01 -4.03876454e-01 2.21938342e-02
4.08438534e-01 1.05190229e+00 -8.29095066e-01 -2.23898977e-01
-3.80395204e-02 9.34347510e-01 4.57482576e-01 2.55065225e-02
-8.21149349e-01 -6.81042790e-01 6.93848848e-01 -9.57837924e-02
3.38854462e-01 6.84533492e-02 -3.37870598e-01 -8.81633282e-01
2.76611328e-01 -8.54359210e-01 4.05155718e-01 -4.62337524e-01
-1.11692834e+00 8.72716367e-01 -1.09993242e-01 -8.57569158e-01
-4.19525385e-01 -8.93761098e-01 -6.08272135e-01 8.18942547e-01
-1.05508518e+00 -5.97669542e-01 2.90069401e-01 1.33957341e-01
1.63751528e-01 5.80476597e-02 9.30469394e-01 -3.96362506e-03
-7.79395401e-01 9.60076749e-01 1.78772599e-01 1.48419082e-01
5.15509069e-01 -1.02194417e+00 7.40762830e-01 9.08929884e-01
1.85462087e-01 7.11410582e-01 1.25540304e+00 -3.82960916e-01
-9.87534046e-01 -1.12659812e+00 1.14766490e+00 -4.54154074e-01
9.10692751e-01 -5.28521836e-01 -1.19798887e+00 6.06756806e-01
-1.94189712e-01 2.33337339e-02 5.54554462e-01 1.45730868e-01
-6.96018815e-01 -7.81982020e-03 -1.01527441e+00 7.65799284e-01
7.69906759e-01 -4.10674512e-01 -7.56374538e-01 1.64297819e-01
1.01987922e+00 -1.18012093e-01 -7.82463908e-01 4.30676728e-01
6.91855967e-01 -9.66176093e-01 5.92842281e-01 -9.04676139e-01
4.81911957e-01 -1.39110982e-01 -3.04569215e-01 -1.39220715e+00
2.79589556e-02 -4.93758291e-01 7.97099173e-02 1.30815756e+00
9.28431809e-01 -7.56247401e-01 4.88641888e-01 5.63821793e-01
1.47510901e-01 -9.25516665e-01 -1.32866156e+00 -8.59926343e-01
2.64954925e-01 -9.83392715e-01 6.38448119e-01 7.86074162e-01
2.17297226e-01 2.94395506e-01 -9.29061919e-02 1.30553782e-01
1.38318585e-02 -2.51844257e-01 5.51819563e-01 -1.12396502e+00
-2.63142347e-01 -6.38764262e-01 -5.87734580e-01 -4.78734076e-01
4.46586609e-01 -9.96816635e-01 -1.84563547e-01 -1.02074397e+00
-3.27817380e-01 -3.10350340e-02 -4.90973681e-01 5.35851717e-01
1.65215060e-01 1.38222739e-01 7.11173937e-02 -5.61134927e-02
-2.12008223e-01 4.56314772e-01 2.93534994e-01 1.79247975e-01
-5.58047667e-02 5.20114508e-03 -3.86270612e-01 9.74442601e-01
9.05241191e-01 -7.88353443e-01 5.49088679e-02 -2.25397497e-01
1.16917968e+00 -5.43398976e-01 3.41424882e-01 -6.67703271e-01
5.07246591e-02 2.71360457e-01 2.35872954e-01 -2.88355708e-01
-6.10795841e-02 -8.17240238e-01 6.03703894e-02 7.94533908e-01
-8.75253141e-01 2.22529843e-01 4.19561237e-01 5.02410650e-01
7.09631965e-02 -7.76357949e-01 8.76327753e-01 3.08475912e-01
-2.73449093e-01 -2.80957997e-01 -5.14032066e-01 -6.57461435e-02
7.31957853e-01 4.48761433e-02 -2.65108943e-01 -8.54315236e-02
-8.27778637e-01 -2.93573993e-03 4.15103346e-01 6.00077689e-01
3.28752726e-01 -1.15001428e+00 -5.64936519e-01 3.15934330e-01
2.70120770e-01 -4.07684177e-01 -5.49556538e-02 6.12859309e-01
-6.21820033e-01 1.88300341e-01 1.75499007e-01 -1.52851537e-01
-1.40092874e+00 6.40277684e-01 6.64218605e-01 -3.36903334e-01
-4.41877127e-01 6.09909177e-01 -2.78204381e-01 -5.95678091e-01
3.57684046e-01 -9.06310916e-01 -1.45026356e-01 3.08366299e-01
6.43667459e-01 2.81337529e-01 5.40164411e-01 -4.32927042e-01
-5.61534703e-01 3.28188211e-01 -2.61278480e-01 -1.04630552e-01
1.19667268e+00 5.38295088e-03 -1.28175303e-01 6.59027576e-01
1.34443724e+00 3.43030551e-04 -8.30659091e-01 -2.99338251e-01
4.47296739e-01 -9.90093686e-03 9.96043012e-02 -8.21141660e-01
-6.77522123e-01 7.76977479e-01 6.24532044e-01 4.68186349e-01
8.05582464e-01 -1.18950821e-01 3.78813624e-01 6.85480237e-01
-2.27913409e-01 -1.21674371e+00 -4.55860019e-01 8.13443899e-01
9.18516159e-01 -7.89263725e-01 -1.22672981e-02 -6.91807568e-02
-3.95879924e-01 1.32459164e+00 9.88879502e-02 -1.32824540e-01
7.05554307e-01 3.05848390e-01 -5.94778694e-02 1.67032704e-01
-9.80834723e-01 2.21843883e-01 2.97121346e-01 -1.52946590e-03
4.75802213e-01 -1.83682457e-01 -4.87491041e-01 6.79658413e-01
-6.24238670e-01 -2.21628100e-01 4.92522091e-01 5.87839246e-01
-4.15735900e-01 -8.51838410e-01 -1.91996887e-01 3.96269321e-01
-7.27685928e-01 -3.31322819e-01 -4.73817706e-01 7.69667208e-01
4.86313254e-02 7.12546229e-01 2.70131439e-01 -4.56056178e-01
5.86240888e-01 5.37898362e-01 2.18503997e-01 -4.64798272e-01
-7.95252979e-01 -3.99567306e-01 3.71096849e-01 -3.05363715e-01
-3.60622965e-02 -6.37663186e-01 -1.29679084e+00 -3.69237870e-01
-5.33887386e-01 5.78234904e-02 9.67156112e-01 1.18429494e+00
2.78017938e-01 4.17457581e-01 3.18069160e-01 -4.12819356e-01
-8.76932621e-01 -1.08098447e+00 -3.97039860e-01 3.77413243e-01
3.30431730e-01 -2.73336828e-01 -8.13799918e-01 -1.52961910e-01] | [10.302454948425293, 8.317198753356934] |
af9d3253-65c3-4301-9c24-e5dbfc597bb7 | deepscalper-a-risk-aware-deep-reinforcement | 2201.09058 | null | https://arxiv.org/abs/2201.09058v3 | https://arxiv.org/pdf/2201.09058v3.pdf | DeepScalper: A Risk-Aware Reinforcement Learning Framework to Capture Fleeting Intraday Trading Opportunities | Reinforcement learning (RL) techniques have shown great success in many challenging quantitative trading tasks, such as portfolio management and algorithmic trading. Especially, intraday trading is one of the most profitable and risky tasks because of the intraday behaviors of the financial market that reflect billions of rapidly fluctuating capitals. However, a vast majority of existing RL methods focus on the relatively low frequency trading scenarios (e.g., day-level) and fail to capture the fleeting intraday investment opportunities due to two major challenges: 1) how to effectively train profitable RL agents for intraday investment decision-making, which involves high-dimensional fine-grained action space; 2) how to learn meaningful multi-modality market representation to understand the intraday behaviors of the financial market at tick-level. Motivated by the efficient workflow of professional human intraday traders, we propose DeepScalper, a deep reinforcement learning framework for intraday trading to tackle the above challenges. Specifically, DeepScalper includes four components: 1) a dueling Q-network with action branching to deal with the large action space of intraday trading for efficient RL optimization; 2) a novel reward function with a hindsight bonus to encourage RL agents making trading decisions with a long-term horizon of the entire trading day; 3) an encoder-decoder architecture to learn multi-modality temporal market embedding, which incorporates both macro-level and micro-level market information; 4) a risk-aware auxiliary task to maintain a striking balance between maximizing profit and minimizing risk. Through extensive experiments on real-world market data spanning over three years on six financial futures, we demonstrate that DeepScalper significantly outperforms many state-of-the-art baselines in terms of four financial criteria. | ['Bo An', 'Jian Li', 'Wanqi Xue', 'Rundong Wang', 'Junlei Zhu', 'Xu He', 'Shuo Sun'] | 2021-12-15 | null | null | null | null | ['algorithmic-trading'] | ['time-series'] | [-6.27973557e-01 -1.85119584e-01 -4.12206054e-01 -1.29897982e-01
-9.10519361e-01 -6.58610761e-01 7.09765255e-01 -1.64472058e-01
-4.72466648e-01 5.81101835e-01 2.52122015e-01 -6.75977767e-01
-1.52291864e-01 -9.26220894e-01 -6.76568568e-01 -3.60223800e-01
-5.53948462e-01 6.99612856e-01 -1.66193619e-01 -3.88585180e-01
2.40927264e-01 2.34234199e-01 -8.32788289e-01 -2.73137372e-02
5.59092283e-01 1.81828260e+00 -3.54307622e-01 3.29958081e-01
-6.80506676e-02 1.19144237e+00 -4.25338387e-01 -6.18258476e-01
9.43158090e-01 -5.47241449e-01 -2.35133871e-01 -5.87599725e-02
-4.44523990e-01 -9.92901802e-01 -3.86500537e-01 9.42648709e-01
4.29218024e-01 1.71443403e-01 5.22471428e-01 -1.08732045e+00
-7.63790846e-01 8.26252878e-01 -7.69544065e-01 4.16706651e-01
-3.72598380e-01 7.25911200e-01 1.65223110e+00 -5.86342096e-01
2.27326676e-02 1.21989548e+00 1.54879898e-01 1.93079486e-01
-1.10916793e+00 -8.04970682e-01 2.26550922e-01 -2.70019889e-01
-3.53138477e-01 1.83812901e-01 9.18324888e-01 -3.40060502e-01
1.12913060e+00 -1.72022060e-01 8.70737314e-01 1.03938270e+00
6.50929451e-01 9.55479860e-01 1.26405621e+00 1.61859602e-01
3.81063491e-01 -1.65672079e-01 -2.36747548e-01 1.82061538e-01
-3.25129628e-01 1.13491392e+00 -3.24940085e-01 -2.59713382e-01
1.15772760e+00 2.93161929e-01 4.94605064e-01 -2.31134161e-01
-1.21759880e+00 1.43779528e+00 5.83894253e-01 -4.85321507e-02
-7.18165457e-01 4.50376004e-01 5.00404537e-01 1.02409554e+00
5.18487155e-01 8.34722340e-01 -9.65690970e-01 -3.83098781e-01
-7.44260609e-01 6.93941355e-01 1.03699398e+00 6.10040426e-01
4.43647116e-01 5.32388151e-01 -2.31392860e-01 5.66137433e-01
4.05562192e-01 5.65151930e-01 5.15531719e-01 -9.45451915e-01
8.37473273e-01 3.65633845e-01 3.77411276e-01 -3.96938503e-01
-3.92699540e-01 -4.36398953e-01 -5.70657372e-01 5.37793219e-01
5.15986323e-01 -5.48980713e-01 -5.60884178e-01 1.57112467e+00
6.85636103e-02 5.19576184e-02 1.82343498e-01 7.82806396e-01
-9.53305289e-02 8.11798453e-01 -1.55005455e-01 -2.42858469e-01
1.50631952e+00 -1.17916119e+00 -4.07587528e-01 -1.89527601e-01
2.19761431e-01 -5.62573671e-01 9.20177460e-01 3.65497887e-01
-1.17995262e+00 -2.89266884e-01 -9.35290098e-01 4.16862488e-01
-2.85977840e-01 -9.50542763e-02 8.91633451e-01 1.81441054e-01
-5.33874094e-01 6.79236352e-01 -7.79879749e-01 7.91548312e-01
3.26608151e-01 3.86859119e-01 4.90140915e-01 6.98225617e-01
-1.69676256e+00 8.42754900e-01 2.89211154e-01 1.93546683e-01
-1.19079375e+00 -8.13652217e-01 -6.45765305e-01 1.92652941e-01
1.04309261e+00 -3.53177905e-01 1.76973689e+00 -9.37477112e-01
-1.95108080e+00 3.21473628e-01 8.81053925e-01 -1.13314259e+00
9.47380841e-01 -5.05805135e-01 -5.86514831e-01 -6.55940175e-02
-1.24337025e-01 4.47498918e-01 1.08893144e+00 -3.07562679e-01
-7.42163420e-01 -2.80591011e-01 7.03702681e-03 4.20874864e-01
1.65823430e-01 -1.05010875e-01 1.91983938e-01 -1.46611834e+00
-4.78672802e-01 -8.34529936e-01 -3.65818977e-01 -3.62405509e-01
-5.75173125e-02 -2.30187073e-01 5.17086923e-01 -8.16906869e-01
1.10419583e+00 -2.00799084e+00 -1.08309858e-01 1.92952588e-01
1.48566484e-01 -3.98420841e-02 7.77832000e-03 5.68116426e-01
-1.43880084e-01 -1.38143644e-01 -2.10254148e-01 1.67451590e-01
6.69497728e-01 -3.71804610e-02 -1.05878603e+00 5.95396757e-02
4.07426327e-01 1.55878222e+00 -7.62188971e-01 1.35600641e-01
1.36374563e-01 -2.72000283e-01 -3.26987624e-01 7.01446950e-01
-8.89599741e-01 4.53568459e-01 -7.88208902e-01 8.82968724e-01
3.91871631e-01 -2.86991000e-01 -2.67498612e-01 5.04852772e-01
-4.33991924e-02 3.70631963e-01 -1.17334473e+00 1.09000170e+00
-3.23235631e-01 -8.01889226e-02 1.83322374e-02 -9.38865304e-01
9.47714031e-01 2.30364457e-01 5.95812082e-01 -1.37776983e+00
4.56835283e-03 5.75575650e-01 -2.64608320e-02 1.42526329e-01
4.38252240e-01 -7.43318379e-01 -5.35205543e-01 1.30759764e+00
-3.70972812e-01 -1.83495536e-01 -1.59495533e-01 -2.19155073e-01
1.00416648e+00 2.48114362e-01 1.41117468e-01 7.57678300e-02
4.71813902e-02 -3.18796426e-01 8.30594122e-01 5.60772479e-01
-3.37527186e-01 6.84643071e-03 1.19359255e+00 -7.65452385e-01
-9.40050662e-01 -1.30735660e+00 1.08205274e-01 1.08938420e+00
-2.52593428e-01 3.71588111e-01 -3.05414706e-01 -9.03992891e-01
7.75872409e-01 7.27224112e-01 -6.49078846e-01 -4.79841121e-02
-5.36177874e-01 -9.73676145e-01 2.59061217e-01 6.02899492e-01
7.22917616e-01 -1.71768558e+00 -9.43886995e-01 6.97311521e-01
5.96564353e-01 -6.58269882e-01 -8.78833950e-01 5.61934471e-01
-7.79553473e-01 -9.56879079e-01 -9.04728293e-01 -2.29839802e-01
-6.32782057e-02 -3.27357203e-01 1.30186927e+00 -4.65698689e-01
-2.33960431e-03 1.47116661e-01 -1.33526977e-02 -5.63050210e-01
-2.65634596e-01 -2.46013418e-01 -2.05242336e-01 1.04960181e-01
4.37006086e-01 -2.51206666e-01 -8.57127130e-01 3.68889630e-01
-1.03619432e+00 -4.06662852e-01 8.86311948e-01 1.18712795e+00
5.64181685e-01 1.08635023e-01 1.12040448e+00 -8.55475664e-01
9.71176505e-01 -5.75629413e-01 -1.44732141e+00 2.57483929e-01
-1.21273303e+00 4.52499837e-01 6.27728641e-01 -5.16848624e-01
-8.97611260e-01 -5.11300743e-01 1.39350325e-01 -4.19902295e-01
7.44981587e-01 5.47161698e-01 2.04594657e-01 4.96008426e-01
-9.38229337e-02 2.16973349e-01 3.67078990e-01 -3.77846271e-01
3.78906339e-01 2.21430153e-01 1.07750006e-01 -6.05330586e-01
7.42121935e-01 4.50362964e-03 -7.87219033e-02 2.26503965e-02
-7.33166873e-01 2.81038657e-02 -8.36324692e-02 8.55579823e-02
7.54280448e-01 -1.04993176e+00 -1.12519145e+00 8.77979636e-01
-4.10400033e-01 -7.88571537e-01 -6.95158839e-01 7.79446185e-01
-1.06890297e+00 -8.44385326e-02 -1.23166800e+00 -1.00400305e+00
-4.62382048e-01 -1.18637836e+00 8.97904813e-01 7.05524161e-02
1.23477384e-01 -1.13319278e+00 3.94453347e-01 3.15629005e-01
4.74577129e-01 4.56140429e-01 1.13335097e+00 -7.35077322e-01
-8.73299003e-01 2.22386166e-01 -1.68712422e-01 7.23501801e-01
1.11553647e-01 -4.03907984e-01 -6.21157706e-01 -3.96702379e-01
3.20006669e-01 -7.92132437e-01 1.03542364e+00 4.62861419e-01
6.63638830e-01 -4.37080204e-01 4.29617316e-01 5.45408845e-01
1.05662727e+00 6.24107480e-01 1.93913996e-01 6.15979016e-01
2.50039399e-01 6.05171919e-01 1.04434097e+00 7.82209635e-01
3.60870272e-01 3.02145720e-01 6.79761827e-01 1.82293713e-01
8.94466043e-01 -6.46162450e-01 6.84165716e-01 5.39189100e-01
4.47497666e-01 2.74797171e-01 -5.72058737e-01 1.81956768e-01
-1.93940663e+00 -9.52008843e-01 8.78567100e-01 2.21401834e+00
9.33744609e-01 5.89539886e-01 6.71358764e-01 -4.92148429e-01
1.67511731e-01 7.62830019e-01 -1.52692759e+00 -5.58790445e-01
-1.21394843e-01 1.37181759e-01 6.77879989e-01 2.29386389e-01
-1.20732296e+00 8.43569756e-01 5.56347466e+00 6.93076015e-01
-1.02796280e+00 -3.18334699e-01 1.11471272e+00 -2.80585974e-01
-5.77777147e-01 6.33547036e-03 -7.53381491e-01 7.66694009e-01
1.08556890e+00 -1.90769330e-01 9.66033876e-01 7.35128224e-01
3.72287109e-02 2.59730071e-01 -1.12429309e+00 7.70171106e-01
-7.23323584e-01 -1.40247822e+00 -1.79535076e-01 4.82239842e-01
6.05929375e-01 1.99034467e-01 7.16134608e-01 9.46238518e-01
8.29406917e-01 -1.23348725e+00 8.01137328e-01 5.93138635e-01
3.49675566e-01 -1.20633268e+00 4.82000828e-01 3.09977025e-01
-1.19846582e+00 -6.15316212e-01 -1.48298338e-01 1.63476661e-01
1.37319341e-01 4.43161756e-01 -2.92879969e-01 2.92324334e-01
4.93105233e-01 8.36160421e-01 -1.43430680e-02 1.65968224e-01
-1.06857918e-01 4.43358153e-01 -3.13168794e-01 -6.85208365e-02
1.10560811e+00 -7.96814978e-01 8.46393183e-02 4.51578259e-01
2.75560260e-01 -9.79914591e-02 1.28400430e-01 1.21235132e+00
-3.76386821e-01 -1.52801216e-01 -5.60574114e-01 -8.17380309e-01
1.61651269e-01 9.25769985e-01 -3.90986353e-01 -1.30530382e-02
-7.03238845e-01 7.11990356e-01 9.41590816e-02 3.54682148e-01
-9.15016592e-01 -1.82692379e-01 7.73345768e-01 -1.86287254e-01
7.12286711e-01 -2.28710189e-01 -6.44740760e-02 -1.18909061e+00
3.08652312e-01 -1.32697594e+00 8.65873575e-01 -2.00299546e-01
-1.54300070e+00 2.66849309e-01 -4.59386647e-01 -1.39917743e+00
-9.09878314e-01 -6.65151179e-01 -6.55809641e-01 1.04260790e+00
-2.08365417e+00 -9.05781329e-01 8.87940049e-01 6.23522401e-01
3.84075046e-01 -6.70368016e-01 3.96635652e-01 -1.88684374e-01
-3.27740431e-01 3.48542064e-01 4.53830928e-01 4.32467550e-01
3.87496859e-01 -1.73819578e+00 9.20577109e-01 3.72897029e-01
3.52746308e-01 3.57128054e-01 7.87620768e-02 -7.25167692e-01
-1.56454504e+00 -9.05702770e-01 3.87652516e-01 -4.35254693e-01
1.39767849e+00 -2.58032203e-01 -8.85239244e-01 7.13740349e-01
2.30661735e-01 -2.96804365e-02 4.00549710e-01 -1.49928570e-01
-3.82281870e-01 -4.42076862e-01 -8.45627189e-01 4.66197103e-01
2.22286567e-01 -6.81919038e-01 -8.32509995e-01 1.69734940e-01
9.15282667e-01 -3.74386430e-01 -1.15876222e+00 1.10143475e-01
4.57192242e-01 -8.57600451e-01 7.78645098e-01 -7.01525688e-01
3.04412097e-01 1.27073154e-01 1.28627285e-01 -1.40515554e+00
1.31115153e-01 -1.46371090e+00 -5.77491164e-01 7.84384489e-01
4.99311626e-01 -9.93566155e-01 6.56502724e-01 6.96542621e-01
4.27866340e-01 -1.13781846e+00 -1.13684046e+00 -8.11786294e-01
4.39063191e-01 -2.66455263e-01 1.07017934e+00 7.01487899e-01
-1.31098434e-01 2.48938650e-01 -7.50356853e-01 -3.07025790e-01
7.15953529e-01 8.50835979e-01 5.46391070e-01 -7.93433428e-01
-8.52315366e-01 -8.21904004e-01 2.88326740e-01 -1.20015419e+00
2.01743856e-01 -5.44256866e-01 -3.67294699e-01 -5.97700357e-01
-1.63315430e-01 -2.47729868e-01 -9.04210150e-01 9.45549011e-02
-1.96301509e-02 -6.61619186e-01 2.98120022e-01 2.93042779e-01
-3.64121318e-01 1.09777534e+00 1.43670905e+00 -1.46483749e-01
-3.07317048e-01 4.69407767e-01 -8.36320698e-01 3.09226692e-01
6.14235520e-01 -3.17264766e-01 -3.66593480e-01 -1.44297034e-02
5.97189009e-01 7.99862087e-01 3.27842832e-01 -1.63713515e-01
-9.37608257e-02 -6.18716717e-01 3.33045721e-01 -7.78093159e-01
2.11037442e-01 -7.40093946e-01 -1.80895180e-01 7.87812948e-01
-6.00281298e-01 7.65273631e-01 -5.94798774e-02 9.83681738e-01
-6.70903385e-01 2.78376162e-01 7.01596022e-01 -3.21555912e-01
-7.14047611e-01 8.17267656e-01 -7.95600414e-02 6.34037077e-01
1.22739720e+00 2.92049885e-01 -1.50300652e-01 -6.05615020e-01
-4.95385170e-01 1.05328763e+00 8.04751888e-02 7.10616231e-01
4.54604834e-01 -1.34169877e+00 -7.43097663e-01 3.68941754e-01
-8.99028108e-02 -1.07191563e-01 4.29006703e-02 4.98361230e-01
-3.97762895e-01 4.68057185e-01 -1.93043962e-01 1.52927395e-02
-2.51929858e-04 6.96612597e-01 6.96419239e-01 -1.16329813e+00
-7.87175119e-01 6.10210478e-01 4.13374335e-01 -5.40990889e-01
1.75929174e-01 -7.31444299e-01 9.20370594e-02 5.12105227e-01
5.43690264e-01 1.80892363e-01 -2.35551059e-01 -1.17207281e-02
9.57297161e-02 2.82109559e-01 -2.79883981e-01 -4.51014727e-01
1.67574012e+00 2.48088256e-01 2.51843840e-01 6.34933054e-01
9.82465923e-01 -5.02769053e-01 -2.32991052e+00 -5.47538817e-01
6.19878352e-01 -2.17885062e-01 -3.48707587e-02 -1.00618315e+00
-1.35393155e+00 8.75063121e-01 1.63164005e-01 4.09709245e-01
9.87867177e-01 -3.32967937e-01 1.37543225e+00 3.76076251e-01
3.60273153e-01 -1.79971457e+00 1.79843128e-01 5.96204519e-01
9.17987108e-01 -1.42169642e+00 -2.17699111e-01 7.93564439e-01
-1.41106510e+00 1.11532164e+00 2.04365551e-01 -5.05001068e-01
1.13468707e+00 2.96169162e-01 3.64999592e-01 -1.83011234e-01
-1.19616878e+00 1.87938944e-01 2.14419544e-01 -1.68327793e-01
-7.03261420e-02 4.15793628e-01 2.00360581e-01 8.74341607e-01
-1.49725974e-01 -1.79178193e-01 9.93173644e-02 8.91648054e-01
-1.43273011e-01 -9.85955596e-01 5.11665903e-02 7.51714528e-01
-8.22170436e-01 -7.26249516e-02 -2.77645111e-01 1.00157881e+00
-6.12456679e-01 3.65967572e-01 1.41794860e-01 -4.41237241e-02
4.60916489e-01 -2.24821232e-02 -6.56318367e-02 -3.74293119e-01
-1.13368726e+00 4.82434362e-01 -3.10581207e-01 -8.49205017e-01
1.03397757e-01 -9.43145037e-01 -1.33402348e+00 -2.89213181e-01
2.93473881e-02 1.96887448e-01 2.15549469e-01 8.25212538e-01
1.87417120e-01 2.87269205e-01 1.33586323e+00 -3.94460231e-01
-2.09348702e+00 -7.16797113e-01 -1.44651866e+00 2.34310672e-01
7.80131102e-01 -6.18412971e-01 -5.23166835e-01 -6.50127590e-01] | [4.433291912078857, 3.9462928771972656] |
a7349d5a-6b7e-4c55-95bf-ba2711b36a55 | human-s-role-in-the-loop | 2204.14192 | null | https://arxiv.org/abs/2204.14192v1 | https://arxiv.org/pdf/2204.14192v1.pdf | Human's Role in-the-Loop | Data integration has been recently challenged by the need to handle large volumes of data, arriving at high velocity from a variety of sources, which demonstrate varying levels of veracity. This challenging setting, often referred to as big data, renders many of the existing techniques, especially those that are human-intensive, obsolete. Big data also produces technological advancements such as Internet of things, cloud computing, and deep learning, and accordingly, provides a new, exciting, and challenging research agenda. Given the availability of data and the improvement of machine learning techniques, this blog discusses the respective roles of humans and machines in achieving cognitive tasks in matching, aiming to determine whether traditional roles of humans and machines are subject to change. Such investigation, we believe, will pave a way to better utilize both human and machine resources in new and innovative manners. We shall discuss two possible modes of change, namely humans out and humans in. Humans out aim at exploring out-of-the-box latent matching reasoning using machine learning algorithms when attempting to overpower human matcher performance. Pursuing out-of-the-box thinking, machine and deep learning can be involved in matching. Humans in explores how to better involve humans in the matching loop by assigning human matchers with a symmetric role to algorithmic matcher in the matching process. | ['Roee Shraga', 'Avigdor Gal'] | 2022-04-27 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [-2.07310185e-01 2.27310330e-01 -2.06970200e-01 -2.87283897e-01
-3.62038523e-01 -3.83532315e-01 7.53759623e-01 3.34504724e-01
-4.71826136e-01 2.80928195e-01 3.87873411e-01 -3.74243706e-01
-4.23346043e-01 -1.14259863e+00 -3.20988357e-01 -3.98090750e-01
4.86673683e-01 9.90466595e-01 -2.43178368e-01 -4.96076763e-01
4.08143610e-01 4.61140811e-01 -1.74403000e+00 3.87037545e-01
9.18871820e-01 7.75247812e-01 -1.25443563e-01 2.38640577e-01
-3.57970685e-01 9.64277565e-01 -3.23052108e-01 -1.06992447e+00
6.07622147e-01 -9.13959891e-02 -7.84334540e-01 -2.69419819e-01
1.86712459e-01 -1.65601864e-01 -1.59727201e-01 1.01496923e+00
4.93708432e-01 -6.87440857e-02 1.84869230e-01 -1.43938100e+00
-9.07709360e-01 5.48812270e-01 -3.61613423e-01 1.29938424e-01
5.06155252e-01 3.54325950e-01 8.80444050e-01 -7.70530820e-01
6.10988379e-01 1.20996678e+00 8.38532627e-01 2.60586619e-01
-9.74061310e-01 -7.42217898e-01 -3.14389884e-01 4.93774921e-01
-1.05494010e+00 -4.05284226e-01 7.44229913e-01 -8.10456336e-01
9.81139362e-01 1.60788283e-01 8.61266494e-01 8.74165356e-01
3.33340973e-01 5.25608361e-01 1.16865408e+00 -6.62715316e-01
3.34900558e-01 4.48130131e-01 1.05886683e-01 4.53714848e-01
5.67451000e-01 3.14025909e-01 -6.12920463e-01 -2.44812965e-01
5.44813097e-01 3.82472843e-01 1.96049750e-01 -2.77751148e-01
-1.42492127e+00 9.56422389e-01 3.97902697e-01 7.64164925e-01
-7.28762388e-01 -2.22340435e-01 3.32118213e-01 6.56147003e-01
3.59890908e-01 7.46519268e-01 -1.58311740e-01 -2.94159442e-01
-8.74822259e-01 4.91083711e-01 9.20008719e-01 6.21944368e-01
7.88109899e-01 -2.83798248e-01 1.98616996e-01 6.04744256e-01
3.67583670e-02 1.47605956e-01 6.43646657e-01 -1.19180381e+00
3.54078591e-01 1.29374170e+00 -6.04692623e-02 -1.41914523e+00
-5.70723712e-01 -5.60601234e-01 -1.01247549e+00 3.22857946e-01
4.74903077e-01 1.86180577e-01 -3.50942969e-01 1.50982249e+00
5.37695825e-01 -4.64521080e-01 -1.02420807e-01 9.19293284e-01
2.38150939e-01 1.63827658e-01 1.19128324e-01 7.67783821e-02
1.49100006e+00 -8.94141972e-01 -5.65120161e-01 -3.86167616e-01
8.12433898e-01 -8.47480595e-01 1.13266087e+00 4.05934721e-01
-1.33546102e+00 -7.97430634e-01 -8.75012994e-01 -2.42772952e-01
-6.68434024e-01 -6.03421330e-01 1.06030941e+00 7.08582103e-01
-1.05323577e+00 6.67256296e-01 -4.99362290e-01 -4.95131999e-01
5.54697394e-01 3.52391124e-01 -4.89556670e-01 -1.90164700e-01
-1.26509786e+00 1.34781146e+00 1.44751474e-01 2.03283772e-01
-1.01658903e-01 -7.44763792e-01 -4.47611392e-01 8.15004930e-02
4.39625889e-01 -1.19251740e+00 1.04999650e+00 -1.03759050e+00
-7.27230787e-01 1.21856856e+00 1.11637734e-01 -3.25120926e-01
7.26775765e-01 -4.89815138e-02 -5.78583241e-01 -2.57750094e-01
1.70534700e-01 5.07314980e-01 6.21400058e-01 -8.45740259e-01
-6.91981435e-01 -8.78212273e-01 -3.19815315e-02 -1.31508842e-01
-3.20948631e-01 2.38454893e-01 2.34293774e-01 -2.91182756e-01
2.78444231e-01 -6.81034803e-01 -2.51580209e-01 1.46907987e-02
2.55536705e-01 -5.80944538e-01 3.96383882e-01 -3.78357112e-01
1.12881887e+00 -1.84799314e+00 -2.70499326e-02 7.57277105e-03
8.58006656e-01 3.05006325e-01 2.20508464e-02 7.54049659e-01
-1.58914831e-02 9.11947563e-02 4.15115625e-01 -1.27094820e-01
3.82076412e-01 1.50103733e-01 -2.17262194e-01 2.10132971e-01
-1.01432785e-01 1.20588362e+00 -8.46390128e-01 -4.58440214e-01
-5.84623814e-02 8.83775130e-02 -5.52602232e-01 2.75681615e-01
-1.35824392e-02 1.85586438e-01 -3.22219044e-01 8.23149979e-01
4.63229746e-01 -4.26848978e-01 6.55042455e-02 7.95808509e-02
-2.29861513e-01 7.78237283e-02 -8.65576863e-01 1.39099586e+00
-2.32113793e-01 5.50586522e-01 7.83434510e-02 -1.06294799e+00
1.01025867e+00 2.38695011e-01 4.31759953e-01 -1.30157888e+00
1.59457013e-01 5.95099866e-01 3.48594606e-01 -7.81282902e-01
6.56882048e-01 -4.66421038e-01 -6.96473643e-02 7.58509696e-01
-3.22482854e-01 7.37594143e-02 1.36243343e-01 1.69424623e-01
1.10772395e+00 -2.02071577e-01 2.58072138e-01 -1.07839786e-01
1.11805722e-01 3.06240350e-01 5.36136270e-01 7.06877768e-01
-4.56178606e-01 -8.27528164e-02 3.97359192e-01 -9.88675475e-01
-1.38006234e+00 -5.96992671e-01 1.56465918e-01 1.22660589e+00
-2.30359733e-01 -2.14423224e-01 -5.22020161e-01 6.04137022e-04
3.02758485e-01 4.76977229e-01 -4.27474707e-01 -2.92985499e-01
-5.47538340e-01 -4.08291250e-01 5.03443897e-01 5.08832753e-01
4.25948709e-01 -1.16626644e+00 -9.13902760e-01 3.55016559e-01
-4.66686822e-02 -8.52895319e-01 2.82367915e-01 5.46583533e-02
-8.46025705e-01 -1.01871276e+00 -3.55476201e-01 -4.40608710e-01
4.53820586e-01 3.17702979e-01 1.32788110e+00 3.48502070e-01
-4.03657258e-01 2.79163212e-01 -3.84941846e-01 -6.72893465e-01
-5.79652250e-01 1.32378131e-01 2.17489809e-01 -3.65437925e-01
1.15244341e+00 -7.71404088e-01 -5.09910583e-01 3.20052236e-01
-8.47551227e-01 3.24242920e-01 7.77311444e-01 6.40825093e-01
-4.99371625e-02 1.11412108e-01 7.81326592e-01 -9.29388940e-01
8.51667047e-01 -7.33374894e-01 -4.24691826e-01 2.08722532e-01
-1.04986417e+00 6.39475137e-02 4.90488738e-01 -3.09515119e-01
-7.21698999e-01 -6.39111280e-01 1.08835936e-01 -3.44151646e-01
-1.40696377e-01 4.98698175e-01 -8.74384940e-02 3.17237526e-02
8.12207818e-01 -3.42818111e-01 2.41378099e-01 -4.00249720e-01
4.60827589e-01 9.48896229e-01 4.58094031e-01 -5.81300795e-01
8.87078106e-01 5.31130791e-01 -2.49590933e-01 -1.28858015e-01
-6.01878166e-01 -4.59184378e-01 -8.45014215e-01 -2.87511915e-01
7.16741145e-01 -6.48549259e-01 -9.81823444e-01 4.83126432e-01
-1.11063218e+00 -4.34092134e-02 -5.09429574e-01 2.70884156e-01
-3.73394161e-01 -2.62632016e-02 -5.63490570e-01 -6.56471312e-01
-4.17429596e-01 -8.79978240e-01 6.49562418e-01 4.46268886e-01
-7.21763313e-01 -9.79143322e-01 1.20670348e-01 1.15250075e+00
8.89892697e-01 4.47443761e-02 1.34937263e+00 -1.09989440e+00
-6.92908585e-01 -6.62259579e-01 -3.32363099e-01 -6.63084686e-02
-9.27981362e-02 -3.84407401e-01 -9.07341182e-01 -7.68162757e-02
3.54023457e-01 -3.58915836e-01 4.35067385e-01 -3.15773308e-01
6.40804648e-01 -3.47374588e-01 -2.45474279e-01 3.28717113e-01
1.08483171e+00 3.08511585e-01 5.68213642e-01 6.50393486e-01
4.55423743e-01 1.04440248e+00 3.71551454e-01 4.84799892e-01
6.87425971e-01 4.50646549e-01 5.12927622e-02 -2.37089336e-01
2.08306089e-01 -3.53058040e-01 -2.61000395e-01 1.03483331e+00
-1.16684265e-01 2.56218642e-01 -1.31641746e+00 3.10617387e-01
-2.15541935e+00 -1.35767591e+00 -6.79145604e-02 2.07961512e+00
6.22227550e-01 2.21312493e-01 8.60354602e-02 2.92517930e-01
5.11801481e-01 -3.90597247e-02 -4.74851340e-01 -5.58159113e-01
7.63112381e-02 8.01505819e-02 4.56277244e-02 9.93769988e-02
-4.42535400e-01 7.97418058e-01 6.21155691e+00 3.54829073e-01
-9.63383138e-01 1.65374190e-01 5.62574565e-01 1.45674914e-01
-6.81213558e-01 2.74763316e-01 -5.47851205e-01 4.05390292e-01
1.00239313e+00 -3.43701720e-01 6.00668311e-01 7.37238288e-01
1.76159635e-01 -1.02450982e-01 -1.34897816e+00 1.01375103e+00
1.05523750e-01 -1.45500386e+00 -1.48530751e-01 3.52377146e-01
4.46763992e-01 -1.60982117e-01 -1.58117339e-01 5.65532684e-01
1.62731826e-01 -9.83800650e-01 5.86855471e-01 7.61482835e-01
7.75461122e-02 -3.19945604e-01 8.15647542e-01 6.54231191e-01
-7.19631732e-01 -4.09923464e-01 -4.16498959e-01 -8.38737726e-01
-1.64612215e-02 7.18655646e-01 -6.21025145e-01 2.84116000e-01
4.04307306e-01 1.24305636e-01 -5.83606243e-01 8.95167291e-01
1.59565434e-01 -6.92274496e-02 9.25019830e-02 -3.53754684e-02
2.71880813e-03 -2.80291647e-01 1.54357478e-01 6.99642301e-01
3.87464091e-02 1.64337441e-01 2.56847054e-01 1.14120865e+00
-9.24876891e-03 4.19327021e-02 -5.96602917e-01 -4.02240276e-01
7.10290968e-01 1.17734218e+00 -6.41340613e-01 -3.58170897e-01
-5.55265486e-01 5.06306231e-01 4.35233325e-01 -7.59448763e-03
-4.33813572e-01 -1.50518000e-01 5.54502547e-01 4.75911856e-01
-4.22334492e-01 -2.04688355e-01 -6.58872128e-01 -9.60337579e-01
1.13035038e-01 -1.52417457e+00 4.01504010e-01 -7.81179070e-01
-1.63143885e+00 3.23056400e-01 -1.00735039e-01 -8.29560161e-01
-4.47927892e-01 -4.10289615e-01 -5.09856403e-01 8.92248094e-01
-1.19845188e+00 -1.19459176e+00 -6.23706341e-01 5.20029724e-01
4.42823887e-01 -3.56332868e-01 6.56992197e-01 5.51754415e-01
-1.67151570e-01 3.96347225e-01 -7.73445815e-02 8.02510902e-02
7.16178656e-01 -6.33984804e-01 5.56786418e-01 4.41147357e-01
3.11006028e-02 8.25482011e-01 5.06667256e-01 -5.56092918e-01
-1.70454812e+00 -2.84523070e-01 1.40022790e+00 -6.39736354e-01
8.66205990e-01 -3.76155198e-01 -9.15634573e-01 4.54414159e-01
5.89755997e-02 -3.97918761e-01 9.53361154e-01 5.06406724e-01
-5.95915616e-01 -1.96869120e-01 -1.10032642e+00 6.03454590e-01
9.30799603e-01 -8.49289954e-01 -9.93468225e-01 2.31990233e-01
6.22726262e-01 -3.63533199e-03 -7.61406243e-01 2.67427385e-01
6.80719912e-01 -1.30336797e+00 8.16361487e-01 -9.00364161e-01
4.67871636e-01 1.85724944e-01 -1.38070747e-01 -9.85505760e-01
-6.13896012e-01 -6.56148195e-01 2.68404007e-01 1.01087821e+00
9.07116905e-02 -9.34543073e-01 9.26327765e-01 1.51268792e+00
-9.29325148e-02 -7.52938986e-01 -6.57509506e-01 -5.87107718e-01
2.06414878e-01 -4.18162704e-01 8.92100036e-01 1.26555729e+00
4.63212162e-01 2.92300850e-01 -1.26185521e-01 -4.22329158e-01
6.13136351e-01 3.23888093e-01 9.84514952e-01 -1.78683007e+00
-2.38240093e-01 -7.81530142e-01 -6.14849091e-01 -2.80501544e-01
-2.64881691e-03 -9.03252184e-01 -6.23571038e-01 -1.60048771e+00
3.23610723e-01 -5.64947367e-01 -8.67431462e-02 5.69864929e-01
-9.33174267e-02 -8.59603360e-02 4.71313268e-01 6.04191959e-01
-3.78497213e-01 6.54630661e-02 1.05149508e+00 -1.20545208e-01
-1.74612284e-01 -1.77008390e-01 -1.12384629e+00 8.21341991e-01
6.27636671e-01 -3.44132334e-01 -1.49890363e-01 -3.18921477e-01
9.34307635e-01 4.12447080e-02 3.11220288e-01 -9.52169836e-01
8.51635695e-01 -1.66828543e-01 4.58215833e-01 -2.45773867e-01
1.24927677e-01 -9.76989269e-01 5.78841746e-01 6.05340004e-01
-2.79214621e-01 4.25160706e-01 -8.63321573e-02 8.77269078e-03
-2.92027295e-01 -1.96479976e-01 6.11646831e-01 -4.21902865e-01
-6.68124437e-01 -4.45732698e-02 -2.49916082e-03 8.89618918e-02
9.93354142e-01 -5.28399050e-01 -7.12857485e-01 -2.17451856e-01
-4.63836074e-01 2.87508279e-01 5.14436901e-01 6.38916910e-01
2.71611959e-01 -1.24994612e+00 -5.62030792e-01 4.48514313e-01
3.90384085e-02 -3.32922041e-01 3.51694345e-01 1.00507879e+00
-3.06124657e-01 5.15340686e-01 -5.43413043e-01 -1.37272537e-01
-9.61518705e-01 1.05216825e+00 1.04436681e-01 -2.85353422e-01
-6.45868957e-01 6.67164147e-01 -5.34945056e-02 -4.49731499e-01
2.42224902e-01 8.34857225e-02 2.11551175e-01 4.77194875e-01
5.87399423e-01 7.01641440e-01 1.20356351e-01 -2.57916749e-01
-2.27893040e-01 2.02308595e-01 -7.40717351e-02 1.08259030e-01
1.31114781e+00 6.24679327e-02 -5.05080879e-01 4.64863330e-01
8.70345056e-01 -1.09967791e-01 -4.65018034e-01 -1.49058908e-01
3.25852871e-01 -9.14797187e-01 -2.04557717e-01 -8.39529634e-01
-6.32429183e-01 9.95154560e-01 5.07266462e-01 4.72509503e-01
8.98790300e-01 1.23440966e-01 9.44153428e-01 3.68161023e-01
7.86794662e-01 -1.25621104e+00 8.42638761e-02 1.90093219e-01
6.05876267e-01 -1.48626781e+00 -5.63985556e-02 1.82227805e-01
-4.39916730e-01 1.03556263e+00 6.71005249e-01 1.84030578e-01
5.21010101e-01 3.11851352e-01 1.92935601e-01 -4.39926207e-01
-9.71302092e-01 -6.28995225e-02 -8.36684406e-02 4.11794633e-01
5.26807487e-01 1.21578760e-01 -4.42878962e-01 5.39625466e-01
-5.06928205e-01 4.18420017e-01 2.49380380e-01 8.81948471e-01
-7.32179999e-01 -1.40710676e+00 -6.68304861e-01 4.89589483e-01
-1.58024162e-01 4.11788635e-02 -5.03061950e-01 7.93998539e-01
5.36139667e-01 8.91313255e-01 2.77394682e-01 -4.95636851e-01
4.88279462e-01 2.99873352e-01 1.92243561e-01 -1.32930875e-01
-9.43184078e-01 -3.03371817e-01 -7.62374774e-02 -5.69933295e-01
-1.48919642e-01 -4.42449749e-01 -1.15452802e+00 -1.05628741e+00
-9.48727801e-02 1.50391802e-01 7.58884192e-01 1.03842366e+00
6.05459690e-01 -4.71522398e-02 2.56118536e-01 -7.16154397e-01
-8.49555671e-01 -6.53661489e-01 -3.33350062e-01 5.98493695e-01
-1.72748014e-01 -4.34904754e-01 -3.13620210e-01 -3.45645517e-01] | [9.050870895385742, 6.4189133644104] |
bebb04ba-3d9b-4523-9948-3d54d3a0d10e | machine-learning-predictions-for-local | 2204.05967 | null | https://arxiv.org/abs/2204.05967v1 | https://arxiv.org/pdf/2204.05967v1.pdf | Machine learning predictions for local electronic properties of disordered correlated electron systems | We present a scalable machine learning (ML) model to predict local electronic properties such as on-site electron number and double occupation for disordered correlated electron systems. Our approach is based on the locality principle, or the nearsightedness nature, of many-electron systems, which means local electronic properties depend mainly on the immediate environment. A ML model is developed to encode this complex dependence of local quantities on the neighborhood. We demonstrate our approach using the square-lattice Anderson-Hubbard model, which is a paradigmatic system for studying the interplay between Mott transition and Anderson localization. We develop a lattice descriptor based on group-theoretical method to represent the on-site random potentials within a finite region. The resultant feature variables are used as input to a multi-layer fully connected neural network, which is trained from datasets of variational Monte Carlo (VMC) simulations on small systems. We show that the ML predictions agree reasonably well with the VMC data. Our work underscores the promising potential of ML methods for multi-scale modeling of correlated electron systems. | ['Gia-Wei Chern', 'Ting-Kuo Lee', 'Puhan Zhang', 'Sheng Zhang', 'Yi-Hsuan Liu'] | 2022-04-12 | null | null | null | null | ['variational-monte-carlo'] | ['miscellaneous'] | [ 7.72939622e-03 -7.49129117e-01 -3.26899320e-01 -4.35792834e-01
-9.77672100e-01 -1.77783966e-02 8.03467274e-01 1.37692079e-01
-5.41523814e-01 1.06899929e+00 1.42996818e-01 -2.33666673e-01
-1.26505181e-01 -8.03302228e-01 -6.29740000e-01 -1.37959456e+00
-3.24328214e-01 7.90718019e-01 8.07685703e-02 -2.82430291e-01
7.38564849e-01 5.46613634e-01 -1.19350421e+00 7.11877763e-01
6.14276469e-01 8.96065116e-01 2.19960049e-01 7.05710232e-01
3.45051378e-01 9.90292251e-01 8.42314810e-02 5.37364423e-01
-1.40301540e-01 -6.75354600e-01 -7.78062761e-01 -5.11972964e-01
3.82078961e-02 1.83672875e-01 -5.20337820e-01 1.02955317e+00
3.77841532e-01 2.95545489e-01 1.32849431e+00 -7.66765118e-01
-9.05579925e-01 5.83055854e-01 1.02624349e-01 2.09867552e-01
5.14941365e-02 2.97255069e-01 1.61466646e+00 -1.03560925e+00
6.52319014e-01 9.46929872e-01 8.11740577e-01 4.96186793e-01
-1.49269378e+00 -1.34783208e-01 -6.54905379e-01 6.09766603e-01
-1.64969814e+00 -4.56866473e-01 9.31298256e-01 -4.13646817e-01
1.58451104e+00 7.57590383e-02 6.75292671e-01 9.53081787e-01
1.11843657e+00 2.50342995e-01 1.30723476e+00 -9.75942314e-01
7.23753810e-01 -2.85586447e-01 3.37050885e-01 6.69289589e-01
1.70672655e-01 2.70593226e-01 -4.52056020e-01 -5.40668547e-01
6.62239432e-01 1.26831949e-01 6.47328645e-02 -7.31129825e-01
-1.11155546e+00 9.97915804e-01 7.47160852e-01 4.42764491e-01
-4.66119915e-01 5.68080604e-01 4.77942914e-01 1.87529251e-01
2.63246119e-01 5.08716822e-01 -2.07073554e-01 -9.65918135e-03
-9.70809340e-01 4.60098267e-01 7.38288641e-01 2.33530536e-01
7.95414865e-01 -2.26037055e-01 2.90159864e-04 5.93205273e-01
5.47644615e-01 7.13521957e-01 2.66219646e-01 -8.55787337e-01
8.60211849e-02 1.92221656e-01 3.03516626e-01 -5.98212361e-01
-4.36169237e-01 9.59810615e-03 -1.23992896e+00 1.49230316e-01
1.01298563e-01 4.32200849e-01 -3.43316942e-01 1.37147760e+00
-2.26125598e-01 -3.74447465e-01 -1.17111623e-01 6.84172630e-01
4.70951289e-01 8.84626985e-01 -1.08013459e-01 -3.75925094e-01
6.72467232e-01 -1.02034009e+00 -6.29093647e-01 1.54265895e-01
9.90079582e-01 -1.61469936e-01 8.66378665e-01 4.89213280e-02
-1.14442682e+00 -3.30581248e-01 -9.84383404e-01 -1.05462067e-01
-5.99715889e-01 -2.43878707e-01 7.20262825e-01 2.18021929e-01
-1.20785296e+00 1.22719109e+00 -1.18121660e+00 -1.62168697e-01
1.28950000e-01 5.84101856e-01 -3.14133257e-01 -1.30869418e-01
-1.13730478e+00 1.01211011e+00 2.10532695e-01 5.35274260e-02
-8.60072196e-01 -9.45992172e-02 -6.46090209e-01 1.08486339e-02
-1.05215468e-01 -3.11785549e-01 9.00986731e-01 -4.87603009e-01
-1.14726925e+00 6.33529544e-01 -5.11192203e-01 -3.56850028e-01
-1.29445538e-01 4.67664570e-01 -3.72299552e-01 -1.61136299e-01
2.06623077e-01 2.49518454e-01 3.48634779e-01 -9.87411916e-01
4.52105939e-01 -2.71446109e-01 -5.90855300e-01 -4.26676832e-02
8.95877108e-02 -1.49487689e-01 2.17962146e-01 9.73852426e-02
1.61214858e-01 -9.19969380e-01 -4.99865204e-01 -4.73958731e-01
-5.22464931e-01 -2.78741807e-01 4.28745061e-01 -7.31538907e-02
1.15295553e+00 -1.71154666e+00 3.55119646e-01 5.67261219e-01
3.51798296e-01 -6.32345974e-02 -3.79111655e-02 1.13330424e+00
-2.56372064e-01 -1.37380022e-03 -2.91390300e-01 -1.36720121e-01
1.11716181e-01 9.74685624e-02 1.12484016e-01 7.25240111e-01
-3.26160854e-03 1.18383169e+00 -6.83133721e-01 -1.75727814e-01
2.12761104e-01 4.10730630e-01 -8.63318324e-01 3.37246172e-02
-3.16791743e-01 5.79606950e-01 -3.36520851e-01 7.11209476e-01
5.37386358e-01 -8.40313375e-01 3.80230784e-01 -2.22977251e-02
-4.07926887e-01 6.98132813e-01 -6.43186331e-01 1.36682081e+00
-1.34242371e-01 3.07935983e-01 5.04167452e-02 -1.14704514e+00
6.52726769e-01 2.60907173e-01 3.59870970e-01 -9.36107457e-01
6.74291328e-02 6.10893965e-01 4.43341821e-01 -2.91371346e-01
2.56332964e-01 -3.10104758e-01 -2.51680613e-01 6.03575230e-01
-5.95802218e-02 -8.87990668e-02 -1.12212621e-01 1.30358130e-01
1.31303990e+00 -3.67245287e-01 7.91089952e-01 -9.89531994e-01
6.38849080e-01 -7.71555630e-03 1.63609851e-02 1.12076628e+00
-3.45660776e-01 2.57354707e-01 5.49190998e-01 -8.30233991e-01
-1.76420569e+00 -9.24111128e-01 -5.82483053e-01 8.55103493e-01
-9.15854722e-02 -5.93909919e-01 -7.43067920e-01 4.84855957e-02
-3.63109149e-02 4.17411208e-01 -6.14954889e-01 -3.61397475e-01
-4.98111188e-01 -1.20713079e+00 1.88908905e-01 3.63183171e-01
3.03723276e-01 -1.60190642e+00 -6.77507818e-02 2.15219870e-01
1.84924051e-01 -5.29942811e-01 -1.97997078e-01 8.81728411e-01
-6.65559173e-01 -7.61162579e-01 -4.05436270e-02 -6.43029690e-01
5.35468459e-01 -1.19144954e-01 9.95858014e-01 1.26232430e-01
-5.17284632e-01 -1.57616790e-02 3.71172689e-02 5.64076960e-01
-6.60233736e-01 1.87502280e-01 5.69963396e-01 -3.99106413e-01
7.89939642e-01 -5.85236311e-01 -4.11085367e-01 -6.40510395e-02
-6.10431075e-01 -1.27236620e-01 4.99909729e-01 1.31118119e+00
7.77524531e-01 3.09686083e-02 3.53045553e-01 -7.18082964e-01
8.27107310e-01 -4.99174029e-01 -5.54053307e-01 1.36244362e-02
-4.79024917e-01 5.44979393e-01 9.88909721e-01 5.63876741e-02
-5.13174057e-01 -1.01320930e-01 -1.43756047e-01 9.91825480e-03
-9.69934762e-02 4.88887966e-01 -1.72105610e-01 -4.89321560e-01
5.18261790e-01 3.63762319e-01 -3.32064897e-01 -3.01892579e-01
9.91960242e-02 8.95352244e-01 1.50499716e-01 -8.63715351e-01
2.96562642e-01 4.00109500e-01 5.95952988e-01 -9.30323601e-01
-4.07710850e-01 -3.72128934e-01 -1.27904391e+00 8.74531269e-02
9.60541844e-01 -6.53068483e-01 -1.07318103e+00 5.65311551e-01
-1.24425101e+00 -5.24274111e-01 4.58353199e-02 4.49110687e-01
-7.00700223e-01 1.16727404e-01 -1.00965250e+00 -9.93935108e-01
-1.84176654e-01 -1.28607905e+00 1.04897583e+00 -3.39469045e-01
-1.16299480e-01 -1.47119224e+00 5.34386337e-01 2.22136334e-01
6.27862930e-01 -1.32817607e-02 1.55636132e+00 -7.38611937e-01
-8.88432324e-01 -4.38350648e-01 -1.71354026e-01 2.29136869e-01
-1.28536567e-01 -2.78289378e-01 -6.77003920e-01 -4.79763985e-01
7.36861257e-03 -6.48570120e-01 1.22343946e+00 6.80944979e-01
1.19131613e+00 -3.86536866e-02 -5.20784318e-01 4.25447851e-01
1.64756334e+00 1.34795487e-01 4.87189382e-01 1.56062961e-01
8.94806206e-01 -6.43370375e-02 -1.03331774e-01 5.11237860e-01
1.39194295e-01 7.06000686e-01 2.69807994e-01 3.09440583e-01
3.82857502e-01 -1.59083605e-01 3.97994816e-01 1.37245762e+00
-4.31044370e-01 -1.82568878e-01 -1.21608186e+00 2.49879107e-01
-1.80165923e+00 -1.28764129e+00 -3.74474972e-01 2.21980667e+00
8.02958608e-01 -4.19526510e-02 -2.36771211e-01 -2.53780484e-01
4.98859882e-01 5.71776330e-01 -5.37629187e-01 -5.82276940e-01
-9.36474949e-02 3.19148719e-01 5.14240503e-01 9.11682546e-01
-9.56394076e-01 9.88016903e-01 7.45751619e+00 8.10851038e-01
-1.00300884e+00 2.77016252e-01 5.95581830e-01 -1.09198861e-01
-1.69053063e-01 6.92646503e-02 -8.59979630e-01 5.23468077e-01
1.33629048e+00 3.22133899e-01 9.73030806e-01 6.03333950e-01
3.75407666e-01 -9.10422578e-03 -1.26440513e+00 6.96417987e-01
-1.84731886e-01 -1.86073422e+00 -1.37040168e-01 4.41734493e-01
1.10184574e+00 5.46109736e-01 9.93396342e-02 1.19398154e-01
1.63043454e-01 -1.50282335e+00 1.77080616e-01 7.96593368e-01
9.13227499e-01 -7.51146317e-01 7.46827304e-01 5.12638211e-01
-1.08513033e+00 1.36592716e-01 -7.37250924e-01 -4.72934723e-01
-2.49454123e-03 4.77847338e-01 -5.09626448e-01 -1.95729360e-01
2.98519105e-01 7.79828131e-01 -3.61056387e-01 5.44587135e-01
3.42836678e-01 5.13982832e-01 -8.00787956e-02 -5.71162343e-01
5.30386150e-01 -5.97335160e-01 1.96471542e-01 8.87366116e-01
1.04853868e-01 -5.65219186e-02 7.97668025e-02 1.43457341e+00
-1.69434413e-01 -5.70432059e-02 -9.16706383e-01 -3.90922010e-01
5.00754654e-01 9.79571044e-01 -4.14755464e-01 -4.92611006e-02
-3.74386162e-01 8.42112601e-01 5.95189273e-01 2.77464092e-01
-4.60342884e-01 -2.10982427e-01 2.98834652e-01 4.04834241e-01
5.25821090e-01 -4.73339081e-01 -3.63807529e-02 -1.28762877e+00
-1.13987647e-01 -6.04048848e-01 -3.69178474e-01 -6.18261576e-01
-1.60028803e+00 4.65925485e-01 -2.70177096e-01 -5.32929480e-01
-3.30920428e-01 -1.15681124e+00 -1.03648043e+00 9.95458663e-01
-1.14876306e+00 -9.39825833e-01 3.97955716e-01 6.08565450e-01
-2.27247149e-01 -2.93875158e-01 1.18351865e+00 -4.57867049e-02
-4.43274736e-01 8.05161819e-02 1.38814592e+00 3.60001926e-03
5.31707644e-01 -1.23999047e+00 4.83850986e-01 2.74200559e-01
-1.23496175e-01 9.09627616e-01 5.37428796e-01 -7.70826101e-01
-1.75464499e+00 -9.25069451e-01 1.23751080e+00 -6.89056516e-01
1.02244139e+00 -7.53119051e-01 -8.58399987e-01 5.83746314e-01
5.81938140e-02 7.05573976e-01 7.02048481e-01 1.68311492e-01
-1.84424609e-01 2.24510223e-01 -1.02396405e+00 2.76723117e-01
9.08729434e-01 -1.20532227e+00 -3.03486079e-01 8.13878477e-01
2.37685949e-01 2.72139102e-01 -7.72127509e-01 7.43259564e-02
3.22310686e-01 -1.24830461e+00 9.16510463e-01 -8.52629423e-01
6.08664393e-01 1.30398035e-01 -8.26117814e-01 -1.27127075e+00
-7.23679423e-01 -2.47583196e-01 -1.61514077e-02 4.09659624e-01
4.24007803e-01 -5.21380067e-01 7.00758874e-01 4.02054727e-01
6.84807217e-03 -8.14544678e-01 -1.31045461e+00 -6.00005269e-01
6.28710687e-01 -2.76804209e-01 2.99288541e-01 8.16140234e-01
3.98876548e-01 3.59255344e-01 -5.63724637e-01 1.06951796e-01
6.99508965e-01 7.58742914e-02 -1.81064606e-02 -1.35960162e+00
-4.27101552e-01 -1.90423205e-01 -4.33218807e-01 -7.37825453e-01
5.12209117e-01 -1.27690172e+00 7.95835778e-02 -1.07335269e+00
7.62874126e-01 -2.69030958e-01 -7.29031444e-01 1.37096927e-01
3.86160672e-01 3.07115436e-01 -3.32638323e-01 7.03949690e-01
-1.28159785e+00 7.68154800e-01 1.09093428e+00 5.52504621e-02
2.17819009e-02 -3.18483204e-01 1.73954982e-02 5.42437732e-01
6.33013785e-01 -5.03287852e-01 3.42816085e-01 1.54979855e-01
4.25322652e-01 2.44323120e-01 7.33212173e-01 -1.30178893e+00
3.83529097e-01 -7.91878775e-02 4.85369891e-01 -5.24284482e-01
5.75449765e-01 -4.52905208e-01 -2.87485182e-01 5.23901403e-01
-6.04672015e-01 5.26658334e-02 -3.75745624e-01 6.28515482e-01
-1.70336321e-01 -3.09546351e-01 9.49771821e-01 -5.33059478e-01
-4.26085353e-01 2.90274769e-01 -5.96214890e-01 -1.64154842e-01
5.72442532e-01 3.73289943e-01 -1.78687319e-01 -2.55935311e-01
-5.73080540e-01 -2.96921790e-01 8.34086537e-01 -3.57232898e-01
4.12739635e-01 -1.71325517e+00 -4.26681519e-01 6.39410436e-01
1.46624804e-01 -5.53135514e-01 2.14014292e-01 9.60476041e-01
-7.68935144e-01 1.10145855e+00 -2.24628702e-01 -5.03072381e-01
-8.60382795e-01 4.80423957e-01 5.18196762e-01 -4.97111231e-01
-2.57211447e-01 5.45116901e-01 -3.37353855e-01 -7.25340068e-01
-4.09196228e-01 -2.25002557e-01 3.44848216e-01 -5.57301223e-01
3.73903602e-01 2.30483294e-01 -4.47432091e-03 -8.67272973e-01
-4.26978230e-01 4.20909196e-01 -2.98455119e-01 -1.15047053e-01
1.50981045e+00 1.27405375e-01 -8.81112099e-01 8.76330316e-01
1.46822727e+00 -1.01162814e-01 -9.72428083e-01 -3.42589259e-01
-1.59794893e-02 4.54451703e-02 3.13405395e-01 -4.64201301e-01
-3.09851110e-01 1.16429043e+00 4.43828613e-01 4.60420132e-01
2.85331577e-01 3.01979780e-01 5.42654574e-01 1.09929693e+00
6.12658083e-01 -1.11056399e+00 -1.48322165e-01 9.09109235e-01
3.36450994e-01 -1.58878171e+00 -1.07032903e-01 4.35129493e-01
-4.88703638e-01 1.10738873e+00 2.63007283e-01 -5.02928376e-01
7.24705338e-01 1.44672528e-01 -5.30898035e-01 -5.66755533e-01
-1.00966799e+00 4.67478663e-01 2.58545756e-01 3.58128369e-01
8.06197166e-01 4.86456126e-01 -1.02868430e-01 3.52033675e-01
4.18235548e-02 -3.30431968e-01 3.52361292e-01 6.93896651e-01
-8.13405335e-01 -1.42265856e+00 -2.52392828e-01 5.05418718e-01
-7.70231858e-02 -6.19927168e-01 -4.12991881e-01 7.44579971e-01
1.19218148e-01 5.12708247e-01 1.47044048e-01 -1.73243493e-01
-6.85345829e-01 5.90659320e-01 5.65488398e-01 -5.78735530e-01
-5.58991507e-02 -2.06225201e-01 -3.00793231e-01 -4.55271333e-01
-4.12159145e-01 -6.90608799e-01 -1.30214632e+00 -7.30253816e-01
-3.32228035e-01 5.17893732e-01 3.76694977e-01 1.34347606e+00
1.82456404e-01 2.71321058e-01 5.21852911e-01 -1.24502826e+00
-7.67757535e-01 -9.13751900e-01 -1.24831295e+00 1.98594809e-01
5.59199691e-01 -4.20618266e-01 -4.83779192e-01 -5.94602525e-01] | [5.417576789855957, 5.098823547363281] |
d64be40c-05d3-42a4-84e4-f9c5181ad2b3 | social-norms-grounded-machine-ethics-in | null | null | https://aclanthology.org/2022.coling-1.114 | https://aclanthology.org/2022.coling-1.114.pdf | Social Norms-Grounded Machine Ethics in Complex Narrative Situation | Ethical judgment aims to determine if a person in a narrative situation acts under people’s social norms under a culture, so it is crucial to understand actions in narratives and achieve machine ethics. Recent works depend on data-driven methods to directly judge the ethics of complex real-world narratives but face two major challenges. First, they cannot well handle dilemma situations due to a lack of basic knowledge about social norms. Second, they focus merely on sparse situation-level judgment regardless of the social norms involved during the judgment, leading to a black box. In this work, inspired by previous knowledge-grounded and -augmented paradigms, we propose to complement a complex situation with grounded social norms. Besides a norm-grounding knowledge model, we present a novel norm-supported ethical judgment model in line with neural module networks to alleviate dilemma situations and improve norm-level explainability. Empirically, our model improves state-of-the-art performance on two narrative judgment benchmarks. | ['Daxin Jiang', 'Xiubo Geng', 'Tao Shen'] | null | null | null | null | coling-2022-10 | ['culture'] | ['speech'] | [ 4.98708546e-01 5.08678257e-01 -9.25524831e-02 -5.00684023e-01
-7.63410255e-02 -3.95427883e-01 7.78699815e-01 1.27759829e-01
-5.35332680e-01 6.85961723e-01 9.67990220e-01 3.01148109e-02
-2.75030434e-01 -8.14034760e-01 -1.71263143e-01 -1.92222670e-01
5.39055884e-01 3.74682814e-01 -1.79757625e-01 -7.20896184e-01
5.50103366e-01 -1.36264846e-01 -1.34536409e+00 7.55360126e-01
1.29997170e+00 7.16864526e-01 -4.67507511e-01 3.10877621e-01
4.02622402e-01 1.63705993e+00 -5.48695683e-01 -1.10313630e+00
-1.02952449e-02 -8.86209667e-01 -9.90991056e-01 -1.79902583e-01
-9.81776416e-03 -6.30781054e-01 -2.53955483e-01 1.07569087e+00
4.07429278e-01 1.19475268e-01 1.03192961e+00 -1.33107746e+00
-1.22202027e+00 1.13329363e+00 2.38917153e-02 -1.89208418e-01
7.28719473e-01 4.20751542e-01 1.24152422e+00 -4.23836678e-01
8.47806334e-01 1.08013344e+00 8.27997386e-01 9.29026365e-01
-9.93495464e-01 -1.53300673e-01 2.44488180e-01 8.37326050e-01
-8.06012094e-01 -4.87961531e-01 1.12986720e+00 -6.49090052e-01
9.92271960e-01 3.52956891e-01 1.01139057e+00 1.74321795e+00
1.78999037e-01 3.99343431e-01 1.30725932e+00 -6.37210086e-02
4.75985318e-01 2.83938157e-03 -3.43319625e-02 2.16709211e-01
3.45361531e-01 1.15944386e-01 -8.64978611e-01 2.80309878e-02
2.91754574e-01 -5.60999036e-01 -9.22599733e-02 -1.65928349e-01
-1.61256564e+00 7.85786033e-01 4.05376762e-01 5.06300628e-01
-4.41294611e-01 2.01115385e-01 6.84851587e-01 1.90301061e-01
3.25185061e-01 1.12382400e+00 4.27322090e-03 -6.30921245e-01
-6.63173199e-01 6.17869079e-01 1.09437931e+00 5.03486753e-01
4.58853483e-01 -8.12828541e-02 -4.67982858e-01 6.24348044e-01
4.81520556e-02 2.53596008e-01 2.22023547e-01 -1.45729494e+00
2.67925888e-01 9.82754230e-01 2.98447728e-01 -1.60377204e+00
-6.72223687e-01 -1.66829169e-01 -9.55312908e-01 1.55613758e-02
6.59543276e-01 -1.16516903e-01 2.91377143e-03 1.89390945e+00
-4.32995521e-02 -1.27684465e-02 3.51778179e-01 1.31516898e+00
6.97331250e-01 3.10529530e-01 -3.86602655e-02 -2.03446031e-01
1.38100314e+00 -7.94798553e-01 -9.83225107e-01 -6.08325303e-01
8.98679733e-01 -2.41925851e-01 1.25491512e+00 5.07291079e-01
-1.06666017e+00 -1.13508336e-01 -1.10094464e+00 -4.14239585e-01
-2.66949624e-01 -9.96316671e-02 8.25582027e-01 5.88787913e-01
-8.78901660e-01 6.62627757e-01 -1.40137924e-02 -3.70670378e-01
5.96385181e-01 -2.19446957e-01 -2.81607628e-01 1.37669176e-01
-1.81150496e+00 1.42863107e+00 3.00336242e-01 4.23748046e-01
-7.43378937e-01 -6.12123370e-01 -8.52828264e-01 1.92385346e-01
8.68951023e-01 -8.53184819e-01 1.06094158e+00 -1.07523537e+00
-1.61340773e+00 1.15056145e+00 2.49436885e-01 -3.62168491e-01
1.06255782e+00 -2.83126861e-01 -2.48245716e-01 2.73186341e-02
1.38089061e-02 2.44275972e-01 3.78025860e-01 -1.36807477e+00
-2.68963091e-02 -1.58914756e-02 5.76894581e-01 4.32417154e-01
-5.08318663e-01 1.96299553e-01 4.23516303e-01 -3.85389090e-01
-2.64759153e-01 -7.84963965e-01 -2.73841262e-01 -5.47901914e-02
-4.58316952e-01 -1.46094397e-01 8.20400789e-02 -4.76352841e-01
1.48736823e+00 -1.80664873e+00 4.12947536e-01 1.88003797e-02
5.93453944e-01 2.24380612e-01 3.64692956e-02 6.10304296e-01
1.96163356e-01 4.32151526e-01 -4.16550547e-01 -1.98657885e-01
5.93163669e-01 1.52997255e-01 -2.82207578e-01 2.47198790e-01
3.40663940e-01 1.04903531e+00 -1.19553041e+00 -5.70355177e-01
-2.43874546e-03 1.97183684e-01 -1.07221758e+00 1.81244209e-01
-3.41965944e-01 3.86901349e-01 -1.92840070e-01 1.64282128e-01
4.53847378e-01 -6.94665834e-02 5.95424175e-01 -1.49571419e-01
9.56160724e-02 4.85044003e-01 -9.68213201e-01 1.64078152e+00
-3.55084747e-01 6.16212070e-01 -1.85987443e-01 -7.61038899e-01
7.79600799e-01 2.78298855e-01 1.89923659e-01 -5.45595884e-01
5.05504131e-01 2.14323848e-01 3.92452806e-01 -9.36891019e-01
6.14392400e-01 -4.26662654e-01 -5.84867299e-01 9.76307631e-01
-4.22607869e-01 -3.95682931e-01 2.57988721e-01 3.25851291e-01
1.05509567e+00 2.45025948e-01 5.63620687e-01 -3.41669381e-01
6.69890463e-01 1.70650166e-02 8.84974360e-01 7.06630707e-01
-7.07155943e-01 3.94386560e-01 1.41260338e+00 -7.52190769e-01
-9.09517765e-01 -6.37251079e-01 2.74705917e-01 1.08440280e+00
2.81845689e-01 -3.67182493e-01 -9.57069993e-01 -6.64423466e-01
-2.13926464e-01 1.11683667e+00 -1.02165544e+00 -5.49335122e-01
-6.72996879e-01 -6.97693646e-01 7.46557832e-01 3.81658196e-01
6.80529237e-01 -1.37891984e+00 -9.59823608e-01 1.18490890e-01
-6.42511547e-01 -1.33463395e+00 -2.64494210e-01 -3.80910695e-01
-1.39837846e-01 -1.17835748e+00 9.44722146e-02 -1.99668139e-01
4.47895914e-01 -2.38407761e-01 1.28067803e+00 1.93383604e-01
2.83482641e-01 1.36550302e-02 -4.99900460e-01 -3.05221081e-01
-7.04451621e-01 3.41454409e-02 1.99241087e-01 2.09941402e-01
5.80461442e-01 -7.62975514e-01 -6.27592862e-01 3.74797791e-01
-8.52357388e-01 3.21065485e-01 2.27547675e-01 6.99457347e-01
-2.48211071e-01 -2.42656022e-01 8.89535546e-01 -9.46609437e-01
1.24867702e+00 -5.78441978e-01 8.55641365e-02 1.62590995e-01
-6.94569468e-01 -1.20816939e-01 8.77375484e-01 -2.94173300e-01
-1.28640902e+00 -5.86479664e-01 1.35901332e-01 3.78865063e-01
4.97578317e-03 5.87821424e-01 -2.56862104e-01 4.32567149e-01
1.17651796e+00 -1.78659528e-01 2.05194782e-02 1.50492042e-01
3.90372008e-01 6.44317687e-01 4.77342248e-01 -9.35429633e-01
6.06506646e-01 5.01330733e-01 -2.37971917e-01 -7.73896202e-02
-1.32957149e+00 1.74525321e-01 -6.76377296e-01 -7.02579260e-01
1.08288419e+00 -5.96163571e-01 -1.24232686e+00 5.98566771e-01
-1.28273535e+00 -4.58239466e-01 -2.72967398e-01 4.27368253e-01
-9.60677207e-01 2.96023935e-01 -6.10700786e-01 -9.06314850e-01
-8.12371522e-02 -7.90798306e-01 4.05515969e-01 9.51886177e-02
-8.22149873e-01 -9.19507146e-01 -8.27331096e-03 8.96133959e-01
5.13905048e-01 5.10804474e-01 8.84574533e-01 -6.55600965e-01
-1.97713345e-01 5.07884547e-02 -3.29126626e-01 4.46910560e-01
-2.16101572e-01 -3.54980938e-02 -9.57773805e-01 4.59847718e-01
3.87112170e-01 -7.63785541e-01 5.78923047e-01 -1.33676514e-01
7.68778682e-01 -6.56275034e-01 2.55377859e-01 3.89060527e-01
1.01730919e+00 1.21133178e-02 9.35036361e-01 4.02204514e-01
6.54982626e-01 1.24655485e+00 5.46445608e-01 7.01446474e-01
1.03333271e+00 4.12212223e-01 3.30315232e-01 3.70605141e-01
3.27123612e-01 -2.03749448e-01 4.44724411e-01 5.81130087e-01
-4.75555509e-01 -2.80875415e-01 -1.22844827e+00 5.65833509e-01
-2.31552100e+00 -1.40444899e+00 -1.95994914e-01 1.50505519e+00
1.15355301e+00 1.73861369e-01 -8.24171901e-02 3.52262944e-01
5.23761511e-01 6.51995242e-01 -6.21968269e-01 -9.31370556e-01
-3.76151413e-01 -7.27832139e-01 -2.70743638e-01 5.22204220e-01
-7.70359397e-01 1.03799164e+00 6.04864073e+00 3.32195729e-01
-5.90741694e-01 2.62553897e-02 6.35014653e-01 -2.83793032e-01
-7.06896544e-01 -2.65159141e-02 -1.21502973e-01 2.88927466e-01
6.75241113e-01 -5.60183764e-01 7.05838799e-01 7.21354663e-01
4.43849146e-01 -3.27205747e-01 -1.63289726e+00 7.99033523e-01
3.39573056e-01 -1.32415402e+00 -2.04769298e-01 -2.74507608e-02
7.97010839e-01 -7.69040465e-01 -2.82086488e-02 4.62343603e-01
3.43068838e-01 -1.25854015e+00 1.09867311e+00 6.73433006e-01
5.48008263e-01 -5.02775133e-01 1.04263222e+00 4.94841099e-01
-3.97463948e-01 -1.60060748e-01 -1.64682463e-01 -9.69314814e-01
3.48048627e-01 5.92639267e-01 -4.52828765e-01 2.70748585e-01
3.00537854e-01 9.76571202e-01 -3.09391588e-01 4.37418297e-02
-6.37441754e-01 2.80062586e-01 9.50567722e-02 -4.41097200e-01
2.64557570e-01 -2.27884486e-01 5.41483402e-01 9.63570237e-01
-4.99136150e-02 3.94105732e-01 -3.28812093e-01 1.36320472e+00
-1.52479753e-01 6.14267252e-02 -6.34621918e-01 -2.47930810e-01
3.04607540e-01 1.12082148e+00 -3.49486917e-01 -1.54228404e-01
-1.99519306e-01 8.88834596e-01 4.77467835e-01 1.99262634e-01
-9.97208118e-01 6.88002408e-02 7.96609282e-01 9.87204090e-02
-4.15228844e-01 -1.07924864e-01 -8.45433831e-01 -1.46679533e+00
2.51833200e-01 -1.26362097e+00 2.02677637e-01 -1.06175840e+00
-1.45080388e+00 4.40968186e-01 -1.23235002e-01 -1.01636469e+00
-4.12493765e-01 -5.50655484e-01 -5.99169970e-01 4.70067352e-01
-1.31207180e+00 -1.12747467e+00 -4.27819908e-01 1.75046101e-01
-7.52905682e-02 5.67600206e-02 7.87394464e-01 -1.06048733e-02
-4.92614180e-01 3.75329286e-01 -3.57202291e-01 3.51338446e-01
8.74385834e-01 -1.14087987e+00 2.54559636e-01 7.83794343e-01
-5.57904482e-01 7.23776996e-01 8.52672040e-01 -6.80745602e-01
-1.04076576e+00 -4.56393063e-01 1.26936567e+00 -8.23862314e-01
9.57449973e-01 -5.88776171e-01 -8.86834800e-01 5.09227335e-01
3.60856116e-01 -2.88677365e-01 1.08270884e+00 3.64912212e-01
-7.65450060e-01 -8.29620194e-03 -1.35336673e+00 1.11667609e+00
1.76952112e+00 -6.76823139e-01 -1.03386784e+00 2.00285167e-01
5.42888939e-01 -2.77167946e-01 -8.14993918e-01 1.13765560e-01
6.86732650e-01 -1.53805494e+00 6.59456909e-01 -8.07942450e-01
1.40123653e+00 -1.69734418e-01 -3.86173651e-02 -1.54013741e+00
-3.71765018e-01 -7.55066335e-01 7.72040486e-02 1.17447102e+00
2.73326755e-01 -6.48170292e-01 5.80909371e-01 1.17936635e+00
-9.42213014e-02 -7.56677866e-01 -9.11241829e-01 -4.54683125e-01
2.03244686e-01 -5.83453417e-01 9.70514178e-01 1.27042651e+00
7.02651918e-01 3.55513066e-01 -7.70036697e-01 -3.72423291e-01
4.05076414e-01 -1.44117445e-01 4.18073833e-01 -1.19347131e+00
7.61330426e-02 -7.67960608e-01 -2.57117152e-01 -3.47760409e-01
4.99293953e-01 -8.09224486e-01 9.47392732e-02 -1.93879628e+00
3.50523204e-01 -1.57622248e-01 -3.31959017e-02 2.71276325e-01
-4.08603132e-01 1.22287050e-01 4.96201992e-01 -7.35076815e-02
-8.97285402e-01 7.03492165e-01 1.46376073e+00 1.27316058e-01
2.36313771e-02 -6.83585525e-01 -1.46162283e+00 9.36912239e-01
8.56276512e-01 -3.85641515e-01 -5.00683904e-01 -6.23467088e-01
1.34660065e+00 -1.66846305e-01 5.50784767e-01 -8.32366943e-01
2.14331195e-01 -7.31183648e-01 -9.77358073e-02 1.13584176e-01
1.68049991e-01 -5.98189831e-01 1.62382820e-03 4.72882003e-01
-8.13279092e-01 -2.19577551e-01 -4.11056191e-01 3.74310851e-01
-7.01990798e-02 -2.61146516e-01 6.26092494e-01 -1.52537778e-01
-6.34917378e-01 -4.08995003e-01 -6.97987795e-01 4.94767785e-01
1.13268316e+00 -2.03512758e-01 -1.04670751e+00 -7.36366868e-01
-5.32131672e-01 3.11967224e-01 6.45502150e-01 3.54250312e-01
5.43119311e-01 -1.44928038e+00 -1.18129921e+00 -2.85667688e-01
3.31703544e-01 -2.12335572e-01 6.21388972e-01 7.91933656e-01
-6.33448541e-01 2.06191301e-01 -4.41584051e-01 -3.90484706e-02
-5.99213660e-01 4.49561477e-01 5.91223896e-01 -4.56975549e-01
-1.41039178e-01 7.58271992e-01 -2.70574298e-02 -5.77770948e-01
-2.42568493e-01 -3.13277751e-01 -5.28018832e-01 4.87945408e-01
5.86787641e-01 4.24355745e-01 -4.08547670e-01 -7.35052228e-01
-5.24521708e-01 4.31516379e-01 2.42060274e-01 -2.10361376e-01
1.33918655e+00 -4.38077971e-02 -3.71484220e-01 5.36552608e-01
6.46408498e-01 -3.69068682e-02 -1.25934315e+00 -1.18370451e-01
1.15997225e-01 -5.82656741e-01 -3.16230744e-01 -1.05954742e+00
-5.48621833e-01 8.00399780e-01 -4.49904084e-01 2.03914300e-01
9.40653086e-01 -3.23541075e-01 5.54965317e-01 5.60420156e-01
3.10444891e-01 -1.70619988e+00 4.32720035e-01 7.17659771e-01
1.42900383e+00 -1.36810195e+00 2.05465123e-01 -4.60561097e-01
-1.50839996e+00 9.05384779e-01 9.14927959e-01 -7.67658949e-02
2.35274911e-01 6.72808066e-02 3.43357176e-01 4.56926133e-03
-1.09653938e+00 -4.03353572e-03 2.11306438e-01 7.78300941e-01
3.48129749e-01 5.79953976e-02 -6.95297182e-01 1.31514490e+00
-5.05786002e-01 1.47272483e-01 1.07654309e+00 5.43633997e-01
-6.02045916e-02 -6.67008936e-01 -1.43555030e-01 4.03011173e-01
-1.84469759e-01 -1.30998015e-01 -9.48155224e-01 4.27754223e-01
3.39170516e-01 1.19047809e+00 -2.26717681e-01 -6.16422176e-01
4.82945532e-01 6.59462065e-02 2.52631664e-01 -4.48192924e-01
-1.18867040e+00 -7.50723779e-01 8.17383945e-01 -9.21444297e-01
-6.16807282e-01 -8.21809828e-01 -1.11688268e+00 -8.80014241e-01
2.73550123e-01 -1.14959575e-01 2.07584307e-01 1.08524501e+00
3.62216532e-01 3.67825001e-01 2.24878028e-01 -4.22460437e-01
-6.46728337e-01 -8.81256282e-01 -2.48860300e-01 9.22842920e-01
-4.84801866e-02 -4.79069769e-01 -5.31511128e-01 3.36891972e-02] | [11.634629249572754, 8.17578411102295] |
caab5d72-0ecc-4ef6-a044-9714d83bff45 | using-ontologies-to-improve-performance-in | null | null | https://openreview.net/forum?id=r1g1LoAcFm | https://openreview.net/pdf?id=r1g1LoAcFm | Using Ontologies To Improve Performance In Massively Multi-label Prediction | Massively multi-label prediction/classification problems arise in environments like health-care or biology where it is useful to make very precise predictions. One challenge with massively multi-label problems is that there is often a long-tailed frequency distribution for the labels, resulting in few positive examples for the rare labels. We propose a solution to this problem by modifying the output layer of a neural network to create a Bayesian network of sigmoids which takes advantage of ontology relationships between the labels to help share information between the rare and the more common labels. We apply this method to the two massively multi-label tasks of disease prediction (ICD-9 codes) and protein function prediction (Gene Ontology terms) and obtain significant improvements in per-label AUROC and average precision. | ['Peter J. Liu', 'Ethan Steinberg'] | 2019-05-01 | null | null | null | iclr-2019-5 | ['protein-function-prediction'] | ['medical'] | [ 4.46421802e-01 1.97224200e-01 -4.01219964e-01 -8.32439721e-01
-8.64505649e-01 -4.58464980e-01 9.33018848e-02 6.23979688e-01
-5.06154954e-01 1.05649042e+00 1.02402316e-02 -4.98466730e-01
-3.88877124e-01 -5.96672654e-01 -6.25812590e-01 -8.30229819e-01
-3.72927822e-02 9.10432756e-01 1.95429295e-01 2.77167946e-01
-1.11101650e-01 6.15170300e-02 -1.46321595e+00 6.60548329e-01
4.85697418e-01 1.03208649e+00 -2.68636178e-02 6.19204700e-01
-4.11336012e-02 7.45611608e-01 -4.17068392e-01 -4.30454791e-01
3.96231748e-02 -1.54991761e-01 -8.47666502e-01 -3.50474715e-01
2.67821431e-01 1.09025992e-01 3.77045155e-01 1.10993385e+00
4.93412405e-01 1.45313904e-01 1.10738182e+00 -1.19630182e+00
-3.46778512e-01 5.45381784e-01 -7.53418744e-01 -3.31134796e-02
1.01833075e-01 -3.74751657e-01 1.22862875e+00 -5.65189123e-01
4.90782171e-01 1.45255983e+00 1.06070125e+00 3.82315040e-01
-1.32401037e+00 -7.17167675e-01 2.12872066e-02 -1.20029589e-02
-1.48719871e+00 -1.37171820e-01 1.67570770e-01 -5.63182056e-01
1.02566826e+00 3.46866697e-01 1.23932630e-01 9.35068786e-01
5.94636261e-01 3.06948483e-01 1.06844521e+00 -4.40725088e-01
2.12578401e-01 4.09288585e-01 2.95109749e-01 6.49462342e-01
1.78883091e-01 -3.14891264e-02 -2.26952627e-01 -8.35175455e-01
1.01344354e-01 4.36426699e-01 2.63370395e-01 -1.40846157e-02
-8.93365741e-01 1.02199531e+00 2.35463515e-01 2.19476327e-01
-2.97479331e-01 2.64675081e-01 3.64586353e-01 3.72149944e-01
7.62559533e-01 4.66365427e-01 -8.03586602e-01 2.42984474e-01
-7.41524518e-01 2.73898810e-01 9.70214367e-01 6.46976888e-01
8.45306695e-01 -7.33429790e-01 -8.94843563e-02 1.24152994e+00
4.15599525e-01 -6.71383142e-02 6.27961695e-01 -7.42501199e-01
6.34036660e-02 4.16588426e-01 1.16027333e-01 -8.22240889e-01
-9.07914996e-01 -3.43164355e-01 -6.84591651e-01 -6.07020073e-02
5.91581881e-01 -3.42122495e-01 -8.14767540e-01 1.87162173e+00
5.65641761e-01 2.19138518e-01 -2.15338215e-01 4.50283378e-01
6.35832429e-01 3.29954028e-01 7.83802450e-01 -9.90952328e-02
1.75040901e+00 -6.97618902e-01 -5.69671094e-01 -2.90607452e-01
1.16620612e+00 -5.57049215e-01 6.28430188e-01 3.21306773e-02
-5.71487427e-01 -7.06955716e-02 -6.05289221e-01 -1.62502918e-02
-6.71292067e-01 -3.49139452e-01 8.15088153e-01 4.87747967e-01
-7.54918933e-01 7.78640628e-01 -4.11441028e-01 -3.16816807e-01
5.84596455e-01 5.98939955e-01 -3.01800162e-01 -2.28124514e-01
-1.41318464e+00 1.03812993e+00 7.43255794e-01 -3.94214958e-01
-3.74934107e-01 -9.89979208e-01 -5.86286485e-01 2.04691008e-01
9.36640501e-02 -6.80815160e-01 1.24502099e+00 -9.38771784e-01
-8.78281057e-01 8.76121104e-01 3.45046185e-02 -2.12362334e-01
2.11222842e-01 2.73044527e-01 -3.16095293e-01 -1.85767934e-01
8.46916139e-02 7.95317352e-01 4.25684363e-01 -7.67790675e-01
-1.08617675e+00 -2.77185768e-01 -2.13861614e-01 1.76538348e-01
-3.97660524e-01 1.93684131e-01 2.84460902e-01 -5.07972002e-01
-1.06797293e-01 -1.08294976e+00 -3.18825454e-01 -1.60565171e-02
-2.67099738e-01 -5.28242052e-01 3.97702962e-01 -4.82823789e-01
9.14277017e-01 -2.10732961e+00 -1.80351481e-01 2.36530751e-01
3.48504990e-01 -2.37775117e-01 -1.09087430e-01 2.29900509e-01
-3.16109091e-01 -1.62164252e-02 8.13725889e-02 -9.52778831e-02
-1.14698291e-01 3.71257722e-01 -1.23774037e-02 4.72336203e-01
2.58777916e-01 7.40590692e-01 -1.06530702e+00 -5.43762684e-01
-3.51158500e-01 3.65911812e-01 -4.09742415e-01 1.53621007e-02
-3.84390831e-01 1.87361881e-01 -3.23695987e-01 5.82807660e-01
2.42071301e-01 -1.00969803e+00 4.08278912e-01 2.05294285e-02
3.85258734e-01 2.36540139e-01 -8.53917122e-01 1.20399725e+00
-3.88732493e-01 1.02143370e-01 -2.50792831e-01 -8.25331688e-01
6.86772346e-01 5.01897633e-01 8.45083892e-01 -2.29859516e-01
1.99202880e-01 1.95753708e-01 5.08356933e-03 -3.69480222e-01
7.62542337e-02 -9.13450360e-01 -1.32544130e-01 4.29017186e-01
1.04442038e-01 7.38169700e-02 -1.48458183e-01 -7.29421750e-02
1.30014884e+00 -2.30544880e-01 6.06925786e-01 -2.00978994e-01
-3.31429280e-02 -8.73050839e-02 7.85959005e-01 8.07998955e-01
-1.18602276e-01 2.99752593e-01 6.59096181e-01 -5.84912598e-01
-8.89907956e-01 -7.01284945e-01 -4.49280947e-01 1.72930396e+00
-2.28668556e-01 -2.29498297e-01 -2.35651463e-01 -8.11236978e-01
4.71145570e-01 6.22820914e-01 -6.05166197e-01 -3.10870737e-01
1.60213315e-03 -1.38086271e+00 5.95220149e-01 2.68320590e-01
-2.06279859e-01 -8.50422084e-01 -2.01265141e-01 2.75333107e-01
-1.18797623e-01 -8.55445087e-01 -1.81986839e-01 1.15023720e+00
-7.18694150e-01 -8.73335838e-01 -5.73210180e-01 -8.74906659e-01
4.61873919e-01 -2.36379862e-01 1.20879006e+00 8.35848320e-03
-3.07302028e-01 -2.12445423e-01 -7.08220154e-02 -6.90392971e-01
-5.31106174e-01 4.70010079e-02 6.45527318e-02 -4.37077507e-02
6.69814110e-01 -5.39689064e-01 -3.87907296e-01 4.04303551e-01
-7.50666082e-01 -1.22288212e-01 4.20491248e-01 1.06320345e+00
6.72896147e-01 1.31218791e-01 1.07947743e+00 -1.43170083e+00
5.82323015e-01 -9.62398231e-01 -3.63902926e-01 4.96644646e-01
-7.60148704e-01 2.37953030e-02 4.48824763e-01 -7.92578995e-01
-6.28425896e-01 3.71515393e-01 -3.11152756e-01 1.93363298e-02
-3.24052751e-01 5.30779839e-01 2.88991332e-01 -2.01637462e-01
8.92866850e-01 -3.52072716e-01 -2.37463653e-01 -5.25396466e-01
1.40956298e-01 1.04415894e+00 6.79046810e-02 -3.46882194e-01
3.78013253e-02 1.73504695e-01 3.00316364e-01 -5.13866127e-01
-1.49400759e+00 -7.08263338e-01 -2.72313446e-01 1.36126680e-02
7.10304558e-01 -1.05006087e+00 -1.05544555e+00 1.69905767e-01
-9.28293109e-01 -2.40466043e-01 -1.50196522e-01 5.67016840e-01
-5.26342750e-01 1.62904531e-01 -8.65984619e-01 -5.66221118e-01
-3.49806726e-01 -9.85838056e-01 1.10719192e+00 2.00844154e-01
-5.55583298e-01 -1.11763942e+00 1.50156230e-01 1.97126061e-01
3.28108341e-01 1.63895831e-01 1.56020045e+00 -1.20224738e+00
4.58532162e-02 -4.43082064e-01 -3.67162943e-01 5.92381395e-02
1.05381245e-03 -5.48004806e-01 -1.11716938e+00 -1.84870601e-01
-1.91444278e-01 -7.22303987e-01 8.64406705e-01 5.86854696e-01
1.26288474e+00 -1.44232452e-01 -7.82829404e-01 3.98170918e-01
1.27091217e+00 6.29641116e-02 1.14736475e-01 -3.08049954e-02
5.93076468e-01 8.00049007e-01 5.64596295e-01 3.97558898e-01
5.83631098e-01 7.27158248e-01 2.03053102e-01 -4.44993339e-02
3.28157574e-01 -6.46547005e-02 -7.29574189e-02 3.51838052e-01
4.44817066e-01 -4.57093656e-01 -9.65670764e-01 3.58431399e-01
-1.85328555e+00 -6.44507945e-01 -5.85522577e-02 2.03215075e+00
1.25152457e+00 2.15963498e-02 2.37971723e-01 -2.19303578e-01
7.53930449e-01 -3.01715851e-01 -6.28522992e-01 -4.49536383e-01
7.79793337e-02 -8.58118236e-02 7.09200621e-01 3.60414475e-01
-1.35823083e+00 6.32675350e-01 7.74067593e+00 1.09929299e+00
-8.94311070e-01 2.71651179e-01 1.07612944e+00 -2.09783480e-01
-8.16387758e-02 -2.65811861e-01 -1.01372933e+00 6.38710141e-01
1.50608802e+00 2.26307899e-01 2.56774843e-01 7.97565162e-01
-1.48070917e-01 -2.82063425e-01 -1.35046411e+00 7.66573370e-01
-3.17580998e-01 -1.09846580e+00 -2.02313483e-01 6.58004135e-02
7.47473121e-01 3.65179747e-01 -4.22954448e-02 2.15024561e-01
1.05754781e+00 -1.05799198e+00 1.26867667e-01 4.76208210e-01
8.84933650e-01 -6.55540407e-01 6.81359947e-01 5.50845385e-01
-6.31134331e-01 -3.75106484e-01 -5.49841166e-01 -5.78633696e-02
2.91230064e-02 1.11817861e+00 -1.23442018e+00 7.04345480e-02
5.27578771e-01 5.10309219e-01 -2.53026992e-01 1.09516835e+00
1.86989710e-01 4.49009866e-01 -4.11775261e-01 -2.69906044e-01
7.72404671e-02 2.10038155e-01 -2.86548082e-02 1.19879282e+00
2.69697726e-01 -9.85684320e-02 4.13873225e-01 5.36906779e-01
-3.35143119e-01 2.97721922e-01 -4.70431596e-01 -3.54003459e-02
2.84244329e-01 1.35404432e+00 -9.46877599e-01 -3.80875468e-01
-6.35212064e-01 5.82037091e-01 5.06624937e-01 2.76540369e-02
-5.96146107e-01 -2.83813387e-01 5.93394339e-01 -2.63411641e-01
2.48038899e-02 4.16061401e-01 -3.17147970e-01 -6.39164925e-01
-4.34102625e-01 -7.75136828e-01 9.59155917e-01 -5.52599549e-01
-1.73704708e+00 2.85132498e-01 -1.94846511e-01 -7.79170811e-01
-4.31731284e-01 -6.96286380e-01 -7.42204674e-03 9.92121756e-01
-1.47345483e+00 -1.07492161e+00 2.83252090e-01 1.51968569e-01
2.76167821e-02 5.30725867e-02 1.24905932e+00 6.51739180e-01
-2.07803771e-01 6.08251333e-01 5.83677649e-01 -3.51856083e-01
1.15242255e+00 -1.48799193e+00 -1.70656648e-02 -1.92304581e-01
-6.42776638e-02 2.97162861e-01 6.15939975e-01 -8.62895846e-01
-5.98340690e-01 -1.22390139e+00 1.35465407e+00 -5.31541169e-01
6.61583602e-01 -2.56706297e-01 -8.32344651e-01 6.28509521e-01
-5.02764821e-01 1.48166269e-01 1.42101538e+00 5.34062386e-01
-7.13038981e-01 1.80172011e-01 -1.46628022e+00 2.79400826e-01
6.12201631e-01 -4.45918739e-01 -2.06310168e-01 9.82856750e-01
7.82972455e-01 -8.59072357e-02 -1.21340966e+00 2.05876276e-01
8.91213953e-01 -4.37996536e-01 1.00104892e+00 -9.87515450e-01
4.66723114e-01 3.29355672e-02 -3.72880250e-01 -1.41009867e+00
-5.31123459e-01 -7.74397328e-02 1.48886302e-02 6.77115500e-01
6.94579959e-01 -7.23843813e-01 6.38325512e-01 7.99144447e-01
2.54252493e-01 -9.11415279e-01 -9.08932388e-01 -5.13641477e-01
1.60947591e-01 -5.46500944e-02 5.46178162e-01 1.16206849e+00
1.43077046e-01 4.98939365e-01 -5.56407094e-01 6.60857558e-02
4.13307488e-01 1.62162986e-02 -2.85537038e-02 -1.87834632e+00
-5.96330583e-01 -1.47444725e-01 -3.60137045e-01 -5.32615840e-01
2.09052250e-01 -1.24142754e+00 2.42989570e-01 -1.15574956e+00
6.59297526e-01 -8.43204200e-01 -7.50687838e-01 9.34929311e-01
-3.03801894e-01 5.40951729e-01 -3.39898080e-01 -5.29775880e-02
-9.04778183e-01 -2.23026611e-02 8.25885296e-01 -2.37931516e-02
1.58555042e-02 1.94144621e-01 -9.13216293e-01 7.49153793e-01
4.86087948e-01 -1.13059199e+00 -9.96949151e-02 -2.24155318e-02
7.54687846e-01 2.57088304e-01 1.27205908e-01 -6.19452298e-01
1.15864851e-01 -2.20575944e-01 4.89239126e-01 -4.10548866e-01
2.31215701e-01 -9.94514287e-01 4.61341560e-01 5.36576569e-01
-8.05660129e-01 -1.11885108e-01 7.09664961e-03 8.40156376e-01
1.57316178e-01 -4.73187476e-01 1.05309117e+00 -2.23082855e-01
-2.32204974e-01 7.59771317e-02 -4.19205785e-01 3.89565937e-02
1.01355600e+00 3.13294739e-01 -3.63007724e-01 -1.79491758e-01
-9.95848060e-01 3.63114536e-01 1.52824312e-01 4.23944026e-01
1.19533062e-01 -1.19939852e+00 -6.14306211e-01 2.43730061e-02
3.54768217e-01 -2.20380619e-01 1.69067249e-01 7.88084269e-01
-1.48029029e-01 3.64145756e-01 -1.56120420e-01 -4.73941594e-01
-1.47438335e+00 5.11710048e-01 3.64897251e-01 -5.87130606e-01
-2.50478327e-01 1.10980582e+00 3.01724225e-01 -7.76349008e-01
3.21508735e-01 -1.25653073e-01 -8.86241570e-02 2.96529680e-01
6.23853981e-01 3.48279059e-01 1.22955441e-01 -3.28464657e-01
-4.69129711e-01 2.17841789e-01 -4.19582963e-01 2.10404173e-01
1.54459012e+00 6.31659850e-02 -3.51798624e-01 6.99854314e-01
1.26887965e+00 -5.04851639e-01 -7.51455784e-01 -2.68994898e-01
4.14830536e-01 -5.54939583e-02 -3.03567089e-02 -1.28866827e+00
-5.38290739e-01 6.09995425e-01 7.04119921e-01 2.02972174e-01
6.80358768e-01 2.79733568e-01 4.67904896e-01 5.56840837e-01
1.41729623e-01 -9.87123251e-01 -2.21491650e-01 4.67497349e-01
2.14904100e-01 -1.39429116e+00 -4.63433713e-02 -2.76167214e-01
-3.58662963e-01 9.45673645e-01 2.46194869e-01 3.73804152e-01
9.42254782e-01 4.13614362e-01 4.63737845e-02 -4.00593907e-01
-1.01710868e+00 2.08771244e-01 2.67993182e-01 3.62488419e-01
6.54851556e-01 3.85085404e-01 -3.70256215e-01 4.35441434e-01
2.39554778e-01 2.77299415e-02 7.69986510e-02 6.06523097e-01
-5.77001631e-01 -1.15890813e+00 -2.44941548e-01 1.21850085e+00
-1.00330162e+00 -2.50742137e-01 -1.99123278e-01 3.15097123e-02
2.66827375e-01 8.50958586e-01 -3.68247554e-02 -3.87693465e-01
-1.39739275e-01 7.01887250e-01 1.61528349e-01 -8.51038516e-01
-7.75517583e-01 5.29892035e-02 2.96573579e-01 -2.53552467e-01
-1.25259221e-01 -4.62260723e-01 -1.23386073e+00 -1.09018475e-01
-5.08059680e-01 -3.99299823e-02 7.71570146e-01 9.54788387e-01
3.97653967e-01 6.33966267e-01 4.31222320e-01 -3.05091262e-01
-7.91577876e-01 -9.87087131e-01 -1.00466084e+00 5.57602584e-01
3.39986414e-01 -7.55307794e-01 -2.20826149e-01 -1.49872273e-01] | [9.452596664428711, 4.387707710266113] |
3805a216-c29f-47dc-91d1-10fd1b0128d1 | reler-zju-submission-to-the-ego4d-moment | 2211.09558 | null | https://arxiv.org/abs/2211.09558v1 | https://arxiv.org/pdf/2211.09558v1.pdf | ReLER@ZJU Submission to the Ego4D Moment Queries Challenge 2022 | In this report, we present the ReLER@ZJU1 submission to the Ego4D Moment Queries Challenge in ECCV 2022. In this task, the goal is to retrieve and localize all instances of possible activities in egocentric videos. Ego4D dataset is challenging for the temporal action localization task as the temporal duration of the videos is quite long and each video contains multiple action instances with fine-grained action classes. To address these problems, we utilize a multi-scale transformer to classify different action categories and predict the boundary of each instance. Moreover, in order to better capture the long-term temporal dependencies in the long videos, we propose a segment-level recurrence mechanism. Compared with directly feeding all video features to the transformer encoder, the proposed segment-level recurrence mechanism alleviates the optimization difficulties and achieves better performance. The final submission achieved Recall@1,tIoU=0.5 score of 37.24, average mAP score of 17.67 and took 3-rd place on the leaderboard. | ['Yi Yang', 'Xiaohan Wang', 'Jiayi Shao'] | 2022-11-17 | null | null | null | null | ['action-localization'] | ['computer-vision'] | [-1.30297497e-01 -1.40325457e-01 -5.26167214e-01 -3.18701893e-01
-7.53178537e-01 -5.03740847e-01 5.02914131e-01 -3.38498980e-01
-2.93547690e-01 5.18665254e-01 7.17932522e-01 2.98109204e-01
2.34835520e-02 -3.27980816e-01 -8.16187918e-01 -4.78104591e-01
-3.36646855e-01 1.39821053e-01 4.07680273e-01 2.80031443e-01
1.03419572e-01 2.64183581e-01 -1.47330070e+00 5.87110877e-01
3.34117919e-01 1.29094911e+00 2.88757324e-01 6.24042749e-01
2.54721642e-01 1.27362037e+00 -3.27472895e-01 -1.90749347e-01
6.79282323e-02 -3.11774999e-01 -9.79447722e-01 3.64267111e-01
6.07842147e-01 -7.02500761e-01 -1.08333564e+00 8.60862970e-01
4.54746336e-01 4.01206940e-01 5.03573954e-01 -1.38505375e+00
-2.02724174e-01 4.14857388e-01 -6.40158951e-01 6.50204778e-01
5.42666018e-01 1.09710298e-01 1.00641429e+00 -9.21219826e-01
8.72455955e-01 1.19266605e+00 3.41672003e-01 5.10424674e-01
-6.95787430e-01 -5.69656551e-01 3.79281431e-01 9.96301770e-01
-1.44560277e+00 -6.73167229e-01 5.18889010e-01 -4.44785804e-01
1.29127073e+00 -2.11763024e-01 5.74369013e-01 1.39137900e+00
1.69658825e-01 1.40763402e+00 6.33611619e-01 2.92144179e-01
1.14875630e-01 -4.88035202e-01 -1.28667681e-02 6.61625803e-01
-3.08013380e-01 -3.10440481e-01 -9.27366972e-01 1.83601528e-01
6.38270557e-01 1.12775750e-01 -2.33049408e-01 -2.84987450e-01
-1.49212623e+00 4.13539439e-01 4.06636715e-01 2.66143858e-01
-5.45380712e-01 6.14075422e-01 6.87268078e-01 1.81574434e-01
4.09506679e-01 1.28096059e-01 -5.08419871e-01 -8.07821691e-01
-9.00425732e-01 2.18909591e-01 3.39024425e-01 1.34465790e+00
4.59180981e-01 -3.13796371e-01 -6.46412849e-01 6.85500443e-01
-1.50792561e-02 3.00691307e-01 3.94545317e-01 -1.45477355e+00
7.04810560e-01 4.65000123e-01 1.28658682e-01 -8.25986981e-01
-3.57077718e-01 -3.51916045e-01 -6.05541050e-01 -5.69731414e-01
3.15525144e-01 1.75137609e-01 -7.77507424e-01 1.76082170e+00
3.15364033e-01 6.55833960e-01 -5.30998223e-02 9.90054488e-01
6.18945897e-01 6.06489182e-01 6.84307665e-02 -1.36509374e-01
1.41209114e+00 -1.33879328e+00 -7.84304440e-01 -1.20174021e-01
8.81189406e-01 -4.55671370e-01 8.62258375e-01 1.92953289e-01
-1.13618898e+00 -6.06572568e-01 -9.04308736e-01 -2.91137844e-01
-3.96960452e-02 4.12020981e-01 5.84854484e-01 -1.76482156e-01
-9.20337021e-01 5.81053674e-01 -1.13590670e+00 -4.28418577e-01
5.70712686e-01 4.57819141e-02 -5.65019548e-01 -3.75453860e-01
-1.10149014e+00 6.53694034e-01 2.86972791e-01 -7.45190084e-02
-1.18942177e+00 -5.23996770e-01 -8.35289359e-01 7.95554966e-02
5.49791992e-01 -2.72006452e-01 1.35216653e+00 -5.64074814e-01
-1.17079270e+00 7.73766041e-01 -4.19878900e-01 -7.53212512e-01
5.66473424e-01 -5.32341838e-01 -3.92947078e-01 5.11617899e-01
3.83608609e-01 8.41453433e-01 5.52579522e-01 -1.94609717e-01
-9.32313323e-01 -4.67535317e-01 3.57807100e-01 4.01676148e-01
-2.28114605e-01 -1.12396561e-01 -1.14297354e+00 -5.76635957e-01
8.13808814e-02 -1.21444070e+00 9.13667977e-02 -2.08373621e-01
-1.32192254e-01 -5.68645120e-01 9.54635859e-01 -5.17804503e-01
1.14977181e+00 -2.51582670e+00 3.32663655e-01 -4.51673329e-01
1.30402878e-01 -1.03137955e-01 -1.48975536e-01 2.39194736e-01
1.85890332e-01 -2.52382904e-01 2.88195282e-01 -4.13393617e-01
8.79926980e-02 1.65587574e-01 -3.05727154e-01 6.13564849e-01
-6.12057671e-02 1.04709494e+00 -9.47294354e-01 -5.33923149e-01
1.93940312e-01 4.09842104e-01 -6.36409163e-01 2.45924965e-01
-1.42239779e-01 3.56986552e-01 -6.09997928e-01 7.52393782e-01
2.85770357e-01 -3.88547808e-01 5.97599614e-03 -4.76321965e-01
-1.66946673e-03 4.12954479e-01 -9.10414398e-01 2.29723096e+00
-4.03179139e-01 9.51277316e-01 -3.98696423e-01 -1.00897002e+00
3.69741321e-01 4.10631180e-01 9.77055550e-01 -9.19948995e-01
-2.56163138e-03 -8.89094174e-02 -4.26789403e-01 -6.91560149e-01
3.43289912e-01 4.11907554e-01 -3.68989468e-01 2.42081061e-01
4.03893262e-01 3.85967225e-01 5.28286994e-01 2.52133936e-01
1.62886894e+00 5.57778537e-01 1.13596730e-01 2.91113574e-02
4.61034715e-01 -1.54759571e-01 7.43589103e-01 6.23333752e-01
-5.60366988e-01 5.91911972e-01 7.24088430e-01 -4.58319873e-01
-6.91569746e-01 -8.89968514e-01 1.22826137e-01 1.04295516e+00
3.36764306e-01 -7.51166403e-01 -6.06337190e-01 -8.41900647e-01
-2.68521845e-01 6.53643489e-01 -5.91442227e-01 -3.08036685e-01
-6.07820511e-01 -2.95428813e-01 5.93006551e-01 8.12803268e-01
8.33877921e-01 -9.20226455e-01 -7.10167170e-01 3.15412581e-01
-8.67336273e-01 -1.67871833e+00 -8.87739301e-01 -7.56154433e-02
-8.72610927e-01 -1.34258592e+00 -6.84477448e-01 -5.90453207e-01
2.74587095e-01 4.85261530e-01 1.01212847e+00 -5.49635231e-01
-2.53268450e-01 4.97946024e-01 -5.59841633e-01 1.39594644e-01
3.53412807e-01 1.27823263e-01 1.15834072e-01 2.12656036e-02
7.22809255e-01 -5.91566861e-01 -8.65652978e-01 6.42448664e-01
-4.56721574e-01 -2.74766106e-02 4.76479262e-01 6.46305084e-01
8.25084507e-01 1.39511516e-02 3.20923060e-01 -2.17332572e-01
-3.33177708e-02 -5.07541478e-01 -2.66603887e-01 2.57733226e-01
-1.00090183e-01 -1.47590548e-01 2.93512613e-01 -3.81149292e-01
-8.11459005e-01 2.82306463e-01 1.21151432e-01 -8.93071115e-01
-3.06182206e-02 1.15487866e-01 -2.34976411e-01 4.34778512e-01
2.45552391e-01 4.73019093e-01 -4.26267385e-01 -4.91942674e-01
2.12209612e-01 4.86693501e-01 6.22894228e-01 -3.67510527e-01
1.78426102e-01 4.57037508e-01 -1.67976752e-01 -8.34888697e-01
-1.13435721e+00 -6.52786613e-01 -6.49673820e-01 -3.49547386e-01
1.21882737e+00 -1.54558957e+00 -8.57407331e-01 4.65294570e-01
-1.08529544e+00 -5.03906906e-01 -2.51820683e-01 8.75214636e-01
-9.72385108e-01 2.95301139e-01 -6.58658504e-01 -2.48317510e-01
-1.34572729e-01 -1.26914823e+00 1.36647367e+00 6.67697266e-02
-1.62392125e-01 -4.64329362e-01 -7.68554732e-02 5.42357504e-01
5.67949861e-02 7.44971694e-05 3.44008654e-01 -3.20742607e-01
-8.12802315e-01 -2.39187017e-01 -3.13874096e-01 2.11828113e-01
1.47294775e-01 -4.09089983e-01 -6.94379151e-01 -2.68169284e-01
-9.51761380e-02 -5.18583596e-01 9.41698372e-01 4.72263008e-01
1.54861188e+00 -1.10926814e-01 -3.42660248e-01 6.05256081e-01
9.20362711e-01 1.37766168e-01 6.83456421e-01 1.88588902e-01
6.00401521e-01 2.54155248e-01 1.17802250e+00 5.79224288e-01
5.51614642e-01 1.11989641e+00 4.21594381e-01 5.05370438e-01
-3.49564880e-01 -3.47258061e-01 6.20281577e-01 4.86692965e-01
-2.36173809e-01 -4.14990187e-01 -5.78733861e-01 5.91820478e-01
-2.07403946e+00 -1.35578156e+00 1.09278947e-01 1.91224325e+00
3.81515056e-01 1.64960489e-01 2.00302199e-01 -2.94689924e-01
5.85556865e-01 6.16007805e-01 -7.22734690e-01 1.83627382e-01
-2.54596677e-02 -3.84733409e-01 5.18098891e-01 2.12677032e-01
-1.47785902e+00 1.08116698e+00 6.10144806e+00 8.46701205e-01
-9.33499753e-01 1.83899492e-01 3.70968461e-01 -6.18430853e-01
3.74212801e-01 -1.76068991e-01 -9.94338334e-01 6.44705534e-01
9.00619507e-01 -1.73901021e-01 3.13179284e-01 9.16190445e-01
3.95356953e-01 -2.20792279e-01 -1.29318476e+00 1.22127748e+00
1.57944649e-01 -1.20287585e+00 -1.54773921e-01 1.12193376e-01
6.36557043e-01 4.48917985e-01 -1.34313911e-01 5.37397146e-01
-9.59146172e-02 -8.00970078e-01 7.25170732e-01 5.46452165e-01
7.95708835e-01 -7.53316998e-01 5.20738542e-01 2.56179690e-01
-1.49640167e+00 -2.51513690e-01 -4.48515087e-01 5.65084144e-02
3.42420131e-01 3.26178074e-01 -6.38321877e-01 2.57803053e-01
9.95423615e-01 1.41202569e+00 -5.00631750e-01 9.17649269e-01
-1.85976639e-01 4.20775473e-01 -2.47068942e-01 2.46024132e-01
4.72771466e-01 8.08726326e-02 5.88967741e-01 1.11645377e+00
5.26263237e-01 3.00124377e-01 1.14191644e-01 2.38933966e-01
-2.61302501e-01 -2.44321883e-01 -5.73059499e-01 -6.95628300e-02
2.59304523e-01 9.40976202e-01 -4.45808530e-01 -4.36513484e-01
-4.04777706e-01 1.44253767e+00 3.94580036e-01 2.60691136e-01
-1.39487958e+00 -1.49008423e-01 8.74128580e-01 3.09620909e-02
6.26701891e-01 -4.18454587e-01 3.95252317e-01 -1.55244648e+00
3.20134103e-01 -6.46065831e-01 5.13660073e-01 -9.01254416e-01
-6.50951147e-01 4.43756610e-01 -2.06315443e-02 -1.53577602e+00
-4.06868190e-01 -2.10510746e-01 -2.73140639e-01 1.43674210e-01
-1.23242915e+00 -9.54858005e-01 -4.93657380e-01 7.66948462e-01
1.15878117e+00 -4.57043350e-02 5.10922790e-01 6.85235560e-01
-4.72425938e-01 5.71663857e-01 -5.79554103e-02 1.19566746e-01
7.03678429e-01 -1.05312037e+00 3.58455926e-01 6.65612936e-01
3.05801392e-01 1.52401388e-01 5.36308646e-01 -5.16885161e-01
-1.56823862e+00 -1.33594573e+00 1.00582087e+00 -5.24899662e-01
6.71595752e-01 -3.03499937e-01 -5.36959529e-01 9.11269724e-01
-1.78740676e-02 2.96322703e-01 3.97363722e-01 -3.68900783e-02
-2.82020062e-01 -1.76145986e-01 -7.40717471e-01 5.37694693e-01
1.65757263e+00 -6.49483263e-01 -5.24908960e-01 5.45951903e-01
6.42859638e-01 -5.17185986e-01 -9.25126255e-01 3.95152092e-01
5.39100468e-01 -8.23763669e-01 9.79694307e-01 -6.36952102e-01
5.43415606e-01 -3.47300500e-01 -4.45758492e-01 -8.59684289e-01
-3.20015222e-01 -4.41035479e-01 -6.94111764e-01 9.70392466e-01
6.47103786e-02 3.29452120e-02 9.10975039e-01 2.96845317e-01
-2.52627254e-01 -7.81316638e-01 -1.20261610e+00 -8.78180265e-01
-6.56591177e-01 -4.93573874e-01 2.91717082e-01 5.68606317e-01
-8.94443411e-03 2.86956608e-01 -8.19934368e-01 1.93512999e-02
7.17170656e-01 2.06698447e-01 8.07326436e-01 -5.41432023e-01
-2.50325382e-01 -9.14633945e-02 -8.15214872e-01 -1.71836960e+00
2.75246084e-01 -7.26213813e-01 -3.92377488e-02 -1.36434770e+00
2.86787003e-01 2.55774669e-02 -5.01447380e-01 4.70888823e-01
1.04120359e-01 3.49438518e-01 2.31785983e-01 1.58507466e-01
-1.30319524e+00 7.80736208e-01 1.12896967e+00 -3.05870235e-01
-5.89597300e-02 1.21406585e-01 -2.04765394e-01 6.30405009e-01
5.06394148e-01 -3.61395061e-01 -6.46420002e-01 -7.08550513e-01
-1.07726239e-01 3.30996454e-01 4.71173644e-01 -1.24226117e+00
3.42163384e-01 -5.72807975e-02 3.54998320e-01 -1.10446119e+00
7.75679588e-01 -6.85436726e-01 9.26513821e-02 3.30633402e-01
-4.27259535e-01 -3.06578521e-02 -1.02780677e-01 8.36970389e-01
-5.05770266e-01 4.29618955e-01 5.39816618e-01 -4.90895696e-02
-1.16201937e+00 7.16136992e-01 -2.90249318e-01 2.50577867e-01
1.33274662e+00 -9.09456387e-02 -2.91574508e-01 -3.75063509e-01
-9.46007490e-01 6.04183376e-01 4.25368756e-01 8.89945626e-01
6.07719839e-01 -1.47899020e+00 -5.00553608e-01 -2.05480069e-01
2.52521992e-01 -1.69922903e-01 7.88697243e-01 1.10084939e+00
-2.71204859e-01 5.99365354e-01 -2.37136051e-01 -7.89794803e-01
-1.23104417e+00 5.12812436e-01 3.62486333e-01 -1.84437558e-01
-9.41322088e-01 9.11743045e-01 3.77951950e-01 2.71935970e-01
5.67668736e-01 -6.93098232e-02 -3.51586759e-01 2.82243907e-01
7.53690541e-01 5.29813111e-01 -1.41635761e-01 -7.56494999e-01
-6.36619091e-01 5.41253090e-01 -1.47302166e-01 9.34672803e-02
1.37538981e+00 -3.21764678e-01 3.40686679e-01 3.82145137e-01
1.60491145e+00 -5.18636882e-01 -1.67081296e+00 -3.24692100e-01
2.77456772e-02 -5.22714615e-01 9.89830345e-02 -4.11874712e-01
-1.20010030e+00 9.00125027e-01 5.60777426e-01 -3.52285087e-01
1.06898427e+00 3.48235190e-01 1.04509139e+00 5.18529356e-01
4.23996478e-01 -1.35646343e+00 3.42406094e-01 6.65107012e-01
8.63915503e-01 -1.13499439e+00 8.34712479e-03 -1.82075158e-01
-8.82535756e-01 8.59949231e-01 7.95315266e-01 -7.03644380e-02
3.99386615e-01 -1.57798216e-01 -4.09032047e-01 -3.80312800e-01
-1.09949565e+00 -1.81980267e-01 2.81145930e-01 2.27381825e-01
2.10964620e-01 -4.34742011e-02 -2.23495081e-01 3.75567585e-01
1.26039073e-01 2.48648018e-01 2.86901146e-01 8.15972745e-01
-3.16014558e-01 -6.27640963e-01 3.17815721e-01 3.33680749e-01
-5.50893068e-01 2.45857313e-01 -1.99889578e-02 5.10977447e-01
-1.70105547e-02 9.02567863e-01 2.81392992e-01 -5.49367428e-01
5.30283809e-01 5.84730059e-02 5.13927996e-01 -2.46510923e-01
6.01387471e-02 1.19527347e-01 2.15604365e-01 -1.42503142e+00
-5.70937276e-01 -9.66656208e-01 -1.15118396e+00 -2.12053239e-01
1.58566982e-01 9.20258686e-02 5.40478766e-01 8.93754184e-01
8.00659180e-01 5.19514084e-01 7.99784124e-01 -8.48225594e-01
-4.01671857e-01 -9.38845873e-01 -6.09285355e-01 3.30170631e-01
2.68930942e-01 -8.99589479e-01 -2.99741924e-01 1.05095372e-01] | [8.362823486328125, 0.40327778458595276] |
1570b094-97ac-43da-a41b-d78c882dad92 | temporal-superpixels-based-on-proximity | null | null | http://openaccess.thecvf.com/content_iccv_2017/html/Lee_Temporal_Superpixels_Based_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Lee_Temporal_Superpixels_Based_ICCV_2017_paper.pdf | Temporal Superpixels Based on Proximity-Weighted Patch Matching | A temporal superpixel algorithm based on proximity-weighted patch matching (TS-PPM) is proposed in this work. We develop the proximity-weighted patch matching (PPM), which estimates the motion vector of a superpixel robustly, by considering the patch matching distances of neighboring superpixels as well as the target superpixel. In each frame, we initialize superpixels by transferring the superpixel labels of the previous frame using PPM motion vectors. Then, we update the superpixel labels of boundary pixels, based on a cost function, composed of color, spatial, contour, and temporal consistency terms. Finally, we execute superpixel splitting, merging, and relabeling to regularize superpixel sizes and reduce incorrect labels. Experiments show that the proposed algorithm outperforms the state-of-the-art conventional algorithms significantly.
| ['Chang-Su Kim', 'Won-Dong Jang', 'Se-Ho Lee'] | 2017-10-01 | null | null | null | iccv-2017-10 | ['patch-matching'] | ['computer-vision'] | [ 3.66423100e-01 -1.97667748e-01 -3.07531655e-01 -2.60948002e-01
-6.25308037e-01 -3.60663503e-01 3.28040048e-02 5.18832766e-02
-3.50697815e-01 6.19920373e-01 -7.92417899e-02 2.14621946e-01
1.59147426e-01 -6.67417347e-01 -4.30317461e-01 -8.46541166e-01
1.73253179e-01 -2.98693366e-02 1.26919961e+00 2.34149843e-01
5.13053238e-01 2.59205788e-01 -1.17682457e+00 2.79736698e-01
1.00848448e+00 9.20195282e-01 4.36687350e-01 4.09564167e-01
1.67812109e-02 5.48830748e-01 -1.83036774e-01 1.95883308e-02
2.64096558e-01 -5.06892979e-01 -8.50656092e-01 6.13785386e-01
9.04404521e-01 -3.58077079e-01 -1.31402284e-01 1.21995461e+00
-6.88286051e-02 3.00769866e-01 2.07600713e-01 -9.07313704e-01
-5.00878990e-01 4.15383697e-01 -1.04908812e+00 4.27685857e-01
-5.53056300e-02 -1.07958168e-01 7.82110155e-01 -8.79560292e-01
9.11630511e-01 1.17891967e+00 5.81613123e-01 1.14937924e-01
-1.21323085e+00 -6.09190643e-01 4.48079467e-01 2.18549222e-01
-1.57027555e+00 9.15186945e-03 6.72556043e-01 -4.86820340e-01
3.21999878e-01 8.46359357e-02 6.98795140e-01 1.63399652e-01
9.27210879e-03 8.16843152e-01 8.33069742e-01 -4.14164752e-01
3.05171132e-01 -3.18562597e-01 3.57085586e-01 7.97384858e-01
-8.50508586e-02 2.46068723e-02 -3.24160814e-01 -1.14944331e-01
1.08907759e+00 -1.02526180e-01 -3.64648104e-01 -3.44361186e-01
-1.41710901e+00 3.41264546e-01 4.05124426e-01 6.71021581e-01
-3.50862414e-01 3.13520968e-01 -5.83562665e-02 -2.60913044e-01
7.25469649e-01 -2.72496849e-01 -2.70444572e-01 4.25632149e-01
-1.53318334e+00 1.28474319e-02 1.04018904e-01 9.65638459e-01
1.08581161e+00 -1.83517843e-01 -5.56441009e-01 6.26374364e-01
3.77356708e-01 2.72319138e-01 1.87083140e-01 -1.64902687e+00
2.96054125e-01 5.39674580e-01 3.96812677e-01 -1.17492783e+00
-2.62296870e-02 -1.36378720e-01 -8.32639039e-01 2.31858358e-01
3.88025522e-01 1.21796377e-01 -1.02506542e+00 1.25359392e+00
8.04607809e-01 1.01224792e+00 -3.18534017e-01 9.76347685e-01
5.34458280e-01 9.76956189e-01 2.08418697e-01 -6.32866442e-01
1.02302420e+00 -1.29480004e+00 -6.81224644e-01 -1.37057051e-01
2.34540969e-01 -1.03869045e+00 4.45622623e-01 1.59990594e-01
-1.31619668e+00 -1.02562630e+00 -8.94280076e-01 -1.02761023e-01
6.41834959e-02 2.81627029e-01 1.93980455e-01 3.25639993e-01
-1.10035443e+00 8.42661619e-01 -1.02469599e+00 -1.06180549e-01
1.60875440e-01 1.77325547e-01 -1.06767863e-01 -5.16986214e-02
-9.40368652e-01 3.70489031e-01 7.47997284e-01 1.16857164e-01
-5.32773674e-01 -6.70680344e-01 -8.26304615e-01 -9.42130312e-02
9.93481278e-02 -5.06295919e-01 9.08004642e-01 -1.23282194e+00
-1.14784777e+00 8.63404989e-01 -5.78427136e-01 -1.35407865e-01
3.70113283e-01 1.87639892e-01 -3.14576119e-01 3.26032490e-01
2.69714892e-01 9.95120823e-01 1.00269568e+00 -1.12062562e+00
-1.38582253e+00 -2.27067828e-01 -3.80672812e-01 9.51597691e-02
-3.41285281e-02 3.87224630e-02 -1.28310382e+00 -9.53242242e-01
7.45918155e-01 -7.08520532e-01 -3.98833603e-01 2.83846706e-01
-4.19115454e-01 -2.84582675e-01 1.00450933e+00 -8.73795629e-01
1.61116695e+00 -2.45604587e+00 3.79629731e-01 2.72311449e-01
9.57148224e-02 2.54340112e-01 -2.26363063e-01 -3.52486521e-01
4.61876541e-02 -1.98825315e-01 -4.43741709e-01 -3.57824028e-01
-4.83273268e-01 -9.20922961e-03 3.97071838e-02 4.42060560e-01
-1.15762606e-01 5.87416768e-01 -1.09691441e+00 -1.09193945e+00
5.04992247e-01 2.63077348e-01 -8.08144286e-02 7.36344140e-03
-1.37251705e-01 3.91386062e-01 -4.32712346e-01 6.64239109e-01
1.04401684e+00 -2.04863444e-01 6.48420602e-02 -6.03897274e-01
-7.18204618e-01 -3.46263230e-01 -1.68082452e+00 1.63444257e+00
1.86123490e-01 4.92153972e-01 1.35943308e-01 -5.82253575e-01
8.32941592e-01 2.53025424e-02 8.07707071e-01 -3.70060593e-01
-8.72480422e-02 1.80036962e-01 -6.35183275e-01 -2.69112676e-01
7.34104156e-01 3.13412338e-01 4.18016493e-01 4.33654964e-01
-2.90324569e-01 -4.63033421e-03 6.24354541e-01 -1.69984736e-02
4.09814626e-01 4.71837610e-01 5.93693033e-02 -3.27594191e-01
8.12258780e-01 2.40396261e-01 1.07572472e+00 2.67017275e-01
-3.84939969e-01 9.76863861e-01 -6.52170600e-03 -4.46241289e-01
-1.09293878e+00 -1.01360321e+00 -1.79465309e-01 9.63033795e-01
8.56536508e-01 -9.74417850e-02 -1.01219940e+00 -3.89387578e-01
-3.06609944e-02 1.80161640e-01 -4.19816792e-01 2.78163165e-01
-8.78358305e-01 -6.19229913e-01 -7.81027153e-02 5.90957522e-01
8.56544793e-01 -5.27835965e-01 -4.62062895e-01 4.60515827e-01
-4.58312541e-01 -1.00221777e+00 -1.21114624e+00 -4.28816289e-01
-1.12887704e+00 -8.97117198e-01 -1.12316239e+00 -1.36588132e+00
7.99173772e-01 8.10306549e-01 7.31950402e-01 1.42332837e-01
-1.58466473e-01 -1.09242275e-01 -3.14594746e-01 6.64181232e-01
-1.96387678e-01 -2.09284291e-01 -3.93144011e-01 3.18115562e-01
-5.50750680e-02 -3.24572355e-01 -9.16287422e-01 8.27127039e-01
-8.34640622e-01 5.00494421e-01 4.57387507e-01 4.91905272e-01
1.33955979e+00 1.33256122e-01 -2.41302922e-01 -5.73715866e-01
-1.66321874e-01 -6.39212206e-02 -9.79894936e-01 8.03944647e-01
-3.79382551e-01 -1.09577656e-01 -6.77055344e-02 -4.99174714e-01
-1.15098250e+00 3.60850811e-01 3.70647579e-01 -4.63462889e-01
1.75104884e-04 -9.89437327e-02 3.66373360e-02 -3.69821072e-01
8.70123059e-02 2.24229693e-01 -4.47594464e-01 -4.02364522e-01
6.57783031e-01 6.11267269e-01 9.05317724e-01 -2.50748008e-01
5.91283262e-01 7.78553843e-01 -1.86891288e-01 -4.79269505e-01
-6.71219170e-01 -8.40535343e-01 -1.06684101e+00 -3.33383977e-01
1.30628860e+00 -7.77465522e-01 -3.36680800e-01 6.56079054e-01
-1.34249914e+00 -4.10316229e-01 -1.35766059e-01 5.02761900e-01
-2.45876968e-01 8.61947775e-01 -8.54498327e-01 -4.24833000e-01
-2.33124033e-01 -9.49285388e-01 1.21981227e+00 4.43243414e-01
-1.01482667e-01 -9.23993945e-01 2.33510181e-01 3.67727667e-01
7.99515396e-02 2.77648598e-01 5.49919069e-01 3.57104778e-01
-1.09300542e+00 2.09656864e-01 -6.42873406e-01 3.43307555e-01
1.23463646e-01 4.38915044e-01 -5.18759847e-01 -1.70038968e-01
-3.09825808e-01 4.18332160e-01 9.80054200e-01 8.30191255e-01
7.40883172e-01 -8.80531594e-02 -6.68485343e-01 7.72222996e-01
1.44901919e+00 2.85611361e-01 5.25138915e-01 2.86027640e-01
8.59661341e-01 5.02599537e-01 1.09221697e+00 2.81658739e-01
3.37139696e-01 7.30248630e-01 1.09292082e-01 -2.25296885e-01
-3.18807870e-01 -1.57517105e-01 1.56030923e-01 6.74419940e-01
-1.07877843e-01 -1.08373240e-01 -5.87993145e-01 6.68178499e-01
-2.14507055e+00 -9.02749777e-01 -6.81394815e-01 2.21637321e+00
7.50790596e-01 -3.99246030e-02 -1.03107519e-01 1.77252311e-02
1.36790013e+00 3.01656514e-01 -4.33247834e-01 2.12706417e-01
-1.97280243e-01 -1.59166828e-01 7.53379881e-01 8.96222115e-01
-1.47917771e+00 1.20726264e+00 6.41813087e+00 1.09783602e+00
-8.20697010e-01 2.84227163e-01 1.00528157e+00 2.91438311e-01
-2.21672565e-01 1.67644769e-01 -8.27986300e-01 6.63450062e-01
-2.55835187e-02 7.97783956e-02 2.59419084e-01 6.97521150e-01
4.20880973e-01 -5.90792060e-01 -6.96703494e-01 8.78475964e-01
2.20132902e-01 -1.40322852e+00 -1.48882911e-01 -3.12775791e-01
1.38221276e+00 -2.68919945e-01 -7.11240247e-03 -4.02488291e-01
1.82290927e-01 -4.41924512e-01 9.49296713e-01 4.43671167e-01
4.09499675e-01 -3.76468241e-01 5.29204130e-01 -3.51921446e-03
-1.80502772e+00 1.48532137e-01 -4.88335937e-01 3.76152545e-01
5.30551493e-01 6.92091405e-01 -9.93612707e-02 4.25479919e-01
7.93485105e-01 1.02502394e+00 -4.20656711e-01 1.11449480e+00
-2.66078919e-01 1.79954752e-01 -2.38415331e-01 4.80283767e-01
3.37026179e-01 -7.41089344e-01 1.24576487e-01 9.68994856e-01
3.81809473e-01 3.85518968e-01 4.80418861e-01 9.62205350e-01
2.94101834e-01 1.08317263e-01 2.40306884e-01 5.42732775e-01
7.64642835e-01 1.37730682e+00 -9.59984720e-01 -6.08099461e-01
-4.06799048e-01 1.47759116e+00 3.93068679e-02 5.30702174e-01
-6.50922477e-01 -1.24818403e-02 3.29328120e-01 -5.37219197e-02
5.73400617e-01 -3.23866069e-01 -2.64646739e-01 -8.40915501e-01
4.04090136e-02 -1.70957908e-01 5.43486297e-01 -9.50229049e-01
-8.11295450e-01 3.81285548e-01 -3.80278490e-02 -1.51417744e+00
1.34568512e-01 -1.13931634e-01 -7.34035194e-01 7.14455187e-01
-1.60058737e+00 -1.25725400e+00 -4.97674078e-01 5.59760392e-01
5.60852706e-01 5.64177215e-01 4.51846085e-02 3.15278709e-01
-6.21297061e-01 1.75324455e-01 2.55381882e-01 1.22872017e-01
6.79344356e-01 -9.12034810e-01 2.64285326e-01 1.14047873e+00
-9.48534831e-02 1.45471841e-01 3.81395012e-01 -8.27138662e-01
-4.82093126e-01 -1.40377867e+00 1.12332165e+00 6.84201941e-02
5.49452364e-01 2.78161377e-01 -1.02120781e+00 5.06849527e-01
1.40161887e-01 1.39499918e-01 2.66780287e-01 -7.22131670e-01
1.92567483e-02 -3.22223902e-01 -9.90526497e-01 3.61252993e-01
9.77821231e-01 -1.75038546e-01 -3.70040178e-01 3.55967790e-01
7.41953254e-01 -4.96841848e-01 -9.97494340e-01 4.05452847e-01
4.05690998e-01 -9.70147610e-01 1.14981461e+00 2.07127765e-01
1.18915372e-01 -1.03950834e+00 6.47400692e-02 -8.08350623e-01
-7.76771605e-01 -3.47557336e-01 2.98588812e-01 1.42692864e+00
1.13423467e-01 5.22403270e-02 9.37474489e-01 5.98858118e-01
-1.14039928e-01 -4.17104065e-01 -1.00991511e+00 -6.24256492e-01
-3.71140987e-01 -1.05662726e-01 3.44069004e-01 9.11196232e-01
-6.45205304e-02 -1.57736331e-01 -3.38352680e-01 4.10710603e-01
1.07184291e+00 5.61350226e-01 2.39736229e-01 -9.03980017e-01
-2.56279349e-01 -5.09948075e-01 -4.12458867e-01 -1.25271261e+00
6.79497942e-02 -4.47932392e-01 2.56535023e-01 -1.51718903e+00
3.60319912e-01 -4.63865280e-01 -4.00009423e-01 3.60016376e-01
-6.21606529e-01 7.37817228e-01 1.36490718e-01 4.97558683e-01
-7.94294894e-01 1.16002977e-01 1.48623073e+00 -4.15031880e-01
-4.27498013e-01 -1.92711785e-01 1.87731698e-01 7.10449696e-01
6.24763191e-01 -2.09348753e-01 -1.62717998e-01 -5.86415887e-01
-3.66575539e-01 1.25557259e-01 2.20053822e-01 -1.01720834e+00
5.64085186e-01 -4.35590833e-01 2.18541563e-01 -8.99719417e-01
1.08370215e-01 -6.15345299e-01 2.80905455e-01 5.42767167e-01
-2.59861737e-01 -1.36434510e-01 -2.05952793e-01 7.28344858e-01
-4.90624696e-01 -2.85604417e-01 1.04886508e+00 -5.70842531e-03
-1.28749931e+00 5.27092874e-01 -2.86963820e-01 -1.33351743e-01
1.44538498e+00 -5.62191784e-01 -6.26153573e-02 2.41216823e-01
-7.62804687e-01 4.52274621e-01 7.23795891e-01 1.17904134e-01
7.38819897e-01 -1.36747754e+00 -6.19513035e-01 1.93720236e-01
4.31399152e-04 1.81911170e-01 6.22869551e-01 9.57534313e-01
-9.19564962e-01 5.20574301e-02 -1.81837037e-01 -7.21795082e-01
-1.68092287e+00 6.66258752e-01 3.38013291e-01 9.82066318e-02
-4.05072063e-01 9.76585686e-01 3.00382286e-01 2.35595107e-01
2.79891878e-01 -6.89495683e-01 -2.03769758e-01 -1.94169283e-01
5.56571305e-01 7.76140034e-01 -3.94645244e-01 -8.71605992e-01
-2.85412133e-01 1.38813722e+00 1.98783010e-01 -1.23123080e-01
7.50243783e-01 -4.58619773e-01 -4.41871136e-01 1.96847856e-01
1.04133821e+00 -3.35576445e-01 -1.55970240e+00 -4.68486220e-01
-3.50341052e-02 -9.35666740e-01 2.62290984e-01 -6.46071211e-02
-1.42361784e+00 5.56318343e-01 8.56649816e-01 -2.91388571e-01
1.12583160e+00 4.90993522e-02 1.09817481e+00 -4.18293208e-01
3.76386017e-01 -1.30124652e+00 9.02084820e-03 9.76983234e-02
1.43923000e-01 -1.03304005e+00 2.09676430e-01 -9.60205972e-01
-4.52939510e-01 1.11596286e+00 6.05312765e-01 -1.69507220e-01
6.89278126e-01 -1.68973580e-02 5.20759113e-02 2.87783682e-01
-2.33466834e-01 -1.77376300e-01 3.78190428e-01 4.17560428e-01
2.01066270e-01 -2.29376964e-02 -5.50180674e-01 -1.28053814e-01
7.22715914e-01 1.80321902e-01 1.12866417e-01 7.33546019e-01
-8.55322897e-01 -1.19827342e+00 -6.91965401e-01 -1.90288827e-01
1.00680552e-01 -1.55240735e-02 -6.36829361e-02 3.27173889e-01
7.61253655e-01 9.75105882e-01 5.78731775e-01 -3.18313539e-01
3.62833850e-02 -3.56048822e-01 4.38659221e-01 -3.94786626e-01
-2.27087468e-01 5.10652661e-01 -3.77923489e-01 -6.61584377e-01
-1.00797272e+00 -6.92860961e-01 -1.35596931e+00 -2.12802276e-01
-3.65826190e-01 5.59774004e-02 6.26387522e-02 7.98374474e-01
2.64866412e-01 1.48128033e-01 7.27977157e-01 -1.04793096e+00
9.80986506e-02 -7.40288675e-01 -7.57537186e-01 4.50061262e-01
6.61526397e-02 -2.55968422e-01 -1.86094522e-01 6.12433612e-01] | [9.328707695007324, -0.4342607259750366] |
98b37f92-c809-4058-80a0-a9c5ba06c4d2 | improving-few-shot-relation-classification-by | null | null | https://aclanthology.org/2022.findings-naacl.34 | https://aclanthology.org/2022.findings-naacl.34.pdf | Improving Few-Shot Relation Classification by Prototypical Representation Learning with Definition Text | Few-shot relation classification is difficult because the few instances available may not represent well the relation patterns. Some existing approaches explored extra information such as relation definition, in addition to the instances, to learn a better relation representation. However, the encoding of the extra information has been performed independently from the labeled instances. In this paper, we propose to learn a prototype encoder from relation definition in a way that is useful for relation instance classification. To this end, we use a joint training approach to train both a prototype encoder from definition and an instance encoder. Extensive experiments on several datasets demonstrate the effectiveness and usefulness of our prototype encoder from definition text, enabling us to outperform state-of-the-art approaches. | ['Dongsheng Li', 'Jian-Yun Nie', 'Yuyang Zhang', 'Li Zhenzhen'] | null | null | null | null | findings-naacl-2022-7 | ['few-shot-relation-classification', 'relation-classification', 'few-shot-relation-classification'] | ['methodology', 'natural-language-processing', 'natural-language-processing'] | [ 3.16574454e-01 6.62963390e-01 -6.38940394e-01 -5.75710297e-01
-4.65451747e-01 -2.11540926e-02 5.75466990e-01 5.16135812e-01
-1.76322848e-01 1.05421734e+00 5.60790859e-02 -4.61923778e-02
-1.93090037e-01 -1.21542037e+00 -6.53538465e-01 -3.97386640e-01
8.27977899e-03 7.10236967e-01 2.40403429e-01 -1.87904030e-01
-1.51592672e-01 9.39145163e-02 -1.66200006e+00 4.69283253e-01
7.97990203e-01 1.02721632e+00 3.25726643e-02 2.46273816e-01
-2.97441721e-01 1.15017974e+00 -6.68198884e-01 -7.07451165e-01
9.90349501e-02 -6.56441510e-01 -1.19230020e+00 2.41308108e-01
-1.30883470e-01 -2.13303745e-01 -4.64342475e-01 8.30568552e-01
1.92961276e-01 4.43099886e-01 6.24064088e-01 -1.09087861e+00
-7.58010983e-01 8.67264152e-01 -3.29990447e-01 1.82941347e-01
5.18928885e-01 -2.95777053e-01 1.28090405e+00 -5.45196235e-01
9.33910072e-01 9.80180144e-01 4.50185239e-01 4.27684844e-01
-1.17099297e+00 -5.14581800e-01 1.20073147e-01 7.56778240e-01
-1.59850025e+00 -6.17852628e-01 8.70143056e-01 -1.73937306e-01
1.37461746e+00 3.46316472e-02 6.82586372e-01 1.11339402e+00
-1.83349818e-01 7.01606095e-01 7.19305575e-01 -7.07565129e-01
2.77969152e-01 2.80094773e-01 5.22179306e-01 5.52164435e-01
3.18239689e-01 7.65638649e-02 -3.27374637e-01 -1.84101492e-01
5.58444798e-01 -6.97836559e-03 -6.29567742e-01 -2.67294079e-01
-6.97999001e-01 7.59708107e-01 6.14557981e-01 6.11860871e-01
-5.40473163e-01 -7.68497065e-02 4.62689131e-01 3.86694193e-01
7.26995766e-01 6.52786374e-01 -5.52907407e-01 -1.87511310e-01
-4.34785694e-01 -9.36689973e-03 1.17508948e+00 1.09249473e+00
1.05107999e+00 -5.82352102e-01 -2.60258228e-01 9.08607483e-01
-1.67822260e-02 -2.25808218e-01 4.82242793e-01 -5.22003829e-01
5.41058183e-01 9.35659289e-01 -1.23369336e-01 -5.46030819e-01
-1.41032606e-01 -3.82347077e-01 -7.61613905e-01 -3.61150861e-01
1.68917343e-01 3.74953449e-02 -8.96387994e-01 1.63902068e+00
3.64115864e-01 6.58008635e-01 4.17752415e-01 6.46489024e-01
9.13745284e-01 5.11946142e-01 -4.48380336e-02 -4.61589873e-01
1.26104355e+00 -9.51955855e-01 -1.07458413e+00 -3.05750847e-01
9.81284142e-01 -2.68500268e-01 7.66412199e-01 -2.85046138e-02
-7.87732184e-01 -6.11048996e-01 -1.30737603e+00 -4.88766730e-02
-5.92766821e-01 -1.03977054e-01 8.01076233e-01 3.63400370e-01
-2.94419318e-01 8.66463482e-01 -7.82274604e-01 -4.35029566e-01
6.47311807e-01 1.20413996e-01 -6.25958323e-01 -3.70027632e-01
-1.62692785e+00 1.07604647e+00 1.01808047e+00 -1.30833969e-01
-3.17976445e-01 -4.50760752e-01 -1.22699499e+00 3.04404616e-01
8.70002449e-01 -5.72899580e-01 1.15835619e+00 -6.08219922e-01
-1.22792077e+00 6.45552397e-01 -2.78494447e-01 -6.65538847e-01
-7.88913518e-02 -5.71205243e-02 -3.62785637e-01 8.63396376e-02
2.21631657e-02 4.41134423e-02 3.52835953e-01 -1.29245019e+00
-5.42564511e-01 -3.24050367e-01 5.30346930e-01 1.76215917e-01
-3.34294289e-01 -2.89546639e-01 -4.54394162e-01 -2.67827123e-01
-1.20955870e-01 -5.03121912e-01 -1.82629917e-02 -2.83706218e-01
-3.71167064e-01 -5.32474577e-01 6.13838077e-01 -4.55576301e-01
1.41522706e+00 -2.09597349e+00 7.57893473e-02 2.00705696e-02
1.54785350e-01 4.74440783e-01 -5.66710904e-02 6.61214292e-01
-2.60036469e-01 8.67475346e-02 -2.18906790e-01 -2.34674618e-01
-1.43537983e-01 7.95437455e-01 1.18435081e-02 6.15137927e-02
5.26153028e-01 1.10248029e+00 -1.18662429e+00 -5.05267501e-01
3.95674706e-02 4.76980180e-01 -1.17365904e-01 4.85321522e-01
-2.55671918e-01 3.86816375e-02 -3.29600841e-01 4.43518549e-01
3.27736109e-01 -4.33604121e-01 4.34301406e-01 -2.53163934e-01
5.88446975e-01 6.76127493e-01 -9.63101685e-01 1.65769553e+00
-4.69324350e-01 4.39318776e-01 -5.92259467e-01 -1.36989760e+00
1.05785215e+00 7.22849488e-01 3.50229949e-01 -5.70516229e-01
1.08598307e-01 9.73837152e-02 2.33884469e-01 -7.17823029e-01
2.27463737e-01 -4.67902601e-01 3.36366683e-01 4.16767240e-01
3.89253169e-01 1.17102861e-01 5.66644073e-01 8.70694071e-02
1.33268619e+00 7.12817311e-02 1.03927219e+00 3.30342084e-01
4.61698085e-01 -1.87695712e-01 7.77294278e-01 5.69084942e-01
9.04688239e-02 3.18856388e-01 6.05839193e-01 -4.27985877e-01
-6.44237816e-01 -7.24732220e-01 -3.35218072e-01 6.39649808e-01
2.18812138e-01 -1.03305542e+00 -3.33871424e-01 -1.17686808e+00
-2.05540031e-01 8.22863221e-01 -7.88953364e-01 -5.66507638e-01
-6.74070776e-01 -5.66481888e-01 3.10553104e-01 7.35190809e-01
5.49291790e-01 -9.91354823e-01 -4.54948097e-01 4.43089634e-01
-2.84339160e-01 -1.25350928e+00 -1.26997018e-02 6.66070521e-01
-6.13451064e-01 -1.37942338e+00 -8.95461589e-02 -6.56118929e-01
4.86161679e-01 5.75221367e-02 1.24855828e+00 2.65870243e-01
-1.86928689e-01 -2.08264008e-01 -9.48868871e-01 -4.57100123e-01
-3.60807776e-01 2.73729980e-01 -3.28184605e-01 -1.05236426e-01
8.27790022e-01 -6.15043163e-01 2.04916038e-02 -3.44019793e-02
-8.02434444e-01 -5.20666763e-02 5.98225355e-01 1.20812166e+00
5.55332422e-01 5.06493092e-01 6.95893168e-01 -1.45699430e+00
5.85378945e-01 -5.53290069e-01 -3.92717216e-03 5.92016220e-01
-6.38176262e-01 1.69180721e-01 6.05347216e-01 -4.19556081e-01
-1.07701373e+00 5.17452024e-02 -1.34567395e-01 -4.30583626e-01
-2.77645797e-01 9.28374827e-01 -4.40684855e-01 2.85607189e-01
6.66336715e-01 7.02562602e-03 -2.52017826e-01 -5.11362910e-01
3.52516174e-01 8.13981950e-01 3.00351053e-01 -4.63696539e-01
4.61823165e-01 -7.48645887e-02 -4.55543362e-02 -5.21004498e-01
-1.30134070e+00 -4.76114959e-01 -9.89403009e-01 3.24715257e-01
6.02467954e-01 -6.57228947e-01 -1.44252196e-01 -1.39830098e-01
-1.26063478e+00 -2.12268487e-01 -8.06846619e-01 4.38735813e-01
-3.69156092e-01 1.89812586e-01 -7.34718859e-01 -7.58089542e-01
-1.54077247e-01 -7.49601901e-01 7.72631228e-01 1.68853119e-01
-2.36839890e-01 -1.00183702e+00 1.51063457e-01 1.36309832e-01
9.79407951e-02 2.18239576e-01 9.57498252e-01 -1.10422099e+00
-3.30416024e-01 -4.90259677e-01 -2.49806628e-01 1.37083799e-01
6.08327270e-01 -4.24782395e-01 -1.08882201e+00 9.63557803e-04
6.87502250e-02 -4.57812190e-01 9.41352308e-01 -1.17831178e-01
1.06322360e+00 -3.21006268e-01 -6.78018689e-01 3.77547413e-01
1.31306338e+00 3.91771466e-01 7.69466877e-01 6.10034578e-02
6.11912191e-01 5.46654940e-01 1.02531111e+00 3.65389973e-01
4.30368692e-01 8.32176208e-01 2.21243612e-02 1.88044965e-01
-1.90183908e-01 -1.79315597e-01 -2.26177543e-01 5.57470560e-01
-2.57164925e-01 -2.41401240e-01 -8.22002470e-01 4.49447960e-01
-2.16484356e+00 -1.12400436e+00 2.01449051e-01 1.91166925e+00
1.29054439e+00 2.14516819e-01 -2.48236433e-01 5.89137375e-01
5.85084200e-01 1.10661574e-01 -2.92923093e-01 -2.75142789e-01
1.02968752e-01 5.87236226e-01 2.09919158e-02 2.89001733e-01
-1.15608084e+00 9.59661245e-01 5.86284733e+00 7.35797107e-01
-7.47636199e-01 1.50264665e-01 3.84533137e-01 2.36798778e-01
-7.01629519e-02 1.98506847e-01 -6.78290725e-01 3.16117883e-01
9.61793303e-01 -3.94700617e-01 1.47549883e-01 8.19155157e-01
-4.72377628e-01 -1.95488636e-03 -1.61293542e+00 8.65493357e-01
3.09857428e-01 -1.28317034e+00 -1.29562169e-01 9.61909518e-02
5.13593316e-01 -4.60081518e-01 -7.44094849e-01 7.47088253e-01
-6.06713556e-02 -9.81549323e-01 5.95276803e-02 4.76312339e-01
6.88916743e-01 -7.54209280e-01 1.21441162e+00 4.53855157e-01
-1.47716641e+00 1.22135460e-01 -6.40654743e-01 -4.72593129e-01
1.58420339e-01 8.17417979e-01 -1.17074883e+00 1.17610192e+00
1.89008355e-01 1.09214306e+00 -5.13123572e-01 8.70239377e-01
-6.30165935e-01 4.02997613e-01 -1.07567236e-01 1.01575084e-01
-1.88488841e-01 1.39765397e-01 1.25784650e-01 1.06455266e+00
-8.93654115e-03 6.09416842e-01 4.36252326e-01 7.63288260e-01
-3.24128419e-01 1.73557818e-01 -9.02318716e-01 -2.02848107e-01
6.71860397e-01 1.06565797e+00 -4.20953900e-01 -6.63400710e-01
-6.27220690e-01 8.31832588e-01 9.03084576e-01 2.39604414e-01
-6.31096303e-01 -6.14692688e-01 4.99359608e-01 -1.51895121e-01
4.39258307e-01 7.93171301e-02 -1.18724018e-01 -1.36344731e+00
1.38216794e-01 -5.01715004e-01 5.61407089e-01 -3.98327798e-01
-1.47400141e+00 7.69475043e-01 1.30567774e-01 -1.13113678e+00
-6.24448299e-01 -3.49348634e-01 -5.89978337e-01 7.46295631e-01
-1.49800324e+00 -1.12773013e+00 -3.39124173e-01 3.30688417e-01
4.16014433e-01 3.22717964e-03 1.18378663e+00 3.33464980e-01
-7.75904357e-01 7.54470646e-01 -5.09894013e-01 3.84367615e-01
5.93905270e-01 -1.33721602e+00 1.66750625e-01 7.23848820e-01
5.47727466e-01 6.60941005e-01 4.65726942e-01 -6.43013120e-01
-1.14738142e+00 -1.02478456e+00 1.14348805e+00 -2.00880334e-01
5.07739007e-01 -1.14268206e-01 -1.40712094e+00 8.73587787e-01
2.99674153e-01 3.08630496e-01 1.13761079e+00 6.68187916e-01
-4.30156291e-01 -1.43682703e-01 -1.11554265e+00 2.18796507e-01
1.23326111e+00 -5.95248878e-01 -1.13462186e+00 1.88272342e-01
7.57690609e-01 -2.42889330e-01 -1.26241672e+00 6.15214467e-01
3.12445164e-01 -5.31958342e-01 8.55812371e-01 -7.29168177e-01
5.86201012e-01 -1.46024436e-01 -1.73869655e-01 -1.48975420e+00
-3.74611706e-01 -8.43038186e-02 -9.81153548e-01 1.50179327e+00
4.78197277e-01 -5.65967619e-01 6.77734315e-01 7.18855441e-01
6.78789541e-02 -1.15541458e+00 -9.00126159e-01 -8.85117233e-01
-3.20485264e-01 -3.72679442e-01 8.92699540e-01 1.15363717e+00
5.92784166e-01 1.07483113e+00 -3.21252465e-01 3.59553769e-02
3.31502020e-01 5.44674635e-01 5.85286796e-01 -1.45653713e+00
-5.93204856e-01 -9.26090032e-02 -7.27041960e-01 -8.21771502e-01
4.85751361e-01 -1.05222452e+00 8.58405009e-02 -1.76033711e+00
4.47346419e-01 -4.53151911e-01 -5.35536885e-01 8.59916389e-01
-7.09319949e-01 7.69850565e-03 1.10895278e-04 -4.82470868e-03
-5.39961994e-01 7.16003358e-01 1.16502702e+00 -2.43029907e-01
-1.50307938e-01 -1.59114331e-01 -6.91295087e-01 3.27889711e-01
7.39106297e-01 -4.51949418e-01 -7.56159544e-01 -1.05437763e-01
-1.41090512e-01 1.72532216e-01 -5.73067740e-02 -8.38034809e-01
1.38873994e-01 -1.87638193e-01 2.95754552e-01 -2.93354452e-01
5.93715847e-01 -7.52441227e-01 1.85322128e-02 2.04521909e-01
-3.60261768e-01 -6.92421615e-01 -2.46990949e-01 5.63039184e-01
-6.29613221e-01 -5.25237203e-01 4.89854455e-01 -1.13709033e-01
-7.88434088e-01 3.61678362e-01 2.79195935e-01 6.60542622e-02
1.07828224e+00 4.60098907e-02 -4.59047139e-01 -1.49915665e-01
-7.48735785e-01 1.03296384e-01 4.42138463e-01 3.45640630e-01
6.95765197e-01 -1.50449443e+00 -5.58847487e-01 2.04414427e-01
6.80105627e-01 1.75915435e-01 -2.13407189e-01 6.45045936e-01
7.58163184e-02 2.40429759e-01 4.16593812e-02 -1.22423656e-01
-1.23785603e+00 1.01837194e+00 1.34976119e-01 -5.26868999e-01
-8.91112864e-01 6.74287379e-01 -2.10577939e-02 -1.34380668e-01
2.50930279e-01 -2.80876458e-01 -4.35344547e-01 1.69177964e-01
7.86846995e-01 -3.98178548e-02 9.36670974e-02 -3.14134985e-01
-3.71591151e-01 2.17119217e-01 -2.82796443e-01 2.26216897e-01
1.35064793e+00 3.98487784e-02 -3.99197601e-02 6.92087352e-01
1.22548151e+00 -4.81997043e-01 -8.09382796e-01 -5.91053486e-01
4.27876472e-01 -7.37948537e-01 -2.27034643e-01 -5.90890884e-01
-9.01383579e-01 6.51246905e-01 -2.14129668e-02 4.74900365e-01
1.18931282e+00 3.79858106e-01 8.30989540e-01 6.45534337e-01
5.91403306e-01 -7.92314649e-01 -8.50777850e-02 6.39076650e-01
6.55755818e-01 -1.54073954e+00 3.01132668e-02 -9.59216714e-01
-5.82590878e-01 1.13595688e+00 7.21485436e-01 2.25716829e-03
7.95486569e-01 3.92668337e-01 -1.66464418e-01 -3.09270084e-01
-1.24695575e+00 -7.69441426e-01 3.59382451e-01 7.05944061e-01
8.40767562e-01 6.88636303e-02 -6.52859390e-01 9.10459936e-01
-1.99607550e-03 2.72464514e-01 3.05585712e-01 1.10541701e+00
-2.19270676e-01 -1.61389649e+00 2.35326961e-01 7.98148274e-01
-3.28086317e-01 -6.21479750e-03 -5.17307341e-01 6.70206845e-01
3.64019811e-01 1.06082642e+00 -9.64178964e-02 -5.73848963e-01
3.85987997e-01 4.49738413e-01 6.54760361e-01 -1.22650206e+00
-3.88786018e-01 -3.49302918e-01 7.36898005e-01 -4.35963213e-01
-4.44293112e-01 -3.78647298e-01 -1.21251059e+00 5.41002862e-02
-6.39863968e-01 3.61790717e-01 3.71588618e-02 1.28327084e+00
1.28290653e-01 7.85063326e-01 5.08963168e-01 -4.42192376e-01
-3.19674015e-01 -1.27498758e+00 -6.26835942e-01 7.02371836e-01
1.64810140e-02 -1.00209939e+00 -5.32492027e-02 -1.43424690e-01] | [9.26545524597168, 8.539090156555176] |
55b36de1-169c-4256-9e41-15c6c013d37f | unified-perspective-on-probability-divergence | 2201.13127 | null | https://arxiv.org/abs/2201.13127v1 | https://arxiv.org/pdf/2201.13127v1.pdf | Unified Perspective on Probability Divergence via Maximum Likelihood Density Ratio Estimation: Bridging KL-Divergence and Integral Probability Metrics | This paper provides a unified perspective for the Kullback-Leibler (KL)-divergence and the integral probability metrics (IPMs) from the perspective of maximum likelihood density-ratio estimation (DRE). Both the KL-divergence and the IPMs are widely used in various fields in applications such as generative modeling. However, a unified understanding of these concepts has still been unexplored. In this paper, we show that the KL-divergence and the IPMs can be represented as maximal likelihoods differing only by sampling schemes, and use this result to derive a unified form of the IPMs and a relaxed estimation method. To develop the estimation problem, we construct an unconstrained maximum likelihood estimator to perform DRE with a stratified sampling scheme. We further propose a novel class of probability divergences, called the Density Ratio Metrics (DRMs), that interpolates the KL-divergence and the IPMs. In addition to these findings, we also introduce some applications of the DRMs, such as DRE and generative adversarial networks. In experiments, we validate the effectiveness of our proposed methods. | ['Kentaro Minami', 'Masaaki Imaizumi', 'Masahiro Kato'] | 2022-01-31 | null | null | null | null | ['density-ratio-estimation'] | ['methodology'] | [ 6.25112280e-02 -7.73188546e-02 -2.08722636e-01 -4.50021505e-01
-9.12725329e-01 -5.44587255e-01 4.82505143e-01 -2.58971393e-01
-1.95317611e-01 1.03389716e+00 -1.83265954e-01 -4.03145492e-01
-5.21332264e-01 -7.65855551e-01 -6.09618247e-01 -7.59999812e-01
-1.70771778e-01 2.38434672e-02 2.25914806e-01 1.23612173e-01
1.09858580e-01 4.72583026e-01 -1.20484710e+00 -6.50300622e-01
1.37318635e+00 1.11535466e+00 -4.62123305e-02 4.02069062e-01
1.15597717e-01 4.96086538e-01 -7.45046556e-01 -7.43687510e-01
3.51573080e-01 -5.63466728e-01 -3.66236567e-01 -1.41132116e-01
8.33065212e-02 -5.96155822e-01 -1.84357285e-01 1.33907247e+00
4.17714030e-01 2.62694061e-01 1.15741861e+00 -1.81564081e+00
-7.14152813e-01 3.74257892e-01 -4.99941409e-01 4.41360064e-02
3.94989818e-01 -3.70662361e-01 8.51567566e-01 -8.35674942e-01
3.27776045e-01 1.39481413e+00 7.56487310e-01 3.25380594e-01
-9.42602992e-01 -6.57981813e-01 -3.30972485e-02 1.45283360e-02
-1.76387334e+00 9.44737494e-02 7.59489059e-01 -4.64957744e-01
1.73717126e-01 2.53473312e-01 2.35923484e-01 9.55544293e-01
4.17053670e-01 8.79167736e-01 1.20131779e+00 -3.95374060e-01
4.35676843e-01 4.55969810e-01 -1.36471510e-01 5.98672271e-01
4.82866973e-01 5.69535121e-02 -5.92499822e-02 -4.22724068e-01
1.08364046e+00 -2.16671042e-02 -3.27024341e-01 -5.36259770e-01
-8.05772960e-01 1.03950846e+00 1.55029029e-01 2.59997874e-01
4.11089361e-02 -9.45740417e-02 -4.49823625e-02 1.87299866e-02
5.99370897e-01 -5.46123534e-02 -2.31956840e-02 9.09832716e-02
-7.67270684e-01 4.06845063e-01 1.05958068e+00 1.04507518e+00
7.90710330e-01 3.60352881e-02 -1.98080152e-01 7.49269545e-01
8.34791839e-01 6.58059537e-01 7.31300190e-02 -8.20838571e-01
4.07283753e-01 1.13023564e-01 3.84958982e-01 -1.06309664e+00
1.99402664e-02 -5.33174455e-01 -7.82215357e-01 -6.85857385e-02
3.51585001e-01 -3.15125674e-01 -4.23309952e-01 2.17534161e+00
3.90110016e-01 5.35949588e-01 5.88131370e-03 9.12430167e-01
1.72095731e-01 6.29708409e-01 -1.53292045e-01 -5.10146737e-01
7.84076810e-01 -4.54091489e-01 -8.67165744e-01 1.88028157e-01
4.69306290e-01 -6.10047340e-01 8.82349432e-01 2.17268988e-01
-1.02362120e+00 -3.20586383e-01 -1.25703502e+00 2.24453986e-01
-7.53779113e-02 -1.40127808e-01 4.57722932e-01 9.62424576e-01
-9.45564985e-01 7.65370369e-01 -8.90886247e-01 -2.36050919e-01
-5.45507930e-02 1.24783744e-03 2.05445122e-02 1.77666247e-01
-1.37369502e+00 8.31125379e-01 1.85183600e-01 1.46994650e-01
-8.56614232e-01 -6.81761146e-01 -8.44582736e-01 -6.80274293e-02
2.96274453e-01 -4.74073559e-01 1.25134921e+00 -5.08509755e-01
-1.74037766e+00 3.25052708e-01 5.09598590e-02 -2.53098518e-01
6.44143224e-01 -2.06855536e-01 -5.30997455e-01 -2.86906753e-02
1.24014549e-01 1.57381669e-01 6.17553055e-01 -1.22690070e+00
-3.05393815e-01 -2.90691435e-01 6.72154948e-02 7.40296692e-02
-2.06084266e-01 -1.85156111e-02 -1.02078944e-01 -8.00927401e-01
1.68806136e-01 -6.88702762e-01 -5.74943647e-02 1.18333004e-01
-5.38619637e-01 -2.00017840e-01 5.71921587e-01 -5.75286806e-01
1.38061464e+00 -2.14306211e+00 1.73253536e-01 3.76597762e-01
-4.38140742e-02 1.35097161e-01 1.58656642e-01 4.59828824e-01
2.48212904e-01 2.90145129e-01 -7.04263806e-01 -2.97152936e-01
3.25719774e-01 1.58713013e-01 -4.41649169e-01 6.88088536e-01
2.65566170e-01 5.35107195e-01 -1.02293742e+00 -5.26474237e-01
1.29665047e-01 4.76453036e-01 -3.46210033e-01 2.78051585e-01
4.65501696e-02 2.71785378e-01 -5.13085961e-01 5.66177070e-01
1.12390137e+00 -3.81883234e-02 8.06498304e-02 -1.64930612e-01
-2.58624628e-02 -1.42902881e-01 -1.43153572e+00 1.40606308e+00
-3.09752971e-01 3.01922917e-01 1.89519804e-02 -9.99605596e-01
1.11275506e+00 1.11926734e-01 1.56127840e-01 -2.13865802e-01
1.69640392e-01 4.40249443e-01 -3.17740619e-01 -2.21463725e-01
3.21699023e-01 -6.31199539e-01 -2.89159596e-01 2.94932663e-01
2.15000853e-01 4.78996038e-02 1.63535729e-01 1.63277060e-01
5.61348617e-01 1.75587580e-01 5.06990373e-01 -2.86468834e-01
6.59063935e-01 -7.29366183e-01 7.42811263e-01 7.08047628e-01
-2.70880699e-01 5.20156920e-01 7.01116860e-01 3.75958234e-01
-8.02998722e-01 -1.90989530e+00 -6.58771694e-01 3.70752186e-01
4.70130444e-01 -2.35272616e-01 -7.33671129e-01 -7.11707294e-01
1.17594883e-01 8.00390184e-01 -1.87676415e-01 -2.84207076e-01
-1.11563496e-01 -8.73323202e-01 8.30241621e-01 5.25734365e-01
6.62795305e-01 -2.11672097e-01 9.40973461e-02 -7.34294280e-02
1.11485953e-02 -8.96102905e-01 -5.08269131e-01 -2.82144874e-01
-7.50105977e-01 -7.57338643e-01 -9.91173148e-01 -4.47752148e-01
3.89099449e-01 -4.74289097e-02 7.84266412e-01 -5.34806490e-01
1.04085550e-01 3.88959765e-01 -2.27211311e-01 -2.48599082e-01
-6.04648829e-01 -3.03090632e-01 3.16152930e-01 1.25367627e-01
2.30218340e-02 -8.76985192e-01 -5.25031209e-01 6.73252225e-01
-1.17322969e+00 -3.73220772e-01 7.94194698e-01 5.31117558e-01
6.74949408e-01 2.31292266e-02 1.05415618e+00 -5.88384628e-01
9.14635062e-01 -7.78029203e-01 -8.45608056e-01 4.68572438e-01
-7.40374804e-01 3.26025844e-01 6.23257697e-01 -5.22382975e-01
-1.25987840e+00 -5.64942181e-01 -2.01196343e-01 -4.70457733e-01
1.98988523e-02 5.68710089e-01 -6.29992366e-01 -5.51147610e-02
2.81142533e-01 4.35865633e-02 -1.42212242e-01 -6.06022894e-01
4.95151132e-01 1.05171382e+00 6.35491729e-01 -6.31722689e-01
7.99927711e-01 3.50850284e-01 9.34177563e-02 -5.90872467e-01
-9.80536222e-01 -3.29459369e-01 -3.81953657e-01 -5.91060035e-02
8.05767894e-01 -5.89883029e-01 -4.65710342e-01 5.73436916e-01
-9.65963662e-01 2.25846499e-01 -1.83147609e-01 8.17633450e-01
-6.75415277e-01 7.51757979e-01 -5.50231993e-01 -1.24298167e+00
-4.53495421e-02 -1.02373075e+00 1.02419937e+00 4.78155762e-01
7.21645728e-02 -1.43714249e+00 3.84200394e-01 -2.04168335e-01
1.35235831e-01 4.22337174e-01 8.25850606e-01 -7.06547916e-01
-5.00596285e-01 -2.51115263e-01 -2.70307630e-01 7.46703386e-01
1.19432345e-01 -9.34709236e-02 -7.29101002e-01 -3.11212450e-01
3.49904507e-01 2.21826043e-02 5.54542243e-01 3.30487311e-01
1.01019180e+00 -2.52109170e-01 -2.76444852e-01 7.32538819e-01
1.54417443e+00 3.56981754e-01 7.80787766e-01 -2.94923391e-02
4.25722212e-01 2.42312521e-01 6.06584609e-01 7.12774158e-01
1.63918361e-01 5.53788900e-01 2.50625104e-01 4.82071966e-01
3.26095790e-01 -5.75790048e-01 5.40692210e-01 1.12446594e+00
2.05664143e-01 -4.59984869e-01 -4.91636217e-01 2.26269081e-01
-1.77266526e+00 -8.21428835e-01 1.81062430e-01 2.56767082e+00
9.02009964e-01 -5.47833741e-02 -2.57034376e-02 8.17561895e-02
1.05839264e+00 3.14186245e-01 -5.94921052e-01 -3.92401099e-01
-1.24613792e-02 1.77281439e-01 4.33158934e-01 7.11947918e-01
-1.02088165e+00 3.84195030e-01 6.75218725e+00 1.25226474e+00
-6.44200623e-01 6.94631487e-02 3.27918649e-01 4.54311728e-01
-5.61371684e-01 3.07729900e-01 -8.89247060e-01 7.26322711e-01
9.70590830e-01 -3.62255424e-01 2.12600306e-01 1.00957191e+00
-8.64690542e-02 -1.11451015e-01 -1.02308631e+00 8.33764017e-01
1.27040952e-01 -7.09288716e-01 5.87888956e-02 1.59297526e-01
7.15587139e-01 -4.96995270e-01 1.82904005e-01 2.49352485e-01
7.00790435e-02 -8.50509882e-01 5.58408499e-01 8.22972357e-01
7.91142166e-01 -8.55398357e-01 7.94194937e-01 4.36352581e-01
-1.05446601e+00 2.19200745e-01 -4.44188178e-01 5.29814176e-02
4.57276791e-01 9.54013228e-01 -6.45410538e-01 9.08846498e-01
7.92265534e-02 6.06142819e-01 -1.22364752e-01 1.30584478e+00
-2.95143396e-01 3.67500454e-01 -3.72747719e-01 -1.80181295e-01
-5.89471608e-02 -5.62214732e-01 7.87825584e-01 9.42813993e-01
7.02514648e-01 -3.29281867e-01 -8.78120810e-02 1.43711555e+00
-7.41436034e-02 -1.67414576e-01 -2.95651436e-01 -7.78999999e-02
8.49054515e-01 1.12920868e+00 -4.46746200e-01 6.33617118e-02
-2.75155872e-01 8.37724447e-01 -8.33975058e-03 4.96050507e-01
-1.08993912e+00 -7.00681448e-01 8.10128927e-01 -1.25363648e-01
4.46877442e-02 -3.19549710e-01 1.25353098e-01 -1.35716653e+00
2.01466799e-01 -3.30144197e-01 1.33368731e-01 -5.53317904e-01
-1.61860383e+00 1.83939815e-01 6.93025768e-01 -1.50970042e+00
-5.21036506e-01 -6.43543601e-01 -6.40970826e-01 9.92750704e-01
-1.40421093e+00 -4.99498546e-01 -4.68283743e-02 3.84658456e-01
2.97955368e-02 1.27456129e-01 5.43756425e-01 4.41956729e-01
-6.78102195e-01 8.11883926e-01 6.08258426e-01 -4.82336953e-02
5.21296978e-01 -1.37084961e+00 3.91794965e-02 8.57865930e-01
-6.46770969e-02 6.38156056e-01 6.98535025e-01 -6.51256919e-01
-1.10554314e+00 -1.02547479e+00 4.45888132e-01 -1.69096962e-01
7.46439993e-01 -2.96482682e-01 -7.72967279e-01 7.03395963e-01
-2.52658725e-01 -1.77683100e-01 6.66272521e-01 -1.91161200e-01
-2.00009868e-01 -1.41770348e-01 -1.41778421e+00 5.06756365e-01
8.71701419e-01 -4.84857082e-01 -5.60432673e-01 2.59512216e-01
7.91252494e-01 -1.58237740e-01 -1.10782087e+00 5.70416510e-01
5.98926604e-01 -1.24159884e+00 9.07889605e-01 -1.39019504e-01
-2.68457029e-02 -3.78270745e-01 -4.67391372e-01 -1.25707638e+00
1.17273135e-02 -7.03680634e-01 -1.35770574e-01 1.50015259e+00
1.65244579e-01 -8.05932403e-01 2.63560086e-01 4.79505301e-01
-7.42129758e-02 -9.62385118e-01 -1.09297872e+00 -1.24216402e+00
3.65545392e-01 -6.44856870e-01 5.15348494e-01 5.78244925e-01
-4.45233546e-02 -8.27825814e-02 -4.87350702e-01 3.21123064e-01
7.83037663e-01 -2.42342785e-01 4.67471331e-01 -1.26437080e+00
-5.45573473e-01 -2.07060337e-01 -6.81240976e-01 -1.25892901e+00
1.82343334e-01 -9.08925891e-01 -2.36968827e-02 -1.36629307e+00
2.03483507e-01 -2.35242620e-01 -3.73782873e-01 -2.84356385e-01
-4.95359227e-02 -2.39353612e-01 9.11618620e-02 3.26412559e-01
-5.66497505e-01 9.40486252e-01 1.03999877e+00 4.08722103e-01
-5.45530096e-02 2.50392854e-01 -5.57755888e-01 1.02683353e+00
6.43131375e-01 -3.66841614e-01 -7.47496843e-01 1.47843128e-02
1.59984827e-01 1.98688224e-01 4.70214069e-01 -9.98510242e-01
-4.22233716e-02 -1.77157670e-01 1.24436535e-01 -5.76591492e-01
1.88665196e-01 -4.58293051e-01 1.56841770e-01 1.20859288e-01
-1.29096374e-01 -2.42138878e-01 -2.65480638e-01 8.78743589e-01
-2.80766904e-01 -5.14307678e-01 7.18135297e-01 2.08343163e-01
-2.74424225e-01 3.06839406e-01 -2.15985253e-01 2.62117207e-01
1.09505117e+00 -7.38869905e-02 -3.73748213e-01 -5.97459614e-01
-5.73232651e-01 8.91860276e-02 3.19312692e-01 7.06034750e-02
7.19112992e-01 -1.70487630e+00 -4.61175412e-01 2.03965306e-01
-1.64428145e-01 -7.82809630e-02 6.42840043e-02 9.62596536e-01
-3.84310842e-01 3.03499460e-01 8.22466910e-02 -4.97736514e-01
-5.42303622e-01 3.25233519e-01 4.59924668e-01 -2.87554026e-01
-7.15557262e-02 5.52085459e-01 3.63539010e-01 -4.63094234e-01
2.46909603e-01 -3.60655487e-01 1.44848779e-01 -2.97247976e-01
4.82109100e-01 7.06158221e-01 -3.60843241e-01 -5.33898771e-01
-3.89563024e-01 5.46593845e-01 1.90804750e-01 -5.22131383e-01
7.30004132e-01 -3.56857359e-01 -2.33328752e-02 6.84612751e-01
1.62083566e+00 4.87237889e-03 -1.36509800e+00 -8.88295844e-02
3.64966765e-02 -5.35932481e-01 -2.34682828e-01 -4.18561459e-01
-7.08915412e-01 8.03896308e-01 5.24785995e-01 4.69500124e-01
1.05291009e+00 7.71752894e-02 8.40917468e-01 8.54731053e-02
3.77135187e-01 -7.43906558e-01 -1.63320765e-01 2.31099963e-01
6.91124916e-01 -7.82262325e-01 -1.66304052e-01 -5.75287879e-01
-4.12528157e-01 8.22161674e-01 3.04237843e-01 -3.24805588e-01
9.17488337e-01 3.83080900e-01 -3.72688562e-01 5.63628435e-01
-2.83145070e-01 -8.21834314e-04 4.23552334e-01 7.18836367e-01
3.08680505e-01 5.79881258e-02 -7.28633225e-01 1.05547667e+00
-1.59033015e-01 4.04974483e-02 3.77786607e-01 8.28587592e-01
-4.29023862e-01 -1.17169380e+00 -2.32452527e-01 2.44762525e-01
-5.11138022e-01 1.41329631e-01 -1.99406371e-01 7.20555842e-01
-2.80483924e-02 8.42268407e-01 1.66722462e-02 -5.34244895e-01
6.17794245e-02 7.14919716e-02 5.13406038e-01 -1.96011573e-01
3.74592781e-01 -2.47418471e-02 -1.74991429e-01 -2.44886905e-01
-4.91677016e-01 -5.40510356e-01 -9.39788342e-01 -2.95798570e-01
-6.84688807e-01 2.92033792e-01 5.76063991e-01 1.03015614e+00
4.95327488e-02 8.61030892e-02 8.71426225e-01 -4.73752499e-01
-1.24229670e+00 -9.39880788e-01 -1.16230226e+00 1.75362021e-01
1.47450030e-01 -8.85856867e-01 -9.71035600e-01 -4.00591761e-01] | [7.231180667877197, 3.9956088066101074] |
fe7484b8-d0dd-46e3-9802-a3b96d4a2aeb | the-surprising-effectiveness-of-visual | 2108.11550 | null | https://arxiv.org/abs/2108.11550v1 | https://arxiv.org/pdf/2108.11550v1.pdf | The Surprising Effectiveness of Visual Odometry Techniques for Embodied PointGoal Navigation | It is fundamental for personal robots to reliably navigate to a specified goal. To study this task, PointGoal navigation has been introduced in simulated Embodied AI environments. Recent advances solve this PointGoal navigation task with near-perfect accuracy (99.6% success) in photo-realistically simulated environments, assuming noiseless egocentric vision, noiseless actuation, and most importantly, perfect localization. However, under realistic noise models for visual sensors and actuation, and without access to a "GPS and Compass sensor," the 99.6%-success agents for PointGoal navigation only succeed with 0.3%. In this work, we demonstrate the surprising effectiveness of visual odometry for the task of PointGoal navigation in this realistic setting, i.e., with realistic noise models for perception and actuation and without access to GPS and Compass sensors. We show that integrating visual odometry techniques into navigation policies improves the state-of-the-art on the popular Habitat PointNav benchmark by a large margin, improving success from 64.5% to 71.7% while executing 6.4 times faster. | ['Alexander Schwing', 'Dhruv Batra', 'Harsh Agrawal', 'Xiaoming Zhao'] | 2021-08-26 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Zhao_The_Surprising_Effectiveness_of_Visual_Odometry_Techniques_for_Embodied_PointGoal_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Zhao_The_Surprising_Effectiveness_of_Visual_Odometry_Techniques_for_Embodied_PointGoal_ICCV_2021_paper.pdf | iccv-2021-1 | ['pointgoal-navigation'] | ['robots'] | [-9.15528983e-02 2.41174057e-01 3.13330591e-01 -1.43111318e-01
-4.52879161e-01 -7.87559390e-01 6.74335539e-01 2.02644482e-01
-9.41852152e-01 8.78233969e-01 -2.31031686e-01 -2.52823383e-01
-7.08975568e-02 -6.71637833e-01 -1.03956330e+00 -5.69951177e-01
-5.82109749e-01 7.22699761e-01 3.01319659e-01 -7.55686879e-01
1.72912762e-01 3.02181154e-01 -1.59127486e+00 -7.65736461e-01
9.11412060e-01 7.29265153e-01 4.35276747e-01 1.14216840e+00
5.97239494e-01 4.77574646e-01 -5.78720033e-01 3.15892279e-01
4.74875510e-01 7.96556249e-02 -3.34403366e-01 -3.85293633e-01
4.79681075e-01 -3.97141069e-01 -4.24242079e-01 1.06939960e+00
8.46867621e-01 3.79840404e-01 4.83295977e-01 -1.50795364e+00
-4.26604599e-01 1.08245291e-01 -3.30044836e-01 -2.33094886e-01
7.49991536e-01 6.74088597e-01 6.58391654e-01 -2.40242958e-01
7.65584886e-01 1.02676165e+00 7.12903321e-01 6.15061224e-01
-1.25645936e+00 -2.73313552e-01 2.11185485e-01 -9.94072780e-02
-1.48469925e+00 -5.57635844e-01 -9.58046690e-02 -3.35375875e-01
1.37905288e+00 -2.40607169e-02 7.98197865e-01 1.26372564e+00
5.94386637e-01 2.78530061e-01 7.58390307e-01 -5.16068265e-02
8.07365477e-01 -3.72232109e-01 -2.38247916e-01 7.50553966e-01
7.78660238e-01 3.69441569e-01 -4.86632973e-01 -3.71106192e-02
6.13181770e-01 -2.14954749e-01 -4.14265424e-01 -1.08958030e+00
-1.30253816e+00 4.64265853e-01 9.86580491e-01 -2.11084351e-01
-5.53474426e-01 7.47000277e-01 -9.73251909e-02 3.30610693e-01
-2.21940503e-01 6.87353671e-01 -2.09417179e-01 -4.07077104e-01
-1.17944069e-01 8.10205996e-01 1.04088855e+00 1.51823008e+00
5.65101504e-01 2.84380108e-01 4.10870880e-01 3.51622641e-01
3.00493270e-01 1.46502101e+00 3.25508863e-01 -1.47858918e+00
5.87519526e-01 3.63168806e-01 1.07585013e+00 -9.97507215e-01
-9.90913272e-01 -5.77910364e-01 -4.07901466e-01 7.81272411e-01
5.78371108e-01 -4.28884894e-01 -1.29962182e+00 2.01162076e+00
3.29230100e-01 -1.70865878e-01 4.15343225e-01 1.25302219e+00
3.29921156e-01 3.81809145e-01 -9.41439718e-02 3.10120195e-01
1.17264688e+00 -6.53814495e-01 -4.58096236e-01 -9.98929024e-01
7.66469002e-01 -2.08025441e-01 1.04918659e+00 4.92608458e-01
-5.48404992e-01 2.59724958e-03 -1.43058145e+00 -4.12538648e-02
-2.48494700e-01 -2.34541580e-01 4.60585386e-01 4.41322386e-01
-1.37205637e+00 1.49787232e-01 -1.27023125e+00 -9.38519061e-01
-1.15839601e-01 5.03670096e-01 -5.35795033e-01 -3.82883310e-01
-8.56636524e-01 1.24352062e+00 -5.04485443e-02 -2.46872470e-01
-1.40354455e+00 -2.48840004e-01 -1.01878583e+00 -2.01884985e-01
4.42331970e-01 -1.08654952e+00 1.66898704e+00 -9.70287994e-02
-1.51753592e+00 5.57039797e-01 -6.65377825e-02 -9.30578589e-01
6.62496626e-01 -6.37420356e-01 1.42930746e-01 -1.43199101e-01
5.04719079e-01 8.94642413e-01 3.32479745e-01 -1.57715607e+00
-8.16702843e-01 -6.84093356e-01 3.44560772e-01 8.93086612e-01
5.17534137e-01 -1.02896762e+00 -4.98722255e-01 2.51970559e-01
4.85760063e-01 -1.47415531e+00 -7.21149027e-01 2.67109007e-01
-1.36170000e-01 3.01093489e-01 1.62590906e-01 -2.67613441e-01
7.63576403e-02 -1.82905304e+00 2.09246263e-01 1.36439651e-01
2.08931655e-01 -3.53617400e-01 2.54880972e-02 3.86705488e-01
8.49972904e-01 -2.86053210e-01 -9.54387039e-02 -6.69475734e-01
3.21666867e-01 5.10393679e-01 -2.06287473e-01 7.63382494e-01
-4.60521132e-01 8.18589091e-01 -1.18038058e+00 2.04690769e-01
2.99000055e-01 5.94242394e-01 -6.09957635e-01 3.41877267e-02
-2.13264674e-01 4.22916263e-01 -3.27241838e-01 6.66051209e-01
3.24386597e-01 1.91417441e-01 3.89328718e-01 5.70749462e-01
-4.38138902e-01 1.84992999e-01 -1.10576141e+00 2.28894258e+00
-5.76261401e-01 5.37538886e-01 4.65808123e-01 -4.40430105e-01
8.37945879e-01 -2.79558331e-01 1.65956050e-01 -1.22596717e+00
1.62130088e-01 4.39015865e-01 -2.87814476e-02 -2.20202073e-01
8.10967922e-01 1.76270112e-01 -5.57983518e-01 -1.93480641e-01
-9.31505784e-02 -7.13716626e-01 -9.41015258e-02 9.39574763e-02
1.51354384e+00 3.18243504e-01 3.91471714e-01 -3.84969682e-01
-1.13250270e-01 7.44028628e-01 3.57365787e-01 1.36313760e+00
-4.57029194e-01 4.44252074e-01 1.00263124e-02 -2.84894824e-01
-1.11984718e+00 -1.39804971e+00 3.26937199e-01 8.69974732e-01
1.01242459e+00 -2.98512191e-01 -5.90915799e-01 4.34347019e-02
4.17815149e-01 7.12056220e-01 -5.62527776e-01 -3.39649945e-01
-5.34776390e-01 -5.53220451e-01 6.23527944e-01 3.87428433e-01
5.10168970e-01 -4.25296932e-01 -1.41468346e+00 2.56898671e-01
-1.30282581e-01 -1.16848218e+00 2.32718080e-01 3.62559408e-01
-4.48863924e-01 -9.69742894e-01 -8.04424584e-01 -5.49998045e-01
5.24613380e-01 7.44094968e-01 9.08841670e-01 -2.33585939e-01
-8.26240778e-02 6.94100916e-01 -3.35733205e-01 -3.24813426e-01
1.78768247e-01 3.17494534e-02 5.42076528e-01 -8.42966616e-01
-8.68260339e-02 -5.15901029e-01 -6.91516578e-01 2.02321336e-01
-1.87484905e-01 -8.70440826e-02 3.90118897e-01 5.90579987e-01
4.62913722e-01 -4.75557476e-01 -1.21871896e-01 7.77478740e-02
5.15593410e-01 -5.54778993e-01 -9.27717626e-01 -3.12935889e-01
-5.18284857e-01 1.54471532e-01 5.23085058e-01 -1.86441377e-01
-4.99941498e-01 2.18129873e-01 4.19175550e-02 5.07186539e-02
-1.52630657e-01 2.30970457e-01 1.14949815e-01 -4.79171485e-01
1.26827264e+00 1.86511710e-01 8.78903121e-02 -1.19684696e-01
4.86143559e-01 4.08940703e-01 8.76060188e-01 -5.07564902e-01
6.84937656e-01 1.08105958e+00 3.41939598e-01 -8.94660413e-01
-9.68768746e-02 -3.20345193e-01 -6.03293516e-02 -1.70680191e-02
7.37997353e-01 -1.15494692e+00 -1.31083167e+00 4.58016098e-01
-7.17297137e-01 -9.54614639e-01 -1.59689635e-01 4.63694960e-01
-1.10872960e+00 1.76323175e-01 -4.64354083e-02 -1.09410834e+00
-1.82926536e-01 -1.16870844e+00 1.16918814e+00 4.89278108e-01
-2.20480546e-01 -3.59429091e-01 2.95732170e-01 -2.46197492e-01
5.53063452e-01 2.27593794e-01 -2.78495960e-02 -1.88716307e-01
-7.36940265e-01 -1.59945086e-01 -1.04222201e-01 -6.66140974e-01
-1.81523949e-01 -7.85646975e-01 -6.00793540e-01 -6.36880577e-01
-3.75643969e-01 -2.20864818e-01 6.52360380e-01 2.57666767e-01
-4.14951704e-02 -2.56881714e-02 -6.99660301e-01 8.62896085e-01
1.44598985e+00 5.75397670e-01 3.15310329e-01 9.81414676e-01
5.28846741e-01 4.34286535e-01 8.57132614e-01 5.43207943e-01
1.17113471e+00 1.02926636e+00 1.28956854e+00 3.92476410e-01
1.83060914e-01 -3.74504089e-01 2.62125373e-01 -1.95375130e-01
-1.17062673e-01 -5.91695130e-01 -1.24264157e+00 7.45205700e-01
-2.10066676e+00 -3.94615084e-01 -3.60235153e-03 2.39848447e+00
1.75970927e-01 3.06821585e-01 -1.51709512e-01 -3.09013218e-01
1.44312486e-01 -5.02942242e-02 -8.61619592e-01 1.55714508e-02
-1.76471903e-03 -3.42223436e-01 1.28283978e+00 9.43105817e-01
-1.06289780e+00 1.05555379e+00 5.75045729e+00 2.52444446e-02
-9.62618232e-01 -1.08200558e-01 -4.75727111e-01 -2.21153423e-01
9.03075561e-02 9.50316712e-02 -9.30135429e-01 2.40863487e-01
8.65421355e-01 5.79418056e-02 7.65157044e-01 1.28127706e+00
5.04157618e-02 -6.69701695e-01 -7.68800199e-01 1.11224627e+00
8.87200311e-02 -6.98022306e-01 -5.95774949e-01 1.57592565e-01
5.37031472e-01 6.34448588e-01 1.00460328e-01 4.27855551e-01
1.08684409e+00 -1.01061773e+00 9.74860072e-01 4.40913051e-01
4.78205353e-01 -5.33484817e-01 7.38486648e-01 6.76326871e-01
-9.55795765e-01 -2.46088430e-01 -4.09397036e-01 -5.50634384e-01
4.41037238e-01 -1.45781830e-01 -1.12405312e+00 5.00588238e-01
9.70508695e-01 3.18555027e-01 -1.95554212e-01 1.41829848e+00
-5.32250762e-01 -1.91579863e-01 -1.06191289e+00 -4.57833946e-01
5.32755554e-01 6.67513534e-02 8.41550708e-01 5.74445188e-01
6.85451210e-01 -1.35776103e-02 2.89643854e-01 3.74389380e-01
4.76887167e-01 -4.79669869e-01 -9.20077980e-01 5.87796211e-01
6.41059756e-01 6.08748376e-01 -3.36881489e-01 -8.11355785e-02
2.34806463e-01 1.09519219e+00 3.52181703e-01 4.73252237e-01
-6.63557708e-01 -6.42622650e-01 1.29914355e+00 4.67670150e-02
5.54881133e-02 -1.11055291e+00 -2.97465265e-01 -9.41443264e-01
5.67837171e-02 -4.75741446e-01 -5.44018708e-02 -1.07453597e+00
-5.30430734e-01 4.72548485e-01 -2.62120157e-01 -1.16027331e+00
-6.02375805e-01 -6.43594861e-01 -5.02356216e-02 7.82171309e-01
-1.37305427e+00 -1.05104184e+00 -9.26677048e-01 3.02213728e-01
1.17413737e-01 1.21561304e-01 9.19354081e-01 -1.50578052e-01
1.89807996e-01 1.59107208e-01 2.13460580e-01 -3.83581966e-01
5.95080376e-01 -1.50235093e+00 1.07148004e+00 8.16321373e-01
-3.85797292e-01 7.02848852e-01 1.44649398e+00 -8.30423832e-01
-1.92807043e+00 -6.49803042e-01 5.30160725e-01 -6.01530731e-01
3.62476766e-01 -6.56193316e-01 -2.33178824e-01 7.92750359e-01
-1.96648821e-01 -1.91101313e-01 -1.47087663e-01 -1.01514645e-01
-8.48253593e-02 9.55756903e-02 -1.27056527e+00 1.31817901e+00
1.60443032e+00 -2.42671948e-02 -5.69891095e-01 9.31775719e-02
8.36097836e-01 -9.91337180e-01 -3.46299201e-01 4.10217613e-01
8.71019304e-01 -7.05611229e-01 1.22880650e+00 -2.02648044e-01
-4.75838125e-01 -8.21093619e-01 -6.76635861e-01 -1.57020354e+00
-2.52806127e-01 -6.84907079e-01 8.11133999e-03 2.58818209e-01
1.04231045e-01 -8.89737129e-01 7.23343551e-01 3.74778211e-01
-1.68140098e-01 -1.47516876e-02 -1.27354336e+00 -9.35680807e-01
-3.28087807e-01 -4.03268874e-01 3.51984710e-01 2.54646301e-01
3.02308500e-02 1.18360586e-01 -5.87062299e-01 7.56423533e-01
8.04345131e-01 -3.07706594e-01 1.35355973e+00 -1.02376974e+00
-3.03088333e-02 -2.18113571e-01 -8.37767720e-01 -1.60568941e+00
-1.28795534e-01 -2.71878183e-01 7.58076549e-01 -1.97951889e+00
-6.03826523e-01 -5.98528206e-01 3.55110228e-01 2.96872526e-01
1.59115896e-01 1.77598506e-01 3.07507932e-01 1.22780107e-01
-8.54121506e-01 6.85342669e-01 8.17807198e-01 -1.55365601e-01
-3.29227984e-01 -1.13120638e-01 -6.60727322e-01 4.55788195e-01
8.51358533e-01 -2.38830283e-01 -4.75392878e-01 -8.00885141e-01
3.36615980e-01 4.89774756e-02 6.69292450e-01 -1.55850959e+00
7.04664528e-01 2.05176827e-02 2.37656668e-01 -2.00534672e-01
1.06916559e+00 -8.11540186e-01 9.14037302e-02 9.77414787e-01
1.28671005e-01 2.32351199e-01 2.07724601e-01 9.13020968e-01
4.76122111e-01 5.12939095e-02 3.18398267e-01 -2.72184193e-01
-1.21452403e+00 -1.69253230e-01 -7.19164312e-01 1.00514488e-02
1.11695278e+00 -3.45959663e-01 -7.84279287e-01 -7.95651972e-01
-7.41014183e-01 7.14248002e-01 1.09617031e+00 4.60271865e-01
4.77023065e-01 -9.72119689e-01 -2.13898763e-01 1.79757938e-01
3.26310188e-01 1.70347914e-01 2.49298334e-01 6.32019877e-01
-8.50524366e-01 5.13931692e-01 -4.57366526e-01 -9.40023959e-01
-8.09790730e-01 2.45849371e-01 5.40287912e-01 3.69140297e-01
-4.36161518e-01 7.43860900e-01 1.01792388e-01 -8.59885097e-01
2.42045432e-01 -4.54654962e-01 9.92150828e-02 -6.93409443e-01
2.69849747e-01 2.34329209e-01 -1.24711573e-01 -5.54412544e-01
-7.13872135e-01 8.34871173e-01 4.12859708e-01 -7.06741989e-01
9.58843887e-01 -4.87870187e-01 6.25982225e-01 2.47128174e-01
3.93669397e-01 -3.73435020e-02 -1.66720498e+00 1.89506933e-01
-1.11747429e-01 -2.89246231e-01 -2.47555733e-01 -8.55746686e-01
-1.41061824e-02 5.22793710e-01 7.76751816e-01 -4.89781499e-02
4.29761320e-01 -2.42270499e-01 3.32728595e-01 1.08995450e+00
1.30462301e+00 -7.42717624e-01 -1.84804514e-01 1.07624030e+00
7.30602741e-01 -1.25106740e+00 -1.67340994e-01 -4.15633395e-02
-5.59387982e-01 3.80537212e-01 7.36777186e-01 -5.21640956e-01
-3.32871638e-02 3.67442548e-01 2.35975638e-01 1.16587460e-01
-6.69399381e-01 -5.18743634e-01 -3.66619736e-01 1.33920097e+00
-3.70467871e-01 2.52107948e-01 1.67534485e-01 2.78305650e-01
-5.56801677e-01 -3.88971031e-01 5.32756627e-01 1.43182099e+00
-1.02238905e+00 -4.15242851e-01 -3.18804473e-01 -1.60848141e-01
1.33567184e-01 6.24988973e-02 -3.13407242e-01 1.06067574e+00
-2.92044640e-01 1.01079392e+00 1.21448494e-01 -4.52376097e-01
7.47688353e-01 -3.51270735e-01 6.00688577e-01 -3.72925013e-01
-2.13716239e-01 -2.88353413e-01 3.05166245e-01 -1.01440632e+00
1.51986316e-01 -4.26055938e-01 -1.61686218e+00 -3.33589941e-01
9.40511376e-02 -2.67803483e-02 1.17642140e+00 6.42747045e-01
7.77124286e-01 3.70656639e-01 -3.09833795e-01 -1.26821613e+00
-7.83319294e-01 -7.35126376e-01 -2.49864861e-01 -4.68744114e-02
6.68101907e-01 -9.72915351e-01 -2.90970325e-01 -5.18845201e-01] | [4.679556369781494, 0.6644973158836365] |
932a39d9-0f2d-432b-a83c-f53e1a4911ad | griddehazenet-an-enhanced-multi-scale-network | 2103.13998 | null | https://arxiv.org/abs/2103.13998v2 | https://arxiv.org/pdf/2103.13998v2.pdf | GridDehazeNet+: An Enhanced Multi-Scale Network with Intra-Task Knowledge Transfer for Single Image Dehazing | We propose an enhanced multi-scale network, dubbed GridDehazeNet+, for single image dehazing. The proposed dehazing method does not rely on the Atmosphere Scattering Model (ASM), and an explanation as to why it is not necessarily performing the dimension reduction offered by this model is provided. GridDehazeNet+ consists of three modules: pre-processing, backbone, and post-processing. The trainable pre-processing module can generate learned inputs with better diversity and more pertinent features as compared to those derived inputs produced by hand-selected pre-processing methods. The backbone module implements multi-scale estimation with two major enhancements: 1) a novel grid structure that effectively alleviates the bottleneck issue via dense connections across different scales; 2) a spatial-channel attention block that can facilitate adaptive fusion by consolidating dehazing-relevant features. The post-processing module helps to reduce the artifacts in the final output. Due to domain shift, the model trained on synthetic data may not generalize well on real data. To address this issue, we shape the distribution of synthetic data to match that of real data, and use the resulting translated data to finetune our network. We also propose a novel intra-task knowledge transfer mechanism that can memorize and take advantage of synthetic domain knowledge to assist the learning process on the translated data. Experimental results demonstrate that the proposed method outperforms the state-of-the-art on several synthetic dehazing datasets, and achieves the superior performance on real-world hazy images after finetuning. | ['Jun Chen', 'Zijun Wu', 'Zhihao Shi', 'Xiaohong Liu'] | 2021-03-25 | null | null | null | null | ['image-dehazing'] | ['computer-vision'] | [ 3.53626072e-01 -2.13835254e-01 6.23302341e-01 -3.80826622e-01
-7.65231371e-01 -1.52293682e-01 3.83709371e-01 -1.95784450e-01
-5.23810029e-01 8.13967645e-01 1.42676011e-01 1.59158930e-02
-1.56150147e-01 -1.14170039e+00 -9.16664362e-01 -1.06276321e+00
1.50931418e-01 8.53034258e-02 5.70419908e-01 -4.44039345e-01
1.24868296e-01 3.67864490e-01 -1.75236619e+00 6.45417213e-01
1.39179432e+00 1.24101877e+00 5.21725655e-01 6.71066940e-01
-1.46337273e-02 6.97995245e-01 -8.75688732e-01 -1.62734404e-01
4.68247026e-01 -4.05215383e-01 -3.26186955e-01 1.01951405e-01
7.98952758e-01 -4.03154373e-01 -1.94354549e-01 1.13495350e+00
6.21243894e-01 3.12736481e-01 5.98240137e-01 -9.94839549e-01
-1.08482313e+00 1.26414910e-01 -6.11923397e-01 3.13762099e-01
-3.08886230e-01 1.04842938e-01 4.61919427e-01 -1.22271740e+00
2.99188942e-01 1.23487496e+00 5.92402339e-01 4.22186881e-01
-1.03770614e+00 -1.11225641e+00 5.25861941e-02 2.71790713e-01
-1.59076834e+00 -2.96429694e-01 7.49840379e-01 -2.21895903e-01
7.36004829e-01 -9.81930643e-02 5.62079072e-01 9.53119516e-01
5.13558149e-01 5.35925508e-01 1.35263062e+00 -4.01837051e-01
2.31481060e-01 2.95219898e-01 -1.28583372e-01 6.82981074e-01
2.57702857e-01 2.11989895e-01 -6.04269922e-01 1.25051364e-01
8.38450372e-01 -1.49787009e-01 -4.43444490e-01 -2.79761493e-01
-1.16617978e+00 8.59013855e-01 7.81416714e-01 5.57869971e-02
-5.38660705e-01 -7.05922097e-02 2.93301754e-02 5.93580306e-01
8.11517656e-01 6.46779180e-01 -3.67204547e-01 5.50794065e-01
-1.09267724e+00 2.20436245e-01 2.98228651e-01 8.51323903e-01
1.17180169e+00 3.55392128e-01 -2.59010613e-01 9.33195114e-01
8.70010629e-03 8.49327326e-01 5.13250530e-01 -7.24041998e-01
5.28541148e-01 3.81128162e-01 1.41673446e-01 -7.53331184e-01
-2.95113802e-01 -5.67777932e-01 -1.14126301e+00 5.13001859e-01
1.33872643e-01 -2.96396017e-01 -1.46665382e+00 1.53733623e+00
2.56680071e-01 4.50298965e-01 2.92571336e-01 9.44036722e-01
9.24602985e-01 1.17849803e+00 -5.50473481e-02 1.46823768e-02
1.06728530e+00 -1.33661723e+00 -7.17887104e-01 -2.74996549e-01
2.29584560e-01 -8.24576139e-01 1.00023401e+00 4.00681496e-01
-8.29266131e-01 -9.10062850e-01 -1.37494993e+00 9.66183022e-02
-7.68550634e-01 9.38004106e-02 2.40209952e-01 2.70818621e-01
-1.27766442e+00 4.92452949e-01 -5.00269473e-01 -1.69145241e-01
4.32529092e-01 2.24944860e-01 -2.93200821e-01 -1.86383635e-01
-1.49079657e+00 9.24062312e-01 6.01162493e-01 9.39259455e-02
-9.36225355e-01 -9.62135732e-01 -9.10800397e-01 3.98041271e-02
2.47110084e-01 -8.94974589e-01 8.43026042e-01 -1.20404375e+00
-1.52209949e+00 4.22352880e-01 1.94165334e-02 -4.56195652e-01
2.44582221e-01 -4.40228581e-01 -7.30937958e-01 2.82097906e-01
8.50713998e-02 8.67511690e-01 1.66785324e+00 -1.44558883e+00
-8.33285749e-01 -1.19125687e-01 -9.72425193e-02 5.01565397e-01
-4.55208510e-01 -3.27253133e-01 -4.54960436e-01 -1.12232280e+00
-2.42030814e-01 -7.36791134e-01 -1.10453457e-01 8.71217847e-02
-7.70667717e-02 1.47687137e-01 8.93631816e-01 -7.62265146e-01
1.22048891e+00 -2.29475331e+00 1.20839318e-02 2.56890148e-01
1.75047845e-01 6.74998820e-01 -4.88415211e-01 3.72449726e-01
-6.37043864e-02 -2.55685031e-01 -4.19137776e-01 -2.02105135e-01
-4.24095333e-01 1.15467608e-01 -5.73324203e-01 2.48797655e-01
5.24743736e-01 7.40361989e-01 -7.36815333e-01 -3.45687866e-01
3.36924076e-01 6.42956734e-01 -5.72912395e-01 4.31373537e-01
-1.29947662e-01 4.43452984e-01 -1.96751773e-01 3.96089613e-01
1.12454343e+00 -3.39542985e-01 -5.24087489e-01 -3.71082723e-01
-2.34855190e-01 4.90521034e-03 -1.19060147e+00 1.61924541e+00
-7.94729114e-01 4.07120615e-01 1.29578337e-01 -7.92412817e-01
8.86155009e-01 1.61647767e-01 5.56754181e-03 -7.59126782e-01
-1.40895963e-01 1.20499939e-01 -1.38469294e-01 -4.60974842e-01
5.28848886e-01 -1.20176822e-01 2.06642821e-01 2.23129317e-01
1.38033375e-01 -4.79440272e-01 -1.23511262e-01 6.15676194e-02
7.21472979e-01 1.27426654e-01 2.37981021e-01 -4.38826084e-01
7.27129877e-01 8.83277431e-02 5.19843638e-01 7.79358089e-01
2.63930019e-02 7.72386968e-01 -5.29619493e-02 -5.95720708e-01
-1.02385974e+00 -1.15952706e+00 -1.46460161e-01 1.08804858e+00
4.54980612e-01 -3.36657241e-02 -7.89686263e-01 -6.19879067e-01
9.10601299e-03 7.86264420e-01 -1.01950157e+00 -4.20573980e-01
-4.07994449e-01 -9.36492920e-01 2.29277492e-01 2.80447960e-01
1.19821012e+00 -1.08686316e+00 -4.56699550e-01 1.87847391e-01
-1.99131802e-01 -1.03028250e+00 -4.84550148e-01 1.30167618e-01
-5.66245615e-01 -6.20363653e-01 -7.81638086e-01 -8.05016458e-01
6.88679934e-01 7.35159338e-01 9.13940132e-01 -7.19075799e-02
5.11000454e-02 -4.90542240e-02 -4.87819403e-01 -7.33397782e-01
-5.00836134e-01 4.93567251e-02 -7.54193068e-02 4.43231881e-01
1.98017329e-01 -7.15855896e-01 -8.45586419e-01 4.05212402e-01
-1.25387335e+00 3.43604296e-01 9.39011455e-01 1.17080736e+00
4.96455967e-01 5.14501214e-01 6.31685674e-01 -8.48604202e-01
5.70573092e-01 -3.54032159e-01 -5.79666615e-01 3.40824306e-01
-6.50782049e-01 2.00057343e-01 9.28340673e-01 -4.38866466e-01
-1.52026784e+00 -2.39545375e-01 -7.23812133e-02 -6.37424886e-01
-1.74358919e-01 2.80280650e-01 5.05320281e-02 -5.11984468e-01
8.59971941e-01 6.00090802e-01 7.86131993e-02 -3.63323957e-01
2.58517146e-01 7.34271288e-01 5.36458611e-01 -3.80914807e-01
1.30784535e+00 7.35802889e-01 -1.61964819e-01 -1.01797950e+00
-1.08697629e+00 -3.41842085e-01 -6.35230184e-01 -9.06892400e-03
9.98588026e-01 -1.42759168e+00 4.57372479e-02 7.17864335e-01
-1.04010963e+00 -5.10621309e-01 -3.25669259e-01 4.58379120e-01
-1.90530822e-01 1.59047589e-01 -4.04933721e-01 -4.28388923e-01
-5.01842737e-01 -8.85214567e-01 1.23066187e+00 3.13760370e-01
4.79114085e-01 -8.22897553e-01 1.07149594e-02 2.13962585e-01
8.69017184e-01 -2.93093529e-02 9.57232893e-01 -2.59899110e-01
-7.62412190e-01 2.03384802e-01 -5.74854076e-01 7.79604971e-01
1.88472047e-01 -2.52331018e-01 -1.16728866e+00 -5.46147346e-01
6.02329895e-03 -3.97595555e-01 1.31904280e+00 3.55241746e-01
1.26569712e+00 -2.03513190e-01 -2.41267458e-02 1.09507143e+00
1.50146198e+00 2.25774888e-02 7.17384577e-01 5.25663257e-01
6.23744249e-01 6.01469755e-01 7.05620885e-01 1.29620478e-01
3.71692896e-01 5.07652938e-01 5.31961977e-01 -5.85882008e-01
-5.40999532e-01 -6.45210594e-02 3.37851465e-01 7.98990786e-01
-1.64352626e-01 -2.41062954e-01 -6.18888617e-01 5.91899395e-01
-1.71917665e+00 -8.21047127e-01 3.50493968e-01 2.05727673e+00
7.21011341e-01 7.79586332e-03 -3.23916405e-01 -3.34312141e-01
5.17319977e-01 3.90697449e-01 -5.35315573e-01 -2.17339948e-01
-2.89573491e-01 4.96288657e-01 5.73499262e-01 4.02835041e-01
-1.28217244e+00 1.18682420e+00 5.62400579e+00 1.06228745e+00
-1.33314919e+00 8.76369253e-02 3.15460891e-01 7.69829676e-02
-1.27146140e-01 -4.82153594e-01 -7.61732638e-01 5.07949352e-01
7.78469980e-01 -3.38717885e-02 5.55777669e-01 6.22872353e-01
9.35511589e-02 1.05060348e-02 -5.85722506e-01 8.28050613e-01
2.95697421e-01 -1.38691151e+00 4.42771763e-01 -1.98015332e-01
1.16969764e+00 1.70513466e-01 3.69530946e-01 3.77905250e-01
2.67213374e-01 -7.84921348e-01 7.46408284e-01 4.60942537e-01
9.21772003e-01 -7.19534874e-01 9.02614176e-01 2.11387992e-01
-1.21005058e+00 -1.72653049e-01 -7.03443110e-01 1.90556750e-01
-1.71294704e-01 6.07566774e-01 -8.66316497e-01 8.20958853e-01
9.06880319e-01 5.77565253e-01 -7.44436443e-01 9.39132988e-01
-3.59355927e-01 5.39908171e-01 -1.47064358e-01 3.42346430e-01
3.33232313e-01 -1.97935134e-01 3.30376446e-01 1.30976057e+00
6.51740730e-01 6.81457967e-02 -7.67351165e-02 5.88829875e-01
8.53426084e-02 -7.56639913e-02 -5.37390888e-01 4.18920994e-01
4.14587140e-01 1.18237209e+00 -3.62748235e-01 -5.42567730e-01
-7.37663507e-01 1.18998480e+00 3.72745663e-01 6.74942791e-01
-9.55413222e-01 -6.70480609e-01 7.65241444e-01 -2.25006044e-02
7.19852865e-01 -1.00325011e-01 -1.46509543e-01 -1.15281343e+00
-1.29968092e-01 -9.43683922e-01 2.81630844e-01 -1.08755660e+00
-1.41826189e+00 1.01541400e+00 6.23452291e-03 -1.36797404e+00
1.38837412e-01 -6.01451397e-01 -5.48020065e-01 1.16811407e+00
-2.17253709e+00 -1.39883959e+00 -8.02264333e-01 8.06548119e-01
7.03043103e-01 -2.12166548e-01 6.85919344e-01 3.12457621e-01
-4.85144883e-01 5.89639604e-01 2.64257938e-01 -1.24265827e-01
1.14295590e+00 -1.20498908e+00 4.47139621e-01 1.12955117e+00
-1.71235770e-01 4.47181374e-01 5.43356240e-01 -5.80674767e-01
-8.34505498e-01 -1.80659497e+00 4.69911814e-01 -2.92334348e-01
5.53522468e-01 -3.15827608e-01 -1.25256383e+00 4.03191805e-01
3.79712075e-01 2.24983647e-01 4.53313589e-01 -4.27072644e-01
-4.66257334e-01 -5.50232053e-01 -1.01155853e+00 4.88819391e-01
9.09907103e-01 -3.76820773e-01 -7.42900431e-01 1.78664565e-01
9.28366005e-01 -4.72831488e-01 -6.50228620e-01 6.24191642e-01
3.46318185e-01 -1.12389719e+00 1.11810744e+00 -2.25158885e-01
4.78741825e-01 -5.28347731e-01 -1.47068366e-01 -1.79594469e+00
-6.69895649e-01 -5.51733792e-01 -1.37941148e-02 8.48587871e-01
4.79504287e-01 -7.38995790e-01 5.75627387e-01 4.00880314e-02
-3.55682373e-01 -7.16335654e-01 -6.62510514e-01 -8.28235745e-01
1.96815833e-01 -8.41570944e-02 9.04409230e-01 9.52194870e-01
-8.18740070e-01 2.63839394e-01 -6.87959909e-01 7.85843730e-01
8.21861327e-01 3.73381376e-01 8.21239233e-01 -1.00249803e+00
-1.57146960e-01 2.66849585e-02 -8.52541178e-02 -1.05734527e+00
-5.45090362e-02 -5.22915363e-01 1.96639061e-01 -1.45525801e+00
-1.07135892e-01 -3.59368712e-01 -5.47669590e-01 4.32587534e-01
-4.53608304e-01 4.92984265e-01 1.02528602e-01 2.73640603e-01
-4.46018636e-01 8.91920924e-01 1.69758379e+00 -1.39843851e-01
-2.92583227e-01 -9.64486450e-02 -7.44773626e-01 5.41554868e-01
7.81825542e-01 -3.84943694e-01 -6.06949985e-01 -5.69022715e-01
-3.63161266e-02 -2.12936595e-01 4.03228670e-01 -1.37245011e+00
1.51677862e-01 -1.63859740e-01 6.88240170e-01 -4.50881392e-01
5.09707570e-01 -8.93398345e-01 -2.06077956e-02 2.20632419e-01
-9.08205435e-02 -8.43400657e-02 2.79878885e-01 6.88969135e-01
-5.08847415e-01 1.72920227e-01 1.15358102e+00 -4.11028787e-02
-1.04137301e+00 3.57391566e-01 -3.32584471e-01 -1.13663599e-01
9.27798688e-01 -3.18224072e-01 -6.49178445e-01 -3.75121355e-01
-4.40023065e-01 2.22763151e-01 4.23746645e-01 4.65269834e-01
7.35607684e-01 -1.23967934e+00 -8.70908737e-01 7.51882493e-01
2.93036610e-01 2.25690544e-01 4.87139344e-01 6.06634259e-01
-6.21825218e-01 2.47785702e-01 -3.87127012e-01 -3.43992352e-01
-9.62971985e-01 6.54409945e-01 2.64757454e-01 -2.46355191e-01
-6.47458315e-01 1.03965199e+00 9.10589635e-01 -3.33640397e-01
-5.11487983e-02 -3.50534111e-01 -1.80657297e-01 -1.34129211e-01
7.37090230e-01 1.80176660e-01 2.76415765e-01 -5.44693947e-01
-1.48621127e-01 7.58305848e-01 -2.23113343e-01 5.45303011e-03
1.44915056e+00 -2.78699726e-01 -5.47308847e-02 1.08573772e-01
1.02076805e+00 -1.27463778e-02 -1.67752326e+00 -6.64397955e-01
-6.02622390e-01 -5.71599722e-01 2.84319460e-01 -9.76030052e-01
-1.06688976e+00 9.46950734e-01 6.70971394e-01 8.85155499e-02
1.69152141e+00 -3.45210969e-01 8.76589775e-01 5.05519927e-01
2.69487917e-01 -1.01592052e+00 2.99835831e-01 6.54029906e-01
1.10497010e+00 -1.28594720e+00 -5.73054515e-02 -4.78696287e-01
-7.26080596e-01 9.36141491e-01 9.88628268e-01 -1.65341705e-01
8.45782399e-01 9.16543007e-02 4.46609199e-01 -2.05602750e-01
-6.90677464e-01 -1.50632918e-01 5.02450168e-01 7.28551447e-01
-1.52571499e-01 -2.41512254e-01 1.64314732e-01 2.91018218e-01
-1.61572605e-01 -9.34341624e-02 3.85818571e-01 6.62984908e-01
-7.58329153e-01 -6.90500975e-01 -6.37507498e-01 4.49930698e-01
-8.76046270e-02 -3.81776482e-01 -6.12115860e-02 8.60927343e-01
4.58361924e-01 8.11544716e-01 8.76721814e-02 -5.81886649e-01
3.88010800e-01 -9.07981098e-02 3.50147188e-01 -6.65440023e-01
-4.05049413e-01 -6.42043538e-03 -1.83625445e-01 -5.24652898e-01
-5.13323307e-01 -1.87068626e-01 -8.60353649e-01 -9.76833701e-02
-4.36984807e-01 3.30208093e-01 4.14334506e-01 8.04067194e-01
6.71470761e-01 9.30034220e-01 8.07423711e-01 -1.16039026e+00
-4.59202170e-01 -1.04151654e+00 -6.42641306e-01 3.75515044e-01
7.83066809e-01 -7.49361336e-01 -3.79909545e-01 1.05146818e-01] | [10.937148094177246, -3.0611257553100586] |
005eb98b-ed3d-4bb2-9edd-ea35f74c6d3d | robustification-of-multilingual-language | 2210.04782 | null | https://arxiv.org/abs/2210.04782v2 | https://arxiv.org/pdf/2210.04782v2.pdf | Robustification of Multilingual Language Models to Real-world Noise in Crosslingual Zero-shot Settings with Robust Contrastive Pretraining | Advances in neural modeling have achieved state-of-the-art (SOTA) results on public natural language processing (NLP) benchmarks, at times surpassing human performance. However, there is a gap between public benchmarks and real-world applications where noise, such as typographical or grammatical mistakes, is abundant and can result in degraded performance. Unfortunately, works which evaluate the robustness of neural models on noisy data and propose improvements, are limited to the English language. Upon analyzing noise in different languages, we observe that noise types vary greatly across languages. Thus, existing investigations do not generalize trivially to multilingual settings. To benchmark the performance of pretrained multilingual language models, we construct noisy datasets covering five languages and four NLP tasks and observe a clear gap in the performance between clean and noisy data in the zero-shot cross-lingual setting. After investigating several ways to boost the robustness of multilingual models in this setting, we propose Robust Contrastive Pretraining (RCP). RCP combines data augmentation with a contrastive loss term at the pretraining stage and achieves large improvements on noisy (and original test data) across two sentence-level (+3.2%) and two sequence-labeling (+10 F1-score) multilingual classification tasks. | ['He He', 'Saab Mansour', 'Jason Krone', 'Sailik Sengupta', 'Asa Cooper Stickland'] | 2022-10-10 | null | null | null | null | ['pretrained-multilingual-language-models'] | ['natural-language-processing'] | [ 9.20017287e-02 -3.59656751e-01 2.62393001e-02 -4.96287614e-01
-1.55435264e+00 -8.45823109e-01 6.16904199e-01 3.33575577e-01
-1.06838262e+00 8.71652782e-01 4.44677353e-01 -4.86099333e-01
3.79768163e-01 -4.79398310e-01 -1.09672880e+00 -4.33111638e-01
1.68844715e-01 3.65950763e-01 -1.41896009e-01 -4.54823941e-01
-4.08284336e-01 1.06196262e-01 -1.15056837e+00 5.83032131e-01
9.18949127e-01 5.79559326e-01 3.39155272e-02 5.08086741e-01
-2.30122358e-01 6.57845855e-01 -7.20781744e-01 -7.43465662e-01
3.39747399e-01 -2.03563914e-01 -7.00753152e-01 -3.11633974e-01
8.12715232e-01 1.94830909e-01 -1.90341666e-01 1.27898908e+00
8.60272944e-01 -6.28547231e-03 2.71270096e-01 -7.73148000e-01
-9.93293941e-01 1.19079316e+00 -3.05792421e-01 4.36746687e-01
1.89003199e-01 4.26999897e-01 1.08556843e+00 -1.15169787e+00
8.26290309e-01 1.54987168e+00 1.01674187e+00 5.41963160e-01
-1.50584865e+00 -5.67726433e-01 3.00766110e-01 1.16907172e-01
-1.10035050e+00 -6.75536990e-01 3.46306235e-01 -2.46287614e-01
1.38927174e+00 7.89949950e-03 -9.13697630e-02 1.85726368e+00
1.92881584e-01 9.87498760e-01 1.29998946e+00 -5.85196793e-01
-6.22204542e-02 1.14285029e-01 3.84302527e-01 3.49778324e-01
1.80830881e-01 1.22773968e-01 -5.75503409e-01 8.62451047e-02
1.50005976e-02 -6.29338741e-01 -2.45614469e-01 1.89329013e-01
-1.40904903e+00 7.13979781e-01 2.78820455e-01 5.35557568e-01
-3.13021421e-01 -4.47345804e-03 9.34581459e-01 7.42138505e-01
7.29795158e-01 6.35196447e-01 -9.39040959e-01 -3.66430581e-01
-7.72267759e-01 1.44533932e-01 7.92858601e-01 7.59701908e-01
4.60102588e-01 4.17275846e-01 -3.24804127e-01 1.27347207e+00
-2.37767145e-01 7.24935353e-01 6.81229651e-01 -4.98997957e-01
9.87715423e-01 8.12101066e-02 -3.66619259e-01 -6.89956248e-01
-3.91897082e-01 -7.05693066e-01 -1.01870394e+00 -1.78991675e-01
5.70260644e-01 -4.80643213e-01 -9.79797304e-01 2.12104940e+00
-1.19911723e-01 -2.08701327e-01 4.08873975e-01 5.34409702e-01
9.49355006e-01 7.30606616e-01 3.07380736e-01 -1.11988083e-01
1.17361271e+00 -1.08671343e+00 -7.41919935e-01 -5.93974411e-01
1.04073799e+00 -9.31905329e-01 1.71127617e+00 6.31520033e-01
-1.06874716e+00 -5.62910199e-01 -9.18076932e-01 -4.63494837e-01
-6.67309046e-01 1.20909743e-01 2.00292990e-01 5.92918336e-01
-1.00341475e+00 5.09270906e-01 -6.62095845e-01 -3.32941264e-01
1.11282825e-01 -9.37205851e-02 -6.35253549e-01 -5.83413541e-01
-1.56347096e+00 1.22452879e+00 4.25054431e-01 8.59538540e-02
-9.32018161e-01 -1.02354765e+00 -1.06368673e+00 -1.49232015e-01
5.23821235e-01 -3.87052059e-01 1.30108440e+00 -9.05810297e-01
-1.11134005e+00 9.24799263e-01 -3.04078944e-02 -6.95104957e-01
6.83030725e-01 -5.20325243e-01 -5.95539033e-01 -5.23116410e-01
1.95990279e-01 6.14423394e-01 3.01787496e-01 -1.16787291e+00
-4.91224498e-01 -1.33440793e-01 -2.20555276e-01 1.82986230e-01
-2.77531803e-01 2.60086238e-01 -3.69980395e-01 -9.29075539e-01
-2.56480634e-01 -7.06740081e-01 -3.23235691e-01 -5.89915156e-01
-5.13613045e-01 -3.55704188e-01 4.03406978e-01 -1.02334702e+00
1.08411109e+00 -2.19089818e+00 -2.07824241e-02 -1.70243308e-01
-2.90087074e-01 4.63162482e-01 -8.31594646e-01 3.08302432e-01
-2.19363153e-01 4.80892420e-01 -4.54705298e-01 -7.30719984e-01
9.15697142e-02 4.03478116e-01 -2.48436034e-01 2.86390990e-01
4.83232915e-01 1.03131318e+00 -8.64511609e-01 -1.91893019e-02
-1.95674822e-01 6.15586996e-01 -4.04143959e-01 -2.06674397e-01
-2.03194335e-01 3.23552102e-01 2.51573920e-01 5.65725029e-01
5.47900259e-01 1.83423206e-01 5.14210016e-02 -9.82479472e-03
-5.04317973e-03 7.83903956e-01 -9.56881702e-01 1.83909404e+00
-8.32361579e-01 6.54038608e-01 1.59188196e-01 -7.93281376e-01
7.25731492e-01 4.22041059e-01 1.02240685e-02 -9.71351504e-01
4.89803888e-02 4.13626581e-01 2.14760885e-01 -4.49175984e-01
5.94174564e-01 -1.84492439e-01 -3.45976502e-01 2.92299002e-01
4.24489856e-01 -4.11892496e-02 5.43404818e-01 1.01311564e-01
1.19326818e+00 -7.37377629e-02 1.19885176e-01 -4.51421112e-01
2.55280823e-01 -6.75509349e-02 6.38082385e-01 1.12673926e+00
-2.66367167e-01 6.24929607e-01 5.01455724e-01 -3.21736008e-01
-1.21511769e+00 -1.01174521e+00 -1.98910668e-01 1.41917765e+00
-5.97975910e-01 -3.35238934e-01 -8.06341767e-01 -8.39435577e-01
-8.18414763e-02 8.09007287e-01 -4.74669486e-01 -1.43112019e-01
-7.35289931e-01 -1.20877028e+00 1.10551226e+00 5.78754723e-01
2.80068278e-01 -1.30951047e+00 1.82144165e-01 3.48627150e-01
-4.26018655e-01 -1.65266621e+00 -3.84193748e-01 7.04752386e-01
-4.97479677e-01 -6.92128479e-01 -3.83338481e-01 -8.75457108e-01
2.04077765e-01 -1.03182375e-01 1.72950554e+00 -1.33694187e-01
-1.28610715e-01 1.65561050e-01 -3.40490252e-01 -4.56888765e-01
-7.66873360e-01 3.55601937e-01 3.15365195e-01 -3.33406478e-01
6.36193812e-01 -3.21360677e-01 1.66232169e-01 -2.12991104e-01
-9.38888490e-01 -3.89294446e-01 6.32025361e-01 1.10074985e+00
6.60519481e-01 -5.57064340e-02 7.27106452e-01 -9.17930663e-01
9.80140805e-01 -6.06293917e-01 -4.13292766e-01 1.95819452e-01
-4.07207608e-01 1.89848095e-01 8.75228584e-01 -4.65491712e-01
-9.69653666e-01 -1.76927999e-01 -5.44741094e-01 -1.09854132e-01
-5.43462098e-01 9.09097314e-01 -3.19035798e-01 3.71907771e-01
1.06518507e+00 1.95606560e-01 -4.70292121e-01 -8.24085474e-01
5.59630513e-01 5.68681180e-01 8.23765576e-01 -8.44800174e-01
5.47252834e-01 -6.71681687e-02 -5.85947216e-01 -9.09988463e-01
-1.14709508e+00 -2.45514274e-01 -6.53898060e-01 3.84288400e-01
6.28521085e-01 -1.20945835e+00 -1.81622878e-01 5.25764525e-01
-1.35697246e+00 -4.51254487e-01 -3.11851829e-01 3.90057772e-01
-1.20881319e-01 3.06345105e-01 -8.87071967e-01 -4.95013416e-01
-4.19869840e-01 -1.33106935e+00 9.66736138e-01 -4.19658899e-01
-1.60346702e-01 -1.08522236e+00 1.03966750e-01 2.44035035e-01
5.52598417e-01 1.22145765e-01 1.01338756e+00 -9.07442033e-01
-9.17448699e-02 -1.48716196e-01 -1.07226020e-03 9.27579224e-01
-1.57558054e-01 -1.85109153e-01 -1.15792525e+00 -2.76405305e-01
-3.90940644e-02 -7.67142475e-01 1.13787532e+00 2.49835953e-01
8.94277632e-01 -3.00471783e-01 9.80016738e-02 4.69550401e-01
1.39336956e+00 -2.62907535e-01 4.73721504e-01 3.80514652e-01
7.34523773e-01 4.97611612e-01 1.27760366e-01 -1.11268498e-01
2.44476661e-01 4.92543429e-01 7.25641474e-02 -1.10419907e-01
-3.90790790e-01 -2.71464378e-01 7.89308846e-01 1.15984130e+00
4.36036766e-01 -4.86176968e-01 -1.20336568e+00 9.32220459e-01
-1.51061189e+00 -7.35078990e-01 -2.11253196e-01 2.09872293e+00
1.41019177e+00 3.56418192e-01 -8.95525739e-02 1.57450557e-01
5.32585025e-01 -1.04885167e-02 -3.86613756e-01 -6.08160138e-01
-8.10745180e-01 2.76283801e-01 4.18642431e-01 6.76124215e-01
-1.27163267e+00 1.28988218e+00 6.50977659e+00 1.04884386e+00
-1.21768856e+00 4.73673701e-01 8.18741620e-01 -2.39046648e-01
-3.12280774e-01 -4.40619230e-01 -1.01675487e+00 3.71880710e-01
1.34054840e+00 -5.45412265e-02 5.59549749e-01 7.26719260e-01
1.98726863e-01 1.55174524e-01 -1.20187330e+00 8.68163466e-01
1.40507817e-01 -8.90316188e-01 3.09800990e-02 -2.96052366e-01
9.74644005e-01 9.55434978e-01 1.58324286e-01 8.68262410e-01
7.03944981e-01 -1.13850558e+00 7.27892756e-01 9.30246338e-02
5.99384546e-01 -8.59573424e-01 9.51085567e-01 5.13151050e-01
-6.70026839e-01 1.38901219e-01 -4.05337036e-01 7.37065747e-02
1.40763521e-01 9.50937271e-01 -6.52328193e-01 3.93677294e-01
9.48923051e-01 6.90750539e-01 -7.45554745e-01 7.43023157e-01
-3.73311490e-01 9.63207364e-01 -4.45229381e-01 1.85187519e-01
5.99030972e-01 7.61483461e-02 6.03955030e-01 1.81840706e+00
1.53365778e-02 -5.12922823e-01 1.61794409e-01 6.56847358e-01
-6.35263324e-01 4.15535927e-01 -7.35721648e-01 -5.79364710e-02
3.22540611e-01 1.01320171e+00 -1.53183222e-01 -3.51251662e-01
-5.55646658e-01 8.78939450e-01 7.03449368e-01 4.10196006e-01
-5.76706171e-01 -2.95525104e-01 7.17032015e-01 -2.44405746e-01
1.26177281e-01 -2.96723664e-01 -5.45235336e-01 -1.34599149e+00
3.76997471e-01 -1.46403039e+00 3.04725528e-01 -3.52175772e-01
-1.81130862e+00 8.67726862e-01 -3.61717433e-01 -6.68474555e-01
-1.49235636e-01 -7.14590013e-01 -1.36340961e-01 9.80022728e-01
-1.73682261e+00 -1.11748350e+00 2.65402913e-01 3.98114264e-01
7.99854159e-01 -2.13068932e-01 8.06921542e-01 7.22718835e-01
-5.94389617e-01 9.96857524e-01 4.33140069e-01 5.02882779e-01
1.00508583e+00 -1.28506649e+00 7.86566794e-01 1.29357374e+00
4.64257210e-01 5.84553361e-01 5.62166572e-01 -5.19354105e-01
-1.07545793e+00 -1.43153453e+00 1.35778964e+00 -7.88392186e-01
9.06476200e-01 -7.14410245e-01 -1.17070663e+00 7.82097816e-01
4.77677256e-01 1.62463903e-01 4.29452658e-01 5.55249214e-01
-7.85086572e-01 7.87790790e-02 -8.72228026e-01 7.17642069e-01
9.57580864e-01 -7.55651712e-01 -5.68148553e-01 5.03976524e-01
1.01743317e+00 -3.61719698e-01 -8.17544758e-01 5.33240736e-01
2.03998029e-01 -6.50264502e-01 7.36078978e-01 -1.08522642e+00
5.10709763e-01 1.09078139e-01 -4.06296283e-01 -1.85632908e+00
-1.78775445e-01 -4.62418407e-01 3.70025605e-01 1.38934386e+00
9.85166371e-01 -3.61392349e-01 2.20267549e-01 2.11668953e-01
-5.01655817e-01 -5.20905018e-01 -9.80271041e-01 -1.08314300e+00
8.43058944e-01 -9.55157518e-01 1.71513230e-01 1.22524059e+00
-2.39597306e-01 5.86768508e-01 -4.09812182e-01 1.00276217e-01
3.33151102e-01 -6.63403273e-01 4.82429653e-01 -7.99438298e-01
-3.06088626e-01 -5.20504177e-01 -9.84894559e-02 -7.40240514e-01
5.33116937e-01 -1.26490188e+00 4.05323446e-01 -1.18437159e+00
4.28979397e-02 -2.64442831e-01 -3.55292261e-01 8.02299976e-01
-4.74877328e-01 3.57648075e-01 2.26081520e-01 -1.63447097e-01
-5.60700595e-01 4.80846822e-01 7.88682818e-01 -2.75203705e-01
-6.13795444e-02 -5.26873052e-01 -6.49389982e-01 6.39240205e-01
7.16478705e-01 -6.40529931e-01 5.59037142e-02 -1.25359130e+00
4.16114479e-01 -5.76112568e-01 8.80891383e-02 -1.09455955e+00
-3.16934586e-02 3.31152022e-01 -1.87842362e-02 -3.03890973e-01
8.48864019e-02 -5.39440393e-01 -4.35088307e-01 3.26206207e-01
-6.26972318e-01 3.20901811e-01 5.59233189e-01 4.21809316e-01
-4.19302046e-01 -1.01975963e-01 8.92809272e-01 -2.68529534e-01
-5.61043561e-01 9.28473920e-02 -4.09714669e-01 7.19187200e-01
2.66215235e-01 4.29208755e-01 -5.29093087e-01 -1.81573719e-01
-7.12849319e-01 2.19956055e-01 1.05408184e-01 8.76363456e-01
8.95289406e-02 -1.25113857e+00 -1.25776625e+00 1.62730396e-01
2.51006663e-01 -1.38327688e-01 9.23761278e-02 8.07964742e-01
-2.53626317e-01 5.63269854e-01 1.65321842e-01 -5.64471483e-01
-1.01583052e+00 4.80852813e-01 4.83214676e-01 -6.16590023e-01
-3.22886080e-01 9.62566733e-01 -1.77117392e-01 -1.19092476e+00
5.36006510e-01 -5.98346114e-01 1.03106894e-01 4.68987562e-02
4.91062671e-01 2.23542914e-01 7.86938310e-01 -6.95741475e-01
-2.07207605e-01 2.49692693e-01 -2.64468193e-01 -1.52789146e-01
1.31990314e+00 -1.67666208e-02 -6.44773361e-04 6.53287411e-01
1.29117489e+00 1.56953543e-01 -8.04128349e-01 -6.89862847e-01
3.28403771e-01 1.05288535e-01 -1.16864294e-02 -1.22762311e+00
-9.10705388e-01 1.07274568e+00 4.49031532e-01 -3.03921681e-02
7.84706652e-01 -2.15186924e-01 8.68671179e-01 7.99686015e-01
1.25213116e-01 -1.22510910e+00 -1.99497551e-01 1.25772107e+00
9.22001839e-01 -1.67070508e+00 -5.05407453e-01 -2.03535572e-01
-7.28912592e-01 6.79933786e-01 5.72710752e-01 3.09761129e-02
5.71361840e-01 6.81415796e-01 5.25151372e-01 2.36994103e-01
-9.48624730e-01 -1.46453649e-01 1.78561330e-01 5.58640182e-01
8.23704422e-01 5.31708263e-02 -3.24397653e-01 8.05926919e-01
-5.06787539e-01 -2.88462698e-01 4.72465456e-01 6.18062079e-01
-2.17318341e-01 -1.16256630e+00 -3.34056348e-01 2.74097949e-01
-9.23040509e-01 -7.88224339e-01 -2.89400816e-01 7.45036066e-01
2.11773843e-01 1.03128409e+00 -7.91674387e-03 -8.99432898e-02
5.82303643e-01 4.56459135e-01 1.43295020e-01 -7.47014105e-01
-9.78638113e-01 -7.32367486e-02 3.98039430e-01 -6.20179415e-01
-1.47766471e-01 -6.38548195e-01 -7.80576527e-01 -1.18438460e-01
-3.46061923e-02 9.69853532e-03 8.58687997e-01 1.06196260e+00
2.88945019e-01 5.14734268e-01 5.98313808e-02 -4.68035281e-01
-7.26368070e-01 -1.35083055e+00 -2.35651329e-01 7.76195943e-01
4.15224105e-01 -1.85595840e-01 -4.25216883e-01 9.18513536e-02] | [10.728277206420898, 9.808436393737793] |
4baff1a4-3d77-449b-b1c5-3abbac54bebc | ftnet-feature-transverse-network-for-thermal | null | null | https://ieeexplore.ieee.org/abstract/document/9585453 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9585453 | FTNet: Feature Transverse Network for Thermal Image Semantic Segmentation | Thermal imaging is a process of using infrared radiation and thermal energy to collect information about objects. It is superior to visible imaging for its ability to operate in darkness and tolerate illumination variations. In addition, it has potential to penetrate smoke, aerosol, dust, and mist, which are critical inhibitors for visible imaging applications, including semantic segmentation. Unfortunately, current state-of-the-art image semantic segmentation methods (i) mainly concentrate on visible spectrum images, which do not adequately capture the context of corresponding pixels, particularly edge details in thermal images, and (ii) accept a trade-off between higher accuracy and lower speed, or vice-versa. Here, a novel end-to-end trainable convolutional neural network architecture, feature transverse network (FTNet), has been proposed to solve the aforementioned problems. FTNet captures and optimizes feature representation at the multi-scale resolution, thereby improving the capability to process high-resolution images and producing quality output with a lower computational cost. Extensive computer experimentations were conducted on publicly available benchmarking thermal datasets, including SODA, MFNet, and SCUT-Seg, to demonstrate the effectiveness of the proposed FTNet compared to state-of-the-art methods. This comparison includes multiple aspects, including the quantitative accuracy and speed of the various approaches. The source code is available at https://github.com/shreyaskamathkm/FTNet. | ['Srijith Rajeev and Sos S Agaian', 'Shreyas Kamath K.M', 'Karen Panetta'] | 2021-10-26 | null | null | null | ieee-access-2021-10 | ['scene-segmentation', 'thermal-image-segmentation'] | ['computer-vision', 'computer-vision'] | [ 5.40685117e-01 -5.84087133e-01 1.64441764e-01 -2.81342208e-01
-3.65724683e-01 -5.09451330e-01 1.27909318e-01 -4.54885542e-01
-2.11940467e-01 3.61845285e-01 -3.85170579e-01 -3.53840560e-01
7.97425807e-02 -7.60960460e-01 -3.55273634e-01 -1.11658919e+00
4.00384367e-01 -3.28278951e-02 3.95241857e-01 3.38218659e-02
2.35111237e-01 4.41122293e-01 -1.47742462e+00 2.57789075e-01
1.25769210e+00 1.34969461e+00 4.85399783e-01 5.60788333e-01
-2.25979745e-01 4.54713196e-01 -3.41634333e-01 -2.34217402e-02
3.90792012e-01 -3.83004069e-01 -6.01439714e-01 4.72299419e-02
5.17416239e-01 -4.20324624e-01 -2.13432238e-01 1.15122032e+00
4.78927732e-01 1.18490033e-01 4.69463557e-01 -9.15399969e-01
-9.31987107e-01 -1.15249328e-01 -8.03055644e-01 3.83425742e-01
-2.34947936e-03 4.92784619e-01 6.10687852e-01 -5.94135582e-01
1.26246482e-01 9.71323013e-01 5.35643816e-01 5.10373950e-01
-8.96787763e-01 -7.16296613e-01 -4.62940857e-02 1.87886655e-01
-1.17104256e+00 -1.72633484e-01 7.33362377e-01 -2.76427954e-01
8.20253611e-01 4.42077786e-01 6.89247847e-01 1.04994643e+00
4.29640293e-01 5.41247129e-01 1.46665621e+00 -1.82881042e-01
2.12947160e-01 4.91551943e-02 6.89302664e-03 8.08500946e-01
2.80807346e-01 2.69879669e-01 -3.47125798e-01 4.31924015e-02
7.46605754e-01 1.63731411e-01 -5.51878512e-01 1.79553837e-01
-7.14706123e-01 5.02242744e-01 6.38668776e-01 2.08736733e-01
-2.46137246e-01 1.88047051e-01 6.19483069e-02 -3.48410308e-02
8.59541059e-01 1.82826176e-01 -3.66335601e-01 1.21898226e-01
-7.85539627e-01 -2.52152801e-01 4.20312136e-01 4.66105789e-01
9.12030518e-01 1.42951086e-01 -9.83188525e-02 8.12118232e-01
4.55362260e-01 7.04453170e-01 3.30913186e-01 -9.66012716e-01
3.63712996e-01 6.55625582e-01 2.08198521e-02 -6.65472567e-01
-4.15628791e-01 -3.40735048e-01 -8.91735911e-01 5.51330030e-01
1.69212922e-01 -1.94411859e-01 -1.52690172e+00 1.24677134e+00
4.62718844e-01 3.62059951e-01 4.29361984e-02 1.22738636e+00
1.08086157e+00 8.90154958e-01 -9.37715471e-02 -1.20243564e-01
1.33570278e+00 -1.00451493e+00 -5.33655763e-01 -5.89232028e-01
1.55873761e-01 -9.67696726e-01 1.04005730e+00 2.54318535e-01
-8.06804001e-01 -6.57475889e-01 -9.99078512e-01 -1.01794966e-01
-4.27288890e-01 2.23270789e-01 7.40935326e-01 8.55300486e-01
-1.01163924e+00 6.19192660e-01 -9.19352531e-01 -3.42811674e-01
6.98307455e-01 1.54204499e-02 1.44051135e-01 -2.02664316e-01
-1.09095132e+00 5.23636878e-01 3.18726420e-01 7.06076503e-01
-7.75882065e-01 -7.87826419e-01 -5.71828604e-01 -4.90753315e-02
5.44036210e-01 -8.83893311e-01 9.63157654e-01 -1.12546778e+00
-1.67484677e+00 6.58056259e-01 -2.01273844e-01 2.61884239e-02
5.32125235e-01 -4.00727719e-01 -3.86898935e-01 5.53665578e-01
-1.95318773e-01 3.70678693e-01 9.28315461e-01 -1.42925215e+00
-5.21443069e-01 -4.29618359e-01 6.43658862e-02 3.90512794e-01
-2.27094278e-01 1.06302798e-01 -4.85946327e-01 -3.01544338e-01
1.31854400e-01 -9.24254179e-01 -1.53464064e-01 3.08416337e-01
-4.63361800e-01 -9.57701877e-02 1.26610243e+00 -5.32130182e-01
7.52771854e-01 -2.20796609e+00 -2.33440295e-01 -4.41990606e-02
1.67396128e-01 6.62245631e-01 8.95133168e-02 1.65428817e-01
1.23465382e-01 2.83739060e-01 -5.85478306e-01 -1.44045949e-01
-3.88122678e-01 8.20284039e-02 1.41296955e-03 6.53690219e-01
1.30525738e-01 9.74528611e-01 -5.96201777e-01 -5.07727385e-01
6.75719619e-01 6.68726087e-01 1.59317002e-01 6.83591366e-02
-3.55506599e-01 6.37241066e-01 -7.59442091e-01 7.76535034e-01
9.40161288e-01 -2.27736846e-01 -1.98682368e-01 -4.15805250e-01
-5.58386683e-01 -1.34728342e-01 -9.31658626e-01 1.33676195e+00
-5.66792905e-01 6.54294908e-01 2.79307723e-01 -5.99470854e-01
7.27439582e-01 2.87464470e-01 5.88639081e-01 -9.71518219e-01
3.85616928e-01 1.50214434e-01 -4.31190044e-01 -7.88628817e-01
3.54028106e-01 -6.57224134e-02 4.82450604e-01 3.57510835e-01
-6.10707104e-01 -1.44352302e-01 -1.07269488e-01 -1.97831422e-01
5.18562496e-01 1.81073815e-01 -2.46834457e-01 -3.12320422e-02
4.54528749e-01 6.23674178e-03 5.14080584e-01 5.64319372e-01
-1.28379509e-01 8.46327901e-01 -2.75349733e-03 -3.11053753e-01
-7.38731384e-01 -1.03310859e+00 -2.03789338e-01 7.77255476e-01
7.73595273e-01 2.64494568e-01 -7.15634763e-01 -3.38014722e-01
-2.10700512e-01 6.05857968e-01 -4.38346684e-01 -9.00907675e-04
-5.79076350e-01 -8.65032196e-01 3.37897450e-01 5.31348705e-01
1.23916340e+00 -1.18021178e+00 -1.00490713e+00 -1.24910779e-01
-2.91659713e-01 -1.24297273e+00 -3.58500034e-01 5.23656458e-02
-9.42981958e-01 -1.18781853e+00 -5.13098955e-01 -5.87273419e-01
5.46507537e-01 8.08065295e-01 8.50874484e-01 1.83442280e-01
-6.36457562e-01 1.19990163e-01 -4.22007442e-01 -4.23444927e-01
-5.60893156e-02 -1.96443070e-02 -5.25851488e-01 5.87075539e-02
1.26348808e-01 -3.34263325e-01 -1.05797970e+00 3.33707392e-01
-1.23675871e+00 2.35197544e-01 8.15432429e-01 6.23799264e-01
5.18404007e-01 3.77577335e-01 4.66020554e-02 -8.84801567e-01
1.89658433e-01 -1.65259659e-01 -6.80379272e-01 2.97290772e-01
-7.11515665e-01 -3.02055508e-01 6.72072351e-01 -6.01511002e-02
-1.60807920e+00 5.03025837e-02 -1.17263317e-01 -5.07397115e-01
-3.71144354e-01 3.03190798e-01 -5.02927043e-02 -3.16085994e-01
5.41337848e-01 4.77749079e-01 -7.39365071e-02 -4.45447177e-01
2.47024372e-01 8.71599615e-01 4.48507786e-01 -3.81405741e-01
9.21188772e-01 9.51868713e-01 -1.24283656e-01 -1.03216088e+00
-1.10307741e+00 -7.54205048e-01 -5.43109655e-01 -3.52087289e-01
1.22313321e+00 -7.23864019e-01 -4.65692520e-01 8.57347190e-01
-9.09108996e-01 -4.50787008e-01 9.49991792e-02 3.81281883e-01
-2.46751994e-01 5.56792259e-01 -6.26743376e-01 -9.58045959e-01
-8.71483028e-01 -1.04314756e+00 1.11945283e+00 9.17275310e-01
4.78399396e-01 -9.52903748e-01 -3.32635045e-01 8.97540748e-01
6.01193607e-01 4.18674618e-01 8.07902455e-01 2.28864357e-01
-8.38464379e-01 2.15832829e-01 -6.20177031e-01 6.55676544e-01
1.21714763e-01 3.52706254e-01 -1.32056379e+00 -2.52533019e-01
1.10258847e-01 -2.91783959e-01 1.21354914e+00 5.73248565e-01
1.42659295e+00 -3.11598019e-03 -2.78876334e-01 1.06231773e+00
1.80845940e+00 2.62098223e-01 8.24746728e-01 3.10198039e-01
9.00564194e-01 4.55213219e-01 6.58487082e-01 7.69801661e-02
4.83191852e-03 4.66526449e-01 7.76196778e-01 -8.16381097e-01
-1.55271560e-01 4.05452967e-01 2.11268723e-01 3.27451229e-01
-3.34475726e-01 -3.71950626e-01 -7.89407074e-01 3.47656876e-01
-1.72057128e+00 -8.93297613e-01 -5.16599357e-01 1.83463418e+00
6.42793655e-01 -5.78715652e-02 -1.60314322e-01 -1.28156915e-01
6.88546240e-01 3.19539100e-01 -7.41042852e-01 -2.95346200e-01
1.71205234e-02 3.88336569e-01 7.74342418e-01 1.69982567e-01
-1.10438836e+00 8.69507909e-01 5.67052794e+00 7.23939657e-01
-1.57491684e+00 1.88389733e-01 7.43185520e-01 1.13785297e-01
-1.26066580e-01 -1.21533208e-01 -4.04580235e-01 4.50814784e-01
6.38908327e-01 2.19707951e-01 6.23271644e-01 5.05027413e-01
5.32772362e-01 -2.89644837e-01 -5.35936356e-01 7.98362255e-01
-6.92063943e-02 -9.37404394e-01 -2.55873740e-01 -8.84331167e-02
8.89762640e-01 4.09920722e-01 3.57006401e-01 -3.27575475e-01
-4.82403487e-02 -1.16702545e+00 6.56384230e-01 3.55562955e-01
9.22455251e-01 -6.28770173e-01 8.17588568e-01 1.83599725e-01
-1.40308630e+00 -1.05241895e-01 -5.17112792e-01 3.01310062e-01
1.34617724e-02 8.14502358e-01 -3.90320718e-01 1.02623153e+00
9.91121948e-01 4.08498615e-01 -5.33919871e-01 1.02628899e+00
-3.57033789e-01 8.50253344e-01 -3.09807628e-01 1.69840410e-01
5.08196414e-01 -6.33055151e-01 2.83843249e-01 1.16361666e+00
2.66000688e-01 1.86342061e-01 3.08076859e-01 1.17207944e+00
1.43947542e-01 -2.17740968e-01 -2.38313839e-01 -6.64700866e-02
2.62121797e-01 1.55854583e+00 -1.00650501e+00 -2.33569250e-01
-5.77626526e-01 1.03188527e+00 -1.40137330e-01 7.17029870e-01
-1.06099451e+00 -2.83463538e-01 7.14760542e-01 -7.46784285e-02
2.75782466e-01 -3.45812768e-01 -5.64895689e-01 -8.43219340e-01
1.21478640e-01 -5.49352229e-01 2.32064858e-01 -1.01253390e+00
-1.23443818e+00 4.77037311e-01 -2.40806416e-01 -1.01071465e+00
5.55611849e-01 -6.84089482e-01 -7.90856063e-01 1.08820796e+00
-1.92702162e+00 -1.18588662e+00 -9.92303610e-01 4.56905007e-01
7.03457713e-01 4.49570864e-01 4.43515241e-01 1.84495851e-01
-7.67359078e-01 4.51415405e-02 4.85710770e-01 -1.07343331e-01
4.02487397e-01 -1.10944700e+00 2.68516690e-01 9.79537964e-01
-1.54506415e-01 4.02996808e-01 4.53374922e-01 -4.95402694e-01
-1.37414026e+00 -1.45406425e+00 -1.01626776e-02 -2.37493694e-01
1.79270685e-01 -1.39299929e-01 -9.69884038e-01 2.60154128e-01
1.76062956e-01 1.10273935e-01 3.71511132e-01 -3.82124841e-01
-2.25320339e-01 -2.07874700e-01 -1.19047928e+00 3.77021760e-01
8.89272273e-01 -4.28831369e-01 -1.65091380e-01 4.50019956e-01
6.46930277e-01 -6.06788099e-01 -6.01734579e-01 4.33328092e-01
6.11798167e-01 -1.25760841e+00 1.08441591e+00 1.16382226e-01
5.88564992e-01 -4.89084870e-01 3.08262080e-01 -1.14747536e+00
-3.33015144e-01 -2.28874043e-01 1.11609593e-01 1.14634967e+00
3.63959849e-01 -9.70239401e-01 7.64782071e-01 3.60257000e-01
-3.33933741e-01 -8.71490359e-01 -8.78725350e-01 -7.35948145e-01
-1.92840174e-01 -3.00459981e-01 3.56111020e-01 8.24985802e-01
-9.61926579e-01 7.97382891e-02 -1.75464734e-01 5.48064768e-01
7.92820871e-01 4.87015843e-01 2.53882527e-01 -1.02065766e+00
-1.39575467e-01 -4.31609690e-01 4.92702276e-02 -7.48231113e-01
-1.71132296e-01 -6.52427912e-01 2.73980498e-01 -2.04098248e+00
3.09694499e-01 -3.30369771e-01 -2.88381040e-01 4.43452150e-01
-4.77214277e-01 5.06576955e-01 -6.79560900e-02 2.89376706e-01
-2.84031779e-01 5.11588454e-01 1.64770174e+00 -1.54123396e-01
-2.34883070e-01 1.43523812e-02 -6.30941212e-01 6.71127200e-01
1.16932821e+00 -3.35500717e-01 -5.35672486e-01 -6.75080061e-01
-1.63770303e-01 -2.03007638e-01 5.23963809e-01 -9.17936087e-01
7.08108321e-02 -5.52101910e-01 6.18372619e-01 -5.16272366e-01
2.83488899e-01 -8.42759788e-01 3.01662892e-01 4.43579793e-01
4.95118499e-02 -4.02086616e-01 1.33875847e-01 6.03558004e-01
4.31116670e-02 -2.63950557e-01 1.14957702e+00 -3.59303117e-01
-8.44618797e-01 4.96565223e-01 -2.25015372e-01 -1.03968583e-01
1.17383802e+00 -6.19817972e-01 -8.59497488e-01 8.86166934e-03
-4.97157574e-02 2.33408347e-01 7.36499250e-01 2.14481845e-01
6.68563962e-01 -7.29363441e-01 -5.21512508e-01 3.71493362e-02
2.89263055e-02 4.02948856e-01 6.13341272e-01 9.01638031e-01
-1.00159729e+00 1.35406867e-01 -7.05547631e-02 -7.39850044e-01
-1.46034479e+00 1.75253868e-01 6.40922904e-01 1.40928596e-01
-9.17501032e-01 7.84567833e-01 4.49780136e-01 -2.00501427e-01
-9.54865217e-02 -4.05363917e-01 -8.17268714e-02 -4.37422931e-01
3.72717917e-01 4.78169769e-01 8.54699761e-02 -2.40871131e-01
-3.77644062e-01 8.04508209e-01 1.75496757e-01 2.64021873e-01
1.31377172e+00 -3.38484466e-01 -1.80359155e-01 1.60773277e-01
9.28963423e-01 -3.47915590e-01 -1.48013747e+00 -7.49753863e-02
-4.80526179e-01 -7.54448056e-01 5.40607274e-01 -8.94681931e-01
-1.44649076e+00 9.17386889e-01 8.92721832e-01 3.06188762e-01
1.49279749e+00 -1.07037097e-01 1.14291453e+00 1.75692830e-02
2.11785398e-02 -1.19527650e+00 1.96304917e-01 2.20995590e-01
5.07214546e-01 -1.18301332e+00 1.35588348e-01 -7.79470563e-01
-6.14081264e-01 1.28681982e+00 8.14148366e-01 2.38349438e-01
2.90288836e-01 2.36831129e-01 5.34310639e-01 -3.78624707e-01
-5.24141073e-01 -3.72577488e-01 2.60503799e-01 5.92155755e-01
4.13089663e-01 -7.91265815e-02 -2.55859103e-02 -1.49409875e-01
1.60799712e-01 -5.67732379e-02 3.49323869e-01 9.09719765e-01
-6.25830412e-01 -4.57738549e-01 -5.86982846e-01 5.23519695e-01
-5.46953857e-01 -1.42973006e-01 -4.91451591e-01 4.45558667e-01
2.63358265e-01 1.27952754e+00 -9.27679092e-02 -2.42067739e-01
-1.97776053e-02 -1.92066520e-01 3.91508698e-01 -4.54227984e-01
-4.93268162e-01 1.03712626e-01 -1.87527034e-02 -5.80635548e-01
-6.92859411e-01 -3.64355713e-01 -1.21994460e+00 -2.73975909e-01
-5.87397337e-01 -1.94743633e-01 9.41744566e-01 9.19071853e-01
1.37063444e-01 8.60689640e-01 7.91835129e-01 -7.88135409e-01
-3.74276608e-01 -9.14637804e-01 -5.65434635e-01 2.21853793e-01
1.67448118e-01 -4.37451571e-01 -4.17552948e-01 -2.96409801e-02] | [10.25008487701416, -2.323323965072632] |
241ae138-e0ef-4069-ba20-8cf8aa1cc12e | gp-guided-mppi-for-efficient-navigation-in | 2307.04019 | null | https://arxiv.org/abs/2307.04019v1 | https://arxiv.org/pdf/2307.04019v1.pdf | GP-guided MPPI for Efficient Navigation in Complex Unknown Cluttered Environments | Robotic navigation in unknown, cluttered environments with limited sensing capabilities poses significant challenges in robotics. Local trajectory optimization methods, such as Model Predictive Path Intergal (MPPI), are a promising solution to this challenge. However, global guidance is required to ensure effective navigation, especially when encountering challenging environmental conditions or navigating beyond the planning horizon. This study presents the GP-MPPI, an online learning-based control strategy that integrates MPPI with a local perception model based on Sparse Gaussian Process (SGP). The key idea is to leverage the learning capability of SGP to construct a variance (uncertainty) surface, which enables the robot to learn about the navigable space surrounding it, identify a set of suggested subgoals, and ultimately recommend the optimal subgoal that minimizes a predefined cost function to the local MPPI planner. Afterward, MPPI computes the optimal control sequence that satisfies the robot and collision avoidance constraints. Such an approach eliminates the necessity of a global map of the environment or an offline training process. We validate the efficiency and robustness of our proposed control strategy through both simulated and real-world experiments of 2D autonomous navigation tasks in complex unknown environments, demonstrating its superiority in guiding the robot safely towards its desired goal while avoiding obstacles and escaping entrapment in local minima. The GPU implementation of GP-MPPI, including the supplementary video, is available at https://github.com/IhabMohamed/GP-MPPI. | ['Lantao Liu', 'Mahmoud Ali', 'Ihab S. Mohamed'] | 2023-07-08 | null | null | null | null | ['autonomous-navigation'] | ['computer-vision'] | [ 1.54626500e-02 9.02425796e-02 -3.05083580e-03 1.19607776e-01
-6.39877796e-01 -5.54953158e-01 3.96946400e-01 1.81891933e-01
-4.84015584e-01 6.26430750e-01 -1.64460376e-01 -3.90753567e-01
-5.84213912e-01 -7.97532678e-01 -6.61162138e-01 -1.02389061e+00
-2.72860199e-01 5.50151765e-01 1.94190294e-01 -2.81801283e-01
6.19248211e-01 5.02360284e-01 -1.29940498e+00 -7.19808996e-01
1.33031356e+00 8.80249858e-01 1.06137049e+00 3.21322381e-01
1.77584723e-01 1.41887203e-01 2.09156245e-01 2.24604636e-01
4.75728989e-01 4.90398407e-02 -3.11714798e-01 -1.40036702e-01
-2.63723373e-01 -5.17063700e-02 -2.02338025e-01 1.27016819e+00
5.08261859e-01 8.28218699e-01 6.84669137e-01 -1.19134247e+00
1.57474816e-01 -6.97026327e-02 -4.52504158e-01 -1.69301152e-01
1.46913707e-01 4.66907889e-01 4.11753535e-01 -8.57640147e-01
4.32379037e-01 1.23900723e+00 5.33755600e-01 2.42507145e-01
-1.01080835e+00 -4.17312533e-01 2.53560305e-01 1.17934875e-01
-1.53769803e+00 -1.87968463e-01 3.89147848e-01 -4.23434317e-01
6.46770895e-01 -2.63339520e-01 3.59888613e-01 8.09559882e-01
9.95609045e-01 2.30102286e-01 6.73366785e-01 6.64578825e-02
5.94009221e-01 -2.15343595e-01 -2.90575922e-01 9.39080417e-01
4.31934863e-01 3.32465351e-01 -3.80185544e-01 -1.60499364e-01
7.34419286e-01 7.28257895e-02 -3.59100372e-01 -9.67217207e-01
-1.18514955e+00 7.44251549e-01 5.93597591e-01 -3.60902756e-01
-7.89457023e-01 1.58137996e-02 1.04978576e-01 -8.30698833e-02
-1.83258608e-01 3.86447459e-01 -3.71650308e-01 -4.38037694e-01
-7.29372576e-02 4.48689699e-01 7.61572838e-01 1.22928679e+00
8.62685621e-01 6.68381825e-02 3.35828424e-01 6.57058179e-01
5.69993377e-01 7.90677726e-01 9.91651565e-02 -1.40661561e+00
7.03283668e-01 4.36357200e-01 6.41371489e-01 -1.34052813e+00
-6.16676569e-01 -3.35412830e-01 -6.78419530e-01 5.90366006e-01
2.39461213e-01 -4.47520375e-01 -6.37293100e-01 1.48836362e+00
6.71297908e-01 -9.45819244e-02 2.76484549e-01 1.07586408e+00
-1.47802662e-02 7.84538150e-01 -1.25127673e-01 3.81312962e-03
9.45128798e-01 -9.27712798e-01 -4.75607216e-01 -7.20555782e-01
5.33898413e-01 -3.99052203e-01 8.56156230e-01 4.39917415e-01
-5.09364009e-01 -3.87505472e-01 -1.05589426e+00 2.39405781e-01
-1.74254224e-01 1.14475071e-01 2.38575250e-01 8.78526196e-02
-9.29132283e-01 5.50916076e-01 -1.16753435e+00 -5.63249648e-01
8.90906900e-02 3.38658571e-01 -3.08447659e-01 -3.42297375e-01
-7.50496686e-01 1.21759820e+00 7.08784163e-01 3.70058298e-01
-1.27667904e+00 -3.35123211e-01 -1.04488254e+00 -2.69023657e-01
8.54774356e-01 -4.97366816e-01 1.02747631e+00 -9.55933258e-02
-1.90048146e+00 -8.28495696e-02 -3.66347194e-01 -3.90051961e-01
3.67023051e-01 -5.62943578e-01 3.79819602e-01 -1.22647077e-01
3.07497770e-01 4.97748792e-01 7.39473939e-01 -1.34811032e+00
-8.17687035e-01 -3.57203096e-01 -1.86916769e-01 7.89220512e-01
2.76756257e-01 -7.43519247e-01 -6.06308281e-01 -7.10556656e-02
5.87356925e-01 -1.31691062e+00 -9.24011946e-01 -1.42272323e-01
-2.67569870e-01 -6.49697660e-03 5.57404459e-01 -4.72398460e-01
6.57815039e-01 -1.99541521e+00 5.16320050e-01 4.39176738e-01
-1.88246936e-01 -2.65742928e-01 -9.84381810e-02 7.19607413e-01
7.29009092e-01 -4.01116163e-01 -3.39517683e-01 -1.86126426e-01
-1.07753575e-01 2.99292743e-01 -2.53268450e-01 6.31479979e-01
-1.24307945e-01 5.09745181e-01 -1.14561796e+00 -7.46235903e-03
3.50890219e-01 2.98817575e-01 -3.97751629e-01 1.21077113e-02
-1.71959102e-01 9.05260444e-01 -8.82051408e-01 5.56116939e-01
5.69155693e-01 2.15060413e-01 5.01735993e-02 8.65815654e-02
-7.41807878e-01 3.31442468e-02 -1.38515103e+00 1.73981428e+00
-5.60725093e-01 2.24897370e-01 6.09877765e-01 -7.57372975e-01
1.22996771e+00 -2.80806243e-01 3.69750440e-01 -4.95998561e-01
3.10825348e-01 5.67212284e-01 -2.89787203e-01 -4.00942385e-01
5.91061473e-01 2.28878483e-01 -1.44894063e-01 4.87839766e-02
-2.71531731e-01 -3.89067411e-01 -2.31203049e-01 -1.36050031e-01
1.06143069e+00 5.26334763e-01 3.70492637e-01 -4.03224558e-01
6.14634991e-01 3.58145744e-01 7.40267396e-01 8.56543779e-01
-2.64021873e-01 2.29973823e-01 -4.79714805e-03 -8.70965123e-02
-6.65054083e-01 -1.03362393e+00 2.05805346e-01 6.62239969e-01
9.83268380e-01 -4.36955737e-03 -2.91213423e-01 -1.85533062e-01
1.11424036e-01 7.61889160e-01 -2.45322794e-01 -2.10650414e-01
-6.41544998e-01 -3.71235192e-01 -1.66635171e-01 2.08588034e-01
4.16957527e-01 -7.35109866e-01 -1.05725014e+00 4.48178411e-01
-1.78536460e-01 -8.65128756e-01 -1.86198041e-01 4.47961152e-01
-8.48344505e-01 -1.06592774e+00 -2.82355726e-01 -8.98136258e-01
8.38475704e-01 5.24067938e-01 1.63836986e-01 -2.59782374e-01
6.54149726e-02 4.45725411e-01 -2.41328821e-01 -2.77463347e-01
-1.88480750e-01 -5.62258102e-02 3.98009360e-01 -2.81106830e-01
-6.64893389e-02 -3.68757784e-01 -5.42779386e-01 5.91921985e-01
-1.45763054e-01 1.26272321e-01 6.61426365e-01 6.33294106e-01
1.00051832e+00 4.50289994e-01 3.70176136e-01 -1.29768737e-02
5.89766681e-01 -7.37945676e-01 -1.09977388e+00 -2.21922100e-01
-5.53871810e-01 7.29401186e-02 5.88234186e-01 -2.05955327e-01
-1.10567796e+00 4.08138692e-01 1.24292076e-01 -3.22203010e-01
-9.83591303e-02 8.36397111e-01 -2.08294347e-01 -2.54030168e-01
3.89595956e-01 3.51290613e-01 4.00296032e-01 -2.80106694e-01
3.09386373e-01 3.15250129e-01 6.43946767e-01 -6.41413093e-01
8.66245389e-01 5.03065169e-01 4.10589963e-01 -7.24927366e-01
-2.17114419e-01 -5.87256193e-01 -5.82862735e-01 -3.64267349e-01
6.85638845e-01 -8.53209257e-01 -8.43121588e-01 2.70282030e-01
-8.66133869e-01 -5.69822669e-01 1.54431447e-01 6.04532719e-01
-1.04832196e+00 2.93919444e-01 -1.20207906e-01 -1.16160893e+00
-2.21165299e-01 -1.19570363e+00 7.19107032e-01 4.94633049e-01
-4.73080687e-02 -7.52622306e-01 4.48554903e-02 -3.39979827e-02
4.05486822e-01 4.45939183e-01 5.11268198e-01 -1.97826385e-01
-8.38727176e-01 -1.87473685e-01 1.35958478e-01 -2.98864961e-01
2.84924626e-01 -4.22642052e-01 -2.09302947e-01 -5.03017366e-01
1.99444875e-01 -3.50701027e-02 4.87443089e-01 4.66251522e-01
4.47036296e-01 -2.69013047e-01 -7.13985920e-01 4.99082297e-01
1.61749434e+00 7.66169131e-01 1.89311400e-01 8.94553542e-01
5.32580435e-01 8.89345825e-01 1.43699849e+00 5.31956136e-01
6.79511189e-01 6.17482007e-01 9.62014139e-01 6.36153460e-01
6.24524415e-01 -5.44736683e-01 6.02677464e-01 4.15051967e-01
1.01257101e-01 -1.76700383e-01 -1.04015481e+00 5.87319434e-01
-2.28706932e+00 -4.04137850e-01 -3.48870084e-02 2.41759038e+00
2.30227858e-01 7.54532516e-02 -3.83431911e-01 -3.44801575e-01
5.99814355e-01 -1.80195361e-01 -8.31541479e-01 -2.46104166e-01
3.58457863e-01 -5.59863269e-01 9.86064315e-01 8.58846188e-01
-1.10498250e+00 9.01888847e-01 3.97896695e+00 6.40629530e-01
-9.29679692e-01 -1.65459633e-01 1.50308628e-02 2.61475801e-01
1.43148050e-01 1.13137759e-01 -9.57164466e-01 3.45816851e-01
8.11405241e-01 -2.39619106e-01 5.81642032e-01 1.16227758e+00
8.83754373e-01 -6.54891133e-01 -5.97524881e-01 6.77760839e-01
-3.41892064e-01 -9.32717383e-01 -3.69527578e-01 2.13723674e-01
5.65680146e-01 3.07046533e-01 -1.80522352e-02 1.92999229e-01
4.13197249e-01 -6.31734431e-01 8.19827855e-01 6.70165896e-01
3.11332326e-02 -9.57060456e-01 8.37426186e-01 8.32959950e-01
-1.28716350e+00 -4.89337593e-01 -5.32896340e-01 -3.01568303e-02
5.51811934e-01 1.58747405e-01 -1.05059385e+00 7.22250402e-01
6.39490008e-01 5.96912026e-01 -4.08624858e-02 1.44395041e+00
-4.06408787e-01 -4.73722294e-02 -5.35604596e-01 -3.42640877e-01
5.75944185e-01 -8.05987418e-01 1.07746255e+00 4.83691901e-01
7.52708554e-01 1.49146587e-01 6.62997961e-01 7.53898978e-01
8.23377371e-01 3.34581695e-02 -7.79592216e-01 1.73526987e-01
7.03393102e-01 1.06332815e+00 -7.79711187e-01 2.15135589e-01
-7.58117810e-02 6.56331360e-01 2.66048700e-01 4.94065613e-01
-5.93068242e-01 -5.33370793e-01 7.35507786e-01 -2.98302323e-01
3.79463106e-01 -9.07080352e-01 -3.73120546e-01 -5.00780404e-01
-8.37857574e-02 -4.70877320e-01 -8.76177941e-03 -6.35245264e-01
-7.74901390e-01 4.08747345e-01 -8.70516673e-02 -1.55815566e+00
-2.33850092e-01 -5.51533699e-01 -6.06339991e-01 1.00692618e+00
-1.35046124e+00 -6.47172213e-01 -5.49779356e-01 2.72023201e-01
6.82568610e-01 -3.08000054e-02 5.57912707e-01 -3.28881353e-01
-3.87267292e-01 -1.75952464e-01 3.70251387e-01 -5.89539647e-01
4.03200567e-01 -9.56447244e-01 5.93902096e-02 8.66944194e-01
-6.53791666e-01 6.19044006e-01 1.08896399e+00 -1.08588839e+00
-1.87181485e+00 -1.11508977e+00 1.79513067e-01 -7.31011108e-02
7.64739454e-01 -5.74186780e-02 -7.85425007e-01 3.85216534e-01
-3.45529556e-01 -5.47894955e-01 1.14127889e-03 -3.20717931e-01
4.55509186e-01 6.19499348e-02 -9.81777251e-01 1.06555557e+00
8.94853652e-01 1.28983289e-01 -1.95472747e-01 1.67056903e-01
7.09438384e-01 -7.71234274e-01 -5.94273984e-01 6.42603397e-01
5.20159662e-01 -4.23825741e-01 7.21015275e-01 1.09386764e-01
-9.62035060e-02 -6.45596623e-01 -1.84841454e-01 -1.51904881e+00
-2.33012199e-01 -6.99574232e-01 2.85400692e-02 8.19171846e-01
3.70234430e-01 -9.77292061e-01 8.60088348e-01 5.19790828e-01
-5.46724319e-01 -9.63749349e-01 -1.11990571e+00 -9.02411282e-01
-1.71846315e-01 -4.21797335e-01 9.28726047e-02 2.20979854e-01
-6.23991266e-02 -1.26302555e-01 -1.80217505e-01 1.00916052e+00
8.63134861e-01 -5.48478635e-03 1.02051353e+00 -8.73169124e-01
1.07654139e-01 -2.14754179e-01 -2.27507278e-01 -1.21339011e+00
1.59699932e-01 -5.79039693e-01 9.94374096e-01 -1.94308889e+00
-4.56162453e-01 -7.83603609e-01 1.13840006e-01 1.94184169e-01
-2.64537763e-02 -5.30470967e-01 1.38514593e-01 3.91328722e-01
-4.13716406e-01 9.74866688e-01 1.17423046e+00 -5.05983233e-02
-7.97856331e-01 4.22325045e-01 -4.02554065e-01 6.96196616e-01
1.06641901e+00 -3.46934080e-01 -7.28107512e-01 -4.24781531e-01
-9.28167719e-03 3.31111401e-01 8.76973644e-02 -1.23664069e+00
7.05381572e-01 -5.32378972e-01 -1.28940493e-01 -8.49424064e-01
4.93972778e-01 -9.81176436e-01 2.01320410e-01 8.49242449e-01
9.61540546e-03 4.77024280e-02 3.64211589e-01 1.16030467e+00
-4.09078365e-03 -4.44842726e-01 5.24965584e-01 8.49588367e-04
-1.20775390e+00 2.77072728e-01 -7.65288055e-01 -3.69750887e-01
1.14272714e+00 -1.72614768e-01 -2.03412458e-01 -3.22258592e-01
-6.30215526e-01 1.02399516e+00 6.60795987e-01 4.06878591e-01
9.25537646e-01 -8.44437420e-01 -2.30632186e-01 1.24590866e-01
-2.65268981e-02 3.97215635e-01 2.97841758e-01 1.02627766e+00
-6.51070297e-01 5.58399856e-01 -1.70007601e-01 -6.57955348e-01
-7.96416938e-01 4.97554600e-01 2.71018177e-01 1.65994734e-01
-7.68775761e-01 6.43360436e-01 1.48855135e-01 -7.70497859e-01
1.20179854e-01 -1.60154179e-01 -1.62916034e-01 -3.99179459e-01
2.03434676e-01 5.93430102e-01 -1.47668913e-01 -6.13599181e-01
-3.81012917e-01 6.50303185e-01 3.79801333e-01 -2.93546289e-01
1.17189717e+00 -8.00314188e-01 1.27316061e-02 3.18402559e-01
7.34902501e-01 -1.12030268e-01 -1.89517117e+00 8.99309386e-03
2.95016050e-01 -4.03076857e-01 4.84755151e-02 -4.40401495e-01
-4.38668638e-01 4.39644545e-01 5.91319859e-01 -3.05074960e-01
7.07155526e-01 -2.49824345e-01 5.97398818e-01 7.32738376e-01
1.21186233e+00 -9.63163137e-01 1.06010633e-02 1.09104848e+00
1.02187991e+00 -1.09891450e+00 -1.38265416e-01 -4.26579773e-01
-7.21641660e-01 1.07453632e+00 7.44218588e-01 -2.21367970e-01
6.32041574e-01 -2.17591505e-02 9.48079377e-02 7.67050404e-03
-5.23674965e-01 -1.34942368e-01 -1.30747214e-01 6.71321630e-01
-5.79566956e-01 6.78760884e-03 -3.11276823e-01 3.47272158e-01
-1.75284803e-01 -3.61172140e-01 5.09155095e-01 1.40586829e+00
-1.05210614e+00 -6.08944714e-01 -5.19310236e-01 1.99517190e-01
3.08740765e-01 3.07678998e-01 3.36926430e-01 6.25679493e-01
-7.86782429e-02 1.04729176e+00 -7.40863904e-02 -4.81387585e-01
3.12348336e-01 -1.92736268e-01 9.66857001e-02 -6.41589522e-01
1.47943482e-01 1.38827056e-01 2.50942018e-02 -9.29715812e-01
-8.21365509e-03 -7.42748857e-01 -1.66038859e+00 1.62480175e-01
-2.61964470e-01 2.20084757e-01 9.70806956e-01 8.69968235e-01
5.96705973e-01 1.83362991e-01 5.70536375e-01 -1.37780476e+00
-6.43487453e-01 -6.40171289e-01 -2.70080596e-01 -5.17952502e-01
3.00042212e-01 -1.15587509e+00 -3.89756054e-01 -3.74830544e-01] | [4.871448993682861, 1.1969724893569946] |
c62c3fbc-c448-4e47-8898-a73b2427d453 | multi-cell-multi-task-convolutional-neural | 1808.10564 | null | http://arxiv.org/abs/1808.10564v2 | http://arxiv.org/pdf/1808.10564v2.pdf | Multi-Cell Multi-Task Convolutional Neural Networks for Diabetic Retinopathy Grading | Diabetic Retinopathy (DR) is a non-negligible eye disease among patients with
Diabetes Mellitus, and automatic retinal image analysis algorithm for the DR
screening is in high demand. Considering the resolution of retinal image is
very high, where small pathological tissues can be detected only with large
resolution image and large local receptive field are required to identify those
late stage disease, but directly training a neural network with very deep
architecture and high resolution image is both time computational expensive and
difficult because of gradient vanishing/exploding problem, we propose a
\textbf{Multi-Cell} architecture which gradually increases the depth of deep
neural network and the resolution of input image, which both boosts the
training time but also improves the classification accuracy. Further,
considering the different stages of DR actually progress gradually, which means
the labels of different stages are related. To considering the relationships of
images with different stages, we propose a \textbf{Multi-Task} learning
strategy which predicts the label with both classification and regression.
Experimental results on the Kaggle dataset show that our method achieves a
Kappa of 0.841 on test set which is the 4-th rank of all state-of-the-arts
methods. Further, our Multi-Cell Multi-Task Convolutional Neural Networks
(M$^2$CNN) solution is a general framework, which can be readily integrated
with many other deep neural network architectures. | ['Zaiwang Gu', 'Kang Zhou', 'Wen Liu', 'Jiang Liu', 'Shenghua Gao', 'Weixin Luo', 'Jun Cheng'] | 2018-08-31 | null | null | null | null | ['diabetic-retinopathy-grading'] | ['medical'] | [ 4.47528362e-02 -3.63779753e-01 -1.52865335e-01 -3.12863231e-01
-7.17521489e-01 3.67372151e-04 -5.66773936e-02 -3.10244948e-01
-3.63275826e-01 8.44162345e-01 7.54875690e-02 -2.22274482e-01
-2.13612810e-01 -6.44499302e-01 -4.18276429e-01 -9.40791965e-01
3.00117671e-01 2.28305057e-01 4.02776033e-01 -2.06704214e-02
1.06366903e-01 5.50451577e-01 -1.72061229e+00 7.84679592e-01
1.19137537e+00 1.29659462e+00 1.56964764e-01 7.30635226e-01
-1.62620451e-02 1.11244142e+00 -2.70541012e-01 -1.79045022e-01
4.01783615e-01 -2.62884289e-01 -7.84058273e-01 4.36375625e-02
8.79538596e-01 -5.96206844e-01 -2.11136386e-01 1.15864658e+00
9.62052584e-01 -3.48122448e-01 6.48092568e-01 -3.24433088e-01
-7.82846093e-01 -1.12593822e-01 -1.05304050e+00 7.30754137e-01
-4.88434941e-01 1.31656721e-01 4.37605470e-01 -6.14531577e-01
2.31241301e-01 1.12005126e+00 5.01495779e-01 6.05985105e-01
-1.00120401e+00 -6.89034462e-01 -2.10049134e-02 2.99892843e-01
-1.48075867e+00 -3.67949486e-01 2.43786737e-01 -7.66108096e-01
7.05584943e-01 1.49367675e-01 6.05039299e-01 6.41620874e-01
2.48538509e-01 4.79707956e-01 1.59106135e+00 -1.03811838e-01
-2.38895744e-01 -5.91728091e-02 2.02719271e-01 9.03794646e-01
3.75310898e-01 1.86594456e-01 8.28068554e-02 2.11484402e-01
1.27054167e+00 1.35774076e-01 -3.21274966e-01 2.22434148e-01
-9.68322635e-01 6.63052440e-01 5.31944633e-01 1.94482297e-01
-2.79002637e-01 1.36454687e-01 4.41831529e-01 2.55728453e-01
2.95401365e-01 9.94047225e-02 -3.91550422e-01 3.74425173e-01
-6.14280224e-01 3.90820652e-02 2.56089680e-03 5.20890594e-01
5.70025325e-01 -5.40775619e-02 -3.49511355e-01 1.24978065e+00
5.01543507e-02 2.76517272e-01 5.40279746e-01 -1.02971208e+00
-1.02716379e-01 1.09653306e+00 -8.32910016e-02 -5.82413495e-01
-5.71359873e-01 -7.66218424e-01 -1.45335019e+00 8.05251896e-01
6.77338660e-01 -3.24948937e-01 -1.20589232e+00 1.03506911e+00
1.28680661e-01 1.21936522e-01 8.89937729e-02 1.08818519e+00
1.11549544e+00 3.92373741e-01 -4.87575941e-02 -5.16344428e-01
1.70537341e+00 -9.95302618e-01 -4.93754476e-01 -1.98759288e-01
7.04538167e-01 -6.23115361e-01 9.14722919e-01 5.95539689e-01
-1.02070451e+00 -6.66091323e-01 -8.08613837e-01 -4.10409212e-01
-4.48964126e-02 7.18683124e-01 8.47703338e-01 2.40501925e-01
-1.21652710e+00 1.41308634e-02 -3.55611891e-01 -2.69258827e-01
8.09719265e-01 4.09535050e-01 -2.96913892e-01 -5.40035009e-01
-7.24077344e-01 9.83201087e-01 1.21700920e-01 4.48641002e-01
-6.55052483e-01 -5.99669635e-01 -1.93053812e-01 -3.27581018e-01
9.15031582e-02 -8.63151252e-01 7.85523474e-01 -9.74945962e-01
-1.21177053e+00 1.17175090e+00 -2.74449557e-01 -2.81745583e-01
4.71237957e-01 -3.77498232e-02 -6.53785646e-01 3.14519107e-01
-1.46973073e-01 6.50815785e-01 6.12565339e-01 -1.05204809e+00
-1.16987491e+00 -5.54508984e-01 3.68993767e-02 1.31256372e-01
-4.44936082e-02 3.48466843e-01 -3.24531406e-01 -9.98882353e-02
5.54197021e-02 -5.45789897e-01 -4.96528804e-01 2.80712396e-01
-7.70243928e-02 -3.49998742e-01 4.46629196e-01 -5.87403238e-01
1.11480951e+00 -2.06621385e+00 -1.62891686e-01 -1.22685291e-01
7.16695309e-01 4.33685988e-01 -1.14539728e-01 -5.65598726e-01
-5.28347790e-02 1.49185240e-01 3.89764994e-01 2.69520611e-01
-8.68061841e-01 -9.52766463e-02 2.74162680e-01 3.58228117e-01
2.24652275e-01 6.66097999e-01 -4.44805086e-01 -6.28908038e-01
9.85655412e-02 5.64175725e-01 -3.19987386e-01 -9.16130841e-02
8.05337355e-02 5.06691277e-01 -5.31670034e-01 9.20921504e-01
6.09504998e-01 -8.49089563e-01 -1.64569423e-01 -6.19958758e-01
-3.90030682e-01 -4.43898112e-01 -1.01368678e+00 1.18325937e+00
-2.34188899e-01 4.71214116e-01 -2.54743665e-01 -9.18690264e-01
1.05348623e+00 2.92068005e-01 3.38226169e-01 -9.67517793e-01
3.33026230e-01 4.28742737e-01 3.45470548e-01 -9.07844722e-01
-2.29825720e-01 -8.86576697e-02 5.78598857e-01 -1.78678691e-01
-2.94372439e-01 5.61101139e-01 1.91229239e-01 -5.22088647e-01
8.86845350e-01 -1.73312768e-01 3.06833506e-01 -7.55593851e-02
6.26681864e-01 -1.61203936e-01 7.23291218e-01 3.73980522e-01
-3.23428363e-01 6.57003582e-01 7.23319530e-01 -9.57280099e-01
-1.13635218e+00 -5.95042586e-01 -6.25905514e-01 4.53858554e-01
1.26175299e-01 3.07475358e-01 -2.46370822e-01 -3.64191353e-01
-3.53633203e-02 -1.29816622e-01 -7.52889454e-01 1.72573254e-01
-4.92551118e-01 -1.28241754e+00 4.72183496e-01 4.99298543e-01
8.36722612e-01 -5.98647952e-01 -4.00306493e-01 2.80703809e-02
-6.88933441e-03 -1.09019089e+00 -1.03360787e-02 -2.68837810e-01
-1.15358531e+00 -1.23879242e+00 -1.07235205e+00 -1.26040208e+00
8.32958758e-01 3.80091339e-01 8.11406136e-01 2.22355321e-01
-6.74974620e-01 -4.97185349e-01 -3.83841842e-02 -2.90941924e-01
1.96540412e-02 -4.40406293e-01 -2.05383822e-01 3.56779784e-01
6.37100816e-01 -4.84262377e-01 -1.18532979e+00 2.20899761e-01
-4.15808588e-01 2.33224496e-01 1.23293447e+00 8.61275554e-01
9.04242337e-01 1.78790748e-01 7.05395818e-01 -7.70730853e-01
2.18086407e-01 -1.80914104e-02 -6.30462050e-01 2.75026858e-01
-7.76418805e-01 -2.82897234e-01 2.28185847e-01 -4.07152534e-01
-9.37125564e-01 -3.32604125e-02 6.08846545e-02 -4.27262396e-01
-2.76798069e-01 3.24769944e-01 2.62782812e-01 -4.32895750e-01
8.40007424e-01 1.93452671e-01 3.32902879e-01 -4.82927769e-01
3.57722603e-02 9.57239389e-01 3.92962605e-01 -4.61097285e-02
1.21440150e-01 3.98463547e-01 4.03112292e-01 -6.13745332e-01
-1.14653027e+00 -4.44176227e-01 -3.60126674e-01 -3.48425984e-01
1.11441922e+00 -1.40332532e+00 -1.04332638e+00 8.55288923e-01
-1.06976426e+00 -2.20309332e-01 1.98914781e-01 7.74662316e-01
-7.01164454e-02 2.56477892e-01 -7.74136245e-01 -5.21240532e-01
-4.11460698e-01 -1.29355025e+00 6.95401728e-01 5.43724239e-01
4.97589290e-01 -7.55185783e-01 -3.38026136e-01 8.05311561e-01
4.38106567e-01 3.74511987e-01 1.44812715e+00 -9.36412066e-02
-9.76955056e-01 -1.36698455e-01 -1.19672132e+00 6.88223124e-01
3.23434523e-03 1.04959942e-01 -9.29160595e-01 -1.84431642e-01
-2.17676774e-01 -4.88440961e-01 1.19079435e+00 1.09042823e+00
1.49025059e+00 8.95311532e-04 -3.69497061e-01 8.43806744e-01
1.89014840e+00 2.91360646e-01 1.12851083e+00 3.77993673e-01
6.29899681e-01 5.51425636e-01 4.41843092e-01 1.54739797e-01
5.91656446e-01 4.11252141e-01 5.65593958e-01 -8.50248933e-01
-5.18375695e-01 5.38005233e-01 -9.28199515e-02 5.30338585e-01
-8.43278944e-01 7.36935344e-03 -7.92781830e-01 6.84877813e-01
-1.64937031e+00 -8.21048021e-01 -6.53363585e-01 2.16452742e+00
1.03172135e+00 -7.87148401e-02 5.11098541e-02 -3.08454067e-01
8.75103772e-01 -1.88911170e-01 -6.47046626e-01 -1.23500168e-01
-2.98767298e-01 3.03724974e-01 6.12955332e-01 1.76182747e-01
-9.81885493e-01 6.28142953e-01 6.01682663e+00 8.66577148e-01
-1.28387582e+00 1.72213912e-01 1.02514172e+00 -4.07322198e-01
2.64699250e-01 -3.28750372e-01 -1.17721665e+00 4.27169353e-01
6.63192987e-01 2.10763842e-01 1.96776271e-01 4.33584630e-01
4.19816256e-01 -1.58412114e-01 -8.89311552e-01 1.24507153e+00
-3.81602608e-02 -1.43500209e+00 3.34207743e-01 2.57694602e-01
7.29462385e-01 2.25929886e-01 7.21747056e-02 2.32171729e-01
4.06862348e-02 -1.39403903e+00 -3.15512717e-01 8.71229768e-01
1.30444181e+00 -5.86121261e-01 1.12521291e+00 5.47488220e-02
-8.55840445e-01 -2.65114486e-01 -6.74876034e-01 9.66252983e-02
-2.07138762e-01 8.30345392e-01 -3.41894895e-01 2.34966278e-01
9.32336807e-01 7.79384315e-01 -6.96744204e-01 1.44938934e+00
9.32100862e-02 2.99373686e-01 1.99487969e-01 8.71225148e-02
-4.77736853e-02 -4.03347880e-01 1.50530696e-01 9.27225649e-01
5.04254699e-01 4.29808587e-01 1.08109042e-01 5.41562796e-01
4.89560291e-02 2.67248303e-01 -3.77474487e-01 2.72118330e-01
4.49751280e-02 1.36980951e+00 -2.66509473e-01 -1.69151857e-01
-7.86110580e-01 5.77797234e-01 2.21270069e-01 5.63497841e-01
-4.50756222e-01 -3.36549222e-01 5.81983626e-01 2.71366686e-01
2.16464981e-01 8.02932978e-02 -5.22656262e-01 -9.80243683e-01
-5.90130426e-02 -8.04883480e-01 3.14798743e-01 -9.00994599e-01
-1.23341727e+00 6.39505446e-01 -8.64422798e-01 -1.46061671e+00
3.63088250e-01 -8.18461835e-01 -2.66151011e-01 1.17308736e+00
-2.28040433e+00 -1.37531638e+00 -6.97757185e-01 8.41992974e-01
3.22645068e-01 -4.81296450e-01 7.90183783e-01 7.00741827e-01
-7.92631626e-01 4.97649610e-01 4.65943366e-02 2.13209778e-01
9.26127672e-01 -1.02576029e+00 -4.69975710e-01 6.20770872e-01
-8.07691514e-01 4.50675756e-01 1.48912165e-02 -4.18111175e-01
-9.54278946e-01 -1.36499107e+00 7.18378127e-01 -1.21425875e-01
4.32206124e-01 3.51640821e-01 -8.96279156e-01 3.97433072e-01
-1.01100408e-01 5.16565263e-01 8.67079258e-01 2.55659342e-01
-1.58789456e-01 -7.23173976e-01 -9.70150530e-01 5.45643568e-01
8.50134492e-01 -2.31942594e-01 -1.72519431e-01 5.47230422e-01
5.23582637e-01 -5.83794355e-01 -1.26912904e+00 7.28598654e-01
5.79403758e-01 -1.07581472e+00 8.75251949e-01 -5.34698248e-01
7.30159521e-01 -4.59318519e-01 -6.07039146e-02 -6.70652211e-01
-5.35268009e-01 -5.01721352e-02 6.84478460e-03 8.79299164e-01
4.62495595e-01 -7.06941187e-01 6.62587404e-01 2.26981580e-01
-2.28463352e-01 -1.10080624e+00 -7.77858615e-01 -3.60681504e-01
9.87033397e-02 3.25592220e-01 8.28978717e-02 8.75716627e-01
-7.48174846e-01 4.13717896e-01 -3.94178808e-01 4.12968218e-01
7.29962111e-01 2.00877622e-01 4.09464151e-01 -1.60500598e+00
-7.50631541e-02 -5.01087725e-01 -5.75792134e-01 -9.98977125e-01
-4.82840657e-01 -7.79887021e-01 -3.82874697e-01 -1.88360548e+00
4.83732253e-01 -4.92275327e-01 -6.60528243e-01 5.80399990e-01
-8.52427557e-02 5.09605587e-01 -5.26080668e-01 4.30134684e-01
-3.54949087e-01 5.56673408e-02 1.90325713e+00 -2.53363457e-02
-2.36073479e-01 2.32111752e-01 -9.89116728e-01 7.94304371e-01
4.06793833e-01 -4.76157339e-03 -2.64853179e-01 -6.21693552e-01
3.93249840e-01 1.78083822e-01 5.98276496e-01 -1.04299295e+00
3.07039499e-01 -1.90316975e-01 7.38074839e-01 -4.36311454e-01
1.70919791e-01 -4.26295340e-01 -1.40942886e-01 3.95955980e-01
-2.11987689e-01 -4.99460220e-01 8.55954662e-02 4.64547962e-01
-3.74457330e-01 1.77441195e-01 1.37687719e+00 -3.84980917e-01
-7.50728786e-01 7.26885676e-01 -8.99364725e-02 -1.29473731e-01
9.94139075e-01 -6.13527834e-01 -8.75918031e-01 2.50956982e-01
-5.59463143e-01 4.08513129e-01 7.07100257e-02 7.38777891e-02
7.39099503e-01 -1.07314444e+00 -1.11562955e+00 5.64043969e-02
2.16457680e-01 3.80464137e-01 5.81577957e-01 1.38848066e+00
-6.60773337e-01 2.19943628e-01 -5.02439976e-01 -7.34610915e-01
-1.42764628e+00 1.74890324e-01 9.21272993e-01 -8.71267393e-02
-7.78084874e-01 9.59174156e-01 3.85561675e-01 2.66881436e-01
7.04888850e-02 -4.44307417e-01 -8.46138239e-01 1.09198615e-01
8.00926566e-01 3.68226379e-01 -1.00882975e-02 -1.33453831e-01
-6.19672239e-02 1.12075102e+00 -7.31956780e-01 6.95354640e-01
1.31358290e+00 -1.18033744e-01 -6.34357154e-01 1.75225243e-01
8.60028505e-01 -3.05149227e-01 -1.09335279e+00 -3.46264631e-01
-4.53755766e-01 -4.22154546e-01 7.10488796e-01 -1.17945659e+00
-1.33539045e+00 1.02315748e+00 1.27856934e+00 -1.53110638e-01
1.39436960e+00 -2.26847664e-01 7.71605015e-01 1.80298790e-01
2.36505225e-01 -8.19296539e-01 1.94286834e-02 1.82319015e-01
7.54616261e-01 -1.51968455e+00 1.32749468e-01 -4.69943047e-01
-7.06041634e-01 1.41447520e+00 1.00377774e+00 -1.63978145e-01
4.40387905e-01 -7.71232834e-03 3.93030941e-01 -2.58989185e-01
-6.99686825e-01 -4.70949143e-01 3.89420509e-01 5.33710301e-01
4.04337406e-01 -1.29465818e-01 -6.59239292e-01 4.90616679e-01
3.67696196e-01 5.63252687e-01 7.36998379e-01 2.42643699e-01
-8.66855800e-01 -9.08037722e-01 2.55761951e-01 9.82866704e-01
-7.72626340e-01 -2.73651898e-01 1.48217469e-01 5.21452963e-01
5.28210640e-01 7.36257553e-01 3.55771780e-02 -1.68388635e-01
2.58567005e-01 -3.55641454e-01 4.43430603e-01 -5.81557870e-01
-2.95003355e-01 3.20457339e-01 1.62626341e-01 -5.10049999e-01
-5.78590453e-01 -3.77465278e-01 -1.21756124e+00 -3.86774093e-02
-1.76249638e-01 -5.19036829e-01 4.30929720e-01 8.27084959e-01
4.42867786e-01 6.61794007e-01 6.14145637e-01 -1.57403827e-01
-3.07298511e-01 -1.06555510e+00 -8.25133264e-01 -2.11815506e-01
6.62702382e-01 -5.62222719e-01 -2.52952814e-01 1.00142255e-01] | [15.810193061828613, -3.9761431217193604] |
70e16d69-afdb-497f-8c40-f507ffaa82b3 | a-self-supervised-approach-to-reconstruction | 2211.00002 | null | https://arxiv.org/abs/2211.00002v1 | https://arxiv.org/pdf/2211.00002v1.pdf | A Self-Supervised Approach to Reconstruction in Sparse X-Ray Computed Tomography | Computed tomography has propelled scientific advances in fields from biology to materials science. This technology allows for the elucidation of 3-dimensional internal structure by the attenuation of x-rays through an object at different rotations relative to the beam. By imaging 2-dimensional projections, a 3-dimensional object can be reconstructed through a computational algorithm. Imaging at a greater number of rotation angles allows for improved reconstruction. However, taking more measurements increases the x-ray dose and may cause sample damage. Deep neural networks have been used to transform sparse 2-D projection measurements to a 3-D reconstruction by training on a dataset of known similar objects. However, obtaining high-quality object reconstructions for the training dataset requires high x-ray dose measurements that can destroy or alter the specimen before imaging is complete. This becomes a chicken-and-egg problem: high-quality reconstructions cannot be generated without deep learning, and the deep neural network cannot be learned without the reconstructions. This work develops and validates a self-supervised probabilistic deep learning technique, the physics-informed variational autoencoder, to solve this problem. A dataset consisting solely of sparse projection measurements from each object is used to jointly reconstruct all objects of the set. This approach has the potential to allow visualization of fragile samples with x-ray computed tomography. We release our code for reproducing our results at: https://github.com/vganapati/CT_PVAE . | ['Vidya Ganapati', 'Juliane Mueller', 'Talita Perciano', 'Vincent Dumont', 'Judith Weng Zhu', 'Minh Nguyen', 'Rey Mendoza'] | 2022-10-30 | null | null | null | null | ['probabilistic-deep-learning'] | ['computer-vision'] | [ 7.09748864e-02 4.81744222e-02 1.82111114e-01 -2.83568650e-01
-7.07076609e-01 -3.06837440e-01 2.15202093e-01 -2.27843657e-01
-4.10248160e-01 6.63036525e-01 2.52500344e-02 -1.79622546e-01
-2.57448368e-02 -1.09109116e+00 -1.07842577e+00 -9.86963093e-01
2.10808009e-01 1.12425709e+00 7.00525865e-02 3.69474888e-01
7.80739486e-02 9.22444522e-01 -1.20597374e+00 1.29443467e-01
1.21671610e-01 6.58915102e-01 5.71223855e-01 6.13956511e-01
5.94147258e-02 3.11521232e-01 -1.95011750e-01 -1.21339492e-01
3.55796665e-01 -4.68289346e-01 -6.48803294e-01 7.59211555e-03
2.17837349e-01 -9.26994383e-01 -4.08776700e-01 9.31119621e-01
4.34115350e-01 -3.10534192e-03 8.47793579e-01 -6.06009781e-01
-8.79766226e-01 4.24765766e-01 -4.90148515e-01 -8.26632828e-02
3.54109883e-01 1.12871997e-01 5.60281813e-01 -9.12342489e-01
9.05986905e-01 9.49159205e-01 5.92969060e-01 7.12552547e-01
-1.47913694e+00 -3.69577438e-01 -8.49416614e-01 1.28486380e-01
-1.00528371e+00 -5.79530261e-02 9.53256845e-01 -6.33140743e-01
7.53687143e-01 2.37326607e-01 9.24652398e-01 1.23579836e+00
5.55330038e-01 3.58514547e-01 1.27480793e+00 -2.83393174e-01
4.19499099e-01 -8.84178951e-02 -1.81216583e-01 7.89309084e-01
3.67092520e-01 3.88889134e-01 -2.46645913e-01 -2.01167300e-01
1.13683808e+00 4.41275895e-01 -5.76816559e-01 -2.95520067e-01
-1.00254786e+00 6.89472675e-01 4.72947836e-01 4.37508672e-01
-5.75741649e-01 8.94023851e-02 5.30660190e-02 -1.38852596e-01
1.34509355e-01 6.72196031e-01 -3.76867145e-01 2.17781901e-01
-1.07605398e+00 1.59865022e-01 7.33035386e-01 2.91487157e-01
8.70291233e-01 1.15083024e-01 5.65265775e-01 3.78246427e-01
5.08905768e-01 6.03370488e-01 5.80002487e-01 -1.36289001e+00
-3.87674481e-01 3.52384567e-01 -1.62496120e-01 -7.56485462e-01
-3.78223330e-01 -7.39850700e-02 -9.01911795e-01 7.46063948e-01
5.63747346e-01 2.72315592e-02 -1.08966196e+00 1.39375401e+00
7.88349211e-01 1.13676205e-01 -1.04750060e-01 1.26607227e+00
8.85106504e-01 8.79177153e-01 -5.02520025e-01 -2.60360688e-01
1.28791595e+00 -2.27270290e-01 -5.81963718e-01 2.60162026e-01
2.14046806e-01 -6.97091758e-01 8.40704978e-01 5.69873273e-01
-1.32980371e+00 -3.43649775e-01 -1.14050603e+00 -3.44699264e-01
-2.83257589e-02 -3.38491380e-01 3.25591326e-01 3.33572060e-01
-7.00144589e-01 9.73881364e-01 -1.50891435e+00 -8.77077281e-02
7.81852782e-01 2.05594629e-01 -6.34650290e-01 -3.62018496e-01
-5.62253058e-01 8.87909651e-01 5.04113026e-02 -1.00734286e-01
-1.04489636e+00 -1.13563311e+00 -6.28172576e-01 4.03335653e-02
-5.70583753e-02 -8.29685092e-01 9.28207397e-01 -2.75305957e-01
-1.81205714e+00 9.09808159e-01 4.13113870e-02 -5.16099930e-02
4.50122714e-01 -1.82261616e-01 2.27885619e-01 4.89825070e-01
-1.15741111e-01 4.90957111e-01 5.84381759e-01 -1.64599299e+00
2.59000987e-01 -7.84405291e-01 -3.05752754e-01 -1.81754827e-01
1.75121590e-01 -2.24890828e-01 -2.42186531e-01 -2.82146573e-01
6.89096570e-01 -9.78553295e-01 -1.60045609e-01 5.94316781e-01
-5.93915284e-01 3.20539623e-01 1.01466787e+00 -9.49162245e-01
-1.06922509e-02 -1.89261389e+00 4.75937039e-01 -1.15086079e-01
4.11252916e-01 7.44169438e-03 2.06225082e-01 4.79043722e-01
-1.09737836e-01 -3.58007960e-02 -6.16760194e-01 -4.30814534e-01
-3.35770011e-01 3.45239371e-01 -1.20787524e-01 9.14762199e-01
-1.57270491e-01 7.81935275e-01 -6.05276406e-01 -4.53389674e-01
3.54905844e-01 9.88901913e-01 -5.30330360e-01 3.79118890e-01
-2.30899632e-01 8.03323805e-01 -2.65828252e-01 7.11126626e-01
9.38224852e-01 -3.51725608e-01 2.36114502e-01 -3.98797095e-01
1.51151884e-02 -1.96481809e-01 -8.59663546e-01 1.71907449e+00
-2.23348811e-01 5.63808560e-01 2.46612728e-01 -1.07238305e+00
5.65449178e-01 4.02207911e-01 7.58256733e-01 -5.55383623e-01
3.02208364e-01 4.33107950e-02 -6.03325143e-02 -6.16281629e-01
3.92485894e-02 -9.60771978e-01 3.85960907e-01 8.53747666e-01
1.48325741e-01 -7.92098880e-01 -3.15255851e-01 6.67630062e-02
1.06198609e+00 2.75379658e-01 2.91186356e-04 -5.34243733e-02
-1.25762209e-01 -7.14808926e-02 4.49922502e-01 3.67277801e-01
6.75681308e-02 1.05761063e+00 3.45299654e-02 -7.48897672e-01
-1.55684984e+00 -1.44962454e+00 -6.95224643e-01 2.37068802e-01
-1.33948907e-01 1.86523572e-01 -6.22018933e-01 -1.39170334e-01
-1.71450917e-02 6.92772269e-01 -7.10406065e-01 -1.99629981e-02
-5.61812639e-01 -8.02200973e-01 2.00424641e-01 2.55267859e-01
8.85354280e-02 -1.00485480e+00 -8.90110493e-01 1.56411901e-01
-8.15780610e-02 -8.86406839e-01 3.86334866e-01 1.80636436e-01
-9.91286159e-01 -1.22356820e+00 -8.18627596e-01 -3.72312307e-01
8.91607463e-01 -5.45516051e-02 9.40357804e-01 1.71830922e-01
-6.88395321e-01 5.09169042e-01 -9.97691303e-02 -3.92769784e-01
-6.86396480e-01 -6.92492247e-01 3.21619451e-01 -4.04215842e-01
-6.43516257e-02 -1.13129747e+00 -6.85013235e-01 9.98655632e-02
-1.09339082e+00 -6.88766269e-03 4.05369461e-01 8.65723193e-01
1.15120244e+00 1.84450001e-01 -1.00948073e-01 -7.33825088e-01
2.06854772e-02 -3.92634511e-01 -6.93868101e-01 -1.88449249e-01
-1.59215316e-01 1.77409932e-01 7.23217964e-01 -4.65009183e-01
-9.65376258e-01 1.96850784e-02 -4.50135589e-01 -7.09600747e-01
-3.20294887e-01 4.76693988e-01 -8.30383133e-03 4.41292487e-03
7.89966583e-01 4.07989174e-01 1.65780768e-01 -5.06018698e-01
9.25884172e-02 7.91894868e-02 5.33638477e-01 -5.15624046e-01
7.13327110e-01 1.04352629e+00 4.68336493e-01 -1.03718197e+00
-5.07067323e-01 1.12883441e-01 -8.28496933e-01 -3.21608543e-01
1.09846270e+00 -5.07799745e-01 -7.67892361e-01 4.42432076e-01
-9.80656087e-01 -3.83278847e-01 -4.11429107e-01 9.63078141e-01
-5.64710677e-01 4.48896646e-01 -7.87263691e-01 -3.83542806e-01
-1.82017475e-01 -1.41798949e+00 9.36766922e-01 8.93976241e-02
-1.62503704e-01 -8.12800527e-01 5.09351015e-01 5.22282660e-01
7.73274153e-02 4.76670772e-01 9.39094722e-01 -7.39736110e-02
-9.19971943e-01 -2.54667550e-01 2.06928000e-01 3.96612257e-01
-1.46469986e-03 2.94124991e-01 -8.99022162e-01 -3.33236992e-01
6.62712872e-01 -6.15422666e-01 7.00716436e-01 9.20642316e-01
1.36454570e+00 -1.72659740e-01 -3.31717432e-01 1.00241613e+00
1.65339434e+00 -2.43740547e-02 7.17824697e-01 4.47509624e-02
7.80982077e-01 4.61550325e-01 -8.10578763e-02 3.22245628e-01
-6.02733018e-03 4.05067563e-01 6.18396521e-01 7.83561841e-02
-2.13181943e-01 1.19004529e-02 -1.24120109e-01 6.93712413e-01
-3.33585054e-01 -9.00003389e-02 -9.03146446e-01 1.82848811e-01
-9.51913953e-01 -1.03972650e+00 -2.41040096e-01 1.99611187e+00
8.74956667e-01 -2.85555512e-01 -3.81428331e-01 2.73558140e-01
3.17277104e-01 -2.27897912e-01 -8.00263822e-01 -2.61658709e-02
6.55236691e-02 3.84552062e-01 9.11231637e-02 6.42860711e-01
-3.57877314e-01 3.56611252e-01 6.39144802e+00 3.03266764e-01
-1.60107791e+00 1.44425124e-01 4.47706908e-01 -2.75917381e-01
-5.42727053e-01 -3.04842684e-02 -2.67110229e-01 5.06483078e-01
8.87431264e-01 1.46616578e-01 3.20701241e-01 7.00092912e-01
1.28597319e-01 -4.01413649e-01 -1.29580796e+00 9.14224863e-01
-5.30543476e-02 -1.66644919e+00 -2.21301652e-02 3.34721386e-01
5.06656170e-01 3.66584003e-01 5.43418527e-02 -3.40348065e-01
3.16844314e-01 -8.78974855e-01 5.68600595e-01 7.25613177e-01
8.14996183e-01 -4.23092693e-01 3.74685287e-01 3.72888416e-01
-2.62241513e-01 2.49524981e-01 -6.28718734e-01 1.62704855e-01
5.68068624e-01 1.05809534e+00 -9.83521342e-01 3.29959422e-01
9.40232396e-01 2.28458479e-01 2.42353529e-02 7.81435668e-01
-1.48006335e-01 6.95534825e-01 -6.80179596e-01 2.54188806e-01
-3.58156919e-01 -5.54197967e-01 7.41244376e-01 5.90436339e-01
6.10122979e-01 5.65256476e-01 -1.97818205e-01 1.31600249e+00
-1.80080429e-01 -3.91489446e-01 -5.84418356e-01 1.18297905e-01
2.73738474e-01 1.19133484e+00 -8.09385538e-01 -1.83885306e-01
-1.46523520e-01 8.24328125e-01 3.05121630e-01 1.92167222e-01
-5.88712037e-01 3.45756352e-01 3.92072916e-01 4.87392932e-01
2.54110157e-01 -4.60716814e-01 -3.66888702e-01 -1.07724941e+00
-4.15123962e-02 -5.25779188e-01 -7.18245432e-02 -1.21833146e+00
-1.12711918e+00 3.35714132e-01 3.66611071e-02 -8.13682556e-01
-1.11867905e-01 -7.20559716e-01 -5.02257049e-01 6.97274446e-01
-9.73493278e-01 -9.38611865e-01 -1.10461928e-01 3.06192011e-01
1.77654311e-01 2.53425598e-01 9.76317406e-01 1.12229712e-01
-4.27889794e-01 -5.32484055e-02 4.66717422e-01 -1.63188539e-02
3.60642701e-01 -1.18646073e+00 -1.52362540e-01 6.33784235e-01
2.59415299e-01 4.30413961e-01 9.96627688e-01 -6.60282195e-01
-1.79489326e+00 -6.49153948e-01 2.77256876e-01 -6.98706686e-01
3.04514766e-01 -2.46496364e-01 -1.25689960e+00 8.68914902e-01
3.96548361e-02 2.09540546e-01 1.04674900e+00 -1.81801394e-01
-1.10490441e-01 3.45456302e-01 -1.54861891e+00 2.51374871e-01
5.21748960e-01 -6.16034687e-01 -6.22193336e-01 5.10840118e-01
3.44903558e-01 -5.14078617e-01 -1.18588495e+00 1.06384404e-01
7.36627519e-01 -1.21028435e+00 1.10276520e+00 -2.59743005e-01
9.42915618e-01 -1.62752464e-01 -2.26490080e-01 -1.21569157e+00
-3.41236502e-01 1.22617818e-02 -1.74925581e-01 4.20041621e-01
1.57794133e-01 -4.76889998e-01 1.03520930e+00 7.93247998e-01
-2.29022220e-01 -8.34460616e-01 -1.05590951e+00 -5.70548356e-01
3.97988319e-01 -4.52496231e-01 3.17381144e-01 1.05915987e+00
-4.10940766e-01 -3.49823013e-02 -1.04076080e-01 5.94680071e-01
9.79596376e-01 3.79972965e-01 6.39090300e-01 -1.22564995e+00
-6.36135697e-01 5.45393117e-02 -5.05028009e-01 -7.71568418e-01
1.10241741e-01 -9.36641932e-01 -2.76934821e-02 -1.55556202e+00
4.03329045e-01 -1.97165713e-01 3.70353818e-01 2.58525401e-01
2.65186429e-01 4.01585966e-01 3.29747908e-02 6.11519516e-01
1.87252015e-01 6.47604287e-01 1.61951888e+00 2.17258297e-02
1.23103805e-01 -2.53832787e-01 -3.14942628e-01 7.45523691e-01
6.11531019e-01 -8.17762375e-01 -6.04877323e-02 -7.24068403e-01
1.94424152e-01 3.46906334e-01 5.45869708e-01 -1.16771948e+00
5.60911112e-02 -1.78228505e-02 1.01138222e+00 -8.06738377e-01
7.98382580e-01 -9.84126210e-01 7.72450745e-01 7.21254349e-01
5.30576892e-02 -2.57015139e-01 6.36428669e-02 4.75147963e-01
1.15960270e-01 -4.03765619e-01 1.14260900e+00 -6.05463088e-01
1.50309339e-01 4.37680900e-01 -4.67517525e-01 -5.07263184e-01
8.77949178e-01 -1.93595499e-01 2.62084175e-02 -1.73606947e-01
-9.47061896e-01 -4.35442299e-01 9.55532432e-01 -3.51690471e-01
9.92495179e-01 -1.12207162e+00 -7.41473198e-01 3.17186415e-01
-4.97161120e-01 6.17194712e-01 6.75157309e-01 5.24925828e-01
-1.08771861e+00 -8.07300583e-02 -5.49806178e-01 -8.80564690e-01
-1.08933902e+00 6.39653087e-01 4.01715785e-01 -3.44185419e-02
-1.05831909e+00 9.34983969e-01 1.19979940e-01 -6.77812099e-01
-4.95879978e-01 -3.17652196e-01 1.57209575e-01 -4.36863065e-01
5.12921751e-01 1.90686792e-01 -4.28829007e-02 -5.12378991e-01
-8.05374794e-03 6.91990018e-01 1.88117623e-02 -3.33151996e-01
2.03227115e+00 9.32843760e-02 -2.93689162e-01 4.62602168e-01
1.46292830e+00 -6.88055158e-02 -1.29853070e+00 -8.55657086e-02
-8.46424818e-01 -4.41714674e-01 5.79330146e-01 -5.62990725e-01
-1.32265389e+00 9.52937067e-01 5.14739692e-01 3.96135412e-02
8.61048937e-01 4.21252072e-01 8.59935760e-01 4.43726331e-01
2.25203842e-01 -5.85456252e-01 1.20245911e-01 7.85689354e-02
1.01865876e+00 -1.28128386e+00 6.14758313e-01 -1.07192345e-01
-6.12169728e-02 1.21816349e+00 2.75081307e-01 -3.50901544e-01
1.02792108e+00 5.04820645e-01 2.27963347e-02 -6.99514985e-01
-3.76484215e-01 5.71031094e-01 -3.81040573e-02 6.84918821e-01
1.88662037e-01 1.17430128e-01 1.66070253e-01 9.48551372e-02
-3.71603340e-01 2.33568382e-02 7.99761117e-01 1.06030846e+00
-1.85900822e-01 -9.77617443e-01 -7.35484421e-01 4.32737201e-01
-4.05159384e-01 2.94782668e-01 -1.75112709e-01 6.24944627e-01
-1.09768562e-01 1.50335729e-01 2.65098244e-01 7.02517433e-03
-1.37518466e-01 -1.18520752e-01 1.02715456e+00 -5.12575686e-01
1.75362736e-01 5.23742624e-02 -4.64690745e-01 -6.50530875e-01
-5.03872693e-01 -8.23504746e-01 -1.52775943e+00 -4.48268563e-01
-2.28322938e-01 -1.61629438e-01 1.11052430e+00 8.13730776e-01
1.77151725e-01 3.83139551e-01 4.35283691e-01 -1.45535266e+00
-2.94498056e-01 -7.69269168e-01 -8.22991431e-01 3.29069138e-01
4.26321298e-01 -6.41819298e-01 -4.96585965e-01 2.88438886e-01] | [13.127585411071777, -2.7547712326049805] |
4f743b7a-f9c4-4e64-a3e2-94a6187ecb3d | on-device-learning-a-neural-network-based | 2203.01077 | null | https://arxiv.org/abs/2203.01077v2 | https://arxiv.org/pdf/2203.01077v2.pdf | On-Device Learning: A Neural Network Based Field-Trainable Edge AI | In real-world edge AI applications, their accuracy is often affected by various environmental factors, such as noises, location/calibration of sensors, and time-related changes. This article introduces a neural network based on-device learning approach to address this issue without going deep. Our approach is quite different from de facto backpropagation based training but tailored for low-end edge devices. This article introduces its algorithm and implementation on a wireless sensor node consisting of Raspberry Pi Pico and low-power wireless module. Experiments using vibration patterns of rotating machines demonstrate that retraining by the on-device learning significantly improves an anomaly detection accuracy at a noisy environment while saving computation and communication costs for low power. | ['Masaaki Kondo', 'Mineto Tsukada', 'Hiroki Matsutani'] | 2022-03-02 | null | null | null | null | ['pico'] | ['natural-language-processing'] | [ 6.28431141e-02 -1.75458163e-01 -7.54569918e-02 -1.10936955e-01
2.70122886e-01 -7.14990571e-02 -2.57604152e-01 3.29118706e-02
-4.78664577e-01 6.42224669e-01 -5.55758417e-01 -2.42351294e-01
-3.14957798e-01 -8.92371118e-01 -8.35545182e-01 -4.32681412e-01
-1.76422730e-01 -7.82167539e-02 3.15827429e-01 -1.44953087e-01
3.11318077e-02 5.29754400e-01 -1.41925704e+00 -2.12417971e-02
2.16541812e-01 1.65802753e+00 -1.65484082e-02 8.41788232e-01
2.17485130e-01 8.30534339e-01 -9.75168705e-01 2.42594555e-01
1.72904521e-01 8.10081512e-02 3.57073434e-02 -4.36213553e-01
-3.00489247e-01 -1.67860538e-01 -3.26391667e-01 9.19857323e-01
8.86437356e-01 2.40638778e-02 5.18802643e-01 -1.51715302e+00
-4.07748550e-01 6.23321772e-01 -3.07038844e-01 3.18825513e-01
2.54533887e-01 -9.98421684e-02 -7.66148092e-04 -5.20054042e-01
9.28663537e-02 3.45393509e-01 1.75115156e+00 3.86145473e-01
-6.96862102e-01 -5.70586801e-01 -2.25120708e-01 4.81585622e-01
-1.27440548e+00 -6.03429079e-02 1.29993951e+00 -1.62614897e-01
1.51401722e+00 1.33880511e-01 6.37180686e-01 1.33267915e+00
7.77772188e-01 3.41324985e-01 4.04106647e-01 -4.68694091e-01
7.48514831e-01 1.16621189e-01 8.97994712e-02 4.73627865e-01
6.47107005e-01 1.33126885e-01 -4.84314173e-01 -1.15703836e-01
6.62640750e-01 3.80190194e-01 -8.34785476e-02 9.86599848e-02
-5.60241699e-01 1.11178920e-01 2.22348824e-01 6.56338334e-01
-6.77412331e-01 5.03844261e-01 4.61021572e-01 3.94400209e-01
9.15573761e-02 2.29038849e-01 -1.21726274e+00 -5.17833650e-01
-4.49444175e-01 -4.88704056e-01 1.06492496e+00 8.74199986e-01
2.31603816e-01 6.38356984e-01 4.04203773e-01 5.78879118e-01
3.27336550e-01 2.48370931e-01 1.09407592e+00 -4.43545699e-01
1.10162366e-02 4.68787134e-01 2.05668658e-01 -1.21111810e+00
-1.13079369e+00 -6.00331366e-01 -1.03102779e+00 4.98297095e-01
4.54859734e-02 -9.75720406e-01 -7.45849848e-01 1.26751709e+00
1.05476059e-01 6.48490429e-01 6.98043779e-02 5.46778202e-01
8.83714080e-01 3.83438110e-01 -7.47366101e-02 -1.21057257e-01
9.16727960e-01 -8.40764523e-01 -9.61518943e-01 -2.22847044e-01
3.88247848e-01 -1.65885359e-01 8.38758111e-01 5.66937327e-01
-6.23069525e-01 -7.84472585e-01 -1.56694269e+00 4.41975892e-01
-8.14561248e-01 2.89058983e-01 5.07747948e-01 1.06070852e+00
-7.87486672e-01 1.14146316e+00 -1.16431820e+00 -3.87275904e-01
3.31918836e-01 8.76222849e-01 8.52001309e-02 8.65866840e-01
-1.00376749e+00 6.81734145e-01 4.58810240e-01 1.83566600e-01
-2.00840622e-01 -4.94258791e-01 -5.60042441e-01 1.13655344e-01
8.75084624e-02 -6.29391968e-01 1.15469313e+00 -1.17399728e+00
-2.35964251e+00 -1.51604768e-02 3.55489731e-01 -5.88606000e-01
1.57206401e-01 -3.86431992e-01 -1.16163969e+00 -1.59685925e-01
-5.64598560e-01 -1.34664485e-02 1.11962020e+00 -5.33455312e-01
-5.52332520e-01 -3.83584082e-01 -4.54054445e-01 -1.71070278e-01
-7.27173746e-01 -5.05175531e-01 3.56015384e-01 -5.45124710e-01
1.08940393e-01 -7.28637457e-01 1.18190035e-01 -3.73565145e-02
3.18983234e-02 -2.58136630e-01 1.32629073e+00 -4.25337732e-01
1.01895356e+00 -1.96979916e+00 -7.28987515e-01 3.04476798e-01
-3.08369875e-01 3.47591639e-01 4.79292661e-01 1.06183223e-01
-1.38870001e-01 -2.69960672e-01 1.49412379e-01 1.65609911e-01
-2.35182464e-01 2.83586890e-01 2.65991300e-01 2.57280827e-01
1.52644396e-01 6.90225959e-01 -6.26358986e-01 2.42070049e-01
4.75170225e-01 6.67547464e-01 1.94972437e-02 -1.95571080e-01
1.77063897e-01 3.31254870e-01 -4.46881741e-01 8.87508154e-01
1.55015931e-01 8.02795440e-02 -3.02400827e-01 -3.28497976e-01
1.85065523e-01 -5.61205707e-02 -1.52351475e+00 1.64441168e+00
-6.17213726e-01 9.49554443e-01 3.64342816e-02 -1.49852145e+00
1.16433215e+00 4.65911865e-01 7.67683327e-01 -7.63941407e-01
8.54758680e-01 1.09705985e-01 -2.13857815e-01 -1.06817973e+00
1.38221636e-01 2.06470057e-01 9.95052680e-02 7.55418465e-02
1.22559503e-01 3.13363343e-01 -6.79699719e-01 -4.57218617e-01
1.51529253e+00 4.19498712e-01 1.18053250e-01 -1.44627795e-01
3.77091855e-01 -2.30613686e-02 8.45521271e-01 7.61543334e-01
-4.39750552e-01 1.65268138e-01 -2.43498638e-01 -7.67580807e-01
-5.84946871e-01 -6.71355426e-01 -1.72832623e-01 9.24465775e-01
2.57828683e-01 -8.65497719e-03 -6.86180770e-01 -2.48062983e-01
1.02979569e-02 6.10636413e-01 -3.28556806e-01 -4.62556005e-01
-4.03132826e-01 -1.01311576e+00 7.78377235e-01 1.13429844e+00
5.84469557e-01 -1.00143516e+00 -1.31592882e+00 7.28295207e-01
4.21776205e-01 -1.13526905e+00 2.77767360e-01 5.82132220e-01
-1.19566011e+00 -8.24062407e-01 -1.22173429e-01 -9.69147682e-01
5.57207465e-01 -3.46245229e-01 1.09775531e+00 -1.21931300e-01
-5.10819733e-01 8.18499029e-01 -2.18128875e-01 -1.23765576e+00
1.16899744e-01 -4.07711715e-02 5.19744158e-01 -2.52123892e-01
8.08154881e-01 -9.96631980e-01 -5.13469577e-01 1.26076996e-01
-3.63174915e-01 -5.18747330e-01 4.55263048e-01 7.45485723e-01
3.97712350e-01 9.16874826e-01 8.37771356e-01 -5.96474826e-01
7.76925921e-01 -6.33225262e-01 -6.02971435e-01 -9.40100700e-02
-9.01877224e-01 -1.98538929e-01 8.46278787e-01 -9.59421039e-01
-7.46032536e-01 3.05281490e-01 -3.85236144e-02 -1.94414124e-01
-3.40741664e-01 2.87088633e-01 -1.47262201e-01 -3.82943273e-01
8.80834758e-01 -2.29986131e-01 -1.46206096e-01 -3.71171534e-01
-4.78374660e-01 1.28393018e+00 8.56100500e-01 -1.98671520e-01
4.36047703e-01 2.72392333e-01 2.52768159e-01 -1.02280939e+00
1.07729621e-02 8.77683982e-03 -2.01773882e-01 -5.70184469e-01
4.42738712e-01 -9.59783077e-01 -1.20550132e+00 7.22256362e-01
-7.34495878e-01 -6.17502928e-01 -2.09201232e-01 6.35296583e-01
-2.55316049e-01 -9.80186611e-02 -5.86897194e-01 -8.18444729e-01
-9.40112114e-01 -5.32883286e-01 5.70703387e-01 9.23498392e-01
-2.97890544e-01 -1.14695895e+00 -1.57141358e-01 -2.57094324e-01
1.06558704e+00 4.63233262e-01 2.55809695e-01 -3.91442239e-01
7.51133785e-02 -8.99032831e-01 5.07091701e-01 3.78218442e-01
1.99721336e-01 -1.47243157e-01 -1.04670978e+00 -4.26450931e-02
4.91744250e-01 2.39939734e-01 9.79529619e-02 6.09595954e-01
1.36617827e+00 -1.06062964e-01 -6.62754416e-01 5.43666303e-01
1.70106149e+00 6.03543222e-01 4.20455277e-01 4.50398475e-01
6.05551898e-01 -2.42908180e-01 1.14277460e-01 3.75407487e-01
6.34820536e-02 1.78210393e-01 5.38761139e-01 1.06425397e-01
2.07292259e-01 1.91168830e-01 4.19817895e-01 5.09624004e-01
-6.32482171e-02 -1.41918719e-01 -7.18676269e-01 2.76053965e-01
-1.83255470e+00 -6.51357889e-01 -1.98317096e-01 1.93449557e+00
2.21204877e-01 5.49986064e-01 1.03944756e-01 7.39638209e-01
7.57786810e-01 -7.37902403e-01 -1.19343305e+00 -5.70876956e-01
6.66078031e-02 4.05776381e-01 9.11640942e-01 1.43054813e-01
-1.00190139e+00 2.17060208e-01 6.58749199e+00 1.82710201e-01
-1.65250683e+00 2.30493322e-01 3.59669566e-01 -1.45029619e-01
4.11042422e-01 -5.93774557e-01 -3.85165453e-01 8.13752592e-01
1.26388276e+00 2.32586309e-01 3.64154577e-01 1.46089423e+00
4.67939582e-03 -1.79237366e-01 -1.05279386e+00 1.49697220e+00
1.51106283e-01 -8.82019997e-01 -9.60350633e-01 -5.84459066e-01
3.93291771e-01 3.59133154e-01 -3.18971813e-01 3.71925503e-01
-1.96372047e-01 -6.36509895e-01 2.18969911e-01 6.01365268e-01
3.47651333e-01 -6.01159215e-01 1.09732139e+00 4.52831149e-01
-1.03077245e+00 -5.37432075e-01 -2.71556795e-01 -8.32193136e-01
-2.36553863e-01 9.06565309e-01 -4.75951612e-01 1.80368021e-01
1.36597729e+00 4.98369545e-01 -2.81922132e-01 7.71361232e-01
5.94742261e-02 7.02386200e-01 -1.01064003e+00 -6.62087500e-01
-3.20092261e-01 1.38573945e-01 4.37159121e-01 8.74628007e-01
6.86942101e-01 1.62763204e-02 -2.08712742e-01 2.80157208e-01
5.59052713e-02 -3.24370623e-01 -6.78701162e-01 3.10347408e-01
5.09136021e-01 1.48119652e+00 -7.82863438e-01 -1.76267192e-01
-5.58442354e-01 9.84624267e-01 -1.99500412e-01 2.98137486e-01
-8.60215068e-01 -9.98808980e-01 3.19985986e-01 1.49723575e-01
4.15465832e-01 -2.69587278e-01 -6.42636478e-01 -7.45136738e-01
3.60550046e-01 -8.87112916e-02 1.50986061e-01 -9.39929903e-01
-1.28793180e+00 4.48096812e-01 -4.93583292e-01 -1.28385568e+00
-2.27181152e-01 -1.13529325e+00 -9.26367998e-01 2.36737922e-01
-1.13013613e+00 -4.95248646e-01 -7.05018580e-01 6.53225780e-01
2.85743088e-01 -5.88599861e-01 1.04442382e+00 5.20873368e-01
-7.95361280e-01 6.43381476e-01 2.78882980e-01 1.96971148e-01
3.24572831e-01 -1.05681813e+00 -1.33071858e-02 7.22534299e-01
-1.66267902e-01 3.77425291e-02 6.39865279e-01 -4.26941872e-01
-1.85924208e+00 -8.71174812e-01 1.67343199e-01 -1.94436401e-01
6.08598292e-01 -1.01999961e-01 -8.39721560e-01 6.43723130e-01
2.62133658e-01 5.49330235e-01 7.00433373e-01 -1.11810215e-01
4.46121961e-01 -7.06402004e-01 -1.63648427e+00 5.07912934e-01
1.07485712e+00 -2.44053662e-01 -3.69666189e-01 2.47840777e-01
3.65427397e-02 -5.83382249e-01 -1.14337373e+00 6.83482468e-01
6.76701427e-01 -5.80565155e-01 7.28376210e-01 -1.98806643e-01
-2.90605307e-01 -3.87789994e-01 1.59932718e-01 -1.15063357e+00
-3.75041455e-01 -8.15611303e-01 -7.61696637e-01 1.03842306e+00
2.49694079e-01 -9.73084152e-01 1.14324307e+00 7.07309425e-01
-1.59811646e-01 -7.73343325e-01 -9.69510615e-01 -7.79934108e-01
-5.62404156e-01 -7.53266156e-01 4.38065469e-01 7.35459924e-01
3.37340355e-01 6.25840008e-01 -7.14750513e-02 5.42527616e-01
2.95682967e-01 -6.11418903e-01 4.37205166e-01 -1.53354394e+00
-5.27842820e-01 -1.69068366e-01 -1.17733204e+00 -3.10959995e-01
-2.89573610e-01 -1.82599321e-01 3.38897854e-02 -1.22509205e+00
-7.87651300e-01 -2.74887294e-01 -7.62043595e-01 4.55948830e-01
3.47233601e-02 3.48565161e-01 -5.85070014e-01 -2.09422097e-01
-3.40362519e-01 1.92116737e-01 1.26818404e-01 -1.98042825e-01
-5.84998846e-01 3.08966219e-01 -2.82579064e-01 1.24006402e+00
1.31929791e+00 -7.10299313e-01 -4.67946500e-01 -6.33376658e-01
5.16012549e-01 -1.62122414e-01 1.61329493e-01 -2.16432333e+00
7.05669045e-01 4.54685718e-01 1.03068292e+00 -2.76770592e-01
9.98670086e-02 -1.83294082e+00 4.29922998e-01 9.35172856e-01
4.11893606e-01 4.15599465e-01 5.72347283e-01 5.40305734e-01
1.93794414e-01 -3.52654129e-01 2.42452100e-01 2.20351294e-01
-7.02023983e-01 -1.36510894e-01 -6.45938814e-01 -6.58573210e-01
1.25462842e+00 -5.13601601e-01 4.51059919e-03 -3.74890417e-01
-7.97464848e-01 -3.96558046e-02 1.70838125e-02 5.21404624e-01
3.58840317e-01 -1.25214124e+00 1.30170891e-02 4.59813148e-01
-2.71610171e-01 4.42535356e-02 -1.00492463e-01 5.49319088e-01
-4.37777102e-01 -2.09964700e-02 -4.42894667e-01 -8.14779282e-01
-1.12313151e+00 3.56203884e-01 8.83094609e-01 3.20558041e-01
-8.05307269e-01 5.20553827e-01 -1.32923031e+00 -1.96567267e-01
7.72256732e-01 -5.63234389e-01 -2.22871974e-01 -4.35783446e-01
4.51046377e-01 7.36879408e-01 4.86912549e-01 1.57349795e-01
-4.56475735e-01 8.96113157e-01 5.61808109e-01 1.49307922e-01
1.33002222e+00 6.81803077e-02 4.46870059e-01 8.41766715e-01
7.82604873e-01 -2.95798004e-01 -1.15892804e+00 4.65678424e-03
1.17773645e-01 2.30927691e-01 3.06454271e-01 -8.92976403e-01
-1.13326061e+00 2.39841461e-01 1.60062206e+00 4.50974792e-01
1.64342320e+00 -6.94814384e-01 1.02927339e+00 1.07657599e+00
4.69360769e-01 -1.73208845e+00 8.22182298e-02 1.84247300e-01
2.57829815e-01 -1.24090612e+00 -1.02228969e-01 -3.31333913e-02
-1.05877563e-01 1.28540099e+00 6.34974003e-01 -5.36442339e-01
1.49712002e+00 1.08038008e+00 1.82830140e-01 3.13082606e-01
-3.71857226e-01 2.61634111e-01 -3.07981044e-01 1.11470985e+00
5.15399203e-02 -1.00786239e-03 -8.34658965e-02 1.20902646e+00
3.57254408e-02 6.81693435e-01 2.99608797e-01 1.26570368e+00
-3.53375763e-01 -6.67436302e-01 -4.70510244e-01 5.65778732e-01
-9.64806855e-01 4.56143469e-01 -2.38833502e-02 5.55638909e-01
5.26192605e-01 1.12869799e+00 5.35947680e-01 -9.88669038e-01
5.67463040e-01 3.49259615e-01 3.02585423e-01 7.14946538e-02
-7.33467042e-01 -4.05355483e-01 -1.70764163e-01 -5.56207180e-01
-3.53797168e-01 -2.38049433e-01 -1.54476678e+00 -2.41637719e-03
-4.47010756e-01 -2.50414133e-01 1.43225217e+00 1.00113952e+00
9.23261762e-01 1.25966358e+00 5.54245889e-01 -9.21469390e-01
-2.89822936e-01 -1.32212174e+00 -4.53839719e-01 2.79819250e-01
1.93972275e-01 -6.38606131e-01 -2.91858912e-01 -1.12189148e-02] | [8.047245979309082, 2.631040573120117] |
ac76eb4b-5e0f-45d7-b873-752ff17b72de | a-novel-decentralized-algorithm-for | 2305.04124 | null | https://arxiv.org/abs/2305.04124v1 | https://arxiv.org/pdf/2305.04124v1.pdf | A Novel Decentralized Algorithm for Coordinating the Optimal Power and Traffic Flows with EVs based on Variable Inner Loop Selection | The electric power distribution network (PDN) and the transportation network (TN) are generally operated/coordinated by different entities. However, they are coupled with each other due to electric vehicle charging stations (EVCSs). This paper proposes to coordinate the operation of the two systems via a fully decentralized framework where the PDN and TN operators solve their own operation problems by sharing only limited information. Nevertheless, the operation problems generally are in mixed-integer programming (MIP) form. To the best of our knowledge, the most existing decentralized/distributed optimization algorithms, such as the alternating direction method of multipliers (ADMM), are not always guaranteed to converge for such MIP problems. Therefore, a novel fully decentralized optimization algorithm is proposed, whose contributions include: 1) it is applicable to MIP problems with convergence and optimality guaranteed for mild assumptions, 2) it only requires limited information exchange between PDN and TN operators, which will help preserve the privacy of the two systems and reduce the investment in building communication channels, and 3) it is fully decentralized so that all the computations are carried out by PDN operator and TN coordinator only. Simulations on the test cases show the efficiency and efficacy of the proposed framework and algorithm. | ['QiFeng Li', 'Santosh Sharma'] | 2023-05-06 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-3.69097888e-01 3.16401601e-01 -2.28514612e-01 -2.43885756e-01
-3.77759129e-01 -7.52745092e-01 1.35506064e-01 1.25275284e-01
-2.08759919e-01 1.08863854e+00 -2.60341287e-01 -2.68516064e-01
-4.70905453e-01 -8.85063529e-01 -4.10946399e-01 -1.20105135e+00
-3.75900388e-01 7.12603748e-01 -3.73364657e-01 -1.37458518e-01
2.92461067e-02 3.58440518e-01 -6.94714606e-01 -3.25875133e-01
1.20997715e+00 1.18733609e+00 3.14487278e-01 -1.70872495e-01
7.18882754e-02 5.18157482e-01 -1.67071879e-01 -2.99543977e-01
9.31378067e-01 1.48061842e-01 -7.19092131e-01 3.12304109e-01
-5.05189657e-01 -6.18130207e-01 -4.73567367e-01 1.42640972e+00
4.53565687e-01 9.92093459e-02 4.64440845e-02 -2.11665726e+00
-5.43537438e-01 7.26428449e-01 -9.58245635e-01 -4.92141813e-01
-2.98607512e-03 -2.24611998e-01 1.12227333e+00 -5.92055082e-01
5.14327288e-01 8.76280606e-01 3.73131961e-01 1.95758954e-01
-1.17715800e+00 -7.98810542e-01 3.12094361e-01 2.10645229e-01
-1.63799322e+00 -1.79859713e-01 8.17437470e-01 3.58751789e-02
4.27826703e-01 4.21907097e-01 7.40436733e-01 8.14801306e-02
-5.13693020e-02 1.05443811e+00 8.58010054e-01 -9.62569714e-02
3.75276327e-01 5.18596053e-01 -9.05891955e-02 2.47520611e-01
6.82102501e-01 -4.64014292e-01 6.68534487e-02 -6.01186574e-01
3.86945009e-01 2.16826931e-01 -5.93735576e-01 -7.69574046e-01
-1.20229387e+00 8.83325815e-01 2.61968344e-01 2.77076244e-01
-7.13430583e-01 -6.85766563e-02 4.69967961e-01 4.99111950e-01
4.19598877e-01 -1.58766255e-01 -4.34661806e-01 3.08972001e-01
-6.92965508e-01 2.97986060e-01 1.18845940e+00 1.51619077e+00
7.90458620e-01 -1.07219229e-02 9.61284712e-02 3.13706130e-01
5.68431258e-01 6.03145778e-01 -2.82638669e-01 -9.56642628e-01
1.01968873e+00 5.16233385e-01 4.70072508e-01 -1.49729729e+00
-3.98590177e-01 -3.66660655e-01 -1.21680593e+00 -6.22545276e-03
9.00970399e-02 -8.28048527e-01 1.51588455e-01 1.57031846e+00
4.46223617e-01 -1.33891359e-01 8.87726173e-02 1.00984335e+00
2.60036707e-01 1.07716894e+00 -2.48838216e-01 -8.45569491e-01
9.97621238e-01 -1.02874279e+00 -1.24554408e+00 -2.94424053e-02
6.57130837e-01 -5.75202286e-01 -2.18245015e-01 5.43043725e-02
-1.35128963e+00 1.82463318e-01 -1.00975335e+00 1.20847225e-01
-3.04631293e-01 2.58912593e-01 7.48298228e-01 5.81055820e-01
-1.19552386e+00 1.14845768e-01 -6.80491507e-01 -1.41513363e-01
4.96477336e-01 9.30396259e-01 -4.09818053e-01 -2.57621437e-01
-1.21461940e+00 7.33280480e-01 4.30335790e-01 8.19221735e-01
-6.03354216e-01 -6.24698877e-01 -6.14469945e-01 2.07926854e-01
4.54417408e-01 -5.67910910e-01 8.54120374e-01 -7.43326008e-01
-1.22550893e+00 2.53544956e-01 -1.30521983e-01 -2.42991671e-01
8.27257931e-01 5.35528183e-01 -2.02606231e-01 6.50235713e-02
2.86657363e-01 2.13329986e-01 1.59730777e-01 -1.32883728e+00
-9.59270775e-01 -3.82833183e-01 1.44310802e-01 3.24638635e-01
-5.40887535e-01 -1.80557668e-01 -3.49382550e-01 -3.64331186e-01
2.35282511e-01 -8.00968826e-01 -5.97874880e-01 4.54003543e-01
-6.06540978e-01 -1.40219122e-01 1.30237675e+00 -7.25008309e-01
9.79767263e-01 -2.13180089e+00 1.66032910e-01 8.92400563e-01
3.76102924e-01 1.13571107e-01 -5.44087514e-02 8.42390358e-01
2.54637431e-02 -1.73828870e-01 -1.63379773e-01 -6.44106388e-01
5.19820809e-01 6.68767393e-01 -4.03146371e-02 1.17668843e+00
-4.84605759e-01 6.96325183e-01 -8.37873161e-01 -3.66174996e-01
2.26718351e-01 1.23749778e-01 -4.65553820e-01 -3.71296071e-02
2.03855485e-01 4.60865170e-01 -9.74094450e-01 6.61054611e-01
1.58193493e+00 1.10559747e-01 5.85102618e-01 -3.22087198e-01
-4.33795094e-01 -3.53274494e-01 -1.76515543e+00 1.42177713e+00
-1.30866960e-01 3.17656904e-01 1.32955420e+00 -1.52326179e+00
6.90411508e-01 8.41738403e-01 1.15381360e+00 -5.44694901e-01
2.46425226e-01 4.82869625e-01 -4.70875084e-01 -3.50604177e-01
2.85911590e-01 1.29237160e-01 -1.17486998e-01 6.88499093e-01
-2.30159014e-01 2.21515015e-01 1.73360571e-01 1.63525447e-01
7.30781257e-01 -8.04500639e-01 1.51469573e-01 -6.89883232e-01
8.11176181e-01 -1.45853788e-01 1.22733974e+00 5.16216218e-01
-2.20448837e-01 -1.46824673e-01 5.63274562e-01 -2.82185704e-01
-8.06008220e-01 -8.09617281e-01 -1.14785627e-01 3.66028160e-01
6.38160765e-01 -2.05878243e-01 -5.44095516e-01 -5.45381010e-01
3.09569240e-01 6.46278560e-01 -1.90903824e-02 3.29241186e-01
-2.39521757e-01 -8.42080355e-01 1.66526139e-01 1.13130145e-01
7.83720315e-01 -1.38207018e-01 2.05535233e-01 3.57860148e-01
-4.04213220e-01 -1.16757762e+00 -6.51513100e-01 -2.20466573e-02
-7.37116098e-01 -9.64736044e-01 -7.41656482e-01 -9.59418774e-01
1.42675328e+00 5.79805851e-01 5.01649439e-01 7.78700858e-02
1.44046411e-01 4.41576719e-01 -6.43853750e-03 -4.07354861e-01
2.23169625e-01 -9.60482880e-02 2.96974510e-01 4.70048219e-01
-4.86641824e-02 -7.74954438e-01 -5.37754536e-01 5.64632714e-01
-8.30423594e-01 -6.01431914e-02 3.79633874e-01 5.57402074e-01
5.02967715e-01 9.31262314e-01 5.42615533e-01 -8.59138548e-01
6.17893755e-01 -7.92142928e-01 -1.15818906e+00 3.26532722e-01
-6.89893782e-01 -3.83375883e-01 7.54720628e-01 1.66860074e-01
-1.09889185e+00 1.86732173e-01 3.74831647e-01 -2.42758796e-01
3.27876419e-01 3.91115546e-01 -7.55520403e-01 -6.53412521e-01
-4.24200684e-01 1.95769519e-01 1.69713080e-01 -2.42732972e-01
2.81112254e-01 7.30362117e-01 5.37153818e-02 -4.69472319e-01
1.34132934e+00 8.25643957e-01 4.23583746e-01 -5.26688695e-01
-1.55106038e-01 -1.86659217e-01 -4.13658619e-01 -1.08137436e-01
4.72867012e-01 -1.22553766e+00 -1.40032017e+00 4.02579069e-01
-1.29866838e+00 5.43290377e-01 -3.16236801e-02 6.59579277e-01
-2.33400583e-01 7.23851621e-01 -3.75991583e-01 -1.03306043e+00
-3.91461641e-01 -1.12754989e+00 4.85290676e-01 3.34061176e-01
3.39846760e-01 -1.32081676e+00 -3.99641126e-01 4.37564135e-01
8.42421949e-01 3.37388396e-01 5.63803375e-01 -3.97004843e-01
-9.54493642e-01 -4.32553202e-01 -2.91532040e-01 2.63019383e-01
2.46545732e-01 -2.95742571e-01 -5.17359734e-01 -9.40789163e-01
2.61517286e-01 1.09300144e-01 -1.37953594e-01 3.41889441e-01
9.89132106e-01 -8.65962088e-01 -6.02478802e-01 5.44622123e-01
1.72344899e+00 1.95424363e-01 4.14382964e-01 1.04433961e-01
5.42787552e-01 6.45064473e-01 7.29321897e-01 9.47298825e-01
9.13689733e-01 5.32107234e-01 7.62065172e-01 -4.56095040e-01
1.05473745e+00 1.63457349e-01 2.74498194e-01 1.12318337e+00
-1.14562176e-01 -3.42784584e-01 -3.16133708e-01 5.65378845e-01
-2.38023829e+00 -7.65863061e-01 -2.83518583e-01 2.02177477e+00
3.76645297e-01 -4.93989110e-01 -2.42769912e-01 -3.16931009e-02
9.91672516e-01 1.11882739e-01 -4.09160137e-01 -2.19642296e-01
-2.38051012e-01 -4.14040834e-01 1.21161616e+00 2.69326955e-01
-9.34043765e-01 3.98189873e-02 5.26133490e+00 8.98199737e-01
-6.77379131e-01 2.63466448e-01 4.81047958e-01 -5.56563810e-02
-5.18823743e-01 3.34309042e-01 -4.02667135e-01 5.63563645e-01
3.68775904e-01 -8.15703213e-01 7.87601531e-01 8.63462627e-01
7.76943386e-01 6.13620766e-02 -8.92511606e-01 1.01024902e+00
-3.80850196e-01 -1.18350184e+00 -3.59664649e-01 4.55410570e-01
1.10086536e+00 -7.32689118e-03 -1.97862566e-01 -2.72157229e-02
1.60569981e-01 -5.01394093e-01 5.73875487e-01 2.91011870e-01
7.66560733e-02 -1.22714293e+00 9.97859359e-01 4.97276783e-01
-1.43572056e+00 -4.64699268e-01 -4.78278965e-01 -3.06683127e-02
6.71816766e-01 8.08989525e-01 -1.65304810e-01 1.30759430e+00
6.38905525e-01 6.77835703e-01 3.20415407e-01 1.10948193e+00
-4.07840498e-02 2.31689047e-02 -5.99317431e-01 3.41896825e-02
4.24943179e-01 -9.24800634e-01 7.26817966e-01 5.07250726e-01
3.87232125e-01 3.16554576e-01 5.83582997e-01 9.07550693e-01
-6.06198534e-02 1.02543585e-01 -4.69385624e-01 1.02627613e-01
7.25807130e-01 1.58319116e+00 -3.22800756e-01 1.33266807e-01
-7.12066770e-01 6.76998138e-01 -1.01346619e-01 5.32504737e-01
-8.16257477e-01 -7.01808870e-01 8.92977536e-01 -1.66356370e-01
3.07744354e-01 -2.57827789e-01 -2.54469424e-01 -1.00504076e+00
4.66621697e-01 -3.39332819e-01 3.87754828e-01 -3.37919116e-01
-1.36952031e+00 1.00893900e-01 -1.16774701e-01 -1.42564499e+00
2.28343368e-01 -2.84679085e-01 -8.09893847e-01 9.41316247e-01
-1.77566290e+00 -1.05602503e+00 -3.21221277e-02 9.40132737e-01
-3.35247546e-01 -7.95700178e-02 6.84310436e-01 9.80030179e-01
-7.60671079e-01 4.90626454e-01 9.53752935e-01 1.89179122e-01
1.11909129e-01 -9.96174097e-01 -6.08608842e-01 8.61958146e-01
-7.49287009e-01 4.15948123e-01 6.28294289e-01 -3.20221126e-01
-2.17629743e+00 -1.15115678e+00 1.03815913e+00 5.02569914e-01
6.99785292e-01 -4.78478611e-01 -3.22592974e-01 7.73441076e-01
6.08375728e-01 2.75866300e-01 5.44197321e-01 -3.99735063e-01
6.06298447e-01 -8.29842031e-01 -1.68817091e+00 3.96204799e-01
4.61986452e-01 -2.36447021e-01 4.31316309e-02 9.27192569e-01
3.08776110e-01 -3.51561576e-01 -9.77966189e-01 9.12564471e-02
-5.02682365e-02 -3.38576972e-01 8.32106471e-01 -8.64913128e-03
-4.34696943e-01 -8.27773273e-01 -3.18271995e-01 -1.24757266e+00
-2.90973097e-01 -1.01578426e+00 7.92526230e-02 1.64789367e+00
3.62621695e-01 -1.26175308e+00 6.23867929e-01 1.24666750e+00
1.75003856e-01 -3.94725889e-01 -1.47500026e+00 -5.39719522e-01
-3.58066976e-01 -1.59225956e-01 8.16287994e-01 1.24173927e+00
3.36967140e-01 -2.69385457e-01 -6.22866035e-01 7.43643463e-01
9.85629141e-01 2.15374321e-01 5.00412166e-01 -9.21439171e-01
2.82419170e-03 2.36354768e-02 -2.78092027e-01 -9.56502140e-01
2.43636146e-01 -1.00521731e+00 -1.18071221e-01 -1.65229499e+00
3.87813034e-03 -8.67008984e-01 -3.86570781e-01 6.34160519e-01
6.12793207e-01 -3.65076542e-01 2.82675624e-01 2.62766123e-01
-7.72594392e-01 8.22385013e-01 1.17294490e+00 -3.93051445e-01
4.05894667e-02 3.85089666e-01 -9.06909883e-01 3.51050556e-01
5.59648812e-01 -5.03789306e-01 -6.21177554e-01 -8.29796016e-01
5.32265231e-02 4.51808482e-01 2.58976370e-02 -4.89213973e-01
7.71228194e-01 -3.01087171e-01 -2.28371292e-01 -7.73839474e-01
1.30703911e-01 -1.86959100e+00 6.02129579e-01 4.45503086e-01
2.49741450e-01 1.55260842e-02 -1.43531576e-01 5.32269180e-01
-3.38452995e-01 -1.90625817e-01 5.10531783e-01 8.31555054e-02
-3.01461369e-01 6.85673058e-01 -5.34977734e-01 -4.53649104e-01
1.73201954e+00 7.50649571e-02 -1.53071642e-01 -6.84571445e-01
-4.75271076e-01 1.40631366e+00 9.71187055e-02 4.48883288e-02
1.99869469e-01 -1.52674663e+00 -7.17986047e-01 5.37526496e-02
-4.52520818e-01 2.33618528e-01 3.98161739e-01 1.37661612e+00
-4.18111861e-01 7.75453806e-01 1.04650864e-02 -3.69240582e-01
-8.59924376e-01 4.14945841e-01 1.09675951e-01 -2.24627316e-01
-3.24703723e-01 3.21310073e-01 -4.26849909e-02 -7.92550147e-01
3.13377529e-01 8.84502307e-02 4.82452847e-02 2.65896678e-01
3.21402043e-01 8.36582243e-01 -1.65608525e-01 -6.71531081e-01
-4.18293178e-01 3.30291599e-01 -3.44520509e-02 2.31265351e-01
1.74774814e+00 -8.42828631e-01 -1.05142677e+00 -4.54106599e-01
1.41701341e+00 -2.33040210e-02 -6.34960473e-01 -3.48054290e-01
-4.57170397e-01 -7.71554112e-01 2.60592252e-01 -1.39745668e-01
-2.06637049e+00 1.70386776e-01 2.38424405e-01 3.44548136e-01
1.43935573e+00 -4.29619551e-01 8.06412995e-01 4.95417506e-01
9.80455518e-01 -1.45424056e+00 -1.07356417e+00 1.71277046e-01
4.09761459e-01 -7.44401276e-01 2.10237876e-01 -6.17397845e-01
-3.83833021e-01 1.23643720e+00 3.83556634e-01 7.51290545e-02
9.27732110e-01 3.69821548e-01 -3.59625161e-01 2.53942385e-02
-6.38899863e-01 5.16904056e-01 -4.16783184e-01 4.47854549e-01
-2.99271256e-01 2.08116814e-01 -8.00372422e-01 8.17648411e-01
3.56595963e-01 1.05581284e-01 6.31441474e-01 1.30775416e+00
5.06067388e-02 -1.26327622e+00 -4.15332437e-01 3.07999790e-01
-3.42605710e-01 2.77030230e-01 2.18713924e-01 6.40395284e-01
2.39419803e-01 1.12058496e+00 -1.69130132e-01 2.13146001e-01
4.03060049e-01 -6.93638623e-01 -2.64183700e-01 7.04335049e-02
-3.03516865e-01 -1.13409579e-01 1.18426196e-02 -5.20422995e-01
-4.08811569e-01 -8.64678204e-01 -1.20005774e+00 -8.48574579e-01
-5.50383151e-01 7.88512945e-01 6.58457100e-01 9.06684220e-01
3.88550282e-01 1.12172544e-01 1.47857690e+00 -7.21994102e-01
-7.14313745e-01 -2.05976695e-01 -1.34814703e+00 -1.84087679e-01
1.00947261e-01 -3.28707010e-01 -5.40481389e-01 -5.93063533e-01] | [6.169833660125732, 4.915868759155273] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.