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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
856449c2-c4a2-457f-864c-c3d92299225e | geann-scalable-graph-augmentations-for-multi | 2307.03595 | null | https://arxiv.org/abs/2307.03595v1 | https://arxiv.org/pdf/2307.03595v1.pdf | GEANN: Scalable Graph Augmentations for Multi-Horizon Time Series Forecasting | Encoder-decoder deep neural networks have been increasingly studied for multi-horizon time series forecasting, especially in real-world applications. However, to forecast accurately, these sophisticated models typically rely on a large number of time series examples with substantial history. A rapidly growing topic of interest is forecasting time series which lack sufficient historical data -- often referred to as the ``cold start'' problem. In this paper, we introduce a novel yet simple method to address this problem by leveraging graph neural networks (GNNs) as a data augmentation for enhancing the encoder used by such forecasters. These GNN-based features can capture complex inter-series relationships, and their generation process can be optimized end-to-end with the forecasting task. We show that our architecture can use either data-driven or domain knowledge-defined graphs, scaling to incorporate information from multiple very large graphs with millions of nodes. In our target application of demand forecasting for a large e-commerce retailer, we demonstrate on both a small dataset of 100K products and a large dataset with over 2 million products that our method improves overall performance over competitive baseline models. More importantly, we show that it brings substantially more gains to ``cold start'' products such as those newly launched or recently out-of-stock. | ['Michael W. Mahoney', 'Ronak Metha', 'Vincent Quenneville-Belair', 'Shankar Ramasubramanian', 'Malcolm Wolff', 'Sitan Yang'] | 2023-07-07 | null | null | null | null | ['time-series-forecasting'] | ['time-series'] | [ 5.01257554e-03 8.81417021e-02 -4.28305209e-01 -7.17948794e-01
-5.28639615e-01 -7.08169341e-01 5.57435751e-01 3.62094283e-01
1.66351765e-01 3.52085471e-01 4.64092731e-01 -7.15582132e-01
-6.89015025e-03 -9.85453069e-01 -1.02033317e+00 -2.20886707e-01
-6.55963659e-01 5.32942832e-01 -2.67660528e-01 -7.46484637e-01
-1.40758485e-01 2.21540004e-01 -1.05474770e+00 1.78693399e-01
5.33334255e-01 1.49534345e+00 1.27059624e-01 5.63140750e-01
-1.40698656e-01 1.09160185e+00 -3.44684571e-01 -6.91133678e-01
5.42364240e-01 6.68445928e-03 -4.54716325e-01 1.47668853e-01
1.39034107e-01 -4.45258111e-01 -5.73091269e-01 8.09749186e-01
2.58071184e-01 7.28099942e-02 2.69785225e-01 -1.28592920e+00
-9.94718909e-01 1.19957590e+00 -5.76114714e-01 3.57258290e-01
-6.11586161e-02 1.99915349e-01 1.30616426e+00 -4.44141656e-01
6.86435401e-01 9.78296220e-01 8.64759147e-01 1.15590386e-01
-1.29273045e+00 -6.27129555e-01 6.07436538e-01 1.95902959e-01
-8.30974638e-01 -2.87654102e-01 1.22226441e+00 -4.02362674e-01
1.29098046e+00 -1.52476281e-01 5.97809315e-01 1.03245544e+00
4.22525555e-01 9.06233907e-01 4.73577559e-01 1.51544720e-01
1.40118256e-01 -2.90633261e-01 1.35555714e-01 2.95935631e-01
-1.30723447e-01 1.79719254e-01 -2.06746712e-01 2.61726715e-02
6.66304469e-01 3.65215033e-01 -8.70467871e-02 6.86157793e-02
-1.11035204e+00 1.16031051e+00 7.88026810e-01 1.15595452e-01
-7.57917106e-01 3.99241179e-01 8.45441401e-01 7.03818917e-01
1.01184452e+00 5.73550940e-01 -9.43182468e-01 -2.68077672e-01
-7.25661278e-01 2.03871682e-01 1.11330295e+00 1.09597266e+00
4.15548772e-01 5.09364128e-01 3.00229371e-01 6.76000416e-01
8.84933397e-02 3.36266786e-01 5.92995763e-01 -3.15548837e-01
7.82960832e-01 4.93091762e-01 -9.37592685e-02 -1.34325266e+00
-7.29734600e-01 -9.61288333e-01 -1.36624789e+00 -3.68940324e-01
1.66786224e-01 -2.62053519e-01 -1.08366525e+00 1.79715872e+00
8.07463005e-03 1.27449557e-01 -4.17388342e-02 8.42188060e-01
5.06376982e-01 1.13202345e+00 -1.58619404e-01 -2.84108788e-01
1.04886734e+00 -1.16927338e+00 -6.99403286e-01 -3.73209149e-01
9.10343587e-01 -5.79407573e-01 7.96788871e-01 5.17782450e-01
-9.39975619e-01 -5.55152535e-01 -9.88632321e-01 -2.87265591e-02
-5.85099101e-01 -1.89498067e-01 1.10640752e+00 -7.68725276e-02
-1.02826691e+00 8.65482628e-01 -6.95533872e-01 -1.10286459e-01
2.75717705e-01 4.01751578e-01 -1.74869686e-01 -1.11010022e-01
-1.58034527e+00 7.83228219e-01 3.24407935e-01 2.36706510e-01
-7.84776568e-01 -8.74445617e-01 -9.30601239e-01 3.01979661e-01
4.14208829e-01 -4.64110225e-01 1.48943532e+00 -1.00703120e+00
-1.26119721e+00 1.70023844e-01 2.58280426e-01 -1.07218564e+00
1.97430000e-01 -1.46917582e-01 -1.14202917e+00 -5.20329297e-01
-2.10935801e-01 1.60434872e-01 7.55600512e-01 -7.02325344e-01
-4.89069164e-01 -3.58451724e-01 -2.75404323e-02 -1.33626416e-01
-3.16487759e-01 -1.24574453e-01 -2.15459436e-01 -1.00995290e+00
-1.90457061e-01 -9.10239458e-01 -6.65452540e-01 -5.51637471e-01
-4.94305998e-01 -3.30991328e-01 6.28561914e-01 -7.88056970e-01
1.44582331e+00 -2.04725909e+00 -8.13520849e-02 2.57500142e-01
5.57955801e-01 5.83588816e-02 -4.76571143e-01 8.65661740e-01
-2.47243732e-01 5.30561022e-02 8.16878602e-02 -2.84827143e-01
2.27898732e-01 2.78041214e-01 -8.05731833e-01 2.00019479e-01
2.37055495e-01 1.26769567e+00 -9.64096606e-01 2.62860358e-01
-5.08274399e-02 4.00809914e-01 -2.22591117e-01 1.67062193e-01
-8.45848978e-01 1.63516209e-01 -2.56556720e-01 5.34372509e-01
4.83087778e-01 -9.88699555e-01 2.41103202e-01 -1.70393750e-01
2.58515090e-01 4.83521342e-01 -6.55856788e-01 1.58456421e+00
-5.77615440e-01 5.87629259e-01 -3.31816852e-01 -1.07666528e+00
1.12935054e+00 8.68443847e-02 5.64120829e-01 -8.61831069e-01
4.15800139e-02 1.71211869e-01 8.36308822e-02 -4.70787026e-02
6.77139878e-01 1.97087172e-02 -2.14732945e-01 5.32662809e-01
2.04270389e-02 1.72127992e-01 2.67566264e-01 2.42809504e-01
1.22456646e+00 -2.65879631e-01 2.75932997e-01 9.40766707e-02
-5.72898276e-02 7.73524046e-02 4.66058999e-01 3.44626844e-01
2.12708950e-01 2.84864277e-01 7.28608608e-01 -9.83222187e-01
-1.38833928e+00 -5.58100104e-01 3.05881143e-01 1.30682611e+00
-3.36089104e-01 -5.83917916e-01 -1.03950724e-01 -6.76407695e-01
4.54132259e-01 7.05861807e-01 -5.72024167e-01 -1.20067656e-01
-6.19130075e-01 -7.70368099e-01 1.23435808e-02 7.87379205e-01
7.70005509e-02 -8.30145836e-01 1.42794266e-01 8.39772642e-01
1.92180678e-01 -1.15052652e+00 -6.84935212e-01 3.39009434e-01
-8.28341067e-01 -6.08222365e-01 -7.80027628e-01 -5.40869594e-01
3.27879101e-01 1.26503438e-01 1.65466011e+00 -3.85307044e-01
2.00969428e-01 -9.40953940e-02 -4.34912294e-01 -4.54070210e-01
-6.05283558e-01 5.66791713e-01 1.83777735e-02 9.97257233e-02
4.51740563e-01 -8.77830267e-01 -6.37689292e-01 1.27733260e-01
-8.43696237e-01 1.55893281e-01 5.50895870e-01 6.59636378e-01
4.66032058e-01 1.11366563e-01 1.12328148e+00 -1.07567561e+00
8.27254534e-01 -1.07749009e+00 -9.18162704e-01 1.04214378e-01
-1.13771367e+00 1.28386512e-01 1.30251861e+00 -5.79411566e-01
-3.90700191e-01 -1.15448341e-01 -3.22266296e-02 -6.39668822e-01
3.28110039e-01 1.36560774e+00 2.48844296e-01 3.07530433e-01
3.78643841e-01 1.87746003e-01 5.47269685e-03 -5.47672808e-01
7.02462196e-01 4.31411982e-01 6.02613032e-01 -8.91216323e-02
7.03306139e-01 -2.30007581e-02 9.48235951e-03 -1.09629326e-01
-9.04836476e-01 -3.11837137e-01 -3.02776277e-01 -1.15213439e-01
2.09534809e-01 -1.17527747e+00 -7.19102502e-01 4.51218933e-01
-1.00714862e+00 -5.67144096e-01 -2.53545940e-01 4.23106909e-01
-3.22272032e-01 -6.08324818e-02 -9.12249744e-01 -5.71218610e-01
-7.06233919e-01 -7.91194975e-01 9.38674927e-01 -1.08749390e-01
-2.12195143e-02 -1.40631115e+00 2.54775789e-02 -2.15671927e-01
8.54787290e-01 5.07588387e-01 1.12391567e+00 -1.05268526e+00
-5.45152128e-01 -5.20295382e-01 -3.04289848e-01 4.03093159e-01
1.65069968e-01 -1.58447817e-01 -6.52867377e-01 -4.46385533e-01
-2.63263226e-01 -1.20603062e-01 8.01549375e-01 4.68492657e-01
1.26920414e+00 -7.36303329e-01 -1.70403078e-01 6.35429025e-01
1.49488521e+00 3.46847147e-01 2.92027563e-01 -4.72302511e-02
1.11819744e+00 4.54005986e-01 3.80338877e-01 7.13712633e-01
9.17703807e-01 5.43748438e-01 6.34860635e-01 -5.41462861e-02
1.66148096e-01 -5.05528033e-01 3.74858826e-01 1.63187921e+00
7.26240426e-02 -7.19758868e-01 -1.06056023e+00 6.42046392e-01
-2.08897996e+00 -6.81573093e-01 8.14242885e-02 1.99236941e+00
7.49260545e-01 4.06445205e-01 3.37048531e-01 -5.86110055e-02
5.18993378e-01 5.53265035e-01 -1.06182921e+00 -4.99569923e-01
-1.21214360e-01 -7.01736733e-02 8.09491217e-01 2.18191788e-01
-1.08586264e+00 6.84107184e-01 5.89095497e+00 5.99893570e-01
-1.46941066e+00 -1.18207686e-01 1.05089724e+00 -1.87592402e-01
-5.88380814e-01 -2.37582341e-01 -7.28336275e-01 7.44121492e-01
1.60663521e+00 -6.34604454e-01 6.54764593e-01 1.15199614e+00
8.80480707e-02 6.50367916e-01 -1.39437163e+00 8.75207782e-01
-1.02863126e-01 -1.71569276e+00 -9.14712921e-02 2.19506204e-01
8.34973574e-01 5.88490427e-01 3.36988658e-01 6.45344079e-01
9.08831060e-01 -1.09346890e+00 5.27619839e-01 4.15869117e-01
7.29649007e-01 -8.09448004e-01 7.59225011e-01 1.63603097e-01
-1.42170370e+00 -2.35116780e-01 -3.30938220e-01 -1.36478454e-01
6.03015244e-01 1.08675194e+00 -8.82360399e-01 6.32483840e-01
3.14586282e-01 1.24349904e+00 -3.36876452e-01 6.82986796e-01
2.96762228e-01 7.62346745e-01 -4.46407378e-01 -9.83170271e-02
5.98958611e-01 -5.09513095e-02 4.49928865e-02 8.64622831e-01
5.33080399e-01 -1.14184164e-01 2.50413865e-01 5.58836162e-01
-4.35827643e-01 1.95482187e-02 -8.27928901e-01 -5.59179008e-01
2.49998301e-01 1.18791604e+00 -4.31531459e-01 -4.11395729e-01
-6.69345081e-01 6.95244670e-01 4.88266319e-01 4.14325893e-01
-9.12836373e-01 -1.63110286e-01 5.21933615e-01 1.67509198e-01
5.70562065e-01 -3.35045487e-01 -2.00284086e-02 -1.14884377e+00
8.53179321e-02 -9.35620606e-01 4.25672084e-01 -7.00146616e-01
-2.00431395e+00 8.93062413e-01 -3.60305309e-01 -1.39019942e+00
-8.16806316e-01 -6.04792297e-01 -3.38216692e-01 8.79366696e-01
-1.61442959e+00 -1.34655869e+00 -5.49932616e-03 2.66408712e-01
4.37985003e-01 -2.62157977e-01 5.03943920e-01 5.04891157e-01
-4.14343596e-01 4.79627341e-01 3.68346989e-01 1.20910935e-01
3.14792782e-01 -1.32540464e+00 1.27941871e+00 7.35654831e-01
1.97839782e-01 2.41676509e-01 6.33937359e-01 -7.41393566e-01
-1.81184340e+00 -1.54291499e+00 1.02812767e+00 -1.24713637e-01
1.22279000e+00 -5.58003962e-01 -9.44887161e-01 1.10788834e+00
2.24619240e-01 1.11204781e-01 5.68637431e-01 4.95821863e-01
-5.08562744e-01 -4.54270065e-01 -5.94174743e-01 3.24961782e-01
9.65356231e-01 -4.86326486e-01 1.57067366e-02 8.21442962e-01
1.27694321e+00 -5.01850486e-01 -1.44561958e+00 3.40309203e-01
3.60175610e-01 -3.71412575e-01 6.41742408e-01 -9.11499262e-01
8.42375755e-01 2.41004303e-01 -1.07710034e-01 -1.78132689e+00
-6.08592808e-01 -1.07379591e+00 -7.12474942e-01 1.09221387e+00
6.76081538e-01 -7.94111729e-01 6.24471366e-01 5.99626780e-01
-2.47571617e-01 -9.87638295e-01 -3.89566869e-01 -9.12387192e-01
-5.03760576e-02 -6.53925002e-01 1.23515201e+00 1.21714556e+00
3.10693402e-02 5.18636405e-01 -7.42122710e-01 3.95291150e-02
1.82684883e-01 5.02859950e-01 7.18848944e-01 -1.20537198e+00
-3.51548016e-01 -5.20311356e-01 -4.24706936e-01 -1.23753393e+00
2.37269774e-02 -1.03424978e+00 -2.41116315e-01 -1.42423165e+00
-3.49598706e-01 -4.34842169e-01 -6.46482587e-01 4.17659104e-01
1.09236695e-01 -2.92498201e-01 2.19978303e-01 1.63983598e-01
-4.94151473e-01 5.69854736e-01 1.22286582e+00 -1.58821374e-01
-5.43452837e-02 1.17882304e-01 -7.85758913e-01 2.81151772e-01
7.67675221e-01 -1.34390116e-01 -6.61429763e-01 -5.79077780e-01
8.53333414e-01 3.93329710e-01 6.45406172e-02 -5.26389122e-01
2.35682622e-01 -5.09824306e-02 1.39806211e-01 -7.70711243e-01
1.05453044e-01 -8.44988763e-01 3.76198173e-01 1.11113787e-01
-4.75892097e-01 7.82749891e-01 1.33993149e-01 8.06264639e-01
-4.17765528e-01 4.07325745e-01 8.09195116e-02 -1.17878988e-01
-1.00503433e+00 9.92814660e-01 8.33527073e-02 -1.01763517e-01
8.49095404e-01 4.18178409e-01 -6.97977424e-01 -1.06743515e+00
-6.79306507e-01 5.55015028e-01 5.67015633e-02 7.70877421e-01
2.83881694e-01 -1.63424957e+00 -8.31630647e-01 1.42823815e-01
2.58219808e-01 3.50844078e-02 1.87397540e-01 6.47901893e-01
-2.92175472e-01 5.78190744e-01 1.45656064e-01 -3.82414490e-01
-4.73123074e-01 1.16356587e+00 4.02634852e-02 -8.03101420e-01
-6.78749561e-01 7.70230234e-01 1.08033195e-01 -3.01166266e-01
1.55273318e-01 -8.89646232e-01 -3.42197232e-02 8.08543488e-02
3.67598742e-01 6.67712092e-02 1.86763406e-01 -3.87920469e-01
-1.52262487e-02 6.89068809e-02 -4.19357955e-01 3.73720735e-01
1.80780292e+00 -1.26326188e-01 8.91363025e-02 6.35678947e-01
1.55511618e+00 -5.04046142e-01 -1.39454627e+00 -6.65979803e-01
1.06689915e-01 -1.58599690e-01 5.56821898e-02 -8.37323368e-01
-1.72909999e+00 7.84309685e-01 1.37908101e-01 1.02486646e+00
1.29242039e+00 4.85290550e-02 1.38710427e+00 2.01541975e-01
3.70789409e-01 -9.61771071e-01 -2.33146250e-01 5.69103897e-01
1.04349864e+00 -1.40306091e+00 -1.71486452e-01 -1.48178130e-01
-8.06713283e-01 1.14551389e+00 2.97612756e-01 -2.17708439e-01
8.92212987e-01 3.25788349e-01 -4.56519723e-02 -3.45241606e-01
-1.49611509e+00 -2.86350474e-02 4.98099327e-01 3.05843294e-01
3.70079041e-01 3.56141269e-01 1.86840266e-01 7.48033822e-01
-5.10143101e-01 1.99757472e-01 4.42283541e-01 5.03311753e-01
1.16985083e-01 -8.24384689e-01 3.82192880e-01 1.03067398e+00
-3.40052873e-01 -2.73154438e-01 7.75846913e-02 7.47714460e-01
-4.67592537e-01 7.48133123e-01 3.84066135e-01 -7.73911536e-01
2.42900237e-01 -9.47876126e-02 -7.15222722e-03 -3.31416130e-01
-6.11452341e-01 1.90483794e-01 3.95152390e-01 -6.73936427e-01
-5.89351878e-02 -4.96374398e-01 -8.09379101e-01 -8.62808168e-01
-1.96601316e-01 -1.64672971e-01 7.75203705e-01 6.76588774e-01
7.88621604e-01 7.56839812e-01 1.07422817e+00 -7.19604611e-01
-8.37603390e-01 -1.06278980e+00 -7.64151573e-01 5.27981937e-01
5.76829553e-01 -2.71665215e-01 -1.50795415e-01 5.87246604e-02] | [6.883018970489502, 2.8020589351654053] |
624f5507-d902-4f56-8e74-5ad67f08cae1 | revisiting-modality-imbalance-in-multimodal | 2302.12589 | null | https://arxiv.org/abs/2302.12589v2 | https://arxiv.org/pdf/2302.12589v2.pdf | Revisiting Modality Imbalance In Multimodal Pedestrian Detection | Multimodal learning, particularly for pedestrian detection, has recently received emphasis due to its capability to function equally well in several critical autonomous driving scenarios such as low-light, night-time, and adverse weather conditions. However, in most cases, the training distribution largely emphasizes the contribution of one specific input that makes the network biased towards one modality. Hence, the generalization of such models becomes a significant problem where the non-dominant input modality during training could be contributing more to the course of inference. Here, we introduce a novel training setup with regularizer in the multimodal architecture to resolve the problem of this disparity between the modalities. Specifically, our regularizer term helps to make the feature fusion method more robust by considering both the feature extractors equivalently important during the training to extract the multimodal distribution which is referred to as removing the imbalance problem. Furthermore, our decoupling concept of output stream helps the detection task by sharing the spatial sensitive information mutually. Extensive experiments of the proposed method on KAIST and UTokyo datasets shows improvement of the respective state-of-the-art performance. | ['Ciarán Eising', 'Martin Glavin', 'Edward Jones', 'Ujjwal Bhattacharya', 'Jonathan Horgan', 'Ganesh Sistu', 'Sudip Das', 'Arindam Das'] | 2023-02-24 | null | null | null | null | ['pedestrian-detection'] | ['computer-vision'] | [ 1.33530766e-01 -1.03231989e-01 -1.28468364e-01 -3.95600438e-01
-6.07945144e-01 -2.97153115e-01 6.70750082e-01 1.90076292e-01
-5.36973596e-01 8.19358647e-01 9.25546214e-02 -1.17708027e-01
-2.46188357e-01 -7.29512811e-01 -6.23377264e-01 -1.22520733e+00
4.37338740e-01 -1.49859069e-02 3.13771725e-01 -2.84476221e-01
1.14191258e-02 3.04714829e-01 -1.97079968e+00 3.60143721e-01
9.52471614e-01 9.96298850e-01 2.37882301e-01 4.74566460e-01
-1.88391260e-03 7.02558219e-01 -4.30070877e-01 -3.83171886e-01
1.95502266e-01 -2.63747483e-01 -5.42830229e-02 -3.84343714e-02
5.85050106e-01 -1.76851913e-01 -3.85791183e-01 9.93699908e-01
8.52421105e-01 2.71283567e-01 6.82761550e-01 -1.46086907e+00
-5.17052263e-02 3.92694503e-01 -9.04310763e-01 1.92749500e-01
-9.22470093e-02 2.35363632e-01 7.57701337e-01 -7.25319088e-01
1.38159901e-01 1.04828691e+00 2.42023438e-01 3.47854584e-01
-8.97716284e-01 -6.02028906e-01 2.85319090e-01 4.86266762e-01
-1.25517523e+00 -5.28085887e-01 9.63856101e-01 -4.02684629e-01
3.52734864e-01 1.69038653e-01 3.77926677e-01 9.48800862e-01
-4.09527384e-02 9.77845252e-01 1.13348663e+00 -3.09630364e-01
-3.65535840e-02 6.59577549e-01 7.40956068e-02 5.16643047e-01
2.11194724e-01 1.03236742e-01 -5.81422985e-01 2.64035970e-01
1.71790257e-01 -1.27768712e-02 -2.01816022e-01 -2.21679062e-01
-9.33036327e-01 4.94019687e-01 4.61096555e-01 1.35665417e-01
-7.62728900e-02 -9.80089605e-02 3.83158177e-01 6.39953390e-02
1.26633421e-01 -1.21620551e-01 -1.14288755e-01 8.48483592e-02
-8.13796401e-01 1.31316856e-01 2.07177058e-01 6.77047789e-01
8.25659335e-01 7.36208484e-02 -2.63478100e-01 9.05834377e-01
4.37609464e-01 6.22345090e-01 1.88040018e-01 -4.85025883e-01
8.70529294e-01 7.29958594e-01 -2.05397978e-02 -8.99010777e-01
-6.10849202e-01 -9.10054207e-01 -9.12369251e-01 3.28213811e-01
7.47929335e-01 -4.11651790e-01 -7.21668959e-01 1.83374369e+00
4.63597506e-01 8.78476351e-02 2.36692742e-01 1.15221334e+00
1.01232398e+00 5.51883459e-01 4.02669728e-01 2.76457518e-02
1.51284695e+00 -7.92889714e-01 -5.50948322e-01 -3.44605863e-01
1.52782723e-01 -6.68445408e-01 6.35459125e-01 2.84341186e-01
-6.84611261e-01 -9.39377487e-01 -1.26889217e+00 8.97508785e-02
-5.51549852e-01 4.68222141e-01 4.13205653e-01 7.25180745e-01
-4.50538665e-01 6.24752752e-02 -3.51662070e-01 -4.72298354e-01
3.05523753e-01 1.64673865e-01 -3.33000362e-01 -1.32632345e-01
-1.30502820e+00 8.78766060e-01 4.47300106e-01 4.38098073e-01
-6.61469579e-01 -4.61191565e-01 -5.91250956e-01 3.16320844e-02
3.25128883e-01 -6.61798954e-01 6.94529593e-01 -9.97373760e-01
-1.22951853e+00 5.26959956e-01 -2.55805373e-01 -2.88322330e-01
7.31855452e-01 -1.50385618e-01 -5.32037854e-01 1.35900214e-01
5.98927448e-03 7.84003437e-01 1.04951704e+00 -1.44325209e+00
-1.20936048e+00 -4.80988175e-01 2.62513936e-01 5.10285139e-01
-4.55220431e-01 -3.22787285e-01 -3.94973993e-01 -2.88702428e-01
-6.78118467e-02 -7.93333113e-01 6.18141666e-02 -1.83469370e-01
-4.30764109e-01 -1.51844546e-01 8.62039983e-01 -4.58695322e-01
1.14099956e+00 -2.48949790e+00 1.41105577e-01 1.95897952e-01
4.54628617e-02 1.10101081e-01 3.90891125e-03 3.71006638e-01
-3.84684503e-02 -3.48745376e-01 -1.24002412e-01 -4.86627549e-01
-1.16315372e-01 1.54827133e-01 -1.81986824e-01 5.78010619e-01
4.69364226e-01 3.74094039e-01 -6.90484464e-01 -6.00765765e-01
4.17074233e-01 5.55147707e-01 -9.37418044e-02 2.16336951e-01
2.11507082e-01 6.85960352e-01 -3.46425682e-01 6.39330089e-01
9.82919157e-01 3.02991092e-01 -7.30667710e-02 -4.89907593e-01
-3.41027558e-01 -2.10372403e-01 -1.31086981e+00 1.38851583e+00
-3.15576673e-01 7.70331919e-01 4.53742631e-02 -1.09173620e+00
6.84080064e-01 2.13737831e-01 2.93929338e-01 -9.47854161e-01
3.81893218e-01 1.05811022e-01 2.27074400e-01 -7.22478688e-01
5.47849774e-01 -1.25167221e-01 1.31787941e-01 -1.88795812e-02
-5.65134287e-02 4.22311038e-01 3.03895503e-01 8.40223208e-02
3.53091151e-01 1.31047860e-01 1.11083761e-01 -3.38457078e-02
7.72465765e-01 -1.38491109e-01 4.96602923e-01 6.65762663e-01
-3.71182144e-01 6.02195680e-01 3.88411969e-01 5.39053902e-02
-6.54985487e-01 -9.75979805e-01 -2.48295963e-01 1.30495572e+00
4.63916719e-01 9.08510461e-02 -5.32119989e-01 -5.46440899e-01
5.27637126e-03 6.54038250e-01 -5.32850206e-01 -3.85175824e-01
-4.94563460e-01 -1.07961261e+00 7.13839412e-01 6.01821125e-01
8.92154455e-01 -5.49142182e-01 -8.53529990e-01 -1.58656329e-01
-4.56942528e-01 -1.05922151e+00 -5.23916632e-02 3.32193702e-01
-5.81437230e-01 -1.04687226e+00 -7.44021595e-01 -5.43423414e-01
5.56172729e-01 5.20125866e-01 6.45019531e-01 -1.67351395e-01
-6.39591739e-02 1.17482118e-01 -2.06683561e-01 -5.61093271e-01
1.00508961e-03 6.37938008e-02 -1.67617157e-01 5.19740880e-01
3.05072814e-01 -3.72566044e-01 -7.23500788e-01 2.42898986e-01
-7.84998298e-01 1.30692288e-01 6.23322129e-01 8.01541924e-01
1.87743396e-01 3.35298538e-01 7.91920841e-01 -5.05444586e-01
1.90542579e-01 -6.29089832e-01 -2.81140208e-01 1.64752126e-01
-2.49238342e-01 5.01490943e-02 6.34597063e-01 -3.70398313e-01
-1.55679393e+00 8.77888799e-02 1.54811726e-03 -8.73550475e-02
-4.73717093e-01 2.10013971e-01 -5.55774391e-01 9.87221226e-02
4.92480069e-01 2.44078726e-01 -4.84008864e-02 -2.38060385e-01
2.66159624e-01 6.99181259e-01 3.95722419e-01 -4.45970923e-01
9.42232192e-01 7.09095359e-01 3.28813702e-01 -9.54509497e-01
-4.82632577e-01 -5.41894376e-01 -4.27172095e-01 -5.53575099e-01
8.01966310e-01 -1.08258891e+00 -7.09968507e-01 5.19245207e-01
-9.15631831e-01 3.00444782e-01 -6.13455474e-02 5.34673035e-01
-1.74374562e-02 2.53280789e-01 -9.99924615e-02 -1.17739606e+00
-2.37256871e-03 -1.30468011e+00 8.81721079e-01 7.24243999e-01
2.84228742e-01 -7.93348432e-01 -1.16492197e-01 4.02052760e-01
3.55720282e-01 2.48360023e-01 9.39261377e-01 -4.84171093e-01
-4.81293917e-01 -9.62760597e-02 -6.10894680e-01 2.61602223e-01
-5.75266257e-02 -4.42345552e-02 -1.73002708e+00 -9.45149884e-02
-2.67445415e-01 -4.76494059e-02 1.21898222e+00 3.72433782e-01
5.78988016e-01 3.37430388e-01 -3.13524395e-01 3.19246262e-01
1.39750552e+00 1.12944998e-01 5.45349121e-01 2.75597662e-01
7.71728396e-01 1.07927728e+00 5.96891463e-01 3.99692833e-01
6.31280065e-01 6.61008835e-01 6.13606811e-01 -4.46290493e-01
-2.49815896e-01 -1.04327248e-02 3.18337530e-01 2.56855369e-01
-1.62126511e-01 -3.66270483e-01 -4.76183057e-01 4.95267421e-01
-2.09215331e+00 -1.03695893e+00 -2.54580170e-01 2.34255672e+00
3.66541207e-01 1.91201493e-01 2.24756643e-01 3.14847976e-01
7.55121589e-01 1.55139014e-01 -4.54438627e-01 -4.37499993e-02
-4.61526006e-01 -3.27446848e-01 5.86164057e-01 1.82058439e-01
-1.18435037e+00 2.75681347e-01 5.00133801e+00 8.62807572e-01
-1.29848695e+00 7.93981701e-02 5.90484560e-01 -2.03852639e-01
-6.24167919e-02 -2.01202184e-01 -8.42333853e-01 6.29649878e-01
5.75763822e-01 2.75009185e-01 2.93622706e-02 5.84571481e-01
4.00042146e-01 -6.23906553e-01 -9.93509054e-01 1.13047314e+00
1.00090116e-01 -4.96048778e-01 -1.20525554e-01 -7.28146434e-02
4.35909122e-01 -1.25052556e-01 3.41081768e-01 3.81064117e-01
-3.94445777e-01 -8.48084509e-01 9.52257872e-01 6.33873105e-01
5.03215134e-01 -9.65746522e-01 9.50643301e-01 4.91627574e-01
-1.35322070e+00 -3.21905524e-01 -2.60776043e-01 9.40476954e-02
1.59963220e-01 6.60884202e-01 -5.83298445e-01 8.41160417e-01
5.62092364e-01 5.08631229e-01 -8.79890740e-01 1.15367520e+00
-1.17310815e-01 4.55710977e-01 -3.52619082e-01 1.57824144e-01
1.40002251e-01 -4.94197495e-02 7.26527274e-01 1.24609518e+00
1.39427349e-01 -2.74079353e-01 5.19437939e-02 5.10642767e-01
2.27611601e-01 1.45325195e-02 -6.55242205e-01 3.45886439e-01
1.61991596e-01 1.37559927e+00 -5.59444189e-01 -1.82406411e-01
-5.23501337e-01 5.89240193e-01 1.05875842e-01 6.84747458e-01
-1.14821064e+00 -4.32433575e-01 5.68507016e-01 1.53030664e-01
3.10716361e-01 -1.11010959e-02 -5.18820107e-01 -9.25118268e-01
1.64134964e-01 -4.99807805e-01 4.37328458e-01 -4.77291912e-01
-1.05875576e+00 4.25855309e-01 1.49186313e-01 -1.45847058e+00
7.28690624e-02 -6.34874761e-01 -5.86539745e-01 9.94133413e-01
-2.05035830e+00 -1.26441395e+00 -5.48843265e-01 6.17385149e-01
4.79415536e-01 -1.96999818e-01 1.44029915e-01 8.82455945e-01
-7.97588527e-01 7.71540821e-01 -7.57073015e-02 -1.71280310e-01
1.10806203e+00 -1.02923465e+00 -5.91001213e-01 1.12128377e+00
-3.53217363e-01 3.70159566e-01 7.12787628e-01 -2.80971944e-01
-1.28271687e+00 -9.69618857e-01 5.66759109e-01 -2.69750953e-01
1.80516198e-01 -1.21531934e-01 -6.05436802e-01 6.48561260e-03
4.08037007e-01 -2.46158719e-01 5.63200295e-01 -1.39910029e-02
-3.11140120e-01 -5.97022474e-01 -1.05688977e+00 4.19158190e-01
7.20301270e-01 -3.66886526e-01 -2.91197985e-01 -1.08233377e-01
1.90428972e-01 -2.36761853e-01 -2.73590922e-01 4.77200001e-01
7.61168182e-01 -1.11150980e+00 7.60835886e-01 -3.56486678e-01
4.37254548e-01 -7.23752201e-01 -2.43703485e-01 -1.18551838e+00
-1.68559223e-01 -3.20975855e-02 -8.72600004e-02 1.46421027e+00
4.49760765e-01 -7.64550209e-01 3.59113365e-01 4.21634823e-01
-6.96156248e-02 -4.87280071e-01 -1.05645049e+00 -2.36303121e-01
-2.75698870e-01 -4.15337920e-01 3.66817981e-01 6.09074771e-01
-1.18890814e-01 3.74496609e-01 -5.28462231e-01 3.86466473e-01
6.91271007e-01 -4.45962586e-02 7.87370861e-01 -1.12293446e+00
-1.91264614e-01 -4.11249310e-01 -4.04236704e-01 -1.01502991e+00
-1.91421777e-01 -6.07707262e-01 1.32251263e-01 -1.41425073e+00
2.98829257e-01 -3.53674054e-01 -4.50222284e-01 3.12796205e-01
-3.55616242e-01 2.95843333e-01 2.66133577e-01 2.62071984e-03
-4.64410514e-01 5.13744175e-01 1.21769845e+00 -3.14381629e-01
-1.06026709e-01 8.29518139e-02 -8.46971929e-01 5.04081130e-01
5.65720320e-01 -1.88444585e-01 -6.41954720e-01 -5.04990280e-01
3.21485341e-01 -2.82339334e-01 5.63612282e-01 -1.18096197e+00
4.52349722e-01 -7.02886209e-02 5.21886170e-01 -5.93376219e-01
3.66488248e-01 -1.18749905e+00 -4.88508493e-02 5.45942672e-02
-1.54847845e-01 -3.28521937e-01 3.66955280e-01 5.93184650e-01
-4.15325284e-01 -1.02383450e-01 8.87504697e-01 2.38717720e-01
-8.75435889e-01 -1.88766822e-01 -3.96069288e-01 -1.19280867e-01
1.00516391e+00 -4.40372616e-01 -6.05044961e-01 -3.22727472e-01
-4.89940852e-01 4.96572196e-01 4.70090881e-02 4.94539142e-01
4.23187554e-01 -1.23704243e+00 -6.71727657e-01 1.76386237e-01
1.93622515e-01 -1.31238237e-01 7.30886221e-01 1.33001566e+00
1.48377061e-01 3.81963551e-01 -2.46914670e-01 -7.73562670e-01
-1.10443473e+00 1.78667396e-01 4.29737478e-01 6.35979772e-02
-2.30071798e-01 6.09336913e-01 5.42362154e-01 -6.37982041e-02
3.66361439e-01 -1.67026788e-01 -7.08509445e-01 6.16487682e-01
4.87993747e-01 6.91292763e-01 1.39549881e-01 -9.48210835e-01
-4.83370006e-01 6.18779957e-01 1.46077260e-01 -1.03632109e-02
8.42505991e-01 -3.95285964e-01 2.13688761e-01 3.76676530e-01
9.75930691e-01 8.75598043e-02 -1.28116894e+00 -1.16243444e-01
-5.11759579e-01 -2.04908296e-01 3.09616793e-03 -7.54944563e-01
-1.12703013e+00 1.05599475e+00 9.98134851e-01 6.24738075e-02
1.28520787e+00 -3.09512377e-01 5.55745125e-01 2.29205891e-01
8.79941210e-02 -1.15291035e+00 -2.78671145e-01 3.17495942e-01
4.51483577e-01 -1.60272527e+00 -1.23344369e-01 -3.22656214e-01
-7.29890525e-01 1.15519261e+00 7.16865420e-01 2.34743088e-01
5.63671172e-01 1.01479143e-01 1.44997910e-01 5.69722317e-02
-3.64045739e-01 -5.73320627e-01 5.08569062e-01 5.24603844e-01
2.97457278e-01 8.35561156e-02 -1.47086218e-01 6.02512002e-01
2.62176692e-01 -3.31150442e-01 1.37216687e-01 7.97492385e-01
-5.15433669e-01 -7.89690316e-01 -5.28464496e-01 1.63251445e-01
-1.93318307e-01 3.33410911e-02 -4.54012938e-02 7.42265940e-01
7.32909381e-01 1.21387482e+00 -1.04376435e-01 -4.56755131e-01
5.04744411e-01 2.07589105e-01 3.63503188e-01 -1.51885018e-01
-5.21091640e-01 1.76484182e-01 1.32902727e-01 -2.75419652e-01
-7.20918179e-01 -6.06163085e-01 -1.16975975e+00 -4.83126082e-02
-3.57994527e-01 1.89611837e-02 6.23447657e-01 1.09217727e+00
2.83394665e-01 8.46524119e-01 5.83455265e-01 -7.90312886e-01
-1.49755687e-01 -9.47461247e-01 -4.17114288e-01 4.38965052e-01
5.01531601e-01 -1.09040737e+00 -3.50747615e-01 -1.13784047e-02] | [9.789265632629395, -1.295159101486206] |
92ea3e1e-0256-4659-a28b-e72470ca3f27 | a-bi-directional-transformer-for-musical | 1907.02698 | null | https://arxiv.org/abs/1907.02698v1 | https://arxiv.org/pdf/1907.02698v1.pdf | A Bi-directional Transformer for Musical Chord Recognition | Chord recognition is an important task since chords are highly abstract and descriptive features of music. For effective chord recognition, it is essential to utilize relevant context in audio sequence. While various machine learning models such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs) have been employed for the task, most of them have limitations in capturing long-term dependency or require training of an additional model. In this work, we utilize a self-attention mechanism for chord recognition to focus on certain regions of chords. Training of the proposed bi-directional Transformer for chord recognition (BTC) consists of a single phase while showing competitive performance. Through an attention map analysis, we have visualized how attention was performed. It turns out that the model was able to divide segments of chords by utilizing adaptive receptive field of the attention mechanism. Furthermore, it was observed that the model was able to effectively capture long-term dependencies, making use of essential information regardless of distance. | ['Dokyun Kim', 'Sungwook Jeon', 'Kyoyun Choi', 'Jonghun Park', 'Jonggwon Park'] | 2019-07-05 | null | null | null | null | ['chord-recognition'] | ['audio'] | [ 2.48042539e-01 -1.18120492e-01 2.23824024e-01 4.78346609e-02
-4.67253953e-01 -6.70229435e-01 4.05101359e-01 2.31212884e-01
-5.46011984e-01 4.28971529e-01 4.49480772e-01 -8.20026174e-02
-3.49973291e-01 -7.31410682e-01 -3.90452236e-01 -5.32759309e-01
-1.76918924e-01 1.82038322e-01 3.77122670e-01 -5.60162306e-01
8.23203743e-01 7.65476584e-01 -1.72549236e+00 3.69079888e-01
3.46737534e-01 9.17351246e-01 5.20788968e-01 9.21886921e-01
-5.23416586e-02 9.68287230e-01 -9.22738969e-01 3.10938269e-01
-4.92083468e-02 -6.05409503e-01 -1.03211236e+00 -3.23505223e-01
9.35832784e-02 2.52919376e-01 -2.79440761e-01 5.57270169e-01
6.95702672e-01 4.82660592e-01 6.07982576e-01 -6.62434697e-01
-5.04826248e-01 8.14558029e-01 -1.05989613e-02 4.73168999e-01
3.44859779e-01 -1.85243294e-01 1.24595296e+00 -8.97355914e-01
5.16462386e-01 5.17799497e-01 7.26835430e-01 4.69608456e-01
-8.73685598e-01 -5.93325555e-01 2.62816902e-03 5.86933553e-01
-1.28433371e+00 -2.61515439e-01 1.24359035e+00 -5.06333232e-01
1.28155458e+00 2.95092374e-01 9.13232505e-01 8.51727545e-01
1.56486556e-01 6.42781019e-01 8.08286428e-01 -7.42967188e-01
1.47568643e-01 -7.21561387e-02 1.41776100e-01 7.03019556e-03
-5.38731098e-01 -8.19593482e-03 -8.67357612e-01 1.98110864e-01
8.93499672e-01 -1.24497131e-01 -3.91430765e-01 8.85180980e-02
-9.22164679e-01 5.82086682e-01 6.82941794e-01 9.90518332e-01
-6.54469728e-01 1.45238757e-01 6.81108177e-01 2.42131382e-01
3.87221910e-02 8.32834184e-01 -2.33699784e-01 -5.87852955e-01
-9.87420857e-01 8.61623809e-02 4.92995828e-01 5.39358675e-01
3.23219657e-01 3.93114686e-01 -2.23604083e-01 8.23528767e-01
-3.68903846e-01 -1.51737005e-01 8.06749821e-01 -5.93091369e-01
2.13694021e-01 6.35127604e-01 -3.55151266e-01 -1.18439424e+00
-4.79363531e-01 -8.42103958e-01 -9.05298054e-01 3.20343047e-01
3.30712229e-01 2.79169291e-01 -7.11369395e-01 1.74288893e+00
-2.81741351e-01 1.20071568e-01 4.67011407e-02 9.57014680e-01
6.09325767e-01 5.90882182e-01 -7.26416260e-02 6.88566118e-02
1.23663008e+00 -7.90111959e-01 -6.72120094e-01 2.48605296e-01
3.05177987e-01 -9.07330692e-01 1.22545898e+00 5.07270455e-01
-1.19753349e+00 -1.06361318e+00 -1.18517578e+00 -1.90298393e-01
-5.15503466e-01 6.87201992e-02 2.02801511e-01 2.84152061e-01
-9.57710862e-01 9.51128125e-01 -4.49739188e-01 -2.01860517e-01
-2.45385151e-02 4.36935425e-01 -2.78924108e-01 6.37985826e-01
-1.28188920e+00 6.15618348e-01 4.26008105e-01 2.66433150e-01
-1.07700086e+00 -5.06347477e-01 -4.50256884e-01 4.63803858e-01
8.94675106e-02 -2.00873613e-01 9.17782247e-01 -1.04726446e+00
-1.63667572e+00 4.68588322e-01 1.95841879e-01 -5.93867660e-01
9.67165157e-02 -3.89836043e-01 -2.34974429e-01 3.42130780e-01
-3.05868119e-01 4.44381237e-01 6.65085554e-01 -8.12436640e-01
-5.02887189e-01 -3.09352845e-01 2.35092610e-01 2.32476979e-01
-5.73170841e-01 1.26265973e-01 -1.13826528e-01 -1.14605939e+00
1.02229528e-01 -1.01033449e+00 5.40802702e-02 -6.61362708e-01
-3.54172498e-01 -2.36418322e-01 8.85510921e-01 -8.53478312e-01
1.50494659e+00 -2.08187270e+00 3.50204766e-01 1.65077925e-01
-1.58020154e-01 4.42916244e-01 -2.33374629e-02 7.51598775e-01
-1.70586795e-01 -2.96348222e-02 -3.07724059e-01 1.66726783e-02
-2.90245831e-01 4.26076986e-02 -4.43627596e-01 2.51635611e-02
5.98347560e-02 6.79957211e-01 -4.51850086e-01 -1.64140478e-01
1.31395563e-01 7.34471202e-01 -5.94462574e-01 3.34449083e-01
-6.91222101e-02 5.73919058e-01 2.06255447e-02 4.00353104e-01
8.85105059e-02 1.73000619e-01 2.07487762e-01 -2.16283068e-01
-4.24138248e-01 5.27000964e-01 -9.47219610e-01 1.89947343e+00
-6.17250025e-01 7.96690643e-01 -1.81730092e-01 -1.20658863e+00
1.22893202e+00 7.49886930e-01 3.77506196e-01 -8.70659471e-01
1.66502818e-01 2.06119716e-01 4.12278682e-01 -3.38572383e-01
7.88813233e-01 -2.74335705e-02 -2.18843501e-02 4.90387619e-01
1.74178407e-01 2.18451872e-01 4.99464087e-02 -2.52906799e-01
9.51565206e-01 9.76227671e-02 1.20900407e-01 -2.16577739e-01
7.84821272e-01 -2.20614567e-01 3.95924985e-01 6.94779038e-01
1.36973947e-01 7.68614471e-01 1.78572297e-01 -5.97521067e-01
-1.04688144e+00 -5.32926857e-01 1.88243955e-01 1.32172596e+00
-2.59234905e-01 -5.79445124e-01 -5.63086569e-01 -3.38983595e-01
-4.51436013e-01 3.19886535e-01 -7.53532827e-01 -3.11109662e-01
-9.03450429e-01 -1.57735780e-01 1.00129163e+00 8.48997653e-01
3.59368056e-01 -1.92036200e+00 -1.19037688e+00 5.89772224e-01
-5.10786623e-02 -7.09728956e-01 -2.04263002e-01 6.46753132e-01
-7.26913214e-01 -9.88981426e-01 -9.24455345e-01 -8.28669071e-01
9.19053033e-02 1.23700565e-02 1.02152097e+00 6.64739162e-02
-5.43340921e-01 1.94338500e-01 -6.05311811e-01 -3.57862353e-01
-5.50400056e-02 6.38819039e-01 -6.87881038e-02 -6.75963834e-02
3.39646498e-03 -1.05426836e+00 -6.27835870e-01 2.10790068e-01
-7.73771822e-01 -7.17699006e-02 4.40022051e-01 8.46062005e-01
5.97631395e-01 -2.15249285e-01 9.99746084e-01 -5.97553611e-01
9.29870546e-01 -8.48179907e-02 -2.37370893e-01 1.40728533e-01
-3.87665272e-01 1.84479705e-03 8.44675839e-01 -4.40223813e-01
-7.40299106e-01 -5.71353696e-02 -4.48221624e-01 -3.91492128e-01
-2.44628370e-01 5.46041965e-01 -1.39775304e-02 -6.79200143e-02
5.11841774e-01 4.26807165e-01 -2.14484319e-01 -6.79120362e-01
-6.39277175e-02 8.22407365e-01 5.58792174e-01 -5.99889934e-01
2.03038827e-01 1.02867261e-01 1.01429885e-02 -1.01116896e+00
-6.07742071e-01 -4.69073236e-01 -8.41822743e-01 -4.20448273e-01
9.23775911e-01 -4.40280199e-01 -8.76242220e-01 3.66026551e-01
-9.57086802e-01 -2.57098734e-01 -4.96904701e-01 5.71968138e-01
-6.38344228e-01 9.77490284e-03 -7.57013261e-01 -1.01977062e+00
-5.72493970e-01 -9.44228768e-01 5.33819914e-01 3.57263237e-01
-3.75737607e-01 -8.41413140e-01 3.40926707e-01 -1.56679004e-02
6.91154540e-01 8.74288306e-02 1.08739293e+00 -8.23824644e-01
-3.45977932e-01 -1.30207270e-01 1.04076058e-01 3.48510742e-01
2.52417356e-01 2.88488083e-02 -1.27372706e+00 -1.38287157e-01
-2.24383011e-01 -2.95206308e-01 9.37375903e-01 2.11997867e-01
1.38909721e+00 -1.33168288e-02 3.68353665e-01 5.33441067e-01
1.17241085e+00 6.60267651e-01 7.45613754e-01 4.82427329e-01
5.17637968e-01 6.24213099e-01 4.33048308e-01 3.48506391e-01
-5.12343757e-02 9.22831118e-01 5.14484465e-01 -5.58739081e-02
-1.59303397e-01 -2.71435261e-01 1.68792546e-01 1.12829757e+00
-7.29480326e-01 1.54231980e-01 -8.91990364e-01 6.50925994e-01
-1.74047434e+00 -1.05433321e+00 1.26657307e-01 2.20357466e+00
5.89447856e-01 3.07692528e-01 1.99922159e-01 8.74769330e-01
4.76850063e-01 1.30334109e-01 -3.18971246e-01 -7.03174889e-01
-1.86533630e-01 5.89637876e-01 -9.45771411e-02 3.13935548e-01
-8.87709081e-01 8.89820993e-01 6.23282957e+00 7.88590312e-01
-1.49404633e+00 -2.12004498e-01 1.56424548e-02 -2.24080253e-02
-1.79878116e-01 1.87488683e-02 -3.85827065e-01 1.81949705e-01
9.23232079e-01 1.30498797e-01 2.52764106e-01 6.11253679e-01
4.29655053e-02 1.95132226e-01 -7.15316236e-01 8.01941991e-01
9.46449116e-02 -1.41303229e+00 2.01226592e-01 -1.26526460e-01
3.25366646e-01 -4.55325067e-01 3.99993211e-02 3.26754719e-01
-4.71873671e-01 -1.18119180e+00 7.27990448e-01 7.71039188e-01
6.69125617e-01 -1.11491644e+00 7.82757580e-01 2.52426982e-01
-1.49145353e+00 -2.31145293e-01 -3.04427177e-01 -4.95970786e-01
1.22887129e-02 7.65625834e-02 -8.07875454e-01 5.92003882e-01
7.94843316e-01 6.49269521e-01 -6.52241051e-01 1.01821470e+00
-1.42517015e-01 7.50654817e-01 -2.34568510e-02 -5.09148911e-02
3.55829626e-01 -3.64417732e-02 5.45496583e-01 1.35903323e+00
4.14480150e-01 -1.47536203e-01 -1.48308471e-01 6.36634171e-01
1.42245159e-01 3.71469498e-01 -5.15017569e-01 -1.16951086e-01
2.53132164e-01 1.19030750e+00 -8.99830699e-01 -1.47473916e-01
1.02192596e-01 8.29370260e-01 3.45941663e-01 1.70212701e-01
-6.61149383e-01 -7.71956861e-01 5.49361706e-01 1.44967884e-01
5.50493479e-01 -2.24561140e-01 -2.19596282e-01 -6.13879502e-01
-1.63085371e-01 -7.15134919e-01 4.14347082e-01 -8.34279418e-01
-6.18673921e-01 1.14201760e+00 -5.18202782e-01 -1.24741423e+00
-3.65151972e-01 -4.09618825e-01 -7.27354467e-01 9.69459236e-01
-1.19206309e+00 -1.12901330e+00 -1.88518003e-01 8.15873146e-01
6.82949007e-01 -4.42687362e-01 1.31019962e+00 4.13847178e-01
-2.49049112e-01 5.10618746e-01 -3.50762457e-01 2.69581825e-01
5.42831719e-01 -1.33622766e+00 1.24643900e-01 5.40493131e-01
7.33592570e-01 7.21284688e-01 5.27105510e-01 -2.96115786e-01
-8.01366687e-01 -5.64971745e-01 7.43647993e-01 2.68904626e-01
3.71369451e-01 -3.13225955e-01 -1.15054095e+00 4.37781483e-01
4.74345088e-01 -2.74927378e-01 9.25340354e-01 3.50621879e-01
-4.43655252e-01 -1.65789530e-01 -3.63021582e-01 3.67690504e-01
8.45783651e-01 -9.80957508e-01 -7.93423176e-01 -3.01495284e-01
4.12777632e-01 -1.52171329e-01 -7.60653019e-01 4.74676371e-01
8.30384135e-01 -1.25043511e+00 8.40415061e-01 -6.79121614e-01
3.19695383e-01 -3.67273092e-01 -5.79374582e-02 -1.14195216e+00
-3.74196231e-01 -4.34016734e-01 -3.51254940e-02 1.26657093e+00
3.81237388e-01 -1.50815114e-01 7.65651047e-01 -1.43417761e-01
-4.56274241e-01 -6.67065024e-01 -8.65690768e-01 -5.74700236e-01
-1.12017818e-01 -4.60263163e-01 3.42559069e-01 8.10767770e-01
2.23854497e-01 4.25384670e-01 -4.58326548e-01 -1.56201646e-01
9.58208963e-02 4.39813495e-01 5.25311530e-01 -1.38476908e+00
-3.78334194e-01 -5.70903242e-01 -6.80056810e-01 -7.86774278e-01
-4.59515229e-02 -7.87082255e-01 -1.24759227e-01 -1.27883577e+00
-1.52721524e-01 -3.57113242e-01 -9.69161510e-01 4.47283506e-01
2.77136385e-01 5.06926537e-01 4.54497278e-01 2.63503850e-01
-2.38050744e-01 5.16774535e-01 9.69555378e-01 -6.61017699e-03
-3.74546885e-01 2.69505024e-01 -3.72853249e-01 6.99675143e-01
1.17480373e+00 -3.81556988e-01 -3.78629953e-01 -1.79700151e-01
3.43696862e-01 2.36390650e-01 2.58376122e-01 -1.37675548e+00
6.68325961e-01 3.28410029e-01 3.72950763e-01 -8.48318636e-01
5.83913326e-01 -6.55082583e-01 1.32199824e-01 4.15497512e-01
-7.36366332e-01 3.91244024e-01 2.88296282e-01 1.91331849e-01
-8.25097024e-01 -2.68860936e-01 4.38490182e-01 -2.07446530e-01
-7.74562478e-01 -2.28871733e-01 -5.92478037e-01 -1.04094937e-01
6.86192870e-01 -2.92243809e-01 3.22698861e-01 -3.86646092e-01
-1.14524508e+00 -4.27120835e-01 2.99868993e-02 4.27289426e-01
7.24435747e-01 -1.28951192e+00 -4.86088127e-01 2.48270735e-01
3.63667198e-02 -2.89088577e-01 5.22671819e-01 6.80775106e-01
-5.03734291e-01 6.48457885e-01 -7.39537120e-01 -4.39376056e-01
-1.44249928e+00 5.25083065e-01 3.12339634e-01 -2.69405067e-01
-7.36790240e-01 8.12725425e-01 -2.20462680e-02 -2.76532948e-01
6.18796468e-01 -4.39341396e-01 -7.65236318e-01 3.34680617e-01
5.66171885e-01 2.06319913e-01 1.33831784e-01 -6.59160733e-01
-2.76502490e-01 9.31866646e-01 6.90255687e-02 -1.67438447e-01
1.42155707e+00 8.76596496e-02 2.40678843e-02 7.57705331e-01
7.44388461e-01 2.54917860e-01 -1.09924221e+00 1.49228096e-01
3.19134235e-01 5.77375386e-03 -2.29991212e-01 -7.75351644e-01
-9.89527345e-01 1.28977323e+00 2.89796054e-01 5.03956676e-01
1.41442144e+00 -3.47554088e-01 6.15440071e-01 2.60766387e-01
2.75748521e-01 -1.17613697e+00 1.11152582e-01 7.42663562e-01
1.11239672e+00 -5.58016896e-01 -4.27430362e-01 1.35449246e-01
-5.34578204e-01 1.46240890e+00 4.32769090e-01 -4.33293015e-01
6.99521422e-01 3.23272765e-01 1.51615337e-01 -1.72632679e-01
-7.22082675e-01 -3.36325645e-01 4.30812895e-01 4.26808417e-01
7.49005497e-01 -1.94822684e-01 -1.58409715e-01 8.42987061e-01
-5.09282470e-01 -6.16465546e-02 3.91021967e-01 8.92123640e-01
-3.64142954e-01 -1.06590772e+00 -1.94660202e-01 -1.47841319e-01
-7.23891616e-01 -1.03598364e-01 -5.32206833e-01 8.45068693e-01
1.61397204e-01 6.64890409e-01 6.49387315e-02 -6.31908655e-01
4.49843049e-01 1.96466431e-01 3.59928995e-01 -5.04940212e-01
-1.37113988e+00 9.37814638e-02 -1.87378004e-01 -3.49993348e-01
-4.78410989e-01 -2.52745956e-01 -1.17179942e+00 1.90395817e-01
-2.56662965e-01 5.43638289e-01 6.25701725e-01 7.51257360e-01
2.67965645e-01 1.16505909e+00 4.99526888e-01 -9.32741880e-01
-1.93605721e-01 -1.22361743e+00 -7.38467038e-01 3.63187492e-01
2.02604860e-01 -4.19888556e-01 -2.05521673e-01 6.67039007e-02] | [15.803908348083496, 5.2602925300598145] |
f754b81f-623a-447d-b8c6-4e945688dd8d | analysis-of-evolutionary-program-synthesis | 2101.03172 | null | https://arxiv.org/abs/2101.03172v1 | https://arxiv.org/pdf/2101.03172v1.pdf | Analysis of Evolutionary Program Synthesis for Card Games | In this report, we inspect the application of an evolutionary approach to the game of Rack'O, which is a card game revolving around the notion of decision making. We first apply the evolutionary technique for obtaining a set of rules over many generations and then compare them with a script written by a human player. A high-level domain-specific language is used that deter-mines which the sets of rules are synthesized. We report the results by providing a comprehensive analysis of the set of rules and their implications. | ['Cassidy Pirlot', 'Rohan Saha'] | 2021-01-08 | null | null | null | null | ['card-games'] | ['playing-games'] | [ 3.22389692e-01 -1.74362451e-01 7.52222314e-02 -2.56565511e-01
4.18720454e-01 -8.52276862e-01 5.42422950e-01 -1.64115295e-01
-4.89762276e-01 8.68861258e-01 -3.88471007e-01 -6.53002739e-01
-4.22552794e-01 -1.02141666e+00 -3.26894850e-01 -4.51204181e-01
-1.54247984e-01 4.35315222e-01 6.26201868e-01 -1.06967008e+00
8.21468651e-01 2.71825403e-01 -2.12091088e+00 2.43753165e-01
8.40403676e-01 6.77053988e-01 1.56682074e-01 1.11700833e+00
1.61324576e-01 1.07073927e+00 -9.08439875e-01 -7.78817892e-01
2.08675131e-01 -1.00893438e+00 -1.00985777e+00 1.42908618e-01
-6.02940917e-01 1.15535632e-01 6.14366569e-02 1.17412400e+00
2.85075456e-01 4.68059659e-01 6.65832996e-01 -1.15603948e+00
-1.58900723e-01 5.84577918e-01 6.63726032e-02 2.11088449e-01
7.30675220e-01 1.92727283e-01 6.65464759e-01 -2.20284760e-01
8.14977765e-01 8.69222760e-01 6.63600624e-01 6.29929304e-01
-1.04993820e+00 -3.27224702e-01 -7.05748126e-02 5.48461974e-02
-1.66272998e+00 -2.62865275e-01 6.30850792e-01 -4.37765598e-01
1.20772541e+00 6.18745327e-01 1.05925775e+00 6.17379248e-01
7.70321786e-01 3.13429952e-01 1.15697348e+00 -9.10875976e-01
7.29495168e-01 5.55607341e-02 -1.58979073e-01 6.67171896e-01
5.66105306e-01 6.22601092e-01 -4.99040365e-01 -2.60965556e-01
7.36211061e-01 -7.35026062e-01 1.22873023e-01 -4.80941057e-01
-6.11213446e-01 9.33598876e-01 -3.09287637e-01 4.88135189e-01
-4.95375186e-01 1.53664738e-01 2.92637378e-01 4.66488212e-01
-1.99364319e-01 8.92440379e-01 -3.31747532e-01 -3.80431712e-01
-7.42837489e-01 5.68675876e-01 1.05673218e+00 6.48009777e-01
2.33409777e-01 2.22158849e-01 1.57212406e-01 4.94336069e-01
2.93061554e-01 -1.70158684e-01 6.42211616e-01 -1.08218682e+00
-5.08063622e-02 5.69274664e-01 2.74302036e-01 -9.15108442e-01
4.33474332e-02 4.47967015e-02 4.33145911e-02 9.02563870e-01
7.39479586e-02 -5.18740952e-01 -5.87966681e-01 1.47153223e+00
2.58460641e-01 -2.33042523e-01 2.94572294e-01 5.17282724e-01
2.90859997e-01 4.53898281e-01 -1.67566314e-01 -4.36515599e-01
1.12541115e+00 -7.51720250e-01 -6.26026750e-01 7.54822092e-03
4.77464825e-01 -5.36351383e-01 5.99287271e-01 9.19961393e-01
-1.13594007e+00 -5.77725053e-01 -1.33599484e+00 7.25095451e-01
-3.82904559e-01 -1.53381094e-01 8.87743950e-01 1.29562020e+00
-1.25037956e+00 8.58924627e-01 -4.10871208e-01 -5.53655922e-01
-2.41065100e-01 5.81855237e-01 9.39638391e-02 4.33263928e-01
-1.09269261e+00 1.10220218e+00 1.14598477e+00 7.12930858e-02
-6.11818373e-01 5.54492585e-02 -6.57883048e-01 -3.38538349e-01
7.80338585e-01 -5.30206919e-01 1.47240531e+00 -1.27490973e+00
-2.26531291e+00 1.13889432e+00 3.34584534e-01 -1.62282750e-01
5.95333159e-01 3.24010223e-01 -8.22872579e-01 -7.29003996e-02
-1.13401644e-01 1.47974789e-01 7.06736028e-01 -1.19198036e+00
-1.23144412e+00 7.71100447e-02 3.53967607e-01 1.75579980e-01
3.23730677e-01 4.48386997e-01 -4.51232165e-01 -7.57912993e-01
-3.58964235e-01 -9.13621783e-01 -5.98275006e-01 -6.95304096e-01
-8.61101598e-02 -1.21591464e-01 4.68567349e-02 -2.54520714e-01
1.88444757e+00 -1.79991710e+00 2.70859152e-01 5.16528428e-01
-1.70173541e-01 2.65158355e-01 2.26842254e-01 7.46041894e-01
-1.28026426e-01 2.27785379e-01 -3.03336680e-01 4.93815988e-01
4.63060811e-02 5.12585461e-01 -1.19793676e-01 -5.33633353e-03
5.95436916e-02 5.66029310e-01 -1.06624556e+00 -3.86896789e-01
-9.77136940e-02 -2.79980183e-01 -8.21281791e-01 2.08959997e-01
-2.62376964e-01 1.12864366e-02 -5.51897883e-01 5.02526879e-01
-1.08468113e-02 3.17621171e-01 8.47888410e-01 6.44281983e-01
-5.76198280e-01 1.18935674e-01 -1.25057256e+00 1.60416353e+00
-9.07239467e-02 4.24303710e-01 -2.56534636e-01 -7.78680027e-01
1.11606908e+00 3.02920133e-01 4.77606691e-02 -2.18625441e-01
7.38130629e-01 2.10877314e-01 5.69059193e-01 -6.70347989e-01
7.48717606e-01 -2.76445776e-01 -6.51424527e-01 8.56222153e-01
-1.52008906e-01 -7.22583890e-01 7.91764140e-01 -1.77206650e-01
1.08328080e+00 4.92663801e-01 8.42285275e-01 -5.19873857e-01
6.71054304e-01 6.30157113e-01 6.48598194e-01 9.87633109e-01
-9.07627270e-02 1.90302521e-01 5.79045236e-01 -7.88788676e-01
-1.00154138e+00 -7.88630545e-01 2.42502198e-01 1.15954781e+00
1.81291461e-01 -5.12276649e-01 -9.26447093e-01 -2.77340591e-01
-1.53714314e-01 1.30127811e+00 -8.64952922e-01 -4.62321252e-01
-4.47377563e-01 -9.28304434e-01 5.61574340e-01 1.63748518e-01
3.17593753e-01 -1.21875262e+00 -1.50234163e+00 5.54991722e-01
2.25682586e-01 -4.64672983e-01 -1.68673769e-01 2.92687207e-01
-7.22918451e-01 -1.07501495e+00 2.36454248e-01 -6.42276347e-01
4.39185798e-01 -2.65547454e-01 1.29490900e+00 4.79561567e-01
-3.46017420e-01 1.29910618e-01 -6.88132405e-01 -5.99598408e-01
-9.69233155e-01 -8.50896910e-02 1.49462730e-01 -1.82287514e-01
3.29437554e-01 -5.34287691e-01 2.94067245e-02 2.52195954e-01
-9.81137872e-01 1.13598034e-01 3.00095201e-01 6.90495014e-01
9.45334285e-02 9.18539166e-01 1.90549120e-01 -8.12363684e-01
1.23067927e+00 -2.44382516e-01 -8.02993119e-01 5.80497622e-01
-6.28382325e-01 3.12918395e-01 4.04272020e-01 -4.24034774e-01
-1.07358730e+00 1.27641201e-01 1.53391764e-01 6.71558529e-02
8.99407640e-02 2.70107180e-01 -1.43842176e-01 -2.51167357e-01
7.43169963e-01 2.64286131e-01 -4.36798893e-02 -1.08649999e-01
1.58931702e-01 5.23139536e-01 7.02272713e-01 -1.05972159e+00
5.88810384e-01 -2.54566688e-02 -3.80813271e-01 -4.00913417e-01
-3.80467288e-02 2.61138082e-01 -5.08113027e-01 -7.71346748e-01
7.17579186e-01 -2.63185464e-02 -8.64029586e-01 5.80712378e-01
-9.78327215e-01 -4.55108434e-01 -5.48150539e-01 1.12785786e-01
-9.96269107e-01 -1.34749576e-01 -2.75200307e-01 -9.20004010e-01
5.53647205e-02 -9.64082897e-01 2.52693951e-01 5.66522002e-01
-6.15826190e-01 -7.55859852e-01 6.83126688e-01 9.87206995e-02
2.73905299e-03 2.79972553e-01 8.23256135e-01 -5.40601552e-01
-5.07415049e-02 -1.11599453e-01 7.79121459e-01 -3.11586522e-02
1.91606864e-01 6.30910933e-01 -3.95757735e-01 -7.32182041e-02
1.38554871e-01 -1.16302408e-02 1.89218409e-02 -1.11081667e-01
8.79860580e-01 6.60593063e-02 -3.05202425e-01 4.25958425e-01
1.52854145e+00 1.24961722e+00 7.76049078e-01 7.50789046e-01
-1.51375249e-01 7.60981143e-01 9.35239136e-01 5.39084375e-01
6.59077242e-02 7.01845586e-01 1.00578003e-01 4.62228149e-01
5.71583748e-01 -3.19137245e-01 1.77306384e-01 7.47688949e-01
-6.96592569e-01 -4.14163828e-01 -1.10238624e+00 4.01845276e-01
-1.83047318e+00 -1.03486240e+00 2.46232018e-01 1.97846270e+00
8.52690279e-01 2.97406912e-01 3.57175767e-01 4.93148088e-01
1.11413741e+00 -2.15325132e-01 -2.86937445e-01 -1.44250858e+00
4.26461510e-02 6.17522836e-01 3.06644380e-01 5.83804846e-01
-6.77702785e-01 1.22334003e+00 8.60608006e+00 8.63327146e-01
-8.43806267e-01 -3.36398542e-01 3.18067402e-01 -1.64681554e-01
-2.50952095e-01 2.06905857e-01 -3.45500201e-01 6.10498965e-01
8.89718652e-01 -6.89027786e-01 7.42007971e-01 5.25261521e-01
2.41272882e-01 -5.93891084e-01 -8.48021746e-01 4.13815618e-01
4.79372963e-02 -1.27923501e+00 1.13269076e-01 1.53436055e-02
7.06847429e-01 -8.19300115e-01 -1.73188925e-01 1.83364406e-01
1.18047285e+00 -1.30990088e+00 1.27159989e+00 3.76187891e-01
4.70380545e-01 -1.25903833e+00 4.53263164e-01 3.34654689e-01
-9.36860442e-01 -4.08113748e-01 -7.94975013e-02 -5.18581867e-01
6.95381090e-02 -1.11467056e-01 -6.97595179e-01 8.37675571e-01
6.62672400e-01 -7.46621042e-02 -3.26043606e-01 8.87850344e-01
-5.01587927e-01 3.57939243e-01 -1.72101557e-01 -5.19732237e-01
2.28886351e-01 -4.79029089e-01 6.10827565e-01 1.05445981e+00
3.86379093e-01 7.78842270e-01 -1.85778141e-01 7.29461014e-01
6.10773802e-01 1.45829082e-01 -5.56041181e-01 -9.93306935e-02
4.48385686e-01 6.29708707e-01 -1.12507749e+00 -3.30421269e-01
1.40184686e-01 1.03355539e+00 -7.95058757e-02 -5.45032956e-02
-8.97897601e-01 -6.97529674e-01 7.34054863e-01 -1.34022355e-01
5.86955011e-01 -1.54950827e-01 -3.05923343e-01 -8.42624903e-01
-4.72686201e-01 -1.33074415e+00 2.44804099e-01 -9.78765488e-01
-7.39164412e-01 8.42476130e-01 2.82187581e-01 -1.08753812e+00
-7.91002631e-01 -6.24261618e-01 -1.03194296e+00 5.85467279e-01
-5.23721635e-01 -2.74479836e-01 3.36102635e-01 3.41667891e-01
3.65573168e-01 -5.96490443e-01 5.26879370e-01 -2.44933262e-01
-7.44178534e-01 3.42096835e-01 -1.08834416e-01 -1.18361637e-01
-8.10411721e-02 -9.59111750e-01 4.69502836e-01 1.09953535e+00
-2.01391205e-01 6.90886021e-01 1.21869040e+00 -7.04389334e-01
-1.22064209e+00 -4.77586299e-01 9.49321568e-01 -3.12542468e-01
7.09174335e-01 -1.50207445e-01 -2.78166771e-01 3.74798387e-01
2.36480251e-01 -5.64763308e-01 7.42705703e-01 -2.92652100e-01
3.61198574e-01 1.26799539e-01 -1.22375000e+00 1.01666903e+00
1.26285434e+00 -1.24072313e-01 -1.21325672e+00 -3.89483809e-01
3.01059157e-01 -3.68281722e-01 -5.13576269e-01 1.85185477e-01
7.72398353e-01 -1.37215745e+00 6.63026214e-01 -1.04615653e+00
2.83240288e-01 -6.69879317e-01 -2.67246272e-03 -1.28562403e+00
-3.84112567e-01 -1.01116550e+00 5.70430577e-01 9.12480593e-01
3.13351363e-01 -6.67061925e-01 7.01614320e-01 3.43597054e-01
1.59693673e-01 -5.90978026e-01 -7.10150898e-01 -8.66177261e-01
1.15666846e-02 -5.38051248e-01 8.66339445e-01 9.17894721e-01
3.81233990e-01 4.38080356e-02 -2.43918940e-01 -1.00921862e-01
3.32429469e-01 2.61261910e-01 4.24672931e-01 -1.09423888e+00
-7.54626691e-01 -5.66841662e-01 -6.40558004e-01 -4.67503071e-01
-2.33062580e-02 -4.89562780e-01 3.14254791e-01 -7.30076253e-01
-2.70954549e-01 -3.69265556e-01 -1.55113503e-01 4.51562293e-02
5.60951307e-02 9.21270400e-02 2.62863576e-01 -9.67654362e-02
-4.71991986e-01 1.65430725e-01 7.39100993e-01 3.38306397e-01
-4.96704161e-01 -3.28148752e-02 -9.51391459e-01 9.02270257e-01
9.85046983e-01 -7.66350925e-01 -5.77507794e-01 -1.02682836e-01
7.64299989e-01 3.08807343e-02 -2.37422287e-01 -1.07220328e+00
2.61108190e-01 -6.70504570e-01 -1.43091902e-01 -1.01555377e-01
-6.31282479e-02 -5.72153270e-01 9.58091259e-01 1.13601255e+00
-2.37704143e-01 6.72788322e-01 1.43052220e-01 1.56082481e-01
-1.87854409e-01 -7.62023211e-01 6.63888931e-01 -3.43684405e-01
-1.10457706e+00 -4.05488640e-01 -1.31366551e+00 -1.26255840e-01
1.59170413e+00 -7.63960481e-01 2.43343636e-01 -1.46450385e-01
-6.59317851e-01 1.44924700e-01 9.42621946e-01 3.59968811e-01
3.96943420e-01 -1.14584160e+00 -6.11166179e-01 1.41810939e-01
-6.47893623e-02 -6.52737260e-01 -3.30817789e-01 1.60447374e-01
-1.28186774e+00 1.38262644e-01 -8.92162740e-01 5.79051822e-02
-1.37369943e+00 5.28294802e-01 7.67869711e-01 -2.01775491e-01
5.60645526e-03 7.37378180e-01 -3.34509820e-01 -3.51895124e-01
-4.59788889e-01 -1.52030904e-02 -4.95846897e-01 -1.03129238e-01
4.33580130e-01 3.10594946e-01 -1.08149789e-01 -4.77276325e-01
-6.17802680e-01 4.61835891e-01 3.63167733e-01 -6.03145301e-01
1.25619137e+00 1.11845210e-01 -4.70323980e-01 3.65688801e-01
1.47054568e-01 -1.00532196e-01 -8.87264550e-01 3.53185713e-01
2.20437840e-01 -5.21793127e-01 -2.34005257e-01 -7.52882063e-01
-6.72307670e-01 1.84258476e-01 2.45276824e-01 2.85774589e-01
1.40014935e+00 -4.34109986e-01 1.07143298e-01 4.93233144e-01
8.89126420e-01 -1.42252290e+00 -2.38837019e-01 6.61790669e-01
5.88371038e-01 -3.95885497e-01 1.78333949e-02 -4.63213354e-01
-8.88527691e-01 1.18656230e+00 2.38626599e-01 -3.44604433e-01
3.06679428e-01 5.74556887e-01 -4.70784046e-02 -1.22188285e-01
-1.21558762e+00 -2.37913236e-01 -2.46002629e-01 7.92485237e-01
2.90164370e-02 1.83472261e-01 -9.72166181e-01 7.51215577e-01
-7.61597514e-01 4.77025747e-01 7.29855061e-01 1.62298810e+00
-5.48537612e-01 -1.49065769e+00 -6.87447548e-01 -7.87643567e-02
-3.62149596e-01 2.82900538e-02 -8.86962235e-01 8.45768988e-01
7.75764227e-01 1.11394465e+00 8.73022452e-02 -9.17445958e-01
4.01524901e-01 -1.30141973e-01 8.81953418e-01 -4.84817863e-01
-1.01677108e+00 -6.34366035e-01 6.05764210e-01 -2.66246557e-01
-3.18538338e-01 -9.64599013e-01 -9.51348364e-01 -7.87946105e-01
-3.50677222e-02 7.25268900e-01 3.08967978e-01 7.66924858e-01
-6.70847446e-02 4.65508848e-01 6.25900984e-01 -4.47278887e-01
-2.62375146e-01 -8.79919827e-02 -6.85225010e-01 -1.29581869e-01
-5.05881667e-01 -5.60855567e-01 1.24067143e-02 1.84330344e-01] | [3.456991195678711, 1.5637187957763672] |
d601a86b-a6d9-4572-a79e-7130246e550d | fakemix-augmentation-improves-transparent | 2103.13279 | null | https://arxiv.org/abs/2103.13279v2 | https://arxiv.org/pdf/2103.13279v2.pdf | FakeMix Augmentation Improves Transparent Object Detection | Detecting transparent objects in natural scenes is challenging due to the low contrast in texture, brightness and colors. Recent deep-learning-based works reveal that it is effective to leverage boundaries for transparent object detection (TOD). However, these methods usually encounter boundary-related imbalance problem, leading to limited generation capability. Detailly, a kind of boundaries in the background, which share the same characteristics with boundaries of transparent objects but have much smaller amounts, usually hurt the performance. To conquer the boundary-related imbalance problem, we propose a novel content-dependent data augmentation method termed FakeMix. Considering collecting these trouble-maker boundaries in the background is hard without corresponding annotations, we elaborately generate them by appending the boundaries of transparent objects from other samples into the current image during training, which adjusts the data space and improves the generalization of the models. Further, we present AdaptiveASPP, an enhanced version of ASPP, that can capture multi-scale and cross-modality features dynamically. Extensive experiments demonstrate that our methods clearly outperform the state-of-the-art methods. We also show that our approach can also transfer well on related tasks, in which the model meets similar troubles, such as mirror detection, glass detection, and camouflaged object detection. Code will be made publicly available. | ['Jian Tuo', 'Xiangui Luo', 'Kai Zhao', 'Qibin Hou', 'Enze Xie', 'Zhengqiang Zhang', 'Yang Cao'] | 2021-03-24 | null | null | null | null | ['transparent-objects', 'transparent-object-detection'] | ['computer-vision', 'computer-vision'] | [ 3.24294358e-01 -2.30513752e-01 1.24496356e-01 -2.33272150e-01
-4.84405458e-01 -4.64436650e-01 1.23324558e-01 -4.96715903e-01
5.28694242e-02 3.93997878e-01 -2.36948412e-02 -7.88491666e-02
3.76249135e-01 -5.90395808e-01 -8.51291060e-01 -9.18778300e-01
2.18629777e-01 4.17345650e-02 8.20576787e-01 -1.54975966e-01
1.67878479e-01 2.17611954e-01 -1.69859827e+00 6.78359330e-01
1.09583032e+00 1.38485312e+00 1.61560088e-01 2.78182358e-01
-3.22512627e-01 5.61091959e-01 -6.36286080e-01 -4.96531934e-01
7.09006429e-01 -1.09105036e-01 -3.72256666e-01 4.51266527e-01
9.41164792e-01 -7.66558051e-01 -2.67378718e-01 1.11918879e+00
4.68125612e-01 -9.51164290e-02 6.08651698e-01 -1.23164749e+00
-9.58461285e-01 1.51318163e-01 -1.06460202e+00 2.12299004e-01
-8.36133026e-03 2.94289976e-01 6.66969836e-01 -9.73901033e-01
3.17441493e-01 1.28515613e+00 7.07225263e-01 6.36666834e-01
-7.64329731e-01 -8.53643239e-01 5.05033553e-01 1.59482032e-01
-9.86813068e-01 -4.49189037e-01 9.21424270e-01 -3.54283065e-01
2.31808126e-01 4.60323066e-01 6.87127888e-01 1.12048161e+00
-1.44065887e-01 1.09119880e+00 1.32195771e+00 -2.65918791e-01
-1.48467124e-01 1.78521499e-01 -6.05225563e-02 7.95616984e-01
4.18708891e-01 2.33476907e-02 -3.31437528e-01 8.21389705e-02
7.27864206e-01 2.23799989e-01 -4.97635067e-01 -2.79415369e-01
-1.00792050e+00 4.38672125e-01 5.31928360e-01 -1.59642532e-01
1.67786628e-01 -2.07727775e-01 2.54150271e-01 1.08494334e-01
6.03565276e-01 1.80188328e-01 -5.65435410e-01 2.39809036e-01
-5.22554219e-01 3.59271877e-02 4.82763857e-01 9.42282617e-01
6.02620125e-01 -3.48605439e-02 -5.24645805e-01 9.99154270e-01
2.30366111e-01 4.50858623e-01 3.95668507e-01 -7.92312682e-01
5.82721353e-01 1.00174546e+00 3.45328599e-01 -7.75145888e-01
-3.10074955e-01 -3.87455672e-01 -8.03822637e-01 3.08776230e-01
5.15136361e-01 -8.98781866e-02 -1.14102137e+00 1.52939200e+00
9.11637306e-01 1.83004797e-01 -2.55634367e-01 1.25950623e+00
1.10947192e+00 4.87810582e-01 -3.34479034e-01 -8.15013945e-02
1.52329123e+00 -1.36404824e+00 -6.63370252e-01 -3.25610012e-01
1.93350986e-01 -9.94338691e-01 1.60744858e+00 6.12272918e-01
-8.12566519e-01 -5.61129212e-01 -9.09964740e-01 -1.68452144e-01
-9.19114202e-02 4.62857068e-01 8.70600522e-01 8.01276207e-01
-6.43498838e-01 2.13277027e-01 -3.97271395e-01 -9.84784290e-02
8.11556160e-01 5.22793606e-02 -7.86737427e-02 -2.65841842e-01
-8.54438961e-01 3.86947930e-01 -2.28977799e-02 3.32412213e-01
-9.29043055e-01 -6.79483950e-01 -6.83195829e-01 -2.28937835e-01
8.08578432e-01 -6.19412363e-01 1.04699242e+00 -1.07672024e+00
-1.48833907e+00 9.38253522e-01 -5.64717390e-02 1.43674567e-01
9.62962389e-01 -4.16413039e-01 -3.73634815e-01 4.60789092e-02
-9.49037261e-03 4.66373205e-01 1.35355818e+00 -1.51636016e+00
-5.35766244e-01 -2.63087004e-01 2.25668803e-01 2.44381949e-01
-8.10997844e-01 1.41077965e-01 -9.36930776e-01 -6.47744179e-01
2.26820141e-01 -7.81681716e-01 -1.09043762e-01 5.60613453e-01
-7.62469471e-01 -1.06742896e-01 1.09510803e+00 -5.22698104e-01
1.03099966e+00 -2.26054215e+00 -3.71078670e-01 -2.76686400e-01
4.54341441e-01 4.28301781e-01 -2.13258266e-01 -1.12110913e-01
1.75110623e-01 6.72433227e-02 -3.66715968e-01 -3.02885741e-01
-1.22755677e-01 4.94741574e-02 -4.54675049e-01 4.23714787e-01
4.11382169e-01 7.11814106e-01 -6.26299262e-01 -6.16049170e-01
6.08415119e-02 2.00552985e-01 -4.12344098e-01 1.87845558e-01
-2.48596847e-01 4.39906597e-01 -4.93255973e-01 1.06011009e+00
1.22756994e+00 -2.98658252e-01 -3.23662609e-01 -3.51847857e-01
-1.53296683e-02 7.86991119e-02 -1.22084796e+00 1.08990359e+00
-1.54491380e-01 5.07877648e-01 9.39824581e-02 -5.70127845e-01
8.72151732e-01 -1.38460442e-01 2.62140423e-01 -7.52578735e-01
1.34422362e-01 8.53465050e-02 2.93257739e-03 -9.73218441e-01
3.42673957e-01 1.72453135e-01 3.14894855e-01 2.67083257e-01
-6.52122796e-01 -4.68435809e-02 -7.31743202e-02 -1.00730151e-01
9.19794261e-01 2.88598239e-01 -3.03243548e-01 1.36388257e-01
3.54747891e-01 -3.52213621e-01 9.62179720e-01 8.07299018e-01
-4.12899584e-01 1.00283098e+00 2.88833112e-01 -4.73222107e-01
-8.86625230e-01 -1.09213066e+00 -3.10388178e-01 9.81013477e-01
6.41028583e-01 -2.48833355e-02 -8.54670823e-01 -7.94301748e-01
8.70117843e-02 -8.31824243e-02 -6.49170458e-01 -1.79469183e-01
-5.67880988e-01 -1.10346997e+00 1.18098550e-01 4.94874656e-01
8.96417558e-01 -8.34275007e-01 -4.22280997e-01 -1.08407952e-01
-3.56755316e-01 -1.35414076e+00 -7.12538838e-01 -2.61222243e-01
-8.47123623e-01 -1.18790054e+00 -8.07061017e-01 -7.07422197e-01
8.30507219e-01 8.36638153e-01 9.91754949e-01 3.46595466e-01
-5.96032560e-01 -1.34433005e-02 -4.06912506e-01 -5.99734426e-01
-8.70683193e-02 -3.29782337e-01 6.80229738e-02 3.72431397e-01
9.13507417e-02 -2.54512966e-01 -1.06971705e+00 7.22960055e-01
-9.30725753e-01 2.21010774e-01 6.17030323e-01 8.06215823e-01
5.77266812e-01 8.96750242e-02 3.17404062e-01 -9.22443032e-01
2.15887785e-01 -1.47756845e-01 -6.06137335e-01 2.62921333e-01
-8.87863189e-02 -3.63191366e-01 5.80723703e-01 -8.40577722e-01
-1.15609133e+00 -8.22823420e-02 2.13042259e-01 -7.03247905e-01
-5.18716574e-02 -2.31813237e-01 -4.91380453e-01 -4.07262981e-01
4.56440896e-01 1.90288693e-01 -2.55801916e-01 -5.77625871e-01
1.30945519e-01 6.83497131e-01 2.83363730e-01 -5.47216117e-01
1.03790450e+00 9.30597723e-01 -3.00216317e-01 -6.15779459e-01
-1.37886143e+00 -4.41984594e-01 -2.10082799e-01 -1.82216048e-01
4.75170135e-01 -9.06513095e-01 -7.09717274e-01 9.54282343e-01
-1.07576013e+00 -4.36945379e-01 -1.29694492e-01 1.73449501e-01
-1.46634653e-01 6.44835591e-01 -6.53477490e-01 -9.71118629e-01
-3.25201750e-01 -1.03793597e+00 1.38672197e+00 5.35771608e-01
5.25581181e-01 -4.90878135e-01 -3.80449057e-01 6.75313652e-01
3.35305542e-01 1.20413385e-01 7.56332219e-01 -2.51418442e-01
-9.50072944e-01 3.02279927e-02 -6.99677110e-01 5.41257322e-01
2.34685928e-01 2.62925833e-01 -1.36540020e+00 -1.48281530e-01
1.57743365e-01 -2.95528471e-01 1.09387457e+00 2.52912313e-01
1.71389210e+00 -3.60353053e-01 -3.80426109e-01 8.53909612e-01
9.85153854e-01 -1.39753371e-01 5.94027519e-01 4.10422742e-01
9.93557215e-01 8.92511129e-01 8.96259248e-01 4.45682645e-01
3.64311934e-01 7.11910367e-01 6.65840983e-01 -5.04918754e-01
-5.37857831e-01 1.72039017e-01 4.65616852e-01 4.28905278e-01
6.04721867e-02 -1.78396121e-01 -5.39254665e-01 3.17209899e-01
-1.68559456e+00 -6.92129791e-01 -5.02050221e-01 2.04721117e+00
9.51720238e-01 2.17691153e-01 1.44356996e-01 -3.41737904e-02
8.63173664e-01 -5.98841980e-02 -9.08880174e-01 1.34324595e-01
-4.11039889e-01 -2.95596421e-01 4.11463916e-01 1.12307388e-02
-1.29397738e+00 9.33689356e-01 6.21047688e+00 8.66702259e-01
-1.13204980e+00 1.79616347e-01 1.02868927e+00 -1.05482250e-01
-3.12950850e-01 -1.87129959e-01 -8.64593983e-01 8.07745337e-01
-4.41718400e-02 4.97556686e-01 3.90855193e-01 8.17196429e-01
1.35842726e-01 -1.17323503e-01 -9.40967977e-01 9.46625531e-01
2.61238217e-01 -1.03894031e+00 2.52075382e-02 -1.51500143e-02
9.63161588e-01 6.65188879e-02 4.20887262e-01 1.31349564e-01
6.99969009e-02 -5.63031495e-01 8.31780314e-01 3.49472165e-01
7.89589226e-01 -1.39382660e-01 5.28670967e-01 9.14791822e-02
-1.17630422e+00 -3.58647794e-01 -6.84033215e-01 7.33918473e-02
-1.40239179e-01 9.99732971e-01 -4.90662068e-01 4.08099651e-01
1.05974996e+00 5.50434411e-01 -7.15102732e-01 1.49205852e+00
-2.56121486e-01 5.91443121e-01 -4.56211090e-01 1.86915994e-01
-1.38351709e-01 -3.05630535e-01 4.89541680e-01 8.86887670e-01
2.06414759e-01 -1.92080036e-01 2.78873324e-01 1.11351991e+00
-2.10599050e-01 -9.18844715e-02 -3.62579226e-01 3.89457673e-01
4.66100484e-01 1.47641742e+00 -6.86883569e-01 -3.61874849e-01
-6.00410163e-01 1.03841710e+00 2.51118779e-01 4.88370091e-01
-1.02014208e+00 -3.02325726e-01 6.16012156e-01 2.67083198e-01
3.82887900e-01 1.01655155e-01 -3.16237628e-01 -1.44721282e+00
7.02709734e-01 -8.98398101e-01 2.77369142e-01 -9.69520986e-01
-1.72522724e+00 4.00430411e-01 -3.97968799e-01 -1.73090589e+00
7.45057523e-01 -7.75611401e-01 -7.94669271e-01 5.52985251e-01
-1.95277584e+00 -1.26577365e+00 -8.29743981e-01 5.80975235e-01
4.29568291e-01 8.54726806e-02 2.71333963e-01 4.44525987e-01
-9.50087965e-01 7.40086079e-01 1.75091282e-01 2.12329298e-01
9.74518061e-01 -9.87629235e-01 4.00168478e-01 9.94044363e-01
-1.92653090e-01 4.49375778e-01 4.60448354e-01 -6.10492289e-01
-1.32103157e+00 -1.19526923e+00 -5.27849607e-02 -6.51448846e-01
5.91139317e-01 -6.60798013e-01 -1.08139455e+00 2.84179717e-01
-2.69495547e-01 4.98846322e-01 4.22021210e-01 -3.63999568e-02
-6.16989195e-01 -4.40056741e-01 -1.08025610e+00 7.73154199e-01
1.32684052e+00 -1.48055479e-01 -2.95360684e-01 6.01824462e-01
9.01878715e-01 -6.59711838e-01 -4.11013931e-01 5.46933770e-01
5.15062809e-01 -1.26719439e+00 1.01335084e+00 -3.56128693e-01
3.90110493e-01 -3.85422170e-01 1.42610684e-01 -8.80764246e-01
-2.04728786e-02 -7.54704177e-01 -3.59494954e-01 1.47662127e+00
4.32219803e-02 -7.63977587e-01 9.08899069e-01 6.47605896e-01
-3.97170335e-01 -8.76536131e-01 -9.29124713e-01 -7.95041978e-01
-3.41493875e-01 -1.49283409e-01 7.13476002e-01 8.56839180e-01
-7.15507269e-01 -4.89856713e-02 -4.68719184e-01 3.51925969e-01
6.52917981e-01 5.21027982e-01 1.01219034e+00 -1.31391907e+00
-3.20312381e-01 -3.73000413e-01 -1.67013735e-01 -1.24425304e+00
-3.10780823e-01 -3.36882651e-01 4.42123786e-02 -1.36670029e+00
5.55817783e-01 -8.79556537e-01 -2.73660898e-01 5.48563600e-01
-6.74202919e-01 5.68229735e-01 1.33642226e-01 4.25569981e-01
-8.15426707e-01 7.54123867e-01 1.98108375e+00 -2.36334771e-01
-1.55845955e-01 1.08566009e-01 -7.87175119e-01 9.25858021e-01
5.83726048e-01 -3.48485231e-01 -1.79796815e-01 -6.22819781e-01
1.76005036e-01 -4.66372609e-01 4.45342720e-01 -9.99913633e-01
-2.24631518e-01 -3.37995529e-01 6.97879195e-01 -4.37737912e-01
3.97891611e-01 -8.08357537e-01 -3.91372055e-01 1.89569607e-01
8.55497718e-02 -4.15081263e-01 2.95252264e-01 7.58450508e-01
1.06533766e-01 -1.74347818e-01 7.45444238e-01 -6.50820285e-02
-5.30071080e-01 7.51755953e-01 6.64339960e-02 1.61934391e-01
1.03876328e+00 -5.88612676e-01 -8.89995098e-01 -1.74224172e-02
-1.55070737e-01 2.97647178e-01 6.84132516e-01 5.32358170e-01
6.18301690e-01 -1.11634839e+00 -5.89049160e-01 2.49290094e-01
3.47973794e-01 4.78374392e-01 5.83962679e-01 1.09900033e+00
-3.43576163e-01 -2.82050818e-01 -1.01163045e-01 -6.51201308e-01
-1.22808957e+00 5.25424123e-01 4.78940845e-01 9.28800404e-02
-8.95205736e-01 1.13980711e+00 1.00516701e+00 -2.75358409e-01
2.79776990e-01 -5.00943422e-01 -1.50500625e-01 -1.39194280e-01
7.00557649e-01 1.60077050e-01 4.63478528e-02 -2.04320773e-01
-2.34665975e-01 7.46207654e-01 -3.75985622e-01 4.72071558e-01
1.06361175e+00 -2.38869771e-01 -1.73464656e-01 3.98505926e-01
8.42938602e-01 2.70738304e-01 -1.79523253e+00 -2.78235734e-01
-5.29205799e-01 -8.54323566e-01 -1.38680801e-01 -7.58526862e-01
-1.30458152e+00 9.42762494e-01 7.59279609e-01 1.71583340e-01
1.23025000e+00 -4.22720984e-02 1.02374601e+00 3.04243028e-01
2.42107004e-01 -1.14838445e+00 5.71663439e-01 2.43399918e-01
6.89538002e-01 -1.66502368e+00 1.19040506e-02 -9.64718580e-01
-6.15413964e-01 9.45862710e-01 1.20006800e+00 2.10811749e-01
2.39411846e-01 3.23178023e-01 2.83301651e-01 -9.80626512e-03
-5.37939548e-01 -2.42689773e-01 2.17315376e-01 8.02795887e-01
5.55035062e-02 -1.20486706e-01 8.50066692e-02 6.87572360e-01
1.79688945e-01 -2.87040591e-01 5.62633812e-01 7.59009838e-01
-4.64073837e-01 -8.21147323e-01 -7.28302538e-01 6.30634785e-01
-3.75570655e-01 -2.50664860e-01 -5.07324100e-01 5.51405966e-01
3.50364268e-01 9.10157144e-01 8.25494677e-02 -2.57658094e-01
1.69718370e-01 -5.00403047e-01 5.12984037e-01 -5.96498907e-01
-3.61162245e-01 1.21500857e-01 -3.77682447e-02 -6.59371495e-01
-2.36830249e-01 -4.22527254e-01 -1.01717079e+00 -3.76718910e-03
-7.14656532e-01 -2.39511952e-01 4.30672944e-01 6.58038557e-01
3.92264873e-01 7.44665325e-01 7.09884763e-01 -8.63782167e-01
-5.65865219e-01 -9.89295840e-01 -4.29740071e-01 6.63336992e-01
5.40119052e-01 -1.02906561e+00 -4.09237742e-01 9.33433790e-03] | [10.252519607543945, -0.8492931127548218] |
d6e24615-5736-49e7-8d17-0a0ff139555c | learning-asymmetric-embedding-for-attributed | 2202.06307 | null | https://arxiv.org/abs/2202.06307v1 | https://arxiv.org/pdf/2202.06307v1.pdf | Learning Asymmetric Embedding for Attributed Networks via Convolutional Neural Network | Recently network embedding has gained increasing attention due to its advantages in facilitating network computation tasks such as link prediction, node classification and node clustering. The objective of network embedding is to represent network nodes in a low-dimensional vector space while retaining as much information as possible from the original network including structural, relational, and semantic information. However, asymmetric nature of directed networks poses many challenges as how to best preserve edge directions in the embedding process. Here, we propose a novel deep asymmetric attributed network embedding model based on convolutional graph neural network, called AAGCN. The main idea is to maximally preserve the asymmetric proximity and asymmetric similarity of directed attributed networks. AAGCN introduces two neighbourhood feature aggregation schemes to separately aggregate the features of a node with the features of its in- and out- neighbours. Then, it learns two embedding vectors for each node, one source embedding vector and one target embedding vector. The final representations are the results of concatenating source and target embedding vectors. We test the performance of AAGCN on three real-world networks for network reconstruction, link prediction, node classification and visualization tasks. The experimental results show the superiority of AAGCN against state-of-the-art embedding methods. | ['Xinghuo Yu', 'Mahdi Jalili', 'Ahmad Asgharian Rezaei', 'Hossein Ghorbanzadeh', 'Mohammadreza Radmanesh'] | 2022-02-13 | null | null | null | null | ['network-embedding'] | ['methodology'] | [-1.59612671e-01 3.05768639e-01 -2.02423528e-01 -3.14758509e-01
3.69457185e-01 -1.86529383e-01 6.78283155e-01 4.12214220e-01
-7.81152844e-02 3.45813006e-01 3.88559908e-01 -1.97185248e-01
-5.93605816e-01 -1.20847952e+00 -8.03301111e-02 -7.83714890e-01
-5.85869431e-01 4.25334662e-01 2.23158881e-01 -1.10384174e-01
-7.75502101e-02 8.00357759e-01 -1.07866621e+00 -1.44302621e-02
4.92732793e-01 7.81633973e-01 -1.68364435e-01 5.25762200e-01
-2.74686486e-01 7.53477454e-01 -2.31656894e-01 -4.62673604e-01
1.02536850e-01 -2.02104107e-01 -6.99190617e-01 -2.21239626e-01
-1.83863062e-02 -7.67697915e-02 -1.34143567e+00 9.96855617e-01
4.27141488e-01 6.58584461e-02 7.00156748e-01 -1.86760569e+00
-1.06435323e+00 6.69588804e-01 -2.45262116e-01 2.04645008e-01
7.78691322e-02 -1.18922330e-01 1.39011872e+00 -6.76849246e-01
7.40094244e-01 1.38227391e+00 5.69750905e-01 2.96419501e-01
-1.52645791e+00 -5.34170568e-01 4.84066606e-02 3.14929903e-01
-1.49248707e+00 -1.24149852e-01 1.26138091e+00 -3.63186002e-01
7.99997091e-01 3.47396910e-01 8.27601016e-01 8.30788910e-01
6.97956160e-02 4.24487293e-01 4.08211976e-01 -2.26195782e-01
-6.77573457e-02 3.12475830e-01 2.42008984e-01 8.76070738e-01
2.15407193e-01 1.24602737e-02 -2.55779237e-01 -3.03696066e-01
5.82085013e-01 4.85421777e-01 -3.24602723e-01 -8.13995302e-01
-1.30800176e+00 9.49982584e-01 1.33078873e+00 4.11541343e-01
-3.16860616e-01 3.49555254e-01 5.38031816e-01 4.65086788e-01
5.55576146e-01 1.55663952e-01 -1.60199516e-02 3.63488793e-01
-3.43796343e-01 -1.13494307e-01 7.69853652e-01 7.89855897e-01
8.40985715e-01 -1.81572556e-01 -7.22732991e-02 8.21251690e-01
7.35743523e-01 1.48059785e-01 3.32491279e-01 -6.02795780e-01
4.19750154e-01 1.19900906e+00 -5.93248427e-01 -2.05473447e+00
-3.60349864e-01 -4.05670971e-01 -1.38636887e+00 3.86544280e-02
-2.44187996e-01 1.85054824e-01 -4.14735973e-01 1.59804046e+00
3.89139682e-01 4.01010185e-01 8.42923895e-02 5.89461029e-01
1.29537773e+00 8.94835949e-01 -1.03200167e-01 1.63948938e-01
1.00678968e+00 -1.02801871e+00 -7.58948147e-01 7.01733604e-02
8.05070817e-01 -3.25524151e-01 4.62409556e-01 -5.44112086e-01
-6.44421637e-01 -3.07224840e-01 -1.09884310e+00 -8.05026367e-02
-7.23000050e-01 -9.36453044e-02 6.87876999e-01 1.24128871e-01
-1.17959082e+00 7.42725492e-01 -6.53961837e-01 -4.04861689e-01
5.22592127e-01 4.54038531e-01 -9.48401392e-01 -4.82765548e-02
-1.37565172e+00 4.93661553e-01 6.41576827e-01 2.91832060e-01
-3.55223626e-01 -5.94486773e-01 -1.17491341e+00 5.28620005e-01
8.62410218e-02 -4.28204715e-01 1.87186718e-01 -4.53886300e-01
-9.13234651e-01 6.29898250e-01 2.29582116e-02 -2.01023206e-01
7.96080530e-02 4.14772868e-01 -7.64834166e-01 1.27117157e-01
5.61703276e-03 5.71726501e-01 5.57037175e-01 -1.09829199e+00
-1.98173091e-01 -4.24244195e-01 -4.07062806e-02 1.02199003e-01
-1.03514194e+00 -3.77641201e-01 -3.04631263e-01 -6.80750489e-01
4.75796431e-01 -8.22846293e-01 -6.86001256e-02 6.51894927e-01
-7.25340903e-01 -3.69270414e-01 1.39772546e+00 -4.65319723e-01
1.39810479e+00 -2.21651673e+00 5.49341798e-01 6.82313740e-01
9.83879030e-01 3.67484957e-01 -4.08041924e-01 8.40987146e-01
-3.61342520e-01 2.79248506e-01 2.67561898e-02 -1.53852478e-01
6.69984100e-03 3.53924453e-01 6.64811255e-03 4.68203932e-01
2.00631395e-01 1.10533893e+00 -1.14246368e+00 -7.51491308e-01
3.12030315e-01 9.34382856e-01 -4.18517053e-01 1.97231084e-01
4.02324468e-01 -9.59532335e-02 -3.47123593e-01 3.60310227e-01
5.40929615e-01 -3.76865685e-01 4.76997167e-01 -6.44564390e-01
2.94196039e-01 2.05777377e-01 -1.10079277e+00 1.30986989e+00
-1.82848677e-01 1.02885878e+00 1.00825904e-02 -1.17314041e+00
1.22872114e+00 1.87492758e-01 5.65168560e-01 -4.96122152e-01
-8.92920569e-02 -7.05323219e-02 2.87579268e-01 -2.74148881e-01
1.14441462e-01 2.78032094e-01 3.35597485e-01 6.25516832e-01
2.11729035e-01 5.11170805e-01 1.00131959e-01 9.16603982e-01
1.18454218e+00 -6.75469697e-01 6.07673042e-02 -2.24075571e-01
6.48593366e-01 -6.61774337e-01 5.08556843e-01 1.89971924e-03
-1.88963279e-01 1.69770598e-01 1.11570668e+00 -5.74908137e-01
-9.11253989e-01 -1.12494028e+00 5.96090443e-02 7.95980930e-01
1.46442488e-01 -7.34639108e-01 -1.67790666e-01 -9.16840792e-01
4.45185333e-01 1.58467174e-01 -9.74111080e-01 -6.61034226e-01
-4.83312666e-01 -4.55313027e-01 5.08583188e-01 5.31033754e-01
4.20671344e-01 -1.04999256e+00 2.52479345e-01 5.71124405e-02
1.71787933e-01 -7.20421195e-01 -5.08099675e-01 -1.08174436e-01
-7.71592915e-01 -1.19117939e+00 -3.25242728e-01 -1.15684140e+00
1.00677490e+00 3.98731351e-01 8.53002846e-01 6.53977513e-01
-4.12860513e-01 6.11071177e-02 -2.59958118e-01 3.36236656e-01
-2.63941765e-01 2.24674985e-01 6.46538511e-02 2.25774691e-01
1.88330501e-01 -8.79934311e-01 -6.19972169e-01 2.31712058e-01
-7.89043427e-01 -2.33851373e-02 6.56888604e-01 1.05673695e+00
2.58965254e-01 2.79702395e-01 5.71325779e-01 -1.00516868e+00
8.40018213e-01 -7.35959888e-01 -1.07666910e-01 3.52853715e-01
-7.63840973e-01 1.92897171e-01 5.66038430e-01 -1.91433370e-01
-3.40430498e-01 -4.00313079e-01 2.81696379e-01 -4.95812178e-01
3.90530646e-01 6.66499972e-01 -4.09883231e-01 -1.81509495e-01
2.84870595e-01 2.45078415e-01 3.80957127e-01 -5.05607963e-01
5.31410396e-01 4.67308760e-01 2.74360001e-01 -7.93862790e-02
1.04763007e+00 2.75768816e-01 3.70367050e-01 -8.20653319e-01
-2.76488692e-01 -4.15659159e-01 -9.67252016e-01 -9.71364006e-02
5.15486240e-01 -4.54065055e-01 -7.73932338e-01 8.44039470e-02
-1.12193370e+00 2.29477525e-01 -1.80180117e-01 2.47591913e-01
1.05917327e-01 4.90925014e-01 -6.76571846e-01 -2.70163655e-01
-4.26366925e-01 -9.42614913e-01 4.59860355e-01 2.93047912e-02
-3.48339304e-02 -1.70292330e+00 1.95500895e-01 -1.53690666e-01
4.04892504e-01 3.93638194e-01 1.37859952e+00 -8.77878726e-01
-3.18090856e-01 -4.61399287e-01 -6.86394453e-01 2.82359570e-01
4.75679427e-01 2.80767411e-01 -4.60914969e-01 -5.83648145e-01
-8.56990337e-01 5.98468892e-02 9.07646358e-01 -1.12236634e-01
1.06581008e+00 -6.56629026e-01 -8.10382009e-01 7.45257139e-01
1.35851216e+00 -3.19715649e-01 4.60710019e-01 2.03328907e-01
1.14002168e+00 7.30336905e-01 1.40293380e-02 1.01985782e-01
2.91109174e-01 6.59103096e-01 7.82455444e-01 -1.04149804e-01
-1.69103593e-01 -5.26815653e-01 -4.63692732e-02 1.07632828e+00
1.82170853e-01 -2.07690939e-01 -8.78106415e-01 6.96768403e-01
-1.90916860e+00 -1.01938832e+00 -1.61780775e-01 1.81456971e+00
3.38589936e-01 1.62521958e-01 1.16334558e-02 3.62096667e-01
9.33819950e-01 6.13026559e-01 -4.08040315e-01 -3.14328223e-01
-6.34947941e-02 -2.58446664e-01 1.24297626e-01 4.91340399e-01
-1.02999020e+00 5.86428821e-01 5.15082359e+00 4.52656925e-01
-8.59213471e-01 -1.56646788e-01 4.14222151e-01 1.16357632e-01
-6.29244328e-01 4.94235046e-02 -2.93474793e-01 5.35355747e-01
6.61092937e-01 -3.91489714e-01 3.39469701e-01 7.08950996e-01
-2.11946815e-01 7.04269588e-01 -1.21996462e+00 8.62207472e-01
-1.59821376e-01 -1.59560466e+00 4.93245393e-01 3.20545435e-01
5.49805164e-01 -2.15731993e-01 1.21734008e-01 1.29801691e-01
1.91086173e-01 -1.34709620e+00 2.35685371e-02 4.89177227e-01
7.42283940e-01 -9.52733040e-01 8.95206153e-01 9.54563171e-02
-1.63372898e+00 -2.23307814e-02 -5.33801496e-01 7.55360126e-02
-5.80954812e-02 5.02641678e-01 -8.78089547e-01 5.82418323e-01
4.81466174e-01 1.41850698e+00 -6.84081256e-01 8.68609548e-01
-1.86231926e-01 2.83234954e-01 -1.57530472e-01 -1.32362619e-01
4.87300932e-01 -4.56488281e-01 6.14235699e-01 1.00030410e+00
-2.38791890e-02 1.92033742e-02 3.15895081e-02 9.47066605e-01
-5.14735043e-01 5.72036356e-02 -9.92399633e-01 -3.51426601e-01
1.08511591e+00 1.61897433e+00 -5.97078919e-01 -6.12594932e-02
-3.06734025e-01 9.10996318e-01 7.70178497e-01 3.58538270e-01
-4.67350900e-01 -9.18462813e-01 9.20612097e-01 1.30250335e-01
1.33873582e-01 -1.19046636e-01 1.43320471e-01 -7.96062291e-01
3.03447153e-02 -2.13448361e-01 3.73102158e-01 -4.37008172e-01
-1.31762266e+00 8.08245122e-01 -2.12279677e-01 -1.15146291e+00
4.26615812e-02 -6.14921510e-01 -1.08156085e+00 9.12707806e-01
-1.26033080e+00 -1.38759887e+00 -5.47593653e-01 5.56164503e-01
-1.23302758e-01 -4.29084182e-01 1.00862932e+00 4.46888924e-01
-8.50084364e-01 7.20485091e-01 5.37005603e-01 8.52111280e-01
2.93542773e-01 -1.28473890e+00 3.69595647e-01 3.95863980e-01
2.31741086e-01 6.93675876e-01 2.13384077e-01 -5.41323543e-01
-1.39908075e+00 -1.36379385e+00 1.04742730e+00 9.75533649e-02
9.01467383e-01 -3.85977834e-01 -1.08066189e+00 9.12045538e-01
-2.61257142e-02 7.62312233e-01 7.27096379e-01 1.39470235e-01
-5.04874170e-01 -4.51390743e-01 -1.08904171e+00 5.25535643e-01
1.09141123e+00 -7.44041145e-01 -1.41213879e-01 2.80247569e-01
9.22272623e-01 2.43590236e-01 -1.19824541e+00 2.51657933e-01
6.35866642e-01 -7.81755090e-01 1.33964562e+00 -8.58614564e-01
4.22940344e-01 -1.67934805e-01 -2.47750413e-02 -1.63835037e+00
-7.13158607e-01 -2.06091270e-01 -4.85117525e-01 1.34734809e+00
2.88341939e-01 -9.32376623e-01 8.90993357e-01 3.48813117e-01
1.94461584e-01 -1.08625245e+00 -7.76841700e-01 -5.74931979e-01
-2.50192374e-01 2.17810452e-01 7.32069671e-01 1.30923545e+00
7.83501044e-02 5.50320983e-01 -3.07637274e-01 1.13969423e-01
8.79806221e-01 -4.08029556e-03 4.92336839e-01 -1.69607449e+00
2.28051335e-01 -5.21176755e-01 -1.23857069e+00 -6.72178626e-01
4.59289134e-01 -1.40898609e+00 -6.71705663e-01 -1.76233542e+00
1.26412228e-01 -7.61396468e-01 -5.01115918e-01 4.76733506e-01
2.26101596e-02 2.30000019e-01 1.82656273e-01 3.31874818e-01
-5.07412314e-01 1.00425816e+00 1.08940959e+00 -4.95596617e-01
-1.06367514e-01 -2.52003253e-01 -5.40745616e-01 4.60823655e-01
6.54948413e-01 -4.61335599e-01 -6.34497344e-01 -4.64262098e-01
2.00371608e-01 -1.13182649e-01 5.08434832e-01 -7.62604237e-01
5.29399335e-01 2.72174954e-01 3.63216251e-01 -5.39669573e-01
4.22915667e-01 -1.13967502e+00 2.52777904e-01 6.41395986e-01
-5.39378583e-01 2.24975511e-01 -1.86492383e-01 1.01472580e+00
-3.26579511e-01 5.35932295e-02 5.05472362e-01 2.82100260e-01
-7.63957441e-01 6.61267817e-01 1.87522218e-01 -2.90393442e-01
1.15119267e+00 -1.56948850e-01 -4.79698241e-01 -3.06799442e-01
-9.19009447e-01 5.28022468e-01 2.93711483e-01 5.04545748e-01
1.19562948e+00 -2.04377222e+00 -7.85012543e-01 3.62220228e-01
2.71540076e-01 -4.25790697e-02 1.41086966e-01 6.73311472e-01
-6.51034176e-01 3.90439719e-01 -3.36815655e-01 -4.19555098e-01
-1.55051601e+00 5.75990915e-01 3.01027417e-01 -2.89493024e-01
-8.63499463e-01 7.94831991e-01 1.02110463e-03 -7.91073203e-01
3.47861797e-01 2.13293642e-01 -4.30846661e-01 1.73910502e-02
3.29140306e-01 6.29519641e-01 -2.36934274e-01 -8.60661387e-01
-4.98907030e-01 4.54032391e-01 -2.90693909e-01 4.37357396e-01
1.45455623e+00 6.28897771e-02 -7.12400377e-01 2.88701028e-01
1.88843727e+00 -4.29743230e-01 -7.19038248e-01 -6.72121465e-01
-7.10693970e-02 -6.81388497e-01 2.79340386e-01 -1.47193819e-01
-1.49540758e+00 1.03897774e+00 3.74308318e-01 4.95336562e-01
6.97683930e-01 1.36708885e-01 6.28815174e-01 4.17136967e-01
-5.52406125e-02 -5.17505467e-01 3.16809475e-01 2.42344812e-01
8.78134787e-01 -1.16147292e+00 1.92954451e-01 -5.83279073e-01
-2.56281763e-01 1.22309124e+00 7.05657184e-01 -2.91263908e-01
1.26943552e+00 -3.28660339e-01 -3.28125358e-01 -6.26372457e-01
-7.86579907e-01 1.43217564e-01 5.68102956e-01 6.71305954e-01
2.92449266e-01 1.60164848e-01 -5.12144703e-04 4.35300320e-02
9.66448709e-02 -6.97050869e-01 9.54593644e-02 6.75467134e-01
-2.94666350e-01 -1.08263600e+00 2.30297893e-01 7.91296840e-01
1.63843513e-01 2.28660796e-02 -6.59766972e-01 8.33896756e-01
-2.40129143e-01 6.14625514e-01 3.41992438e-01 -7.77890980e-01
1.14041932e-01 -2.47209996e-01 -1.56057447e-01 -5.62553763e-01
-2.57806212e-01 -5.99200428e-01 -1.04541607e-01 -4.04615611e-01
-2.24367872e-01 -2.18626887e-01 -1.19374514e+00 -8.08260918e-01
-3.98042440e-01 3.96969855e-01 4.19549406e-01 4.52632368e-01
6.44231737e-01 6.50105536e-01 9.31178987e-01 -7.15841830e-01
-3.25143576e-01 -9.20298755e-01 -7.70430207e-01 5.89527249e-01
3.19752306e-01 -7.21867800e-01 -4.89035219e-01 -6.02419734e-01] | [7.22449254989624, 6.2042694091796875] |
49b2f3fc-08c5-45d9-b280-d3270e3f1f53 | computational-discovery-of-new-2d-materials | 2012.09314 | null | https://arxiv.org/abs/2012.09314v1 | https://arxiv.org/pdf/2012.09314v1.pdf | Computational discovery of new 2D materials using deep learning generative models | Two dimensional (2D) materials have emerged as promising functional materials with many applications such as semiconductors and photovoltaics because of their unique optoelectronic properties. While several thousand 2D materials have been screened in existing materials databases, discovering new 2D materials remains to be challenging. Herein we propose a deep learning generative model for composition generation combined with random forest based 2D materials classifier to discover new hypothetical 2D materials. Furthermore, a template based element substitution structure prediction approach is developed to predict the crystal structures of a subset of the newly predicted hypothetical formulas, which allows us to confirm their structure stability using DFT calculations. So far, we have discovered 267,489 new potential 2D materials compositions and confirmed twelve 2D/layered materials by DFT formation energy calculation. Our results show that generative machine learning models provide an effective way to explore the vast chemical design space for new 2D materials discovery. | ['Jianjun Hu', 'Yong Zhao', 'Edirisuriya M. Dilanga Siriwardane', 'Yuqi Song'] | 2020-12-16 | null | null | null | null | ['formation-energy'] | ['miscellaneous'] | [ 3.18740517e-01 -8.90585184e-02 -3.12703520e-01 -2.24938586e-01
-5.29433370e-01 -3.86475086e-01 7.20674157e-01 1.16450498e-02
2.26006195e-01 1.40445352e+00 3.25240880e-01 -4.79221851e-01
1.45407915e-01 -1.08082533e+00 -7.86773503e-01 -1.07686961e+00
-4.92671877e-02 7.51578867e-01 1.32512853e-01 -1.76983684e-01
3.78310889e-01 6.63880587e-01 -1.79625380e+00 3.62796366e-01
1.21868348e+00 1.09359848e+00 5.29992521e-01 -5.07485382e-02
-4.61054057e-01 1.33176133e-01 -3.42130393e-01 -3.37535381e-01
3.78582686e-01 -4.77466375e-01 -4.03971612e-01 -2.13086396e-01
2.35229284e-01 -3.58052582e-01 -2.33770579e-01 7.76074648e-01
7.56958008e-01 -7.39336461e-02 1.30044866e+00 -8.03318679e-01
-1.01604652e+00 7.30148375e-01 -4.44177151e-01 -4.82831180e-01
6.23203337e-01 4.18615371e-01 1.11546457e+00 -1.06884789e+00
8.17521334e-01 1.02994084e+00 5.47984898e-01 9.30341423e-01
-1.29746366e+00 -9.86090541e-01 -2.51755595e-01 2.08242074e-01
-1.12108803e+00 -3.99024636e-01 1.08077526e+00 -6.12061262e-01
1.54437780e+00 -9.19047464e-03 8.94649446e-01 1.26203489e+00
4.67041016e-01 5.68352938e-01 1.18591309e+00 -2.74534941e-01
4.37210351e-01 -7.66451657e-02 -2.17048690e-01 7.71028996e-01
8.19593489e-01 5.06378338e-02 -7.26992726e-01 -2.35642374e-01
2.44676024e-01 -3.10897995e-02 -5.16980067e-02 -3.82221729e-01
-7.16234148e-01 5.73324025e-01 5.87921143e-01 1.96002707e-01
-6.23919189e-01 5.53859659e-02 1.24302007e-01 -1.99163944e-01
3.00267905e-01 9.22157109e-01 -6.43399835e-01 1.39286965e-01
-4.66915905e-01 5.43607712e-01 4.52039450e-01 7.52179980e-01
1.00164962e+00 2.56908029e-01 1.24121308e-01 5.85899651e-01
3.49308342e-01 3.24818283e-01 3.27726275e-01 -4.15383518e-01
-1.49076516e-02 8.88179123e-01 -4.76683788e-02 -3.31688493e-01
-3.59625250e-01 -1.41231596e-01 -9.09241498e-01 3.71519197e-03
-2.45662734e-01 1.36185676e-01 -1.03490102e+00 1.36362183e+00
4.16526675e-01 -3.83978367e-01 -8.27928036e-02 5.68335235e-01
1.18805885e+00 8.96529138e-01 3.25214528e-02 -4.86575156e-01
8.23688507e-01 -4.66320038e-01 -1.89189270e-01 4.77223516e-01
7.32496157e-02 -3.22084695e-01 7.50427425e-01 3.02895874e-01
-1.17525923e+00 -4.03990179e-01 -1.05023396e+00 9.49440151e-03
-3.81569296e-01 -2.47641116e-01 1.22905350e+00 8.32525551e-01
-6.43834233e-01 7.20119238e-01 -6.03767157e-01 -1.63238436e-01
7.29066610e-01 5.96933663e-01 -2.04212442e-01 -7.90240169e-02
-9.97509539e-01 5.13751924e-01 6.66213214e-01 -2.33988225e-01
-1.14746380e+00 -7.58907557e-01 -3.74522477e-01 -2.38416836e-01
1.86205462e-01 -1.09076202e+00 8.42850685e-01 -8.44951868e-02
-1.52308285e+00 8.47657859e-01 -2.88340837e-01 -2.68884569e-01
3.16476487e-02 3.99795026e-01 -4.42529380e-01 -1.28107980e-01
1.92971081e-01 7.29463875e-01 5.47866762e-01 -1.22611046e+00
-1.79335058e-01 -3.05409282e-01 -3.96873206e-01 1.85691472e-02
-2.19570756e-01 -5.73853910e-01 4.40937370e-01 -3.79246891e-01
3.51571023e-01 -5.59310675e-01 -4.56321210e-01 -2.52552450e-01
-8.83054078e-01 -5.20798385e-01 3.48610580e-01 -1.67532936e-01
7.89892554e-01 -1.51310813e+00 1.68570131e-01 3.62013251e-01
6.57524705e-01 3.00276675e-03 6.82680383e-02 6.67581022e-01
-5.90052195e-02 3.81297052e-01 -2.26264536e-01 2.85370469e-01
4.62731160e-02 -4.81178105e-01 -3.83716412e-02 1.89966455e-01
2.97389984e-01 1.07589126e+00 -7.04212248e-01 2.12517545e-01
1.52464777e-01 3.65439683e-01 -6.99177742e-01 7.33908042e-02
-9.89488721e-01 3.19674611e-01 -7.02055037e-01 1.01655388e+00
9.70685661e-01 -1.85634062e-01 4.13385779e-01 -1.92629278e-01
-4.08414334e-01 1.05730712e+00 -4.52474803e-01 1.50923049e+00
6.21827021e-02 7.95327201e-02 -5.53817093e-01 -8.74441862e-01
1.27824032e+00 -7.63954734e-03 8.25383604e-01 -1.03704119e+00
-1.83213294e-01 5.50702333e-01 3.88913095e-01 -3.67335856e-01
3.89576972e-01 -8.10168207e-01 7.27437362e-02 4.02569145e-01
3.94160971e-02 -4.78644222e-01 1.01831861e-01 -1.51235327e-01
1.10269845e+00 1.90190971e-01 2.24231988e-01 -3.71611714e-01
2.94300288e-01 1.12311259e-01 3.27044368e-01 5.93581200e-01
2.25785747e-01 2.05671057e-01 1.85387343e-01 -6.75022781e-01
-1.36631298e+00 -1.12911272e+00 -2.40418285e-01 4.12505507e-01
-6.01249151e-02 -5.14687836e-01 -3.29019636e-01 -2.74078637e-01
1.28695935e-01 6.08892381e-01 -3.55914861e-01 -3.72072995e-01
-2.71386683e-01 -1.05313504e+00 1.87870383e-01 2.49527872e-01
5.55646181e-01 -7.79517114e-01 9.19617116e-02 4.98938829e-01
3.44077319e-01 -4.79808420e-01 -3.26053705e-03 4.76089746e-01
-7.00384676e-01 -9.61431384e-01 -4.67807680e-01 -8.52748036e-01
4.82039332e-01 4.31484908e-01 8.20043266e-01 -2.23444685e-01
-6.70393050e-01 -2.56184280e-01 -4.17027138e-02 -5.81380188e-01
-9.00448799e-01 2.69620806e-01 5.60673177e-01 -6.63862348e-01
3.30542624e-01 -9.83218968e-01 -7.02594280e-01 -6.43328875e-02
-4.12926823e-01 4.85466272e-01 8.14383268e-01 5.53533912e-01
9.62469161e-01 4.94363308e-01 8.46866250e-01 -8.05033565e-01
6.45047486e-01 -5.47823012e-01 -5.99174082e-01 2.11105183e-01
-7.39787281e-01 4.72181022e-01 7.30444074e-01 -1.99178562e-01
-1.08726466e+00 9.91120115e-02 -5.16191959e-01 1.56178987e-02
-2.44954214e-01 4.22490597e-01 -6.49504066e-01 1.22222729e-01
6.36369407e-01 4.70141441e-01 -2.75768638e-01 -5.78573525e-01
8.45729113e-02 5.45538485e-01 3.57218198e-02 -9.99689579e-01
8.53026509e-01 -4.47248556e-02 4.08798814e-01 -1.08115280e+00
-4.33903515e-01 1.22886755e-01 -3.45720083e-01 -6.45173527e-03
8.63254607e-01 -8.97116661e-01 -1.15397441e+00 3.83836865e-01
-9.61782753e-01 -4.03428972e-02 -5.96758313e-02 2.49769479e-01
-2.68883526e-01 1.38626218e-01 -3.41378868e-01 -8.42058301e-01
-7.12450683e-01 -1.34056520e+00 1.03899455e+00 7.68316304e-03
-3.30009133e-01 -5.13311088e-01 1.62964627e-01 3.74763191e-01
3.41732502e-01 5.25148511e-01 1.91716731e+00 -3.93817723e-01
-1.03171206e+00 1.19204246e-01 3.63987349e-02 -9.77338701e-02
8.05530369e-01 -1.32190287e-01 -6.75408363e-01 1.15394248e-02
-2.66084075e-01 -1.79685757e-01 1.12037492e+00 3.64331067e-01
1.40404141e+00 -2.48475567e-01 -5.44988930e-01 3.55033100e-01
1.20006442e+00 8.11272681e-01 5.65559149e-01 2.33767927e-02
8.44601512e-01 9.87333953e-02 -1.74465459e-02 4.49097782e-01
6.05982877e-02 3.18507731e-01 4.49034065e-01 2.95203447e-01
-1.65344670e-01 -7.10974097e-01 1.86805114e-01 6.39131606e-01
-4.71818119e-01 -4.05795783e-01 -8.38514864e-01 -5.19280843e-02
-1.07598710e+00 -1.18879402e+00 -3.44051003e-01 2.13465333e+00
1.00861168e+00 3.76153052e-01 2.62977421e-01 6.59312010e-02
6.00343764e-01 -4.62519526e-02 -1.35263288e+00 -9.75871757e-02
-4.66071606e-01 7.77813911e-01 1.30557910e-01 -9.78174210e-02
-5.44837594e-01 8.68364811e-01 7.23713160e+00 6.61373556e-01
-1.17807078e+00 -3.99184048e-01 6.89343870e-01 -5.79636097e-02
-1.11560297e+00 5.55018969e-02 -1.08820999e+00 4.93647158e-01
4.20608044e-01 -2.52047598e-01 2.94841647e-01 6.97153389e-01
7.94268176e-02 -2.04344280e-02 -1.01269639e+00 9.83345032e-01
-4.62389112e-01 -2.23259759e+00 5.99665940e-01 4.58368897e-01
1.19518936e+00 -1.20466709e-01 4.41906862e-02 -4.92774099e-02
2.50756424e-02 -9.94044662e-01 5.78239262e-01 4.49258178e-01
1.13287139e+00 -6.49929285e-01 -9.00459066e-02 1.03052624e-01
-9.32223320e-01 2.94645261e-02 -4.39294606e-01 -1.35864004e-01
-1.14107326e-01 9.49598372e-01 -1.31371701e+00 4.12705988e-01
4.71208274e-01 7.03307927e-01 -1.08055919e-01 9.22405481e-01
-4.71181460e-02 4.12941962e-01 -1.90417737e-01 -8.12889278e-01
2.44469661e-02 -4.03623194e-01 3.76752466e-01 2.96449363e-01
5.90216041e-01 6.79101497e-02 3.17107849e-02 1.33048093e+00
-6.63575590e-01 2.85728090e-02 -6.99000478e-01 -4.79381084e-01
5.59262574e-01 7.00693667e-01 -6.13902450e-01 -1.86017647e-01
-2.18036890e-01 6.64064229e-01 3.22939716e-02 2.42447164e-02
-5.55044472e-01 -1.74219504e-01 9.12666142e-01 6.49477422e-01
3.24889094e-01 -4.35219646e-01 -1.17938258e-02 -1.12436795e+00
-1.22944258e-01 -7.43561327e-01 -1.64268911e-01 -8.34001899e-01
-1.61079645e+00 1.87806398e-01 -1.85491219e-01 -7.03660131e-01
2.73311555e-01 -8.74263346e-01 -5.53089321e-01 6.62387431e-01
-1.13738322e+00 -8.27216685e-01 6.27491772e-02 1.28603913e-02
4.80266303e-01 -3.77663344e-01 7.51397431e-01 4.97245155e-02
-5.55025816e-01 2.17003569e-01 5.90337217e-01 -7.60189474e-01
5.34216821e-01 -1.05396688e+00 6.20176673e-01 1.28776655e-01
-1.28394261e-01 6.01036489e-01 6.79537356e-01 -1.09793937e+00
-1.97814286e+00 -9.84698296e-01 5.25068641e-01 -6.00992031e-02
3.88653338e-01 -7.89945304e-01 -6.91420674e-01 -1.28234178e-01
2.10418757e-02 -7.67175734e-01 1.03713477e+00 -1.67964503e-01
-1.61930829e-01 5.53154536e-02 -1.22732472e+00 6.40026033e-01
1.62080932e+00 -2.95303255e-01 -6.90729022e-02 5.92549145e-01
9.69168842e-01 -1.30768567e-01 -7.41569519e-01 5.83015203e-01
7.55453646e-01 -1.00107491e+00 9.92916703e-01 -6.29835486e-01
6.75686836e-01 -2.59476185e-01 -4.53774601e-01 -9.77978349e-01
-4.27383840e-01 -7.11763322e-01 -2.17574567e-01 1.04229474e+00
7.75896311e-01 -5.10878026e-01 9.32815194e-01 7.71531999e-01
-5.89174032e-01 -9.84600842e-01 -8.80422473e-01 -6.71231151e-01
6.02659807e-02 -1.93816841e-01 1.31927323e+00 6.56349957e-01
-3.70724320e-01 7.04861701e-01 -2.34445736e-01 -4.12703395e-01
5.24609447e-01 5.46149552e-01 5.36859214e-01 -1.48936629e+00
-3.47360790e-01 -6.36285722e-01 -3.21384996e-01 -1.00168276e+00
3.22737336e-01 -1.27124083e+00 -3.27221215e-01 -1.82578325e+00
4.51731056e-01 -6.83514833e-01 -1.55257583e-01 3.28760952e-01
2.89589822e-01 -8.31996053e-02 -5.52748024e-01 6.92757368e-02
-1.93995312e-01 1.13261020e+00 1.33739507e+00 -4.25673664e-01
-3.51686209e-01 3.10787171e-01 -1.18435621e+00 1.16890185e-01
7.38015294e-01 -1.51165873e-01 -2.38048255e-01 -5.78138232e-02
6.79034948e-01 -3.24106693e-01 -4.24620025e-02 -8.41961503e-01
-4.67699379e-01 -4.23045516e-01 7.87641406e-01 -1.05737197e+00
3.32092017e-01 -5.30730784e-01 4.97866005e-01 6.57614112e-01
1.58743951e-02 -6.11379385e-01 9.45391133e-02 4.30991799e-01
3.46032143e-01 -1.57231778e-01 5.12784898e-01 -4.21871692e-01
-5.81788421e-01 8.10154378e-01 -7.26846993e-01 -5.79905748e-01
9.36386943e-01 -4.89244640e-01 -1.56041086e-01 9.93398875e-02
-3.90590638e-01 -2.26260483e-01 7.35890210e-01 1.50296926e-01
7.80764699e-01 -1.34156847e+00 -2.01328322e-01 2.53179193e-01
5.70140556e-02 2.99300224e-01 2.50354230e-01 4.89642359e-02
-5.47137141e-01 5.15960753e-01 -4.24138129e-01 -3.89010757e-01
-8.28333795e-01 6.51158631e-01 2.93783218e-01 8.41599703e-02
-2.68648565e-01 8.24533582e-01 1.08126514e-01 -3.64608496e-01
-2.27201164e-01 -2.10366443e-01 3.58180225e-01 -5.94747439e-02
8.21355432e-02 2.59061337e-01 3.52916598e-01 -1.51059017e-01
-4.40332711e-01 2.91105956e-01 -1.94914401e-01 5.06193936e-01
2.05682969e+00 4.02238369e-01 -3.07670921e-01 1.66307181e-01
1.16267276e+00 -1.87519282e-01 -9.89115596e-01 -6.38470724e-02
-7.31499270e-02 -1.49834469e-01 -3.45696002e-01 -6.65675163e-01
-6.14295721e-01 7.66270339e-01 3.29410732e-01 -2.19404371e-03
1.07850003e+00 3.52367193e-01 1.02899647e+00 7.30801880e-01
5.45820475e-01 -7.79590189e-01 1.42147869e-01 2.84210086e-01
7.28242218e-01 -1.16942072e+00 3.20030868e-01 -2.03443661e-01
3.87029983e-02 1.09429479e+00 7.33960450e-01 3.00905734e-01
5.69823027e-01 -3.02020069e-02 -9.54512358e-01 -4.12931979e-01
-9.01632190e-01 -8.94610733e-02 8.09418559e-02 8.23654592e-01
6.26619220e-01 2.13089034e-01 -2.21497685e-01 5.47153413e-01
-4.03444678e-01 -5.22006631e-01 1.76759914e-01 8.01092803e-01
-6.26061618e-01 -1.69223237e+00 1.29057929e-01 8.67431045e-01
3.66996191e-02 -3.03055078e-01 -8.24514329e-01 3.57281923e-01
2.54801929e-01 3.28879237e-01 -1.38210803e-01 -4.97559875e-01
1.21752676e-02 3.58989805e-01 1.00079513e+00 -5.53247690e-01
-9.28508192e-02 3.58542725e-02 8.20074379e-02 9.39858034e-02
-4.79290217e-01 -6.10747755e-01 -1.30318975e+00 -2.86258280e-01
-6.34715080e-01 1.02148287e-01 9.31071818e-01 7.98751831e-01
6.52722239e-01 2.41281360e-01 9.47866559e-01 -9.67983305e-01
1.18774943e-01 -4.55484509e-01 -5.81154525e-01 -9.39432904e-02
-1.57296032e-01 -7.82480896e-01 3.32643121e-01 -2.48750493e-01] | [5.143200397491455, 5.403586387634277] |
0727e262-de8c-4e4a-a675-028e8e13eb52 | survey-on-the-attention-based-rnn-model-and | 1601.06823 | null | http://arxiv.org/abs/1601.06823v1 | http://arxiv.org/pdf/1601.06823v1.pdf | Survey on the attention based RNN model and its applications in computer vision | The recurrent neural networks (RNN) can be used to solve the sequence to
sequence problem, where both the input and the output have sequential
structures. Usually there are some implicit relations between the structures.
However, it is hard for the common RNN model to fully explore the relations
between the sequences. In this survey, we introduce some attention based RNN
models which can focus on different parts of the input for each output item, in
order to explore and take advantage of the implicit relations between the input
and the output items. The different attention mechanisms are described in
detail. We then introduce some applications in computer vision which apply the
attention based RNN models. The superiority of the attention based RNN model is
shown by the experimental results. At last some future research directions are
given. | ['Feng Wang', 'David M. J. Tax'] | 2016-01-25 | null | null | null | null | ['implicit-relations'] | ['natural-language-processing'] | [ 3.24777186e-01 -6.72901943e-02 -1.37436137e-01 -2.08195075e-01
2.72632539e-01 -2.54415244e-01 2.66897529e-01 -2.94453084e-01
-3.90880257e-01 5.15212834e-01 5.03124356e-01 -4.32523131e-01
-1.91720068e-01 -7.25410223e-01 -1.71733737e-01 -6.96128368e-01
1.41269222e-01 3.07777882e-01 1.26198247e-01 -6.32254958e-01
6.34226382e-01 4.96409148e-01 -1.54636371e+00 3.73751372e-01
5.44517875e-01 5.84922135e-01 9.36683834e-01 8.48463535e-01
-5.99045873e-01 1.30655539e+00 -5.96962810e-01 7.58237243e-02
-4.16165404e-02 -7.82647431e-01 -1.04633307e+00 -3.84431668e-02
-5.89481667e-02 -3.42126675e-02 -4.83397752e-01 1.31241119e+00
4.72397566e-01 6.94463909e-01 6.68345153e-01 -9.17001069e-01
-1.32084382e+00 7.82165170e-01 -3.81454647e-01 9.63650405e-01
1.72923818e-01 -4.86614741e-02 1.15215719e+00 -1.09077418e+00
3.79672676e-01 1.33398867e+00 4.13205296e-01 7.06537545e-01
-6.38432860e-01 -5.85956395e-01 6.88365102e-01 8.20777297e-01
-1.11269557e+00 -6.34945929e-02 6.92862689e-01 -3.23620290e-01
1.42529190e+00 3.57125670e-01 6.30235434e-01 9.17608321e-01
1.73246443e-01 9.40215886e-01 6.36995673e-01 -7.87775278e-01
-3.94345284e-01 -5.41452952e-02 8.66016626e-01 1.93261862e-01
-4.78171296e-02 8.02017078e-02 -1.61138996e-01 3.72736990e-01
1.04565406e+00 4.37227219e-01 -3.71316344e-01 1.38292536e-01
-8.69084597e-01 9.86212790e-01 7.78314352e-01 9.25125599e-01
-5.79453290e-01 5.42694964e-02 5.10773718e-01 2.87325710e-01
1.17807597e-01 5.32392085e-01 -5.06954432e-01 3.22338492e-02
-4.11065936e-01 -2.17329755e-01 3.96133512e-01 1.11513340e+00
6.76688612e-01 3.76524210e-01 -3.03129375e-01 9.86424685e-01
3.25320631e-01 -9.80487326e-04 9.96991098e-01 -4.66320336e-01
5.05252898e-01 4.15564865e-01 -3.12771678e-01 -1.11970174e+00
-2.02145711e-01 -4.50805843e-01 -1.00511575e+00 -8.87378380e-02
-1.71633199e-01 -4.94560711e-02 -9.14170384e-01 1.61166918e+00
-1.91844597e-01 2.40644872e-01 9.38738212e-02 9.70627904e-01
1.16850913e+00 9.58737910e-01 1.74057424e-01 -4.21380639e-01
1.39490986e+00 -1.46531248e+00 -1.30472982e+00 -4.88573402e-01
4.52083349e-01 -7.50098050e-01 1.00271904e+00 -6.10090196e-02
-1.08880389e+00 -9.04646099e-01 -7.54004657e-01 -4.39377606e-01
-4.56010759e-01 -8.55333582e-02 3.57621193e-01 1.07762784e-01
-1.05118585e+00 5.53221285e-01 -4.09842491e-01 -5.23773491e-01
-2.52831783e-02 5.10028422e-01 1.01335235e-01 1.60288647e-01
-1.66623926e+00 1.15340257e+00 8.87060642e-01 7.04549193e-01
-6.47146583e-01 -2.31960326e-01 -8.90511036e-01 3.93022060e-01
3.46916020e-01 -5.36571085e-01 1.52212691e+00 -1.47603500e+00
-1.23152256e+00 3.82764161e-01 -3.02667946e-01 -2.94373751e-01
-8.60732719e-02 -5.40490866e-01 -3.80932868e-01 -1.61941737e-01
-2.02871278e-01 3.80212456e-01 3.91415417e-01 -6.70278668e-01
-7.07391381e-01 -1.88071921e-01 -3.82114649e-02 5.35470247e-01
-3.10720861e-01 4.43826109e-01 -5.75491726e-01 -1.01586401e+00
-5.56607842e-02 -7.02912033e-01 -4.76779878e-01 -5.55532694e-01
-2.34621316e-01 -5.07659137e-01 9.44759488e-01 -5.48382938e-01
1.57122684e+00 -2.10247040e+00 4.96297181e-01 -1.10907108e-01
-7.14399740e-02 6.21243238e-01 -3.00574005e-01 5.99130630e-01
-6.97335482e-01 1.07675508e-01 -5.50731234e-02 1.49487287e-01
-5.38338721e-01 4.49165255e-01 -4.72130895e-01 5.44745363e-02
-2.76510175e-02 1.25480640e+00 -9.66148496e-01 -2.55575895e-01
9.03746337e-02 3.73328686e-01 -1.34097561e-01 4.91775751e-01
-7.89142475e-02 2.08421662e-01 -3.68014336e-01 7.76604041e-02
1.26941830e-01 -2.87603736e-01 2.76177496e-01 5.07491194e-02
-1.25846922e-01 6.49814248e-01 -7.94219792e-01 1.20834970e+00
-8.93315524e-02 1.09369564e+00 -2.95808583e-01 -1.11718559e+00
1.15541267e+00 6.25676870e-01 -1.55705675e-01 -5.93366146e-01
3.48756045e-01 -3.33619006e-02 6.99039340e-01 -8.68437946e-01
6.28047466e-01 -5.66163659e-01 3.17007750e-01 5.81027269e-01
8.33354145e-03 4.83835042e-01 1.20331801e-01 8.64137784e-02
4.55943525e-01 -1.52195588e-01 8.39650571e-01 -4.14676517e-01
9.89700019e-01 -4.11305964e-01 6.51053488e-01 6.92394674e-01
-1.28532365e-01 4.54872042e-01 3.51694405e-01 -6.92274094e-01
-1.19306827e+00 -3.33638489e-01 3.92401308e-01 1.37017095e+00
1.99187584e-02 -2.10693389e-01 -5.03540337e-01 -6.09552324e-01
-5.38873196e-01 7.70810008e-01 -9.62797701e-01 -1.37089700e-01
-7.48829544e-01 -3.10112357e-01 2.25692332e-01 1.17375064e+00
3.18314910e-01 -2.05401707e+00 -4.64274824e-01 2.38733366e-01
-2.87857234e-01 -6.08051419e-01 -7.84421325e-01 2.80927986e-01
-1.21255970e+00 -7.49824524e-01 -7.90043712e-01 -1.22512031e+00
6.38510764e-01 5.14724433e-01 9.44406152e-01 5.47899008e-01
-2.39461064e-02 1.12144902e-01 -5.96538007e-01 -4.83506799e-01
-1.24512084e-01 4.05388027e-01 3.19788978e-02 -1.62924439e-01
6.32506192e-01 -5.42564094e-01 -2.64014304e-01 3.03931922e-01
-9.10105824e-01 1.02133825e-01 7.26954401e-01 8.60039234e-01
2.44362354e-01 -1.59264773e-01 6.78567648e-01 -9.87774909e-01
9.31465328e-01 -7.01330960e-01 -2.66338289e-01 3.31480682e-01
-2.61053711e-01 2.49838218e-01 5.58785021e-01 -6.47086322e-01
-1.12480605e+00 -2.85802871e-01 -3.02114546e-01 -5.78408837e-01
-4.63750921e-02 8.41592014e-01 -2.24725023e-01 4.18335974e-01
3.61067772e-01 4.20715094e-01 8.26346576e-02 -6.98668957e-01
3.65829468e-01 6.58341706e-01 1.37210101e-01 -2.18303293e-01
3.12783688e-01 -2.09263191e-01 -4.50080335e-01 -8.97971809e-01
-9.59148347e-01 -6.33032858e-01 -8.58415067e-01 -5.31354249e-02
7.77811766e-01 -3.33089083e-01 -7.53429651e-01 1.95513278e-01
-1.58722198e+00 -1.90830350e-01 -3.23108792e-01 5.86913168e-01
-5.34492314e-01 5.82676411e-01 -1.00537515e+00 -9.73954797e-01
-5.84734440e-01 -1.11289752e+00 2.25866586e-01 5.16189396e-01
-4.96977597e-01 -1.30777693e+00 3.17901745e-02 -1.08916298e-01
4.51638967e-01 -6.42051041e-01 1.08857799e+00 -1.11815619e+00
-4.62519765e-01 1.94058508e-01 -2.51758903e-01 2.58113742e-01
5.31464636e-01 -2.13026166e-01 -7.40414202e-01 -5.43688834e-02
4.98772353e-01 1.72499160e-03 1.02303743e+00 3.16222191e-01
1.22599816e+00 -5.46632230e-01 -2.45088413e-01 2.09988177e-01
1.23733175e+00 9.12752748e-01 1.14779305e+00 3.76069605e-01
7.54738986e-01 9.05105948e-01 5.46822846e-01 -8.40150043e-02
-1.77214548e-01 3.19534630e-01 3.32302928e-01 1.25870541e-01
1.83212720e-02 -1.85810745e-01 3.02410841e-01 1.62841356e+00
-3.80525649e-01 -4.57872212e-01 -7.62304127e-01 7.37558126e-01
-2.12884188e+00 -1.30827773e+00 -3.27490032e-01 1.72937322e+00
6.10139966e-01 2.74260584e-02 -1.16298206e-01 4.69310284e-02
1.12044108e+00 1.78568780e-01 -6.28580391e-01 -8.37975800e-01
-5.40678836e-02 4.54908572e-02 4.55671549e-02 6.96530700e-01
-7.55585015e-01 1.08553958e+00 7.60260773e+00 8.37092698e-01
-9.20202136e-01 -2.09940508e-01 4.01879698e-01 8.40056762e-02
-2.66413391e-01 -1.03801653e-01 -1.03992975e+00 2.68699378e-01
9.16237950e-01 -1.19239494e-01 4.09358501e-01 8.21475327e-01
-1.04900692e-02 2.44815350e-01 -9.87522364e-01 8.93906593e-01
1.87975898e-01 -1.26939332e+00 5.03376663e-01 -3.45387198e-02
4.69555080e-01 -1.34483278e-01 7.36210048e-02 4.22152162e-01
2.91851550e-01 -1.33570087e+00 2.74080157e-01 7.34414339e-01
3.42327684e-01 -8.69445443e-01 9.89808857e-01 6.60157561e-01
-1.33076668e+00 -1.55264974e-01 -7.13322878e-01 -6.50918901e-01
8.88208449e-02 8.26200396e-02 -7.09869385e-01 5.44688344e-01
7.35118091e-01 1.06642175e+00 -2.25919306e-01 9.32049215e-01
-8.02957237e-01 3.47477198e-01 2.83457309e-01 -6.19313836e-01
4.78528142e-01 -4.07646269e-01 5.20994663e-01 1.39862251e+00
2.52326876e-01 4.94482219e-01 -6.73515052e-02 9.13482189e-01
-1.83140617e-02 1.49425536e-01 -7.95952559e-01 -6.29524067e-02
2.57936895e-01 1.12280285e+00 -6.90703452e-01 -4.65138793e-01
-4.86139327e-01 8.55759442e-01 6.60136461e-01 6.04226470e-01
-5.51158190e-01 -7.00922906e-01 7.90743828e-01 -2.34750718e-01
6.14072263e-01 -1.87675983e-01 -1.74272701e-01 -8.89056921e-01
-3.49092931e-01 -1.01029360e+00 6.11564517e-01 -1.05839455e+00
-1.41608036e+00 8.54980111e-01 -2.86175191e-01 -9.06604230e-01
-2.82946467e-01 -6.51364684e-01 -7.65705168e-01 1.42478204e+00
-1.32825708e+00 -5.03719866e-01 1.55568555e-01 3.30719382e-01
1.28619921e+00 -2.09207177e-01 7.38208532e-01 7.98492786e-03
-6.68741822e-01 2.55967498e-01 -8.11189711e-02 6.13345623e-01
6.04604296e-02 -8.92260849e-01 4.89704311e-01 8.00592124e-01
2.59508699e-01 1.37613416e+00 5.93847752e-01 -8.39494467e-01
-8.33342373e-01 -8.58933866e-01 1.31709146e+00 -2.62968004e-01
4.28250700e-01 -1.09487258e-01 -1.47610259e+00 1.30875480e+00
9.28223431e-01 -3.17425609e-01 8.91706467e-01 4.36113298e-01
-1.34860948e-01 2.75755733e-01 -4.66439992e-01 8.18738937e-01
9.69400823e-01 -6.30661607e-01 -1.30261850e+00 -1.81355700e-01
1.03079963e+00 -1.16589749e-02 -3.30023795e-01 3.11623961e-01
4.18693066e-01 -7.54719317e-01 8.21014404e-01 -1.25140178e+00
5.14302552e-01 -4.49666083e-01 -5.48477471e-02 -1.24863279e+00
-1.07308829e+00 -4.51464295e-01 -1.26171470e-01 1.11557853e+00
4.47918564e-01 -4.12826777e-01 3.12155336e-01 4.41707760e-01
-1.85457602e-01 -6.63626969e-01 -3.72140080e-01 -8.17010045e-01
4.25734892e-02 -4.78177041e-01 4.76284474e-01 8.54808092e-01
2.52036899e-01 1.11819386e+00 -5.18943012e-01 -1.65229052e-01
-1.71559587e-01 1.34396115e-02 1.16555542e-01 -1.12414384e+00
-2.43572205e-01 -6.75410867e-01 -2.68679500e-01 -1.57027411e+00
1.17892556e-01 -7.91066885e-01 1.62610903e-01 -1.67564619e+00
3.45200509e-01 6.99248211e-03 -7.77451932e-01 4.35213268e-01
-5.68640053e-01 -2.10695878e-01 4.55590367e-01 4.66054857e-01
-5.67399979e-01 7.84031272e-01 1.23885274e+00 7.46060163e-03
-2.67154098e-01 1.27479464e-01 -8.48461807e-01 7.94281840e-01
9.76102293e-01 -3.93164128e-01 -6.20181262e-01 -6.42681003e-01
1.98437080e-01 -4.00240421e-02 -2.30425969e-01 -5.10253072e-01
5.84786117e-01 -3.22703779e-01 3.47625643e-01 -9.93967474e-01
8.60545635e-02 -9.32815552e-01 8.60074162e-03 6.50593698e-01
-7.56645560e-01 4.30439144e-01 2.84748465e-01 4.77789879e-01
-2.88351476e-01 -8.42561245e-01 7.11707473e-01 -6.18660271e-01
-8.71806681e-01 1.37983531e-01 -8.33789945e-01 -8.06614831e-02
9.17308688e-01 -2.48376176e-01 8.08461010e-02 -5.79061091e-01
-8.20709586e-01 2.80011833e-01 -2.36881986e-01 7.17862010e-01
8.71210515e-01 -1.47862148e+00 -4.84855354e-01 2.42979214e-01
-2.25702241e-01 -2.28924304e-01 3.73425603e-01 5.15752435e-01
-2.89428413e-01 9.13803697e-01 -4.99846071e-01 -8.11026767e-02
-1.57761514e+00 1.42927980e+00 5.01252890e-01 -3.01249534e-01
-4.75744426e-01 9.49591637e-01 7.85332084e-01 -2.78706938e-01
4.18113649e-01 -2.23447144e-01 -1.07623017e+00 2.08817068e-02
1.06215441e+00 3.85173798e-01 -3.25544357e-01 -5.64919412e-01
-2.55109251e-01 6.07994556e-01 -5.18011093e-01 1.33333704e-03
1.25408435e+00 -3.26741964e-01 -4.81929958e-01 8.77919853e-01
1.02616692e+00 -5.33457816e-01 -7.08368838e-01 -6.15655839e-01
3.98011893e-01 -1.90938205e-01 -3.66574258e-01 -4.43305641e-01
-1.22573423e+00 1.36656117e+00 1.15698352e-01 3.40303302e-01
1.24927545e+00 -1.21113680e-01 6.33262753e-01 3.42889458e-01
-6.87232837e-02 -9.23519194e-01 1.81795314e-01 1.28861082e+00
1.23084736e+00 -6.73389554e-01 -1.76708102e-01 -4.04043615e-01
-6.20998740e-01 1.41770995e+00 9.40976977e-01 -2.96820432e-01
5.46735525e-01 1.88107938e-02 3.55565213e-02 -1.28445268e-01
-1.04786730e+00 -3.99068117e-01 3.30085814e-01 5.94229579e-01
8.60630095e-01 -3.54419380e-01 -3.49205852e-01 6.41990542e-01
2.84820814e-02 -1.37324423e-01 3.89364660e-01 7.11285651e-01
-7.73950338e-01 -1.22257447e+00 -4.56908405e-01 1.06035568e-01
-4.64004695e-01 -4.92004007e-01 -4.88671601e-01 6.03597939e-01
-8.63885507e-02 7.36580551e-01 2.70309538e-01 -3.11208516e-01
3.48998576e-01 1.61751181e-01 9.79844779e-02 -9.61665452e-01
-9.39010262e-01 1.53544098e-01 -2.67716706e-01 -4.29294705e-02
-3.94360453e-01 -5.57173133e-01 -1.27981532e+00 -8.87652040e-02
-3.81710380e-01 3.93794000e-01 1.89441979e-01 1.06730056e+00
3.34784269e-01 9.56740737e-01 3.09453011e-01 -4.72669542e-01
-3.21680576e-01 -1.36071837e+00 -4.24005330e-01 9.65864584e-02
4.32609260e-01 -4.05535787e-01 2.88734259e-03 -6.71792477e-02] | [11.03646183013916, 6.974472999572754] |
efa35633-10ad-4fd6-9fb7-6d90d4e3291d | self-learning-symmetric-multi-view | 2305.07307 | null | https://arxiv.org/abs/2305.07307v2 | https://arxiv.org/pdf/2305.07307v2.pdf | Self-Learning Symmetric Multi-view Probabilistic Clustering | Multi-view Clustering (MVC) has achieved significant progress, with many efforts dedicated to learn knowledge from multiple views. However, most existing methods are either not applicable or require additional steps for incomplete MVC. Such a limitation results in poor-quality clustering performance and poor missing view adaptation. Besides, noise or outliers might significantly degrade the overall clustering performance, which are not handled well by most existing methods. In this paper, we propose a novel unified framework for incomplete and complete MVC named self-learning symmetric multi-view probabilistic clustering (SLS-MPC). SLS-MPC proposes a novel symmetric multi-view probability estimation and equivalently transforms multi-view pairwise posterior matching probability into composition of each view's individual distribution, which tolerates data missing and might extend to any number of views. Then, SLS-MPC proposes a novel self-learning probability function without any prior knowledge and hyper-parameters to learn each view's individual distribution. Next, graph-context-aware refinement with path propagation and co-neighbor propagation is used to refine pairwise probability, which alleviates the impact of noise and outliers. Finally, SLS-MPC proposes a probabilistic clustering algorithm to adjust clustering assignments by maximizing the joint probability iteratively without category information. Extensive experiments on multiple benchmarks show that SLS-MPC outperforms previous state-of-the-art methods. | ['Jieping Ye', 'Chen Shen', 'Yaowu Chen', 'Rongxin Jiang', 'Junlong Liu', 'Junjie Liu'] | 2023-05-12 | null | null | null | null | ['incomplete-multi-view-clustering', 'self-learning'] | ['computer-vision', 'natural-language-processing'] | [-1.86710045e-01 -1.84905365e-01 -2.65491903e-01 -3.23361695e-01
-9.73190188e-01 -5.99746704e-01 2.79048562e-01 2.52151549e-01
4.78520542e-02 3.18601131e-01 2.22550482e-01 1.38719216e-01
-2.58129984e-01 -5.99095583e-01 -6.81770921e-01 -9.00447011e-01
1.98418975e-01 7.83365905e-01 6.66390955e-01 2.35419229e-01
8.48343770e-05 1.03766592e-02 -1.31518209e+00 4.41701710e-01
8.90838683e-01 4.39592391e-01 3.47242534e-01 4.94205058e-01
1.87486783e-02 2.44011536e-01 -3.50358129e-01 -4.37869877e-01
1.18639283e-01 -3.66974115e-01 -5.31781137e-01 4.51507241e-01
2.52569437e-01 9.86649171e-02 -4.23214994e-02 1.14189327e+00
4.17960584e-01 -1.13313913e-03 7.22902298e-01 -1.44471967e+00
-4.25289452e-01 8.02222610e-01 -1.09489679e+00 -1.53790906e-01
2.24527255e-01 4.49951534e-04 8.85192633e-01 -7.98546255e-01
5.42958260e-01 1.22530127e+00 6.30696356e-01 2.97136605e-01
-1.42769349e+00 -4.02672946e-01 6.55069172e-01 3.89705449e-01
-1.74971390e+00 -1.22025259e-01 9.14482594e-01 -3.26395839e-01
3.89863998e-01 1.13209426e-01 5.98488033e-01 8.48673046e-01
-1.18399180e-01 7.85064936e-01 1.16238081e+00 -1.31775588e-01
3.34621727e-01 1.96851909e-01 -2.19896302e-01 6.56246781e-01
3.24228913e-01 -4.66088891e-01 -4.13272381e-01 -3.14593345e-01
4.92469668e-01 2.22738922e-01 -4.32561219e-01 -1.04811156e+00
-1.42743850e+00 5.67712903e-01 1.49085388e-01 1.39078438e-01
-2.57160723e-01 -2.09293872e-01 3.11626196e-01 -1.59026325e-01
2.31527358e-01 -1.07399821e-01 -5.29214203e-01 7.78099149e-02
-1.00720394e+00 -1.00623786e-01 7.07056403e-01 1.26991367e+00
9.29167569e-01 -2.87909746e-01 -7.74093438e-03 9.12119091e-01
3.97263169e-01 6.14266634e-01 7.75846243e-02 -1.26511145e+00
4.65682626e-01 9.15977061e-01 -1.71462476e-01 -1.24321282e+00
-2.98199445e-01 -5.61104715e-01 -1.24460602e+00 -4.75250520e-02
2.21763104e-01 4.44963574e-03 -7.25933969e-01 1.65901363e+00
6.88419342e-01 3.63452762e-01 -6.62477165e-02 8.52434576e-01
7.04363644e-01 5.03558517e-01 -3.09579819e-01 -5.55370569e-01
1.08553326e+00 -9.18795764e-01 -4.93571371e-01 9.95181352e-02
2.36201704e-01 -7.80208349e-01 8.15076172e-01 5.55279911e-01
-9.32402313e-01 -5.04871309e-01 -9.31875169e-01 5.86684167e-01
-9.27931517e-02 -1.65182114e-01 2.68796265e-01 6.31290436e-01
-9.53936934e-01 2.54112184e-01 -8.83535564e-01 -2.65552610e-01
3.60619694e-01 3.18609029e-01 -4.20086145e-01 -5.81553459e-01
-5.89067578e-01 3.45498621e-01 5.64473510e-01 -1.08477101e-01
-6.63908601e-01 -7.60081351e-01 -8.19293559e-01 1.26854377e-02
8.68042767e-01 -9.89382565e-01 5.66172481e-01 -5.35671651e-01
-1.30487573e+00 4.38805580e-01 -4.31266904e-01 7.32273012e-02
3.53537500e-01 1.07895739e-01 -5.51376224e-01 2.76535779e-01
2.47565269e-01 6.27723038e-01 8.94141972e-01 -2.15666938e+00
-7.42919803e-01 -4.84064132e-01 -2.65594929e-01 5.04842341e-01
-1.15469284e-01 -4.32120234e-01 -1.50240684e+00 -6.04986846e-01
6.30843580e-01 -1.00338542e+00 -4.73927468e-01 -4.01357979e-01
-5.94844878e-01 -9.24162716e-02 9.31190372e-01 -2.91482031e-01
1.33170259e+00 -2.06058431e+00 6.15507364e-01 5.46951830e-01
3.59117478e-01 -1.08922459e-01 -5.63446246e-02 3.98211956e-01
2.15472028e-01 9.04881582e-02 -5.60935259e-01 -6.81855440e-01
-8.69368836e-02 4.79691118e-01 3.83767933e-01 6.34095252e-01
-3.79065305e-01 3.61830205e-01 -8.90028656e-01 -8.90852094e-01
4.48882729e-01 4.99520838e-01 -8.24885249e-01 9.11640301e-02
-9.03047994e-02 6.35461271e-01 -6.96159378e-02 7.82518327e-01
1.05183601e+00 -5.42840302e-01 5.67854404e-01 -4.45426941e-01
2.16518059e-01 -6.11704111e-01 -1.69437206e+00 1.78839874e+00
-2.39570618e-01 -1.79266289e-01 2.66609132e-01 -9.37770188e-01
6.22660935e-01 1.35711044e-01 7.92269170e-01 1.21688163e-02
-2.17766032e-01 -7.37793222e-02 -2.53223777e-01 -2.48597145e-01
1.77704975e-01 -3.53487320e-02 -4.98318784e-02 3.23381960e-01
1.09019153e-01 9.38897058e-02 1.67861417e-01 5.60331047e-01
8.20024550e-01 7.16565177e-02 2.49733925e-01 -3.00244689e-02
9.29580748e-01 -1.15638234e-01 1.13103402e+00 5.87119102e-01
-2.43446186e-01 1.10622144e+00 3.14641446e-01 1.41309932e-01
-7.06421435e-01 -1.36754370e+00 1.09551519e-01 8.13897431e-01
5.32510877e-01 -8.40269685e-01 -7.59760499e-01 -1.14169490e+00
-1.60519347e-01 4.88396466e-01 -3.17180663e-01 -4.80826087e-02
-2.92504698e-01 -8.31119537e-01 -6.14395700e-02 4.12204146e-01
3.25140953e-01 -4.40249771e-01 2.45353311e-01 1.32761791e-01
-6.47344232e-01 -1.29607308e+00 -8.57307494e-01 -2.31782168e-01
-9.18501139e-01 -1.29947615e+00 -6.30575001e-01 -6.98232293e-01
9.26783264e-01 5.92790842e-01 1.02615714e+00 -5.85841648e-02
2.38509737e-02 8.54228258e-01 -4.04733062e-01 1.40257508e-01
-4.07835364e-01 8.95973518e-02 1.53152362e-01 3.75128090e-01
3.73577684e-01 -9.21103477e-01 -8.18469405e-01 7.15451717e-01
-6.89266801e-01 9.40851960e-03 6.36838257e-01 5.52221835e-01
1.19447696e+00 5.05090773e-01 3.85907143e-01 -1.18497860e+00
-7.63626695e-02 -6.53642416e-01 -5.49664676e-01 4.54127342e-01
-8.43855917e-01 -1.22508891e-01 7.33711421e-01 -2.24305391e-01
-1.16490126e+00 3.31034362e-01 1.85193479e-01 -9.06965911e-01
-5.05322158e-01 4.25247490e-01 -7.89657056e-01 1.74750164e-01
6.93616122e-02 4.56414491e-01 -9.24411416e-02 -4.68149900e-01
5.35036027e-01 3.39760244e-01 7.51105070e-01 -5.01685262e-01
1.00532174e+00 8.22119176e-01 -2.71782745e-02 -5.41416705e-01
-8.42204630e-01 -9.93833899e-01 -1.06738544e+00 -3.67170215e-01
8.34639907e-01 -1.08602667e+00 -6.81752563e-01 4.86596555e-01
-6.67840779e-01 2.71834936e-02 1.71276197e-01 5.95396340e-01
-4.75691855e-01 8.09133708e-01 -3.33746701e-01 -5.40254474e-01
1.68904327e-02 -1.18935061e+00 9.64980960e-01 9.03535560e-02
1.14700615e-01 -1.16641915e+00 2.62224246e-02 8.21297526e-01
-9.22820643e-02 2.03995883e-01 7.94129431e-01 -5.17304301e-01
-7.34769285e-01 -1.93812270e-02 -1.11346014e-01 3.64792049e-01
2.71632940e-01 1.16547294e-01 -6.03877366e-01 -5.79299927e-01
-1.90313786e-01 1.35930374e-01 7.53963232e-01 5.46037436e-01
1.26438999e+00 -1.25707045e-01 -6.17252648e-01 8.57401073e-01
1.69588876e+00 -8.72653797e-02 3.70821536e-01 -4.71012667e-02
1.21980035e+00 4.66258079e-01 5.46310425e-01 6.81313336e-01
1.06315017e+00 5.11582792e-01 6.14491343e-01 1.78729534e-01
-7.52024651e-02 -2.82072246e-01 2.87166387e-01 1.36789978e+00
-1.58831716e-01 -2.63000488e-01 -7.85942256e-01 6.12169147e-01
-2.07971025e+00 -9.49434161e-01 -4.59106952e-01 2.38078237e+00
4.72956359e-01 1.07272610e-01 2.55197644e-01 2.56372523e-02
1.17627370e+00 2.21400663e-01 -5.37678421e-01 3.65011126e-01
-2.12542266e-01 -4.31178838e-01 4.72855717e-01 4.56897348e-01
-1.23834300e+00 7.29451597e-01 5.44524384e+00 8.33743811e-01
-4.25561905e-01 2.37041697e-01 4.06481713e-01 -2.63486840e-02
-5.19765198e-01 2.37145692e-01 -7.97554433e-01 5.57526946e-01
4.04104590e-01 2.01560616e-01 4.40195858e-01 6.91793919e-01
4.23912657e-03 -2.64860064e-01 -8.80833268e-01 1.25727153e+00
4.97598857e-01 -1.17245162e+00 7.93168768e-02 6.43638521e-02
1.15626562e+00 -1.37682751e-01 -1.63325042e-01 2.54449278e-01
4.99488205e-01 -4.32103246e-01 3.03982139e-01 6.35670483e-01
5.40051758e-01 -1.10835683e+00 7.40465939e-01 6.01767361e-01
-1.52599645e+00 5.61863966e-02 -3.42855871e-01 6.01389647e-01
3.19199681e-01 7.92669594e-01 -5.18717647e-01 1.10710573e+00
9.31087017e-01 9.42176878e-01 -7.58256078e-01 1.02375436e+00
-1.11615479e-01 7.68058896e-01 -4.35281157e-01 4.54832196e-01
-2.04769187e-02 -3.76342952e-01 8.21819305e-01 1.08866429e+00
2.07016304e-01 8.63779187e-02 7.48401999e-01 4.88927186e-01
3.61921974e-02 7.50903934e-02 -2.81732976e-01 6.75730824e-01
8.12971115e-01 1.38095963e+00 -1.05226469e+00 -3.97770464e-01
-6.94559455e-01 1.21892798e+00 3.02594930e-01 5.00541210e-01
-8.09591711e-01 1.50225028e-01 4.27617937e-01 1.51038677e-01
6.19832516e-01 -1.23991609e-01 -2.59513289e-01 -1.37261963e+00
1.30282259e-02 -7.58202314e-01 9.25970256e-01 -4.47439939e-01
-1.60205245e+00 1.65395528e-01 1.98840961e-01 -1.47993374e+00
1.45440131e-01 3.63928378e-02 -5.87197602e-01 4.22023565e-01
-1.27669573e+00 -1.33927667e+00 -3.22663367e-01 9.17496920e-01
6.23596609e-01 -2.18873292e-01 5.38322151e-01 2.99098164e-01
-6.84455693e-01 6.74410820e-01 1.35983124e-01 -2.08599597e-01
1.10041332e+00 -1.45004690e+00 -1.44380689e-01 8.79402637e-01
6.69856966e-02 4.54804540e-01 4.79120076e-01 -7.07984626e-01
-1.35278201e+00 -1.36804652e+00 3.53422880e-01 -4.74663794e-01
2.62948930e-01 -2.54411250e-01 -1.06642902e+00 5.01402557e-01
2.57611841e-01 1.50279909e-01 9.97103810e-01 2.22764939e-01
-4.71232176e-01 -3.90863508e-01 -1.03283441e+00 5.33053935e-01
1.01501215e+00 -2.88004011e-01 -3.55432928e-01 2.26689026e-01
6.58398032e-01 -2.09197998e-01 -1.19543409e+00 6.26579344e-01
1.30702823e-01 -1.36684513e+00 1.08826721e+00 6.21858016e-02
6.68104962e-02 -8.70852828e-01 -4.33695674e-01 -1.36059833e+00
-6.20919347e-01 -4.37082887e-01 -4.11392331e-01 1.71295547e+00
3.42875570e-01 -4.44178462e-01 8.97229373e-01 1.83342218e-01
-2.98358291e-01 -6.58751190e-01 -6.42244220e-01 -7.13424861e-01
-1.77787125e-01 -4.58491206e-01 6.05185270e-01 1.16168892e+00
-2.37236217e-01 2.35475913e-01 -3.77336890e-01 8.93277049e-01
1.32091105e+00 3.89027715e-01 9.08715904e-01 -1.18748975e+00
-5.90888202e-01 -2.17367321e-01 -2.99472511e-01 -8.30439091e-01
7.56431147e-02 -9.36834097e-01 -1.92384585e-03 -1.87398887e+00
6.47416294e-01 -3.69010359e-01 -3.50509852e-01 5.79339638e-02
-6.53001666e-01 1.11262441e-01 2.22065181e-01 3.61887187e-01
-1.23642886e+00 6.04395866e-01 1.22238982e+00 -3.12569737e-02
-2.05998600e-01 2.93219447e-01 -5.97593248e-01 9.92814779e-01
6.36883974e-01 -5.06187916e-01 -5.77445388e-01 -6.00958839e-02
3.97735089e-02 1.84514374e-01 2.46356487e-01 -1.13118529e+00
6.22851074e-01 -1.90692768e-01 4.63712066e-01 -1.28384781e+00
2.30203316e-01 -1.04529023e+00 2.99342871e-01 1.59218922e-01
2.71289140e-01 2.30643079e-02 -3.10373276e-01 1.28298783e+00
-3.58662724e-01 1.22860625e-01 8.86106133e-01 -2.97886968e-01
-6.99913383e-01 4.80398357e-01 -2.97288418e-01 2.34348685e-01
1.16253841e+00 -2.79951781e-01 1.72347017e-02 -4.35970277e-01
-1.03682983e+00 7.99774647e-01 8.42253745e-01 3.45048100e-01
6.77420795e-01 -1.44147158e+00 -6.50431454e-01 -6.04275838e-02
3.27134967e-01 4.12881941e-01 8.73611212e-01 1.01674640e+00
-1.44432291e-01 -1.07633963e-01 3.85274231e-01 -1.10289252e+00
-1.40347111e+00 9.38465893e-01 -7.02359527e-02 -2.32328817e-01
-5.68080068e-01 5.96279144e-01 1.58244118e-01 -7.69291401e-01
1.34061024e-01 2.62150794e-01 -1.63289562e-01 1.11099724e-02
1.67267382e-01 5.53027213e-01 -2.35453814e-01 -7.50375926e-01
-4.72932696e-01 8.21843922e-01 -2.55197715e-02 3.73766087e-02
1.25806618e+00 -7.30130613e-01 -2.69191384e-01 6.30879879e-01
1.07674634e+00 1.67024851e-01 -1.29904258e+00 -3.69646400e-01
-1.69704944e-01 -4.16901529e-01 -2.15504900e-01 -5.95063686e-01
-1.39377928e+00 5.94886184e-01 2.89290845e-01 -1.71829566e-01
1.20632946e+00 2.85028517e-01 6.60062313e-01 7.03627989e-02
5.00755191e-01 -1.20148599e+00 7.73814693e-02 2.25603014e-01
4.19400692e-01 -1.46650326e+00 2.25648940e-01 -8.31512332e-01
-9.78179276e-01 7.15301514e-01 9.94209766e-01 1.69104531e-01
9.57598031e-01 1.11724712e-01 -1.99926235e-02 -2.41547599e-01
-7.85411239e-01 -1.86699495e-01 3.21012616e-01 8.48952770e-01
-1.05830356e-02 1.27673641e-01 1.27395108e-01 5.89383006e-01
2.74850428e-01 -5.34273863e-01 4.36461002e-01 6.21962965e-01
-2.36833498e-01 -1.06343853e+00 -6.40688062e-01 4.25344408e-01
-2.51498699e-01 7.91827887e-02 -1.66057870e-01 6.38764739e-01
3.50092202e-01 1.03989112e+00 -1.57212198e-01 -4.21164691e-01
2.16127276e-01 -4.63822857e-02 3.49432737e-01 -7.17914641e-01
-2.38879964e-01 6.46794140e-01 -2.53790587e-01 -6.00014985e-01
-7.91059613e-01 -9.88865316e-01 -1.17612469e+00 -1.95195332e-01
-4.25290972e-01 1.59552425e-01 2.63866484e-01 7.03134894e-01
5.17409384e-01 4.98000473e-01 8.21984649e-01 -6.05463922e-01
-1.34356782e-01 -6.06398106e-01 -6.23046875e-01 4.92966086e-01
-8.20950195e-02 -5.54367721e-01 -3.85602951e-01 1.56930014e-01] | [8.25517749786377, 4.625959396362305] |
f87fee91-d137-480c-a879-54a946d4c783 | dycsc-modeling-the-evolutionary-process-of | 2210.12690 | null | https://arxiv.org/abs/2210.12690v2 | https://arxiv.org/pdf/2210.12690v2.pdf | DyCSC: Modeling the Evolutionary Process of Dynamic Networks Based on Cluster Structure | Temporal networks are an important type of network whose topological structure changes over time. Compared with methods on static networks, temporal network embedding (TNE) methods are facing three challenges: 1) it cannot describe the temporal dependence across network snapshots; 2) the node embedding in the latent space fails to indicate changes in the network topology; and 3) it cannot avoid a lot of redundant computation via parameter inheritance on a series of snapshots. To this end, we propose a novel temporal network embedding method named Dynamic Cluster Structure Constraint model (DyCSC), whose core idea is to capture the evolution of temporal networks by imposing a temporal constraint on the tendency of the nodes in the network to a given number of clusters. It not only generates low-dimensional embedding vectors for nodes but also preserves the dynamic nonlinear features of temporal networks. Experimental results on multiple realworld datasets have demonstrated the superiority of DyCSC for temporal graph embedding, as it consistently outperforms competing methods by significant margins in multiple temporal link prediction tasks. Moreover, the ablation study further validates the effectiveness of the proposed temporal constraint. | ['Zhan Bu', 'Shanfan Zhang'] | 2022-10-23 | null | null | null | null | ['dynamic-link-prediction', 'network-embedding'] | ['graphs', 'methodology'] | [-2.60649294e-01 -1.19757734e-01 -3.90886813e-01 9.91214663e-02
5.14933765e-01 -6.32293701e-01 8.77812445e-01 4.74919192e-02
5.06142639e-02 3.57934684e-01 3.49220008e-01 -3.30704898e-01
-8.58805180e-01 -8.03496897e-01 -1.46624193e-01 -7.36539066e-01
-7.45968878e-01 2.54946738e-01 4.60063964e-01 -2.90044576e-01
-4.98180911e-02 5.26209116e-01 -7.35659719e-01 -3.18956375e-01
5.72849751e-01 6.91139102e-01 -1.71221018e-01 2.09395349e-01
3.97828966e-02 6.29418731e-01 -1.99984610e-01 -2.93063372e-01
1.88949943e-01 -3.38754117e-01 -6.17027044e-01 -1.57711625e-01
-2.31380448e-01 4.15259302e-02 -1.09349048e+00 8.57713521e-01
5.39811738e-02 1.82371095e-01 5.27779162e-01 -1.78493643e+00
-7.87108302e-01 7.14191496e-01 -4.18954670e-01 3.97739947e-01
1.74162641e-01 1.06928319e-01 1.30270398e+00 -6.63167655e-01
8.87152731e-01 1.30297959e+00 8.20211411e-01 3.11579973e-01
-1.55366015e+00 -4.78428513e-01 6.58833325e-01 3.12338561e-01
-1.32535696e+00 -7.32065588e-02 1.25571954e+00 -6.79402411e-01
6.06645703e-01 1.87178507e-01 1.04017806e+00 1.16130626e+00
2.23760396e-01 3.47522169e-01 7.07585096e-01 1.22541778e-01
-1.12737350e-01 -2.94801563e-01 1.96373448e-01 7.16161370e-01
-8.72876309e-03 2.34384611e-01 -4.02123719e-01 -3.33762914e-01
6.97397470e-01 3.32180172e-01 -2.92703062e-01 -6.36901855e-01
-1.61885345e+00 8.35483313e-01 8.14191341e-01 6.53321326e-01
-2.17886209e-01 3.59116107e-01 6.37786567e-01 6.33206248e-01
7.21387804e-01 4.49158922e-02 -3.18144053e-01 -6.09846786e-02
-4.79321033e-01 -2.16296479e-01 5.23792207e-01 7.92714179e-01
4.30840552e-01 1.82006732e-02 -4.00740467e-02 4.86040175e-01
3.51980567e-01 1.33670032e-01 3.30222815e-01 -6.62262797e-01
4.85929281e-01 1.05797470e+00 -2.30420589e-01 -1.78029215e+00
-3.81652325e-01 -3.60211134e-01 -1.33279157e+00 -3.73744130e-01
6.47934005e-02 8.28355104e-02 -4.41734046e-01 1.92328048e+00
2.56085962e-01 6.18397951e-01 -2.82219380e-01 4.55790371e-01
5.71698725e-01 8.14944625e-01 -2.05114052e-01 -5.56011081e-01
7.92686641e-01 -8.17632437e-01 -1.06127095e+00 1.90911725e-01
5.69701076e-01 -2.46553868e-01 7.55594730e-01 -3.41584384e-01
-6.55796051e-01 -3.36404681e-01 -9.71268594e-01 3.53039026e-01
-4.73916322e-01 -3.10937762e-01 8.51097405e-01 2.42481217e-01
-1.27778220e+00 8.33589733e-01 -9.91140425e-01 -5.88486373e-01
-2.98250876e-02 2.53059953e-01 -3.65366161e-01 9.16933864e-02
-1.43880987e+00 4.28310245e-01 4.36045945e-01 4.39966381e-01
-5.56026816e-01 -7.85430491e-01 -8.34304571e-01 2.47847810e-01
4.52676743e-01 -3.93315315e-01 3.72314960e-01 -5.72398305e-01
-1.29662383e+00 3.01321179e-01 -2.11336553e-01 -2.40735397e-01
5.84089458e-01 2.32923582e-01 -8.81688118e-01 7.20921084e-02
8.44127610e-02 2.53799319e-01 7.43875623e-01 -1.01244915e+00
5.16216531e-02 -6.24894165e-02 -5.43351332e-03 -1.42002285e-01
-8.14693868e-01 -3.48422289e-01 -7.42193341e-01 -8.50081921e-01
4.93897527e-01 -1.17780483e+00 -2.99971342e-01 2.06833065e-01
-5.17086744e-01 -4.74809378e-01 1.35194981e+00 -2.78490305e-01
1.89526224e+00 -2.22333312e+00 6.36532187e-01 5.09703517e-01
5.53961694e-01 7.43157342e-02 -2.69814461e-01 9.37335968e-01
-3.57014567e-01 5.43495536e-01 -7.95679763e-02 2.11276393e-02
2.96583027e-02 2.98637569e-01 -4.54288810e-01 4.89314497e-01
-2.04274710e-02 1.11379290e+00 -1.19627130e+00 -5.48642933e-01
2.61799484e-01 5.34995079e-01 -2.55301505e-01 -2.12879866e-01
9.01883990e-02 3.09394658e-01 -3.76396507e-01 3.36047411e-01
2.91479766e-01 -6.29807055e-01 7.83252060e-01 -4.14593011e-01
-3.04845050e-02 5.32178394e-02 -9.25551593e-01 1.22821355e+00
1.23839401e-01 7.30212212e-01 -3.63519967e-01 -9.76526737e-01
7.60435164e-01 5.64494193e-01 1.01342165e+00 -6.04517639e-01
-2.16910794e-01 -1.52240977e-01 1.10709183e-01 -3.31331551e-01
7.37141818e-02 2.09887445e-01 -1.08155914e-01 7.01181531e-01
-1.11148991e-01 5.53283632e-01 4.08644617e-01 6.83503449e-01
1.27044058e+00 -4.34983015e-01 -5.45808636e-02 -3.25268656e-01
3.78793120e-01 -4.71444339e-01 9.84593868e-01 1.82600990e-01
-4.85079795e-01 -1.15164556e-01 1.19413412e+00 -5.94221950e-01
-1.08724678e+00 -1.05093372e+00 1.01773985e-01 5.58441281e-01
2.35120445e-01 -7.57814884e-01 2.46389881e-02 -8.60242069e-01
2.40141422e-01 1.75880436e-02 -1.05344224e+00 -4.33565974e-01
-6.53946996e-01 -7.33943820e-01 1.94957376e-01 4.55671906e-01
3.04769844e-01 -6.92627966e-01 2.26944566e-01 2.39025369e-01
-2.91080803e-01 -1.04440475e+00 -8.94449413e-01 -2.91636467e-01
-1.18846738e+00 -1.34631717e+00 -3.24627191e-01 -7.41610944e-01
8.47244918e-01 4.14374143e-01 7.48865306e-01 3.75594944e-01
-3.12378127e-02 3.82594943e-01 -4.26509947e-01 5.12042046e-01
-8.44241902e-02 1.55630618e-01 3.73749048e-01 2.45765030e-01
-5.08381538e-02 -1.10381508e+00 -6.50888920e-01 5.69237828e-01
-7.92522192e-01 2.23806538e-02 3.33225876e-01 8.43359768e-01
2.67797977e-01 5.87757885e-01 6.83505237e-01 -7.30839968e-01
7.80883431e-01 -6.88309848e-01 -4.81361777e-01 4.53154773e-01
-1.18695152e+00 -3.36298272e-02 5.10138631e-01 -6.99253201e-01
-4.17658031e-01 -3.95938516e-01 7.04502106e-01 -6.15104139e-01
8.33261013e-01 9.00367141e-01 2.11028099e-01 7.65778646e-02
1.04359671e-01 3.69080931e-01 2.54097164e-01 -3.94976020e-01
3.85597467e-01 -1.49999440e-01 1.25291646e-01 -2.00056359e-01
1.30142283e+00 6.78272188e-01 3.85880440e-01 -5.56562841e-01
-3.08111519e-01 -4.20758963e-01 -1.08193839e+00 -4.42457199e-01
4.72469836e-01 -5.41465104e-01 -8.93612504e-01 2.30864897e-01
-8.22232604e-01 -3.43441278e-01 4.63253744e-02 3.69592905e-01
1.10655390e-02 5.95111072e-01 -8.56133521e-01 -5.15286088e-01
-1.19554073e-01 -4.81144130e-01 2.08931148e-01 -2.38299921e-01
-9.68459323e-02 -1.67881107e+00 2.81300128e-01 -2.31295437e-01
3.35375607e-01 7.21605420e-01 1.23497891e+00 -2.98208203e-02
-6.20858908e-01 -3.75550836e-01 -2.49560878e-01 -9.01856646e-02
4.39275920e-01 6.25082791e-01 -1.06916934e-01 -6.08315766e-01
-4.38121617e-01 3.09745759e-01 6.74603701e-01 2.23866776e-01
8.43495011e-01 -5.91378212e-01 -7.63495684e-01 5.55769742e-01
1.38288975e+00 1.06252030e-01 3.39154631e-01 1.61695570e-01
8.59985769e-01 6.69799387e-01 4.53478694e-01 2.55883634e-01
3.01839322e-01 8.70656073e-01 4.86328632e-01 4.54580113e-02
5.18598966e-02 -4.69645381e-01 4.01528925e-01 1.54710758e+00
-7.27896243e-02 -2.12080449e-01 -9.94067729e-01 8.43690634e-01
-2.27428532e+00 -1.08557010e+00 -2.55343974e-01 1.84349203e+00
4.97339368e-01 3.61203313e-01 2.45086461e-01 1.35091990e-01
8.78400326e-01 6.38853848e-01 -6.73371375e-01 2.51847543e-02
-7.61823207e-02 -6.00219190e-01 2.71592110e-01 3.90617490e-01
-8.02363753e-01 7.78433263e-01 6.50811625e+00 4.02276963e-01
-1.02446580e+00 1.10593110e-01 1.01002552e-01 -2.03136522e-02
-4.51493919e-01 3.41263175e-01 -1.48503348e-01 7.79364407e-01
8.13814104e-01 -6.41429961e-01 4.11139190e-01 4.13071185e-01
3.39131355e-01 5.88909805e-01 -1.06593955e+00 6.70947254e-01
-3.57889414e-01 -1.35509264e+00 1.41504005e-01 3.38743240e-01
7.30932832e-01 -2.56760772e-02 1.74471602e-01 2.11751312e-01
3.55063587e-01 -8.44242632e-01 3.76243919e-01 7.36334622e-01
8.13083589e-01 -7.10617185e-01 5.21745622e-01 6.29634261e-02
-1.88121057e+00 -1.88822240e-01 -2.12915510e-01 6.63528740e-02
3.43570203e-01 6.97242618e-01 -6.08935356e-01 6.83695316e-01
6.34248257e-01 1.52837718e+00 -7.49446869e-01 7.87958384e-01
-2.12989897e-01 5.62554598e-01 -1.71716154e-01 1.24623626e-01
3.07481408e-01 -5.93499303e-01 8.40320289e-01 7.55967915e-01
1.44629568e-01 -9.11886841e-02 6.10597506e-02 7.84939408e-01
-1.38208708e-02 -2.44599193e-01 -8.48155499e-01 -5.11323094e-01
8.84413362e-01 1.14805067e+00 -8.60752165e-01 7.04545602e-02
-2.38108829e-01 8.68250668e-01 3.77421975e-01 7.00029433e-01
-7.72737861e-01 -3.33124220e-01 6.37418270e-01 1.31409213e-01
2.30385005e-01 -6.95606709e-01 2.59467959e-01 -1.14113176e+00
2.46968448e-01 -2.16638282e-01 4.84989226e-01 -4.61698890e-01
-1.42079151e+00 6.10716164e-01 -1.10825375e-01 -1.54661822e+00
7.47025711e-03 -2.25683436e-01 -8.87817085e-01 4.30030018e-01
-1.27635288e+00 -1.24244237e+00 -1.89876765e-01 7.46694744e-01
3.48954648e-02 -3.58345583e-02 6.54965758e-01 4.93439674e-01
-1.04697645e+00 4.40661341e-01 3.86328846e-01 4.23524976e-01
3.31595212e-01 -1.12517810e+00 3.97955328e-01 9.26019192e-01
-8.88862740e-03 8.93046975e-01 4.78083044e-01 -8.26944232e-01
-1.38003051e+00 -1.23850048e+00 1.01956332e+00 -4.01843846e-01
1.28382194e+00 -5.07283926e-01 -8.85301352e-01 9.78247583e-01
-6.65927306e-02 3.54731560e-01 5.69608331e-01 3.95574272e-01
-5.59030473e-01 -3.39521021e-01 -6.03886366e-01 7.84315944e-01
1.29633069e+00 -8.33062828e-01 -1.08900554e-01 4.84092325e-01
1.10695505e+00 2.47798830e-01 -1.23753083e+00 5.39800167e-01
4.32041496e-01 -6.13184631e-01 1.04620409e+00 -7.08327055e-01
1.51761353e-01 -3.77619654e-01 1.15560792e-01 -1.22867930e+00
-8.28306913e-01 -8.63536477e-01 -6.85090959e-01 1.41785967e+00
2.24265307e-01 -8.68474424e-01 6.02652609e-01 3.01538765e-01
3.10722023e-01 -9.09453452e-01 -1.06932652e+00 -1.25902760e+00
-1.94570124e-01 1.29634356e-02 6.67766333e-01 1.64917445e+00
2.29657769e-01 2.48626351e-01 -4.92889464e-01 1.01505123e-01
6.09031856e-01 1.46068946e-01 4.57374394e-01 -1.52866745e+00
1.02854028e-01 -7.14052379e-01 -6.16174161e-01 -6.85225785e-01
2.25020871e-01 -1.03989136e+00 -6.70579612e-01 -1.49675810e+00
2.77641803e-01 -6.00178957e-01 -5.39941669e-01 2.92756706e-01
-5.80861941e-02 -2.22419634e-01 -3.39123830e-02 6.99252784e-01
-7.10538089e-01 1.03446972e+00 1.15406811e+00 -1.17495999e-01
-1.75463632e-01 -1.69621691e-01 -3.14720601e-01 3.21623713e-01
5.42290330e-01 -4.47758913e-01 -8.27894390e-01 -1.08409107e-01
6.15548193e-01 2.22925708e-01 4.86542016e-01 -6.75038278e-01
3.32468808e-01 -2.16482103e-01 -1.04255736e-01 -6.24116123e-01
1.95136711e-01 -1.09209275e+00 4.79986578e-01 7.56405115e-01
-9.59675685e-02 4.59134042e-01 -1.04142264e-01 1.33675694e+00
-2.10323170e-01 4.46264237e-01 4.07151520e-01 3.90317857e-01
-6.30934834e-01 8.83072555e-01 -1.79330751e-01 -1.77156225e-01
1.13547158e+00 -2.13226393e-01 -4.00650769e-01 -3.53057623e-01
-8.81383657e-01 7.21066177e-01 4.08580095e-01 7.30256796e-01
5.75175822e-01 -1.98247695e+00 -3.33526284e-01 -1.18814111e-01
9.32392702e-02 -4.34862137e-01 3.20009589e-01 1.26851249e+00
-1.91416606e-01 5.80363393e-01 4.94053848e-02 -5.94722986e-01
-1.24798584e+00 8.73505473e-01 3.39415073e-01 -5.23988724e-01
-9.81196821e-01 4.73003209e-01 -1.47014514e-01 -5.52553236e-01
1.43574357e-01 -2.72258632e-02 -3.92229497e-01 4.13059413e-01
-5.02347201e-02 4.29666251e-01 -3.90800536e-01 -6.63516581e-01
-4.36294198e-01 6.47218525e-01 1.65763032e-02 1.53685668e-02
1.58885229e+00 -3.71257812e-01 -6.01089001e-01 8.66654515e-01
1.47274172e+00 -3.09127659e-01 -1.20018625e+00 -6.90187871e-01
2.66762346e-01 -5.35447836e-01 -1.03043675e-01 -1.88888192e-01
-1.33606434e+00 5.02199888e-01 4.00796890e-01 7.18725681e-01
9.13231194e-01 -1.14374802e-01 7.77544975e-01 2.91677266e-01
1.82351619e-01 -9.14970994e-01 5.76171041e-01 5.69128335e-01
8.69792044e-01 -9.66921210e-01 1.59638047e-01 -5.09024322e-01
-3.51192206e-01 1.11934710e+00 4.45674777e-01 -3.83138866e-03
1.25217605e+00 -4.78264600e-01 -4.08162355e-01 -7.56800354e-01
-1.11161065e+00 1.43068671e-01 2.91840643e-01 3.15316796e-01
2.00252727e-01 4.51767929e-02 -3.24166268e-01 -5.79202808e-02
1.76251028e-02 -5.32318175e-01 3.14548254e-01 6.45621419e-01
1.34228542e-01 -1.23311067e+00 3.99721831e-01 3.23905200e-01
1.31069720e-02 1.44598067e-01 -6.65959358e-01 1.01781809e+00
-2.71322638e-01 7.56767333e-01 7.37699401e-03 -7.31449842e-01
1.35689557e-01 -1.09913319e-01 1.73115805e-02 -4.04500604e-01
-7.93967023e-02 -7.25049451e-02 -1.81863949e-01 -6.44206643e-01
-4.81046975e-01 -8.10467124e-01 -8.68955553e-01 -7.39712894e-01
-3.85110825e-01 2.29033992e-01 2.60993421e-01 6.02551162e-01
6.03816032e-01 8.16788912e-01 1.04551017e+00 -4.48673129e-01
-2.67941784e-02 -9.09560263e-01 -6.94654465e-01 5.84486067e-01
3.99183542e-01 -8.35500240e-01 -7.23636270e-01 -1.48892790e-01] | [7.199100017547607, 6.0484795570373535] |
18523426-79fd-4d13-858c-53d9eb328df5 | adapting-descriptions-of-people-to-the-point | null | null | https://aclanthology.org/W18-6540 | https://aclanthology.org/W18-6540.pdf | Adapting Descriptions of People to the Point of View of a Moving Observer | This paper addresses the task of generating descriptions of people for an observer that is moving within a scene. As the observer moves, the descriptions of the people around him also change. A referring expression generation algorithm adapted to this task needs to continuously monitor the changes in the field of view of the observer, his relative position to the people being described, and the relative position of these people to any landmarks around them, and to take these changes into account in the referring expressions generated. This task presents two advantages: many of the mechanisms already available for static contexts may be applied with small adaptations, and it introduces the concept of changing conditions into the task of referring expression generation. In this paper we describe the design of an algorithm that takes these aspects into account in order to create descriptions of people within a 3D virtual environment. The evaluation of this algorithm has shown that, by changing the descriptions in real time according to the observers point of view, they are able to identify the described person quickly and effectively. | ['Daniel Ruiz', "Pablo Gerv{\\'a}s", "Raquel Herv{\\'a}s", "Gonzalo M{\\'e}ndez", 'Ricardo de la Rosa'] | 2018-11-01 | null | null | null | ws-2018-11 | ['referring-expression-generation'] | ['computer-vision'] | [ 9.71709266e-02 1.97249223e-02 4.82656568e-01 -3.92311960e-01
-7.82319531e-02 -5.97204447e-01 9.08055902e-01 4.31760967e-01
-5.30920923e-01 5.85419714e-01 2.43849456e-01 2.63965040e-01
9.17978305e-03 -8.88017714e-01 -6.92942813e-02 -2.99243271e-01
7.39918351e-02 8.50829303e-01 5.74141979e-01 -4.94062781e-01
2.61868924e-01 1.00506103e+00 -1.77410746e+00 -5.75390309e-02
3.95019576e-02 2.71437883e-01 3.19716334e-01 8.96113157e-01
-3.39799613e-01 5.34210086e-01 -1.03677392e+00 1.86145958e-02
-2.53221858e-03 -6.51557088e-01 -9.21708524e-01 6.90734863e-01
2.07995608e-01 -3.72733921e-02 7.11803064e-02 7.64179409e-01
6.03186548e-01 7.15722501e-01 4.52106327e-01 -1.04313397e+00
-2.64531404e-01 1.97078556e-01 -1.09870676e-02 1.85319066e-01
1.39452815e+00 -1.84434950e-01 2.51558661e-01 -4.73784745e-01
1.08540010e+00 1.26469278e+00 4.49042022e-01 8.55749190e-01
-1.01493001e+00 -4.89707552e-02 2.97384441e-01 1.01678841e-01
-1.68548644e+00 -5.99035323e-01 7.46623695e-01 -5.66400707e-01
8.71498883e-01 5.27064741e-01 9.19398189e-01 6.43855810e-01
-1.55728295e-01 1.15742572e-01 5.85595131e-01 -9.75721717e-01
4.64842647e-01 7.72925258e-01 1.75867323e-02 3.67205620e-01
-1.29973829e-01 -1.61515772e-01 -3.93309683e-01 -2.27523461e-01
8.98198307e-01 -3.00590426e-01 -2.64632523e-01 -6.42284691e-01
-1.20586300e+00 3.26406121e-01 1.98745206e-01 7.84048438e-01
-3.19490641e-01 -3.19809392e-02 2.98134506e-01 -1.01746015e-01
2.67079830e-01 4.93538290e-01 -3.02469265e-03 -4.42225426e-01
-3.59801531e-01 5.55994511e-01 9.75782573e-01 1.14754450e+00
5.67049503e-01 -3.39053929e-01 -2.03108303e-02 4.30746794e-01
1.46910861e-01 2.34307900e-01 3.11598420e-01 -8.74054730e-01
1.45012319e-01 8.65265787e-01 5.60453773e-01 -1.31166947e+00
-3.99548560e-01 3.47269475e-02 -8.85650516e-02 4.53315198e-01
2.82336414e-01 -4.51368317e-02 -4.24364001e-01 1.66422820e+00
9.12061632e-01 -2.72500277e-01 -9.79503244e-02 7.39202738e-01
8.45531166e-01 4.26988244e-01 4.76360954e-02 -4.25476372e-01
1.41351044e+00 -2.82198399e-01 -1.13736868e+00 -1.17186032e-01
7.22976506e-01 -8.20902467e-01 7.41695344e-01 -8.29846561e-02
-1.19745648e+00 -8.26923907e-01 -8.43341351e-01 5.30958995e-02
-5.75815976e-01 -3.40882540e-02 1.99123085e-01 5.60887277e-01
-1.17979050e+00 3.97337735e-01 -6.52685106e-01 -1.03562486e+00
-5.61691463e-01 5.18273413e-01 -3.48820657e-01 4.12704170e-01
-9.56873655e-01 1.24777472e+00 2.45602086e-01 1.32770389e-01
-2.54174918e-01 1.02055147e-01 -8.89146090e-01 -9.34304819e-02
2.93810576e-01 -8.34205449e-01 1.31530547e+00 -1.17933238e+00
-1.38426173e+00 1.16993427e+00 -5.32369316e-01 1.61861897e-01
7.99725771e-01 4.23760526e-02 -5.18923044e-01 7.08876103e-02
4.00050193e-01 4.29307699e-01 2.43949935e-01 -1.46828067e+00
-7.45019853e-01 -5.83667696e-01 4.63752359e-01 6.09461486e-01
2.56761312e-01 4.04390484e-01 -6.96096420e-01 -2.20141143e-01
1.51636302e-01 -7.37144172e-01 -2.56581485e-01 6.07214160e-02
-1.36567295e-01 -1.68291420e-01 9.24119890e-01 -1.47372231e-01
1.12118483e+00 -2.41076517e+00 1.42938435e-01 3.73597294e-01
1.00071646e-01 2.06306875e-01 3.12843204e-01 6.47490978e-01
-1.90343469e-01 -1.94727942e-01 1.89538524e-01 -4.05885696e-01
-1.65666804e-01 1.78651869e-01 2.94333786e-01 3.99590999e-01
-3.43137741e-01 2.87628412e-01 -1.01273310e+00 -6.65936410e-01
5.21297216e-01 7.55962074e-01 -5.00882976e-02 3.65768939e-01
2.91578434e-02 3.91256154e-01 -6.64218724e-01 -1.24419995e-01
3.41238409e-01 1.93861142e-01 1.89368784e-01 3.59766155e-01
-4.83109325e-01 1.10852662e-02 -1.52058804e+00 1.20327306e+00
-3.83834839e-01 5.12643516e-01 -6.20785803e-02 -2.57351488e-01
1.11756313e+00 7.49085546e-01 2.32348576e-01 -4.49028611e-01
8.59555826e-02 5.42146005e-02 -2.59914577e-01 -7.39020586e-01
6.25224769e-01 -2.80225396e-01 -5.30768856e-02 6.34196341e-01
-5.27001917e-01 -2.40968406e-01 3.49678338e-01 1.90735668e-01
7.06860065e-01 2.21557260e-01 8.42731714e-01 1.67497978e-01
8.96524489e-01 -2.28527673e-02 1.20253630e-01 5.47171772e-01
-1.47966087e-01 3.35788041e-01 2.39248797e-01 -8.87121677e-01
-8.31176758e-01 -7.48054683e-01 1.92303777e-01 8.82857978e-01
4.18681562e-01 -5.77182531e-01 -8.82350326e-01 -3.31791043e-01
-3.96326900e-01 8.68918180e-01 -7.93482840e-01 1.95080657e-02
-6.68049157e-01 -1.07711487e-01 -2.83437595e-02 2.94195771e-01
4.48734820e-01 -1.18878412e+00 -1.46421480e+00 3.51430237e-01
-8.36051479e-02 -1.04354107e+00 -2.76097208e-01 -3.48930925e-01
-6.63561881e-01 -9.26010191e-01 -5.58444142e-01 -7.70134747e-01
1.19204676e+00 1.34368435e-01 8.09799910e-01 8.55878443e-02
-2.09445536e-01 9.28044498e-01 -5.33041835e-01 -4.32071090e-01
-7.59369254e-01 -2.99144894e-01 -1.54245362e-01 -6.17377041e-03
2.79475123e-01 -1.98858321e-01 -5.25956042e-02 4.55050737e-01
-6.69966578e-01 7.90479630e-02 -3.06987643e-01 8.20992067e-02
4.44250017e-01 2.16215700e-01 -8.61469284e-02 -8.78470004e-01
7.27777421e-01 1.44222319e-01 -4.18045580e-01 2.85127729e-01
1.03012063e-01 -1.40756115e-01 4.00980145e-01 -4.10830915e-01
-1.10505939e+00 3.18451673e-01 -4.08164971e-03 1.45979345e-01
-7.06890285e-01 3.02194189e-02 -4.30180550e-01 -5.53854555e-02
7.59398937e-01 8.10721293e-02 -1.29492983e-01 -2.09029689e-01
1.64228424e-01 5.19450366e-01 5.20963848e-01 -3.08420748e-01
6.54396832e-01 4.87822533e-01 1.15782134e-01 -7.72278607e-01
-4.39708591e-01 -6.46740437e-01 -1.19731414e+00 -7.36798942e-01
7.46275008e-01 -4.51153040e-01 -6.33707225e-01 3.18451583e-01
-1.38894749e+00 -1.59166068e-01 -6.06690645e-01 1.95924878e-01
-6.42085731e-01 2.97972292e-01 1.63393557e-01 -1.11898565e+00
1.99196726e-01 -8.63292754e-01 1.00386691e+00 3.54301184e-01
-7.69644856e-01 -1.35999584e+00 5.48393913e-02 -3.78208309e-02
3.02728951e-01 6.32712007e-01 7.19768226e-01 -5.14271915e-01
-2.17860684e-01 -6.13904893e-01 4.49814886e-01 -3.09793115e-01
5.72720468e-01 2.24137306e-01 -7.47689426e-01 5.27001284e-02
6.57467842e-02 6.25335455e-01 -2.07062855e-01 1.27500832e-01
4.07434314e-01 -1.84380949e-01 -6.86057866e-01 1.31753683e-01
1.21510911e+00 7.05058217e-01 5.28398514e-01 5.58200896e-01
5.46599329e-01 8.28039706e-01 6.73589766e-01 5.39270580e-01
5.36785185e-01 1.39845181e+00 8.55216607e-02 -1.17434807e-01
3.19959559e-02 2.50023697e-02 4.41559553e-02 2.33091768e-02
-4.14003611e-01 -2.83547640e-01 -9.09781516e-01 4.54939872e-01
-1.84455478e+00 -1.24714208e+00 -3.73958558e-01 2.50665140e+00
2.49842942e-01 -1.52445868e-01 2.75091350e-01 2.69051909e-01
1.02114773e+00 7.23454356e-02 -2.17969529e-02 -7.64528513e-01
3.01997900e-01 -3.03521305e-01 -5.50412126e-02 8.13849032e-01
-8.53038967e-01 7.19716132e-01 6.68643427e+00 -2.71010902e-02
-9.04406071e-01 -8.72851163e-02 6.61975592e-02 3.28949094e-02
-2.58368161e-02 2.26538293e-02 -7.37637639e-01 1.66265205e-01
5.85633457e-01 -4.86955881e-01 3.71468902e-01 5.61953723e-01
6.58868730e-01 -5.12699664e-01 -1.27702415e+00 9.79875326e-01
5.46694100e-01 -7.32116461e-01 5.59863076e-02 -9.47173387e-02
3.48242909e-01 -8.32649469e-01 -2.89522082e-01 -2.73804218e-01
-1.27568096e-01 -6.17768228e-01 9.96787965e-01 6.48986876e-01
4.31579739e-01 -8.20222795e-01 8.78477156e-01 4.05400991e-01
-1.15829134e+00 1.33508742e-01 6.82630613e-02 -5.27320087e-01
6.07692957e-01 -2.61959936e-02 -1.22074056e+00 2.65289396e-01
3.96439463e-01 4.02944759e-02 -4.73875850e-01 1.01747966e+00
-3.43813896e-01 -3.19263786e-01 -3.40902627e-01 -3.46499622e-01
5.87750152e-02 -4.56008911e-02 7.92636156e-01 1.09713709e+00
3.84982347e-01 3.02396119e-01 4.83999737e-02 7.81112731e-01
6.61789119e-01 3.28838319e-01 -8.93979192e-01 5.31469643e-01
4.00232315e-01 8.65112484e-01 -8.63314509e-01 -5.59447229e-01
5.51003776e-02 9.98034656e-01 -1.72389939e-01 3.09774995e-01
-7.03045964e-01 -4.25024629e-01 3.15268636e-01 8.64783525e-01
-4.55116853e-02 -2.74359614e-01 2.58194208e-01 -5.43447971e-01
1.52022243e-01 -5.23991168e-01 3.65334541e-01 -1.26223636e+00
-4.85927463e-01 9.56034482e-01 4.82910544e-01 -1.21415889e+00
-6.51701987e-01 -3.03023994e-01 -6.33399189e-01 9.86204207e-01
-6.65013194e-01 -1.04401720e+00 -5.83563507e-01 5.46950281e-01
4.56867486e-01 -2.80794874e-02 1.09242344e+00 -7.74867758e-02
-2.23427311e-01 1.32111264e-02 -5.04485846e-01 3.28991860e-02
3.82861108e-01 -1.16161728e+00 4.08003807e-01 8.46406877e-01
1.26385167e-01 7.32134581e-01 1.08925450e+00 -5.35415471e-01
-7.10485637e-01 -4.90620106e-01 1.69385004e+00 -6.16472781e-01
1.93306118e-01 -3.02345872e-01 -7.07616568e-01 8.12577367e-01
-9.73684061e-03 -1.13233894e-01 5.14075220e-01 7.47059882e-02
1.49148047e-01 -4.72427756e-02 -1.27251089e+00 6.56935096e-01
1.01235235e+00 -5.13190925e-01 -7.44913936e-01 2.99823463e-01
1.57991484e-01 -8.74028087e-01 -3.05433959e-01 -1.43899500e-01
3.88101995e-01 -1.30334234e+00 8.73251021e-01 -2.50668645e-01
-3.51653218e-01 -6.28523588e-01 2.73050100e-01 -1.21943533e+00
-2.55130380e-01 -6.54584706e-01 4.02045369e-01 1.28826618e+00
9.96754915e-02 -7.68863201e-01 3.92173499e-01 1.06445026e+00
2.97639936e-01 -8.88891667e-02 -1.05378127e+00 -3.94387633e-01
-7.04583406e-01 -8.37356150e-02 7.37611711e-01 6.73955083e-01
4.93607372e-01 3.12840611e-01 -6.92717060e-02 2.85182267e-01
7.01313987e-02 -2.94667125e-01 1.07070947e+00 -1.35378957e+00
-9.27647203e-02 -1.26818269e-01 -1.07841158e+00 -7.92092919e-01
-1.67646974e-01 -2.81746417e-01 6.27671257e-02 -1.81728458e+00
5.17058074e-02 -2.29588598e-01 5.18407762e-01 1.87969595e-01
-2.01428339e-01 -4.19343024e-01 4.00672823e-01 2.63941437e-02
-5.58730066e-01 1.00642674e-01 1.02245569e+00 3.27568203e-01
-5.92093825e-01 4.56586301e-01 -3.03361803e-01 8.33445430e-01
5.55706799e-01 -4.50084299e-01 -4.26231980e-01 -2.91978657e-01
1.51465505e-01 7.18997186e-03 4.54744250e-01 -1.26170814e+00
3.81080478e-01 -7.05295727e-02 4.30875421e-01 -3.88922691e-01
6.66755676e-01 -1.15632200e+00 9.41022635e-01 4.13773894e-01
-2.73588270e-01 5.53393006e-01 1.45535171e-01 1.17247790e-01
-1.34410650e-01 -5.30864060e-01 5.91900051e-01 -6.88553572e-01
-1.01435804e+00 -2.28321955e-01 -7.36661315e-01 -4.17557567e-01
1.48693728e+00 -8.43066990e-01 5.10734133e-02 -7.84879446e-01
-1.16975665e+00 4.36995737e-02 8.01244020e-01 2.13845178e-01
5.81338704e-01 -1.21993375e+00 -2.03164801e-01 1.60821348e-01
2.18349725e-01 1.17001511e-01 1.27050921e-01 3.80199730e-01
-7.76506424e-01 2.99492300e-01 -2.33993456e-01 -3.56511533e-01
-1.84225917e+00 7.17984021e-01 7.34536409e-01 1.46954596e-01
-6.28462434e-01 5.12699366e-01 1.61827996e-01 -1.64723814e-01
1.04648255e-01 -1.41458556e-01 -8.34675431e-01 1.91307247e-01
7.63393104e-01 3.15640002e-01 -1.94416255e-01 -1.31517124e+00
-4.10880268e-01 1.08231044e+00 3.11909407e-01 -5.34409285e-01
1.08359873e+00 -5.71892500e-01 -9.89219695e-02 8.45916152e-01
6.83272243e-01 4.10224468e-01 -8.49947035e-01 -1.86150838e-02
-3.76549214e-02 -7.66164541e-01 -4.53692466e-01 -5.21658003e-01
-4.67019498e-01 5.96146822e-01 4.28349227e-01 5.55647254e-01
1.06966984e+00 1.87471092e-01 7.46177286e-02 1.82059228e-01
8.59885454e-01 -1.05088532e+00 -9.08884034e-02 3.15310985e-01
9.42340672e-01 -5.32419562e-01 8.28931928e-02 -7.32918382e-01
-7.96892226e-01 1.28964388e+00 5.16388416e-01 2.40335986e-01
2.66175777e-01 1.94702238e-01 5.17708242e-01 -4.70054537e-01
-5.63578665e-01 -3.31244916e-01 -1.27811834e-01 1.03143430e+00
3.10133696e-01 8.04157928e-02 -3.28037858e-01 -2.45920941e-01
-2.03583181e-01 1.68595556e-03 9.34561014e-01 1.06983232e+00
-4.94630039e-01 -1.05501497e+00 -8.89461875e-01 -3.75171423e-01
8.93558934e-03 5.74511886e-01 -8.45250845e-01 1.01667202e+00
6.11894667e-01 9.54839170e-01 3.93031090e-01 2.06062812e-02
1.16320539e+00 -2.31557600e-02 5.28774440e-01 -8.44547868e-01
-7.03434527e-01 -4.43324000e-02 3.07103574e-01 -3.58830601e-01
-8.00247967e-01 -1.00914717e+00 -1.45920217e+00 -1.48387760e-01
-7.68648535e-02 4.74328995e-01 5.23494244e-01 9.37663138e-01
7.10659474e-02 3.67151558e-01 5.92563927e-01 -1.01834655e+00
1.28862727e-02 -5.54591656e-01 -6.07963204e-01 7.09288418e-01
2.46292502e-01 -5.32437563e-01 -3.94962847e-01 4.00858760e-01] | [5.093184947967529, 0.49617475271224976] |
582c4af0-de76-4ae2-8143-c92df3fe179d | bridging-active-exploration-and-uncertainty | 2305.12240 | null | https://arxiv.org/abs/2305.12240v2 | https://arxiv.org/pdf/2305.12240v2.pdf | Bridging Active Exploration and Uncertainty-Aware Deployment Using Probabilistic Ensemble Neural Network Dynamics | In recent years, learning-based control in robotics has gained significant attention due to its capability to address complex tasks in real-world environments. With the advances in machine learning algorithms and computational capabilities, this approach is becoming increasingly important for solving challenging control problems in robotics by learning unknown or partially known robot dynamics. Active exploration, in which a robot directs itself to states that yield the highest information gain, is essential for efficient data collection and minimizing human supervision. Similarly, uncertainty-aware deployment has been a growing concern in robotic control, as uncertain actions informed by the learned model can lead to unstable motions or failure. However, active exploration and uncertainty-aware deployment have been studied independently, and there is limited literature that seamlessly integrates them. This paper presents a unified model-based reinforcement learning framework that bridges these two tasks in the robotics control domain. Our framework uses a probabilistic ensemble neural network for dynamics learning, allowing the quantification of epistemic uncertainty via Jensen-Renyi Divergence. The two opposing tasks of exploration and deployment are optimized through state-of-the-art sampling-based MPC, resulting in efficient collection of training data and successful avoidance of uncertain state-action spaces. We conduct experiments on both autonomous vehicles and wheeled robots, showing promising results for both exploration and deployment. | ['Seongil Hong', 'Beomsu Kim', 'Junwon Seo', 'Jungwi Mun', 'Taekyung Kim'] | 2023-05-20 | null | null | null | null | ['model-based-reinforcement-learning'] | ['reasoning'] | [-3.15153636e-02 4.04342949e-01 -6.08749211e-01 5.84497489e-02
-6.57764494e-01 -2.93547034e-01 6.31804705e-01 -5.97303025e-02
-5.61262667e-01 1.16550243e+00 -2.49550074e-01 -1.92136630e-01
-5.79346299e-01 -7.28967905e-01 -7.26070881e-01 -1.20225275e+00
-4.29954678e-01 6.05857849e-01 1.75998032e-01 -2.64976978e-01
8.39370042e-02 3.62856805e-01 -1.61596680e+00 -6.70467257e-01
1.10701525e+00 1.04983342e+00 5.75725615e-01 3.25989425e-01
3.01863432e-01 7.27251053e-01 -1.37349665e-01 1.34947509e-01
2.32548445e-01 7.15064481e-02 -4.47634310e-01 -1.55066233e-02
-5.57743311e-01 -1.12073652e-01 -4.65311855e-01 1.06416571e+00
3.85714769e-01 5.65155864e-01 5.94111204e-01 -1.44433236e+00
-1.47228792e-01 5.46303928e-01 -2.61981249e-01 -3.44352365e-01
-2.75761694e-01 5.48922420e-01 7.21743941e-01 -5.25911629e-01
4.11702812e-01 1.30563235e+00 8.14471245e-02 5.46950817e-01
-1.11743188e+00 -4.65353191e-01 2.25930005e-01 4.72630531e-01
-1.13811600e+00 2.29829201e-03 5.52431226e-01 -6.46622181e-01
8.02457094e-01 -3.72493595e-01 6.81487799e-01 9.75291133e-01
7.28073418e-01 8.12898159e-01 1.03268778e+00 -1.36356905e-01
8.24639857e-01 -1.62274018e-01 -5.17688811e-01 4.72080261e-01
5.44777572e-01 7.53970265e-01 -1.81968004e-01 1.23099640e-01
5.63967109e-01 -2.10111332e-03 -1.40457109e-01 -9.98147786e-01
-1.17902970e+00 1.09686017e+00 4.83150750e-01 -3.67929727e-01
-6.44694149e-01 4.78265762e-01 2.72330135e-01 -4.11279239e-02
1.03426397e-01 7.88627744e-01 -4.41270202e-01 -2.88599163e-01
-1.60335004e-01 6.70262694e-01 8.34856927e-01 9.25384402e-01
7.39759743e-01 3.30397069e-01 1.29380584e-01 6.06963634e-01
5.17405987e-01 6.86046183e-01 1.67870745e-01 -1.55212760e+00
4.33490366e-01 4.79154587e-01 6.66515112e-01 -7.25036621e-01
-4.39438879e-01 -2.84040451e-01 -7.33158708e-01 9.40867841e-01
1.84311897e-01 -5.09620309e-01 -6.57048643e-01 1.91055024e+00
5.58958530e-01 3.08989305e-02 3.85488063e-01 9.05292571e-01
-3.26225460e-01 6.41761065e-01 -5.94989918e-02 -3.51954192e-01
9.09140646e-01 -7.81864226e-01 -9.54449415e-01 -3.83194417e-01
3.03721040e-01 -2.36443873e-03 5.75688958e-01 6.56537592e-01
-6.79803252e-01 -1.18565179e-01 -1.26004481e+00 4.71774161e-01
-2.26356462e-01 -1.54978245e-01 3.67645890e-01 1.88957945e-01
-6.22200191e-01 6.60232425e-01 -1.18017840e+00 -1.10743567e-01
4.30867344e-01 3.68998915e-01 -4.12860215e-02 -2.70746611e-02
-1.20709276e+00 1.63352907e+00 7.64989853e-01 2.27119923e-01
-1.34286654e+00 -2.81926364e-01 -8.75579059e-01 -3.51203769e-01
1.21230173e+00 -4.00500834e-01 1.43960762e+00 -2.97035247e-01
-2.09597993e+00 -9.01669115e-02 2.91050375e-01 -8.18693638e-01
4.77180481e-01 -5.27101636e-01 1.41847236e-02 -4.70185243e-02
-5.48702851e-02 5.84308743e-01 8.54240596e-01 -1.25483012e+00
-8.67527783e-01 -3.15556854e-01 1.53763026e-01 3.69256824e-01
-1.38350964e-01 -6.76490009e-01 -6.78640045e-03 -3.47711742e-01
-3.46692540e-02 -1.53834748e+00 -6.00432694e-01 -2.09705587e-02
-8.35360810e-02 -4.61086035e-01 8.93027186e-01 -2.85386950e-01
9.38219488e-01 -1.63031626e+00 8.49092841e-01 3.07278428e-02
-2.04249159e-01 7.29106516e-02 1.86679035e-01 4.91429418e-01
6.10392511e-01 -1.56749383e-01 -3.98614556e-01 -1.36214226e-01
2.34096661e-01 8.05213034e-01 -5.95117927e-01 4.77098286e-01
3.98762375e-01 5.20465195e-01 -1.21399534e+00 -3.05436045e-01
4.73580569e-01 1.85292244e-01 -3.44078928e-01 2.30253458e-01
-6.02246702e-01 7.13252544e-01 -7.61909008e-01 5.05398512e-01
2.51266420e-01 7.81878009e-02 2.44808465e-01 3.80600631e-01
-4.26331997e-01 -1.62826389e-01 -1.21989894e+00 1.43399251e+00
-7.94230640e-01 4.59471285e-01 4.24845576e-01 -1.10103679e+00
8.04557741e-01 1.27293691e-01 5.72971165e-01 -3.46207321e-01
3.25522810e-01 4.03763771e-01 -6.99553937e-02 -5.99331021e-01
6.55864298e-01 4.50321808e-02 -2.92548805e-01 -1.80573203e-02
-8.58664215e-02 -9.01403606e-01 1.91219434e-01 -1.71558812e-01
9.91408646e-01 5.69394946e-01 3.55475128e-01 -2.94005930e-01
4.12845969e-01 1.96811736e-01 8.67174864e-01 4.44125414e-01
-2.95027912e-01 -1.81144699e-01 2.86695957e-01 1.10931426e-01
-7.68813133e-01 -9.80605125e-01 -9.13396031e-02 7.83033907e-01
6.93273306e-01 1.61973685e-01 -4.64035392e-01 -2.96228737e-01
2.58737415e-01 1.06829786e+00 -5.51459193e-01 -4.52622384e-01
-6.54625297e-01 -4.35352027e-01 -3.13918404e-02 3.42770696e-01
2.54630983e-01 -9.63883698e-01 -1.19742930e+00 5.15523255e-01
-1.67276673e-02 -9.60660219e-01 8.02058652e-02 5.10642469e-01
-7.82700539e-01 -1.27040446e+00 -5.35846412e-01 -4.35380369e-01
4.08928066e-01 4.47636982e-03 4.64405090e-01 -4.90422368e-01
-2.01174527e-01 5.41805029e-01 -3.35725188e-01 -7.83123791e-01
-5.27192473e-01 -1.24735124e-01 5.54674327e-01 -1.97919905e-01
-2.63363838e-01 -3.57755721e-01 -2.84247458e-01 5.50227642e-01
-6.93734646e-01 -1.20499767e-01 6.74922764e-01 1.01980293e+00
7.27060854e-01 3.98404866e-01 7.30074644e-01 4.16043960e-03
4.94175881e-01 -7.61339545e-01 -1.24810493e+00 -2.60061305e-02
-7.46299744e-01 5.17048776e-01 5.76607108e-01 -5.63669503e-01
-1.08143616e+00 9.60189477e-02 4.34415877e-01 -5.33897758e-01
1.09314650e-01 7.55426049e-01 -4.98401336e-02 1.19104996e-01
3.58205974e-01 6.79944381e-02 5.31097949e-01 -5.01924800e-03
4.41259533e-01 4.93755400e-01 5.02012610e-01 -8.21199358e-01
6.36030376e-01 3.18899751e-01 4.92843777e-01 -6.71968699e-01
-5.31515121e-01 -2.15179965e-01 -3.19138050e-01 -4.55057174e-01
8.10381830e-01 -7.91984260e-01 -8.98359478e-01 5.44795513e-01
-8.62047374e-01 -5.37018061e-01 -4.91493046e-01 7.01311767e-01
-1.31994891e+00 1.95116494e-02 7.55975954e-03 -1.39396393e+00
6.06536940e-02 -1.57909310e+00 7.21501350e-01 4.32886690e-01
1.93993077e-01 -7.55086124e-01 2.28003696e-01 -7.91163892e-02
3.24752897e-01 5.11964679e-01 6.70978308e-01 -2.22737461e-01
-7.63979852e-01 -1.25888601e-01 4.40175861e-01 3.02158475e-01
-9.90310032e-03 -1.72334507e-01 -5.42632043e-01 -3.47277671e-01
5.36191976e-03 -6.60739303e-01 6.38525665e-01 4.51930195e-01
7.90479302e-01 -3.32991600e-01 -5.63551426e-01 -1.11912815e-02
1.40836048e+00 5.44954002e-01 2.81154871e-01 5.85672915e-01
1.79265872e-01 9.31131303e-01 1.27307177e+00 6.71023428e-01
4.67367917e-01 7.96675622e-01 1.32728803e+00 8.81154001e-01
5.10419548e-01 -1.31684989e-01 5.48579395e-01 3.45299125e-01
-2.86400337e-02 -1.59538269e-01 -7.90473938e-01 6.68841064e-01
-2.46248651e+00 -8.65256727e-01 2.03579769e-01 2.44071794e+00
7.74390757e-01 1.41133055e-01 -9.58693698e-02 -3.98015194e-02
7.81223536e-01 -9.54945013e-02 -1.07148361e+00 -1.00599952e-01
1.75057739e-01 -3.79060924e-01 7.40736902e-01 4.95205402e-01
-1.14941692e+00 6.16984785e-01 5.10311031e+00 8.29864442e-01
-1.01540935e+00 -1.67658746e-01 2.95685172e-01 -9.32545066e-02
1.23525530e-01 -3.32414918e-02 -8.00116420e-01 5.37258625e-01
8.74033988e-01 -2.04037711e-01 6.03828013e-01 1.27664816e+00
5.98166645e-01 -5.75746179e-01 -9.71793413e-01 6.97818398e-01
-5.84827006e-01 -1.14729941e+00 -4.58268255e-01 3.44865501e-01
9.98669088e-01 2.37893730e-01 1.64945930e-01 4.50846702e-01
7.75104702e-01 -7.74877667e-01 1.16935456e+00 8.11070085e-01
3.61026525e-01 -9.67017531e-01 7.79797256e-01 7.61542380e-01
-9.48990941e-01 -7.54908204e-01 -2.37664759e-01 -5.94272986e-02
5.76272726e-01 6.35334373e-01 -7.85728216e-01 5.57995856e-01
5.39370954e-01 6.36899054e-01 3.19410443e-01 1.11224306e+00
-6.21497035e-01 2.86186159e-01 -5.05107105e-01 -7.37565815e-01
2.82391131e-01 -5.18848479e-01 9.88975644e-01 3.30435216e-01
4.86471832e-01 -1.15781926e-01 5.36653042e-01 9.89886820e-01
4.07612115e-01 -6.12498462e-01 -6.42506778e-01 -4.73719276e-02
6.57150030e-01 1.04939842e+00 -3.91552895e-01 -7.36348471e-03
9.68838781e-02 2.11500868e-01 2.97407091e-01 1.93929300e-01
-9.13370788e-01 -4.50955957e-01 9.16583002e-01 -3.82290125e-01
3.47372890e-01 -7.72446871e-01 -2.74134278e-01 -6.97230518e-01
-1.26781628e-01 -3.86133522e-01 -3.46214175e-02 -3.57095987e-01
-9.78708446e-01 1.72366902e-01 2.75931239e-01 -1.39800954e+00
-7.97049522e-01 -8.36172998e-01 -4.12084520e-01 6.45117164e-01
-1.50095117e+00 -5.64685524e-01 7.94719607e-02 8.59044865e-02
6.38110459e-01 -2.31135935e-01 5.77644229e-01 -3.70911539e-01
-6.47940576e-01 -1.38853788e-01 6.33442581e-01 -5.10401368e-01
4.14255768e-01 -1.34960103e+00 -3.02743405e-01 6.70758426e-01
-6.25004411e-01 2.15353459e-01 1.16138339e+00 -7.21402526e-01
-1.95366931e+00 -1.26007831e+00 4.99777682e-02 -2.92710423e-01
9.57314909e-01 4.39859480e-02 -5.87053120e-01 2.53372759e-01
-1.07846558e-01 -5.76575249e-02 -1.81894496e-01 -3.07247251e-01
2.86824048e-01 -1.26373962e-01 -1.07726908e+00 8.56528103e-01
6.11326933e-01 5.60460091e-02 -5.12668252e-01 3.26626785e-02
8.99588406e-01 -3.98860931e-01 -8.43461990e-01 6.90266252e-01
4.37089801e-01 -3.64391029e-01 6.68555558e-01 -3.64189684e-01
1.48345962e-01 -4.51071680e-01 -2.17228726e-01 -1.76640165e+00
-6.47201538e-02 -7.80535042e-01 -6.90789759e-01 7.14620233e-01
2.86980242e-01 -8.00490320e-01 5.54918647e-01 3.97366673e-01
-4.98924285e-01 -1.17639017e+00 -1.25254667e+00 -1.11480904e+00
2.53493220e-01 -4.98216629e-01 2.21569449e-01 2.38472015e-01
2.64443427e-01 -8.14794824e-02 -4.16278183e-01 3.96587312e-01
7.49550641e-01 -9.67277065e-02 8.29455435e-01 -1.12484670e+00
-1.04014426e-01 -5.10403037e-01 -2.06153929e-01 -8.30504298e-01
4.73274082e-01 -4.51544374e-01 9.44263756e-01 -1.69676721e+00
-3.16109866e-01 -5.91073275e-01 2.82986052e-02 1.99929923e-01
-4.55353372e-02 -5.52884996e-01 1.05275355e-01 1.83046788e-01
-6.14005327e-01 1.29743266e+00 1.20058763e+00 -2.27652177e-01
-3.53242218e-01 3.11244458e-01 -2.27116019e-01 7.38728166e-01
9.40389693e-01 -2.52887100e-01 -6.11473858e-01 -2.22511858e-01
3.10829580e-01 3.60835224e-01 2.01378196e-01 -1.21354043e+00
2.80555040e-01 -7.93766856e-01 -4.41592067e-01 -5.55566669e-01
6.67427659e-01 -1.04294097e+00 -1.64282266e-02 8.83695126e-01
-3.28884095e-01 -3.01217943e-01 5.31702451e-02 1.36185193e+00
-1.12654641e-01 -2.88820446e-01 9.40343738e-01 3.60572599e-02
-9.83019054e-01 4.14093316e-01 -7.58398533e-01 4.47957963e-02
1.45568442e+00 1.04791239e-01 -5.37669919e-02 -3.36259067e-01
-6.58716917e-01 8.22401404e-01 2.31551334e-01 5.75650334e-01
4.79149550e-01 -1.08252215e+00 -4.27555829e-01 -3.25546116e-01
7.41388053e-02 4.01830703e-01 3.82605530e-02 7.21484959e-01
-5.17286770e-02 4.54249620e-01 -1.60663903e-01 -7.70794213e-01
-4.17830020e-01 4.53612983e-01 2.78974354e-01 -2.53599882e-01
-2.60913819e-01 4.20759320e-01 -9.81779471e-02 -6.86350644e-01
2.42879078e-01 -4.87820983e-01 -1.44113854e-01 -1.15272075e-01
2.42748827e-01 9.36651707e-01 -1.51701123e-01 -3.44569176e-01
-6.71691298e-02 2.55636841e-01 3.97613734e-01 -4.11143541e-01
1.27624190e+00 -3.05937558e-01 1.29746899e-01 6.13361537e-01
7.06843317e-01 -6.06796205e-01 -2.00069928e+00 -8.76836032e-02
3.54984373e-01 8.52519879e-04 2.47539565e-01 -7.25724995e-01
-5.46336055e-01 7.60715365e-01 4.68252361e-01 1.18336000e-01
5.51224887e-01 -2.29859501e-01 3.97664517e-01 7.81817138e-01
1.15198135e+00 -1.62868774e+00 3.32292914e-01 1.03393984e+00
1.04693723e+00 -1.33779967e+00 -1.68510422e-01 1.23370430e-02
-7.00533390e-01 1.08904028e+00 8.48806322e-01 -3.40491593e-01
5.52918553e-01 2.67669320e-01 -3.96146655e-01 3.13042939e-01
-9.54708338e-01 -1.63476288e-01 -2.90886667e-02 6.38993561e-01
-4.32792872e-01 3.46537262e-01 -3.00740778e-01 4.81841326e-01
1.35062367e-01 -6.66402504e-02 6.21849418e-01 1.32148504e+00
-9.49186921e-01 -8.34067404e-01 -5.53812683e-01 3.51365119e-01
3.28088403e-02 6.22079253e-01 3.43564749e-01 8.17797422e-01
-9.15818512e-02 1.12604630e+00 -1.18355423e-01 -2.49657899e-01
2.23039970e-01 -1.66824579e-01 3.55826706e-01 -5.56194067e-01
3.94726098e-01 -1.18451692e-01 6.00114726e-02 -7.42959559e-01
-3.28590661e-01 -7.26490796e-01 -1.58188677e+00 3.02054077e-01
-6.99564040e-01 2.95242220e-01 1.10723758e+00 1.13437760e+00
4.38544214e-01 5.70525050e-01 7.87849307e-01 -1.33305061e+00
-1.52772427e+00 -8.93254995e-01 -5.01301706e-01 -4.52473074e-01
3.94929856e-01 -1.52659750e+00 -3.89526546e-01 -3.46837848e-01] | [4.701535224914551, 2.0122108459472656] |
d6722d93-4636-4827-85d6-42d606730398 | adversarial-contrastive-estimation | 1805.03642 | null | http://arxiv.org/abs/1805.03642v3 | http://arxiv.org/pdf/1805.03642v3.pdf | Adversarial Contrastive Estimation | Learning by contrasting positive and negative samples is a general strategy
adopted by many methods. Noise contrastive estimation (NCE) for word embeddings
and translating embeddings for knowledge graphs are examples in NLP employing
this approach. In this work, we view contrastive learning as an abstraction of
all such methods and augment the negative sampler into a mixture distribution
containing an adversarially learned sampler. The resulting adaptive sampler
finds harder negative examples, which forces the main model to learn a better
representation of the data. We evaluate our proposal on learning word
embeddings, order embeddings and knowledge graph embeddings and observe both
faster convergence and improved results on multiple metrics. | ['Yanshuai Cao', 'Avishek Joey Bose', 'Huan Ling'] | 2018-05-09 | adversarial-contrastive-estimation-1 | https://aclanthology.org/P18-1094 | https://aclanthology.org/P18-1094.pdf | acl-2018-7 | ['learning-word-embeddings'] | ['methodology'] | [-9.99851823e-02 3.34603548e-01 -4.10817355e-01 2.35677939e-02
-7.43069410e-01 -7.44338989e-01 9.11332011e-01 2.35910892e-01
-8.20724666e-01 6.58185780e-01 1.35515809e-01 -1.18108846e-01
-5.29236607e-02 -1.11735535e+00 -5.88129818e-01 -7.18713999e-01
-1.77397981e-01 7.09031820e-01 4.23997611e-01 -2.22642168e-01
1.15045838e-01 6.81737661e-01 -1.09343934e+00 -2.33127270e-02
6.95645332e-01 6.17031991e-01 -2.04214215e-01 9.59608674e-01
-4.08817798e-01 5.11900246e-01 -6.51833236e-01 -9.40786421e-01
2.97621161e-01 -2.58541077e-01 -7.12416530e-01 -2.25664631e-01
5.28031707e-01 2.38725260e-01 -2.80695379e-01 1.25582337e+00
5.84443927e-01 3.16530019e-01 1.15645707e+00 -1.50904906e+00
-9.49978590e-01 4.98828769e-01 -5.01701057e-01 2.74828702e-01
2.32414797e-01 1.79322839e-01 1.26094913e+00 -1.08411205e+00
6.45896018e-01 1.33177924e+00 7.54230142e-01 5.70369124e-01
-1.59257591e+00 -4.78668094e-01 1.11132428e-01 3.43199730e-01
-1.31612027e+00 -1.74206614e-01 8.55297744e-01 -5.17343342e-01
7.19702780e-01 2.50017434e-01 9.28852379e-01 1.46490431e+00
-1.42880240e-02 9.44402635e-01 1.00708628e+00 -5.19256830e-01
4.60077316e-01 5.02553225e-01 2.29210839e-01 6.42785132e-01
6.16391659e-01 2.71801233e-01 -4.08257246e-01 -4.48965311e-01
4.62347269e-01 -1.18675185e-02 -2.82562822e-01 -1.05061758e+00
-6.21593356e-01 1.01822066e+00 3.53066921e-01 1.81010231e-01
-2.48912767e-01 3.68277967e-01 4.52965319e-01 4.82800901e-01
8.26258481e-01 5.39363444e-01 -2.85799950e-01 -5.77476108e-04
-6.89881861e-01 3.62742811e-01 1.18346298e+00 7.36372352e-01
8.61646175e-01 1.13487005e-01 -1.91034332e-01 8.04135263e-01
4.29464072e-01 6.33415639e-01 4.72818464e-01 -6.22610927e-01
8.26085061e-02 4.54661280e-01 3.07927839e-02 -7.52274096e-01
-2.67987043e-01 -5.65375447e-01 -6.22222662e-01 3.92082334e-01
5.40396869e-01 -1.36954069e-01 -1.12701368e+00 1.67586076e+00
2.90569752e-01 3.05313528e-01 8.85179080e-03 5.17526090e-01
3.63354623e-01 5.66680849e-01 2.56111741e-01 -6.96921945e-02
1.15861189e+00 -8.12445462e-01 -8.46771240e-01 -2.15056986e-01
6.34550571e-01 -5.94204783e-01 1.10496700e+00 4.89543110e-01
-9.62660670e-01 -3.10862958e-01 -1.04325807e+00 -4.15873304e-02
-9.12964106e-01 -2.23438546e-01 3.54474664e-01 9.42675829e-01
-1.09177184e+00 7.00100303e-01 -6.92092896e-01 -3.65862757e-01
6.95702016e-01 1.58831224e-01 -2.86163568e-01 -2.19084859e-01
-1.29369032e+00 1.03108132e+00 3.86826724e-01 -3.02523542e-02
-7.38479972e-01 -7.39209950e-01 -9.84369099e-01 -6.90718740e-02
3.99573982e-01 -7.37196863e-01 7.98627853e-01 -7.94203103e-01
-1.41366220e+00 8.60895455e-01 8.29937682e-02 -4.29135710e-01
8.29595566e-01 -1.81586131e-01 -3.24186295e-01 -1.90183312e-01
-2.65656531e-01 4.84520137e-01 1.29406106e+00 -1.36255562e+00
-7.30971992e-02 -4.10604000e-01 -4.85871620e-02 -5.09540215e-02
-5.27679622e-01 -4.49347556e-01 -1.38740361e-01 -6.29525840e-01
-2.48194963e-01 -7.08447516e-01 -2.25888833e-01 2.58578509e-01
-2.64928550e-01 -6.18791938e-01 6.53886199e-01 -3.64190340e-01
1.13770974e+00 -2.14338756e+00 2.91164100e-01 5.47783315e-01
5.55598557e-01 5.09439766e-01 -4.21059430e-01 5.95833719e-01
-1.54364154e-01 2.94002771e-01 -1.36929333e-01 -4.19713944e-01
3.76039654e-01 4.66590911e-01 -1.16131775e-01 6.97445512e-01
5.45136690e-01 1.02223563e+00 -1.33120179e+00 -3.23587298e-01
9.02071893e-02 4.90935564e-01 -7.24269450e-01 2.60466188e-01
-1.99063495e-01 -3.86538327e-01 -2.54950166e-01 1.93784237e-01
7.26302385e-01 -1.12939551e-01 2.90347904e-01 -1.24115773e-01
3.20627004e-01 -2.05443297e-02 -1.53383720e+00 1.21911573e+00
-4.12991792e-01 5.29633224e-01 -1.57398462e-01 -1.21285605e+00
8.96201730e-01 1.04893342e-01 -1.70235604e-01 -2.22737148e-01
1.66092098e-01 2.37554893e-01 2.76863091e-02 -5.63513041e-01
5.52687287e-01 -5.84595680e-01 1.23821892e-01 3.73714000e-01
4.39605296e-01 -2.66395569e-01 8.77218321e-02 2.83717543e-01
1.11128151e+00 7.86893442e-02 4.31976438e-01 -2.91490167e-01
3.77820700e-01 -3.56486082e-01 2.47523725e-01 9.55955863e-01
-4.33476239e-01 3.70861024e-01 8.30558479e-01 -2.96192527e-01
-1.15538657e+00 -1.49185276e+00 1.50545433e-01 1.01579678e+00
-1.58853814e-01 -5.76967776e-01 -5.35223901e-01 -1.22133255e+00
9.14786458e-02 6.66187346e-01 -8.72136652e-01 -4.24292386e-01
-5.14174819e-01 -6.45965874e-01 4.96651828e-01 7.07497954e-01
-3.01427633e-01 -9.09471512e-01 2.03314140e-01 -1.40398905e-01
2.69537866e-01 -6.54016793e-01 -3.31472576e-01 4.11497593e-01
-6.20457590e-01 -1.20777464e+00 -7.40040004e-01 -9.02070045e-01
6.03475153e-01 -2.58160263e-01 1.31732273e+00 -1.98657453e-01
-2.50819683e-01 5.90113342e-01 -2.29153857e-01 -4.62110132e-01
-6.06183648e-01 7.86571652e-02 1.14345342e-01 -6.35303780e-02
5.73603690e-01 -5.95675886e-01 -1.45909384e-01 -3.41748089e-01
-1.17449284e+00 -5.30218780e-01 5.34001470e-01 9.22588825e-01
4.81105834e-01 -2.46249884e-01 5.59850574e-01 -1.15336573e+00
9.78178024e-01 -5.54260015e-01 -5.81205964e-01 1.62530988e-01
-6.70883656e-01 4.59787130e-01 7.65636921e-01 -8.70266736e-01
-4.58727300e-01 -2.32295275e-01 -2.54843533e-01 -6.40824080e-01
8.74775052e-02 1.63450435e-01 -2.01246455e-01 -3.03472430e-02
9.52150047e-01 1.24383889e-01 2.99610812e-02 -3.63729537e-01
9.98742521e-01 3.55873287e-01 2.18278229e-01 -5.94489872e-01
1.32579982e+00 4.29262310e-01 -1.46661624e-01 -9.82556105e-01
-6.90607071e-01 -4.08779353e-01 -4.06650931e-01 -1.33869722e-01
7.84370959e-01 -5.78610539e-01 -4.53676522e-01 8.34833235e-02
-1.14933550e+00 -3.63596320e-01 -9.08643782e-01 5.98492265e-01
-4.54776078e-01 3.55157167e-01 -2.84961373e-01 -1.05931354e+00
-2.47975532e-02 -7.68395424e-01 8.03246677e-01 -3.72020826e-02
-3.64556193e-01 -1.46685982e+00 6.49273157e-01 -1.59608707e-01
4.32749718e-01 -3.58920591e-03 1.07197511e+00 -1.32197654e+00
-3.17454726e-01 -4.15254295e-01 -2.21307538e-02 8.53087187e-01
-1.98213562e-01 3.36308638e-03 -1.17494929e+00 -3.17833900e-01
-1.23916283e-01 -4.51533616e-01 1.05447543e+00 7.10067302e-02
9.54183578e-01 -3.35137397e-01 -2.87937135e-01 3.10690254e-01
1.70306873e+00 -3.60204428e-01 7.15489805e-01 6.62950128e-02
7.19683886e-01 6.77726567e-01 2.49483541e-01 1.12427197e-01
-5.79739623e-02 1.95592389e-01 4.06806737e-01 4.76255678e-02
-7.20013585e-03 -4.17637765e-01 4.25153792e-01 8.67771089e-01
1.98977590e-01 -3.36900264e-01 -8.37200880e-01 9.55645442e-01
-1.51668060e+00 -1.04474699e+00 -1.10456850e-02 2.13527894e+00
9.37164247e-01 3.00811648e-01 2.59681970e-01 2.17531666e-01
7.42881477e-01 4.95492339e-01 -3.51243287e-01 -4.15979564e-01
9.67046767e-02 6.72187865e-01 4.37097162e-01 8.05171311e-01
-8.92888308e-01 6.86364889e-01 6.79212332e+00 1.14276540e+00
-7.42210150e-01 3.15142900e-01 1.25374064e-01 -1.51050419e-01
-7.95047820e-01 -5.44746630e-02 -4.77738857e-01 1.71884000e-01
7.23829687e-01 -1.65975943e-01 2.45379344e-01 9.35790360e-01
-3.12860072e-01 2.33774722e-01 -1.37148786e+00 8.95747364e-01
2.40466923e-01 -1.20557094e+00 1.18855558e-01 1.57909140e-01
7.60251462e-01 6.26464635e-02 1.09107159e-01 5.85513592e-01
5.82591116e-01 -9.50421333e-01 4.62323517e-01 6.42548800e-01
3.98994327e-01 -8.67648959e-01 8.49506080e-01 2.91508049e-01
-8.85353386e-01 1.22309677e-01 -4.27576035e-01 1.12623654e-01
5.04728518e-02 7.48648226e-01 -6.12516165e-01 3.53607744e-01
2.53311366e-01 5.57271659e-01 -5.79503953e-01 8.94202769e-01
-3.65555853e-01 7.11200595e-01 -2.95967817e-01 -4.16030467e-01
1.48924142e-02 -4.47386980e-01 7.23666906e-01 1.37580574e+00
-9.17964429e-02 -5.07434726e-01 1.52974740e-01 1.02760649e+00
-4.02020872e-01 1.83450162e-01 -9.62894499e-01 -3.15419465e-01
4.13096189e-01 1.29641843e+00 -4.23468381e-01 -5.83164632e-01
-4.62812990e-01 8.57284546e-01 7.17132688e-01 4.19154793e-01
-7.28686154e-01 -5.02264142e-01 7.18565941e-01 6.77031949e-02
4.48399425e-01 -2.06064694e-02 3.38054113e-02 -1.09242070e+00
9.32564866e-03 -6.58330977e-01 3.26614082e-01 -4.19032723e-01
-1.97982526e+00 3.51720273e-01 -1.23972110e-01 -1.04452395e+00
-1.30599871e-01 -9.02934134e-01 -5.03586113e-01 9.06107485e-01
-1.72290659e+00 -7.48515904e-01 -5.92475720e-02 2.97634095e-01
9.93822888e-02 -5.41488081e-02 7.59214997e-01 1.53086931e-01
-4.18577373e-01 6.65999532e-01 2.98779845e-01 3.35073918e-01
7.26617217e-01 -1.84053218e+00 3.12047184e-01 6.17887378e-01
4.11999345e-01 5.88562548e-01 9.26322877e-01 -4.57096130e-01
-1.23779905e+00 -1.05875623e+00 7.83554792e-01 -7.13615716e-01
1.15617168e+00 -5.79984188e-01 -1.13227212e+00 6.15598559e-01
2.62010753e-01 3.91144186e-01 8.16933393e-01 2.32185543e-01
-6.58263743e-01 -4.77674510e-03 -1.11800754e+00 6.90386951e-01
8.41537237e-01 -7.34603345e-01 -7.51243711e-01 3.56654137e-01
5.88530540e-01 3.57311964e-02 -9.15919840e-01 5.17331623e-02
3.90771151e-01 -8.06812108e-01 9.38188493e-01 -9.44059730e-01
1.26451418e-01 -1.04593389e-01 7.59811848e-02 -1.76869869e+00
-1.51653811e-01 -6.37218833e-01 -4.73501295e-01 1.13826561e+00
3.74218762e-01 -9.52781022e-01 9.00927544e-01 -3.62903699e-02
2.66576707e-01 -7.64371753e-01 -7.35445976e-01 -1.07717299e+00
4.83838946e-01 -6.65020466e-01 1.41805157e-01 9.83328223e-01
-3.43581252e-02 5.35658538e-01 -7.86443055e-02 6.56200871e-02
7.27977574e-01 -3.24094951e-01 6.49278820e-01 -1.29762709e+00
-3.65926355e-01 -5.19074261e-01 -6.78457797e-01 -9.30815697e-01
4.26505148e-01 -1.30133700e+00 -1.26995578e-01 -1.39584994e+00
1.49874434e-01 -7.00718090e-02 -1.46689460e-01 1.46063447e-01
-2.82631695e-01 2.64822185e-01 2.23979838e-02 -3.30710858e-01
-6.32949829e-01 7.14592636e-01 1.06119251e+00 -1.16598435e-01
1.02693893e-01 -2.28435084e-01 -5.10526180e-01 7.18209803e-01
7.15769351e-01 -7.53262758e-01 -4.22673106e-01 -2.27514431e-02
5.90750039e-01 -5.63463032e-01 5.68549573e-01 -5.94785213e-01
8.47264230e-02 1.02846399e-01 3.93356353e-01 -2.64899522e-01
3.46185386e-01 -7.72505939e-01 -2.85419732e-01 6.22967005e-01
-4.84817564e-01 1.00114681e-01 -4.53819335e-02 1.09500575e+00
-5.35222366e-02 -5.37769198e-01 9.00572002e-01 -3.97960544e-02
-3.68123531e-01 2.73874134e-01 -2.48855203e-01 7.38656282e-01
6.83889151e-01 -1.49200767e-01 -2.31749073e-01 -2.38349587e-01
-1.05798745e+00 8.71575400e-02 4.23183590e-01 2.82582700e-01
6.16730392e-01 -1.58218932e+00 -6.73556566e-01 2.22898513e-01
3.29892248e-01 -2.21282795e-01 -1.45023718e-01 8.15707445e-01
-3.39053005e-01 -2.04684615e-01 3.10331643e-01 -2.34980419e-01
-9.97773588e-01 7.26787269e-01 3.71256858e-01 -4.60220754e-01
-3.14939857e-01 8.73041749e-01 6.07954059e-03 -6.83461607e-01
3.46829772e-01 -6.09698817e-02 -2.18959019e-01 4.42545205e-01
4.46194828e-01 6.07605994e-01 -1.75838202e-01 -2.38421727e-02
-2.83832788e-01 2.07737043e-01 -6.92999363e-02 -1.19551018e-01
1.33453763e+00 3.15193266e-01 -1.66276947e-01 7.37217963e-01
1.67173052e+00 4.22820210e-01 -9.21159148e-01 -4.16386336e-01
2.26064026e-01 -3.11633915e-01 -1.80494979e-01 -3.11241448e-01
-8.52141917e-01 9.94483232e-01 7.66797304e-01 6.82222307e-01
6.01951420e-01 1.97246537e-01 3.93517166e-01 2.74457186e-01
-3.84203009e-02 -1.19937432e+00 3.89584571e-01 3.65454018e-01
6.96639776e-01 -1.10108697e+00 -1.33846132e-02 -3.62040818e-01
-3.16533417e-01 1.13173771e+00 4.14113820e-01 -7.22561598e-01
8.80502343e-01 3.41704607e-01 -3.27138938e-02 -2.59549171e-01
-7.07685769e-01 -5.01202464e-01 3.60117942e-01 9.55925524e-01
2.31817767e-01 2.36185305e-02 -5.35715520e-01 3.90529752e-01
-8.94285552e-03 -2.88170755e-01 4.50905085e-01 6.78508520e-01
-3.76787007e-01 -1.30540144e+00 -1.44313917e-01 5.19487023e-01
-2.16937453e-01 -2.37991065e-01 -4.47643995e-01 1.09352124e+00
-1.80026833e-02 5.39602101e-01 -2.76660994e-02 -2.66678154e-01
4.47420388e-01 4.76650327e-01 7.31929600e-01 -7.49423921e-01
-5.00968575e-01 -3.09180707e-01 2.43876632e-02 -3.87366951e-01
-2.24601611e-01 -3.67230505e-01 -7.34537721e-01 -1.41026348e-01
-5.67534208e-01 3.45711946e-01 4.54694211e-01 8.75023544e-01
-6.91900924e-02 5.72779000e-01 4.72027153e-01 -5.80597341e-01
-1.12922454e+00 -1.05671942e+00 -7.23024249e-01 5.45077920e-01
3.55632693e-01 -4.54527438e-01 -9.05628502e-01 -3.20354134e-01] | [10.382803916931152, 8.499197959899902] |
31771b06-3613-4a44-9149-f86634ddda86 | fchd-a-fast-and-accurate-head-detector | 1809.08766 | null | https://arxiv.org/abs/1809.08766v3 | https://arxiv.org/pdf/1809.08766v3.pdf | FCHD: Fast and accurate head detection in crowded scenes | In this paper, we propose FCHD-Fully Convolutional Head Detector, an end-to-end trainable head detection model. Our proposed architecture is a single fully convolutional network which is responsible for both bounding box prediction and classification. This makes our model lightweight with low inference time and memory requirements. Along with run-time, our model has better overall average precision (AP) which is achieved by selection of anchor sizes based on the effective receptive field of the network. This can be concluded from our experiments on several head detection datasets with varying head counts. We achieve an AP of 0.70 on a challenging head detection dataset which is comparable to some standard benchmarks. Along with this our model runs at 5 FPS on Nvidia Quadro M1000M for VGA resolution images. Code is available at https://github.com/aditya-vora/FCHD-Fully-Convolutional-Head-Detector. | ['Aditya Vora', 'Vinay Chilaka'] | 2018-09-24 | null | null | null | null | ['head-detection'] | ['computer-vision'] | [-5.87918699e-01 2.10854754e-01 7.49382675e-02 -5.51172674e-01
-6.69703662e-01 -2.94623584e-01 2.53453374e-01 8.55032131e-02
-7.14795768e-01 4.52885419e-01 1.93105340e-01 -6.45223558e-02
4.68436420e-01 -7.04488575e-01 -7.58654237e-01 -5.23831427e-01
-2.95609504e-01 1.58696085e-01 5.18487692e-01 1.85177028e-01
-2.78153140e-02 6.41859829e-01 -1.80287552e+00 1.64503045e-02
4.27404970e-01 1.05823028e+00 1.51247576e-01 1.10289478e+00
6.85888469e-01 7.12103009e-01 -4.53700155e-01 -2.99764127e-01
2.59648353e-01 2.29033560e-01 -5.92346728e-01 -1.50767937e-01
8.51204634e-01 -9.44561183e-01 -6.44907534e-01 7.26281464e-01
1.11569166e+00 -5.51417097e-02 4.37260568e-01 -1.35342598e+00
-2.05568522e-01 4.45274323e-01 -6.39242887e-01 3.94835293e-01
2.43705615e-01 3.64088863e-01 6.67069674e-01 -9.65476394e-01
1.99246421e-01 1.36424065e+00 7.03042984e-01 6.84459329e-01
-5.78625262e-01 -9.34216022e-01 3.13376561e-02 3.74229774e-02
-1.74579346e+00 -5.34413099e-01 8.58324021e-03 -2.57373154e-01
9.96485472e-01 1.32358953e-01 5.52920640e-01 7.34102428e-01
4.71796900e-01 9.32076573e-01 7.19464481e-01 -1.53768450e-01
2.30127916e-01 -2.12703720e-01 2.59906232e-01 8.57809544e-01
5.12305737e-01 5.11695370e-02 -4.03796971e-01 -3.30315307e-02
7.78912365e-01 -5.05844951e-02 -1.91396117e-01 -5.45628555e-02
-6.98044300e-01 9.45891798e-01 7.43221700e-01 -2.27590606e-01
-4.13084000e-01 3.73605728e-01 5.67088962e-01 -3.83318454e-01
1.95773333e-01 -2.88958549e-01 -3.15042228e-01 1.77191019e-01
-9.73778486e-01 4.71808136e-01 6.74470127e-01 1.20815074e+00
2.52917171e-01 1.28002599e-01 -4.63376969e-01 4.66484994e-01
4.72616047e-01 9.34388041e-01 2.72124499e-01 -8.61539423e-01
7.23732412e-02 4.85884368e-01 -7.16618150e-02 -4.43957865e-01
-9.03838873e-01 -4.22853529e-01 -7.19550073e-01 4.84616488e-01
4.08428133e-01 -5.58736145e-01 -1.10639656e+00 1.46115220e+00
6.11596823e-01 5.15451878e-02 -2.66622186e-01 1.19971836e+00
1.20885038e+00 3.07633668e-01 3.58444870e-01 3.07761610e-01
1.67414427e+00 -1.04402602e+00 -4.19194579e-01 -4.31538612e-01
7.19584823e-01 -5.33411443e-01 8.10144186e-01 2.63349295e-01
-1.01184952e+00 -4.46156174e-01 -1.09426391e+00 -4.64569628e-01
-2.33767718e-01 4.45203036e-01 5.51208735e-01 6.31592691e-01
-1.41023958e+00 1.72135726e-01 -1.05561042e+00 -5.72400868e-01
4.46468890e-01 8.23776484e-01 -2.69811183e-01 1.73694059e-01
-6.85676515e-01 6.43281400e-01 3.97684813e-01 1.06455266e-01
-9.79293525e-01 -4.79380012e-01 -8.51567805e-01 3.51881534e-02
-7.97955394e-02 -8.44319344e-01 1.62141800e+00 -3.22198957e-01
-1.21904910e+00 8.27959836e-01 -1.04064196e-01 -9.19135511e-01
7.32312381e-01 -5.24746656e-01 -1.30719105e-02 -3.32258106e-03
2.07556970e-02 1.14832175e+00 7.34595060e-01 -7.00470448e-01
-1.13288236e+00 -7.29393363e-01 -9.97970998e-02 1.03047639e-01
2.13745460e-02 2.11850226e-01 -5.76000690e-01 -1.68470621e-01
-2.93361247e-01 -1.13548553e+00 -8.08376223e-02 2.35647619e-01
-4.69412267e-01 -2.88718730e-01 9.07418847e-01 -5.99160969e-01
1.17731881e+00 -1.91848755e+00 -4.83031332e-01 -1.17804848e-01
4.02576506e-01 4.02833998e-01 1.85625836e-01 1.26528868e-03
1.20401874e-01 -3.78352910e-01 8.16590190e-02 -6.76571429e-01
-2.25153053e-03 -1.26449347e-01 -1.09983265e-01 9.24909949e-01
-1.12982117e-01 8.20790589e-01 -5.43850183e-01 -6.01457775e-01
5.32078147e-01 1.09919178e+00 -5.79094887e-01 4.39392805e-01
3.58041823e-01 2.77365148e-01 -2.17969716e-01 6.82847738e-01
1.00776935e+00 -7.33636767e-02 -1.13992237e-01 6.36621639e-02
-2.64243096e-01 1.24619469e-01 -1.14106500e+00 1.26301634e+00
-1.99288145e-01 7.36493886e-01 1.57674432e-01 -1.70483842e-01
7.23777175e-01 4.61693406e-02 -6.96896687e-02 -6.83702886e-01
6.15388334e-01 2.35664882e-02 -9.83996391e-02 -1.23765811e-01
4.82223928e-01 5.00881314e-01 4.49510403e-02 2.12058818e-04
7.01995241e-03 4.50828969e-01 2.06954852e-02 6.93860054e-02
1.08508313e+00 -1.87348798e-01 3.97963583e-01 -5.09124517e-01
4.61125165e-01 -2.80739427e-01 3.70590985e-01 6.77372515e-01
-6.05115235e-01 7.88210332e-01 2.00931311e-01 -5.30036151e-01
-9.74623919e-01 -9.26049292e-01 -4.96069372e-01 1.48816693e+00
-1.62314534e-01 -4.09005612e-01 -1.21045375e+00 -3.77432853e-01
5.77612668e-02 4.60616738e-01 -7.83413351e-01 1.00192644e-01
-7.46570706e-01 -4.85923558e-01 9.03975248e-01 1.10610020e+00
6.29993081e-01 -1.02146792e+00 -1.27295244e+00 2.11416278e-02
2.84025699e-01 -1.23750949e+00 -3.94870907e-01 2.15781957e-01
-6.35653853e-01 -9.70493555e-01 -8.81695032e-01 -7.50316143e-01
4.25176233e-01 2.23384351e-01 1.06023681e+00 2.05811128e-01
-4.64116484e-01 1.17455892e-01 -4.99574021e-02 -7.63482153e-01
1.95727453e-01 3.29663962e-01 1.12435922e-01 -5.75065792e-01
5.08289158e-01 -1.46946222e-01 -1.24188972e+00 2.29901262e-02
-4.62166399e-01 -1.12489037e-01 3.97906899e-01 4.58202899e-01
5.56569755e-01 -5.60911953e-01 1.38501763e-01 -4.11384255e-01
2.25478619e-01 -3.29167306e-01 -1.00476193e+00 -1.31411031e-01
-3.03159773e-01 -2.20532998e-01 4.37132806e-01 -1.87813595e-01
-6.81494176e-01 4.67752069e-01 -3.82856250e-01 -4.71871257e-01
-5.00637829e-01 -2.69931614e-01 6.77014068e-02 -1.82898287e-02
6.21618509e-01 7.31228888e-02 -1.20910513e-03 -2.85162419e-01
1.13618255e-01 8.15873027e-01 7.89775252e-01 -5.98786809e-02
2.48710588e-01 6.33183777e-01 2.29842626e-02 -1.00329280e+00
-7.12014019e-01 -5.79537451e-01 -8.00890088e-01 -4.59671170e-02
1.03801215e+00 -1.46948016e+00 -1.03152132e+00 8.39824796e-01
-1.00012708e+00 -6.25402927e-01 3.38489115e-01 3.64552706e-01
-1.80963367e-01 8.52973666e-03 -9.19665813e-01 -9.62409198e-01
-1.16070378e+00 -9.91572976e-01 1.63974881e+00 5.27291656e-01
-8.33353400e-02 -8.33202720e-01 -9.39024389e-02 -1.42966397e-02
3.24972004e-01 2.82889485e-01 1.10805351e-02 -4.30508137e-01
-3.96673054e-01 -3.74505013e-01 -4.82399344e-01 -1.64895747e-02
-4.49889809e-01 1.15582317e-01 -1.31588411e+00 -8.57130110e-01
-6.34316266e-01 -2.35982195e-01 7.64942467e-01 9.39957559e-01
1.25624645e+00 -2.52801776e-01 -4.19450939e-01 7.32778251e-01
1.35057223e+00 -1.33489430e-01 6.98454976e-01 4.14595097e-01
8.06081116e-01 2.95160174e-01 5.52987397e-01 8.81045163e-01
7.45550394e-01 6.46667659e-01 6.04317546e-01 -2.94969112e-01
-3.75936031e-01 -1.58285946e-01 3.27793777e-01 8.24124739e-02
1.89332992e-01 -3.78520578e-01 -1.15277934e+00 5.95010996e-01
-1.70865476e+00 -6.18942797e-01 -3.88996601e-01 2.22216988e+00
2.19801992e-01 1.73841655e-01 8.71558130e-01 4.50728014e-02
8.34988892e-01 -7.03890175e-02 -6.17928624e-01 -4.37643945e-01
3.09345871e-01 2.22425714e-01 9.74818468e-01 5.01894474e-01
-1.39274740e+00 1.04721904e+00 5.67762852e+00 4.10271913e-01
-1.43833268e+00 1.13141701e-01 4.89475399e-01 -4.16884243e-01
7.50615656e-01 -5.70043385e-01 -1.61913872e+00 5.10943115e-01
1.27950788e+00 3.38592790e-02 -1.71967167e-02 1.25943136e+00
1.04346260e-01 -2.25580245e-01 -1.07335067e+00 1.03600037e+00
8.82027112e-03 -8.53950202e-01 -2.20680878e-01 3.90467942e-01
1.48074254e-01 6.85435653e-01 2.54873447e-02 4.87330109e-01
2.04345256e-01 -1.17159367e+00 9.85433578e-01 -2.02925831e-01
8.15542579e-01 -1.11607242e+00 9.11958873e-01 3.19235086e-01
-1.48165131e+00 -2.05762386e-01 -5.36617160e-01 -1.27354816e-01
-4.76484559e-02 1.85363352e-01 -1.06028545e+00 -2.90243775e-01
1.14792597e+00 6.63596988e-02 -8.04414153e-01 1.34850645e+00
-3.54599327e-01 4.40075010e-01 -6.62021279e-01 -4.89803702e-02
2.05388248e-01 5.42306066e-01 1.69825643e-01 1.70449460e+00
1.82546645e-01 3.18281382e-01 9.23214555e-02 3.62209141e-01
-6.89013153e-02 -4.29706741e-03 -5.37713051e-01 7.15656102e-01
6.65302575e-01 1.45621908e+00 -8.37531447e-01 -2.93716580e-01
-2.22906917e-01 7.52769530e-01 4.63187009e-01 -1.64646536e-01
-1.39640510e+00 -4.60570812e-01 6.74880922e-01 3.99752498e-01
4.92826790e-01 -6.68836907e-02 -1.60256401e-01 -8.25971305e-01
-9.72081125e-02 -4.31556433e-01 4.16945368e-01 -5.58637261e-01
-7.80503869e-01 8.42263818e-01 -3.58978771e-02 -9.28491235e-01
-1.99533477e-01 -7.38032401e-01 -7.07852304e-01 5.13327062e-01
-1.33506203e+00 -1.46724403e+00 -8.30547690e-01 7.62085140e-01
5.33432305e-01 4.45020318e-01 6.66316450e-01 4.16335791e-01
-9.08476591e-01 1.17644262e+00 -3.92049849e-01 5.87463796e-01
6.67092204e-01 -1.28971577e+00 8.49979877e-01 8.10996830e-01
-5.46393394e-01 5.24847865e-01 7.15355933e-01 -5.05750239e-01
-1.22923207e+00 -1.47541451e+00 5.59532881e-01 -4.24121320e-01
2.16938987e-01 -4.94993657e-01 -8.93437326e-01 8.79585326e-01
2.95156628e-01 4.14457947e-01 4.77991402e-01 -4.30166908e-02
-3.43700886e-01 3.68972309e-02 -1.29537570e+00 2.78504997e-01
1.10578597e+00 -5.42222597e-02 -2.59220034e-01 3.66667807e-01
4.45910245e-01 -9.36117530e-01 -6.98443711e-01 4.36589807e-01
8.03346336e-01 -1.06806660e+00 7.45860279e-01 -1.90904826e-01
9.80495960e-02 -2.37601653e-01 -7.71770775e-02 -8.08055282e-01
-4.10736889e-01 -1.47002488e-01 -3.89444739e-01 9.18847919e-01
2.79665917e-01 -6.17246032e-01 8.98266852e-01 6.65075421e-01
-1.32570818e-01 -8.68014395e-01 -7.95321047e-01 -6.58947229e-01
2.89792210e-01 -2.07683042e-01 5.65594971e-01 2.18956798e-01
-2.71277338e-01 3.66890162e-01 -1.15361944e-01 5.91648698e-01
9.89823043e-01 -2.18745098e-01 9.66165543e-01 -1.11582196e+00
5.25429100e-02 -1.77284852e-01 -6.62140489e-01 -8.89893889e-01
1.40019476e-01 -4.17350769e-01 6.00247160e-02 -1.41626894e+00
3.65319788e-01 -1.60605386e-02 -1.85699277e-02 8.08543026e-01
2.42955214e-03 4.29569453e-01 2.30423704e-01 -5.76434471e-02
-7.73382783e-01 2.89218694e-01 8.06329787e-01 3.17290246e-01
-1.65964931e-01 -5.16596846e-02 -2.89706767e-01 1.01082134e+00
1.13787508e+00 -2.51936525e-01 -5.89133874e-02 -6.10648394e-01
-4.79751259e-01 6.80816593e-03 5.12063146e-01 -1.60951841e+00
5.14430106e-01 3.93596023e-01 8.92855525e-01 -8.93486559e-01
3.68277341e-01 -4.02382016e-01 -1.68937162e-01 8.09109628e-01
3.59817706e-02 3.00809443e-01 5.01097441e-01 1.32686585e-01
2.99473792e-01 3.90664071e-01 1.24552953e+00 1.27821907e-01
-9.67227042e-01 5.01181304e-01 -2.69759089e-01 -3.85769270e-02
1.11548960e+00 -8.65723565e-02 -6.10778213e-01 -3.40233952e-01
-4.28686827e-01 5.28729439e-01 5.26860297e-01 4.10540104e-01
7.01156080e-01 -1.00530910e+00 -8.80415797e-01 3.91911417e-01
1.12359121e-01 1.77454576e-01 2.20989719e-01 7.51089036e-01
-8.55279565e-01 7.45667636e-01 -2.81374663e-01 -7.80563414e-01
-1.67120814e+00 2.83936501e-01 3.86169106e-01 1.25913963e-01
-8.41104507e-01 9.68064427e-01 3.09416920e-01 -2.35256553e-01
8.03894758e-01 -5.09962678e-01 -1.91844150e-01 -2.94023186e-01
1.03398669e+00 5.79562843e-01 1.22029237e-01 -9.83454466e-01
-8.01368415e-01 3.41792792e-01 -1.71160176e-01 1.18935481e-01
1.06979859e+00 -3.15420814e-02 2.70869613e-01 -3.04371387e-01
1.23950207e+00 -2.81897366e-01 -1.40243983e+00 4.01150912e-01
-2.98541516e-01 -1.66567087e-01 4.99650896e-01 -5.15508473e-01
-1.23886216e+00 8.66749942e-01 1.34515238e+00 -3.77260894e-01
9.46835577e-01 2.90385578e-02 9.78090346e-01 2.80532241e-01
5.16203046e-01 -7.50806212e-01 -3.22032452e-01 6.01095021e-01
7.84219384e-01 -1.41931057e+00 1.20154932e-01 -2.55377740e-01
-5.85556030e-01 8.20045114e-01 1.12346399e+00 -4.61819053e-01
5.94148099e-01 7.10244954e-01 -2.09700363e-03 -2.46127293e-01
-6.05921328e-01 -5.51155150e-01 1.57829329e-01 5.30797422e-01
6.39110506e-01 3.78371865e-01 -2.16385350e-01 3.38519871e-01
-5.70459366e-01 1.93205491e-01 2.50660777e-01 9.94678855e-01
-6.74948335e-01 -4.64730144e-01 -3.96801353e-01 1.97446465e-01
-7.64539421e-01 7.84084052e-02 -2.27848470e-01 9.74033356e-01
1.31394016e-02 9.39310133e-01 4.29032177e-01 -4.07590508e-01
4.09398913e-01 -3.55013371e-01 4.00829613e-01 -5.68065584e-01
-5.73218107e-01 2.55756173e-02 -1.12853795e-01 -7.97443509e-01
1.46191657e-01 -3.90844077e-01 -1.64668941e+00 -9.09981072e-01
-3.04641694e-01 -1.81214839e-01 6.73549056e-01 4.59203571e-01
5.20546794e-01 2.46429518e-01 2.38613546e-01 -1.20569265e+00
-3.98256183e-01 -1.07369447e+00 -5.72378278e-01 -1.29392818e-01
5.10246933e-01 -7.05273986e-01 -2.19950736e-01 -2.57529646e-01] | [13.576828956604004, 0.3143201172351837] |
a87bfbb4-e957-47b7-bfb1-ed6fad99c94b | fsganv2-improved-subject-agnostic-face | 2202.12972 | null | https://arxiv.org/abs/2202.12972v1 | https://arxiv.org/pdf/2202.12972v1.pdf | FSGANv2: Improved Subject Agnostic Face Swapping and Reenactment | We present Face Swapping GAN (FSGAN) for face swapping and reenactment. Unlike previous work, we offer a subject agnostic swapping scheme that can be applied to pairs of faces without requiring training on those faces. We derive a novel iterative deep learning--based approach for face reenactment which adjusts significant pose and expression variations that can be applied to a single image or a video sequence. For video sequences, we introduce a continuous interpolation of the face views based on reenactment, Delaunay Triangulation, and barycentric coordinates. Occluded face regions are handled by a face completion network. Finally, we use a face blending network for seamless blending of the two faces while preserving the target skin color and lighting conditions. This network uses a novel Poisson blending loss combining Poisson optimization with a perceptual loss. We compare our approach to existing state-of-the-art systems and show our results to be both qualitatively and quantitatively superior. This work describes extensions of the FSGAN method, proposed in an earlier conference version of our work, as well as additional experiments and results. | ['Tal Hassner', 'Yosi Keller', 'Yuval Nirkin'] | 2022-02-25 | null | null | null | null | ['face-reenactment', 'facial-inpainting'] | ['computer-vision', 'computer-vision'] | [ 4.75982428e-01 2.95343429e-01 3.84090066e-01 -5.65226674e-01
-7.38802493e-01 -6.65022075e-01 5.83520889e-01 -5.72827160e-01
-2.12915719e-01 7.38219917e-01 -2.24823225e-02 1.23549551e-01
2.42509589e-01 -4.77087617e-01 -9.62082982e-01 -5.73093057e-01
1.09824747e-01 3.96859944e-01 -2.16607880e-02 -2.32263073e-01
-1.10661360e-02 9.25545752e-01 -1.54787743e+00 3.18906963e-01
5.12899637e-01 9.61427748e-01 -2.89720982e-01 7.02468455e-01
1.30180463e-01 2.32593551e-01 -3.79043400e-01 -7.49681294e-01
8.17102671e-01 -5.04993618e-01 -6.95612907e-01 3.89817029e-01
1.31417120e+00 -5.55627763e-01 -1.03857815e-01 6.72270358e-01
6.54566705e-01 1.09764047e-01 3.99183422e-01 -1.69524860e+00
-5.94100296e-01 3.80402606e-04 -9.76099074e-01 -4.26538736e-01
5.49872577e-01 9.08765011e-03 5.11114359e-01 -1.03407288e+00
7.95002759e-01 1.62377071e+00 9.25379097e-01 9.40794170e-01
-1.53707635e+00 -9.49948013e-01 8.12366791e-03 -2.07856581e-01
-1.43153679e+00 -8.59002054e-01 9.21791494e-01 -3.14467549e-01
5.53147793e-01 2.82004267e-01 7.14139938e-01 9.70080376e-01
1.67484537e-01 4.07613814e-01 1.16706395e+00 -5.54547429e-01
7.98725039e-02 5.04204892e-02 -6.65265203e-01 9.12685454e-01
-2.49510974e-01 -4.61459048e-02 -5.14779866e-01 -3.51072133e-01
1.11884487e+00 -1.39259258e-02 -3.68669659e-01 -7.70172417e-01
-7.03137636e-01 6.56401217e-01 4.45817769e-01 -8.58200863e-02
-2.09715754e-01 3.33622724e-01 2.21757248e-01 3.39705408e-01
6.95818663e-01 2.33441349e-02 -2.45854646e-01 3.61057073e-01
-1.11543000e+00 4.33874309e-01 7.55284488e-01 8.59738171e-01
8.68030071e-01 -1.80760473e-02 -1.72105163e-01 8.97428513e-01
2.97830790e-01 2.54928917e-01 -8.98707286e-02 -1.62775385e+00
-9.45943445e-02 1.78794771e-01 1.43938018e-02 -9.31130052e-01
-6.47749528e-02 1.27195165e-01 -6.77702248e-01 7.03846276e-01
1.15747131e-01 -2.38724992e-01 -1.08422244e+00 1.96230805e+00
7.61583209e-01 4.60811853e-01 -2.24880919e-01 6.14598632e-01
7.11458385e-01 2.72220373e-01 -1.80974290e-01 -2.89023370e-01
1.24631584e+00 -9.45827544e-01 -6.80516064e-01 1.19540930e-01
-6.14920259e-02 -1.03530133e+00 7.36884713e-01 3.86781126e-01
-1.51819777e+00 -3.01021963e-01 -8.32317829e-01 -3.06477070e-01
-2.71286778e-02 -1.35404512e-01 4.02313769e-01 9.51262057e-01
-1.89681244e+00 7.21193135e-01 -7.18726873e-01 -6.26555145e-01
6.35944366e-01 7.49702215e-01 -7.38576472e-01 -4.06938717e-02
-6.54796302e-01 6.28136933e-01 -1.65218577e-01 9.41479281e-02
-7.82579780e-01 -1.01000667e+00 -9.43416059e-01 -2.13803053e-01
2.49805793e-01 -9.66870129e-01 1.24525595e+00 -1.49940646e+00
-1.85155785e+00 1.14466596e+00 -3.54056329e-01 -1.25811443e-01
8.66328716e-01 -2.84587219e-02 -7.71041214e-02 1.88959822e-01
4.59723100e-02 1.25116563e+00 1.23498106e+00 -1.65684640e+00
-2.82946885e-01 -3.17308813e-01 -2.03553345e-02 4.02557909e-01
-1.21732637e-01 1.58365920e-01 -7.87438691e-01 -7.62512565e-01
-8.90353844e-02 -1.00896919e+00 3.52256373e-02 7.62222290e-01
-2.70033628e-01 2.07096219e-01 1.26772463e+00 -9.13838387e-01
5.24976909e-01 -2.23373866e+00 2.18266860e-01 2.42380083e-01
2.49795049e-01 1.45319834e-01 -4.16187942e-01 3.02494884e-01
-3.64193171e-01 -5.39565384e-02 -5.23916543e-01 -1.30680501e+00
-2.05939010e-01 1.13674872e-01 -1.26050651e-01 7.33244956e-01
1.10117793e-01 6.58625484e-01 -6.30618274e-01 -3.78803939e-01
2.80198365e-01 9.86421287e-01 -8.32007527e-01 3.96612287e-01
-6.95794150e-02 5.29648900e-01 3.91660124e-01 6.79403841e-01
1.34223068e+00 4.13843960e-01 2.37711206e-01 -3.72567952e-01
2.88133568e-04 -2.68790781e-01 -1.03335738e+00 1.75152719e+00
-5.39758146e-01 4.68083858e-01 6.25374138e-01 -3.62420440e-01
7.16280937e-01 3.29589784e-01 5.84402740e-01 -1.95009977e-01
7.67559092e-03 1.09225005e-01 -5.63298583e-01 -1.33531898e-01
2.44526818e-01 -2.92686194e-01 5.23231924e-01 4.07935679e-01
9.32875499e-02 -5.33532679e-01 -1.53336346e-01 8.34187046e-02
7.05517411e-01 4.20554757e-01 7.50521347e-02 -3.20413828e-01
4.37734514e-01 -7.08988547e-01 5.40766239e-01 1.90535992e-01
-1.10726438e-01 1.16502988e+00 6.44988000e-01 -4.37032759e-01
-1.10357332e+00 -1.18053102e+00 -1.41659468e-01 8.72542679e-01
-9.79953632e-02 -2.37303481e-01 -1.22230721e+00 -6.41235411e-01
-4.14374024e-02 3.60429972e-01 -8.44937980e-01 2.10728601e-01
-6.59292161e-01 -4.71574277e-01 4.50464994e-01 4.21161532e-01
6.58324003e-01 -9.11827862e-01 -2.68342495e-01 -1.05739072e-01
-4.84237932e-02 -9.71049309e-01 -9.57337916e-01 -2.19010830e-01
-6.32323265e-01 -1.12573230e+00 -8.89507353e-01 -1.04798007e+00
9.96862173e-01 1.76365554e-01 9.66424763e-01 3.12632084e-01
-3.16060454e-01 6.05417252e-01 1.15358771e-03 -9.65104401e-02
-3.25053066e-01 -1.80685028e-01 2.13764265e-01 3.02179456e-01
-1.74831435e-01 -8.87109160e-01 -7.82578230e-01 2.45721653e-01
-1.04533851e+00 1.13114297e-01 7.39748925e-02 6.84420228e-01
4.17382181e-01 -2.48560503e-01 1.48902416e-01 -9.12692726e-01
2.78257132e-01 -1.19753145e-01 -5.91256320e-01 2.09658608e-01
-3.35091650e-01 -3.10884118e-01 3.50034177e-01 -2.34904528e-01
-1.23272467e+00 3.60533178e-01 -2.42419302e-01 -8.36744785e-01
8.75521302e-02 -4.89193261e-01 -3.98801357e-01 -8.42377901e-01
3.46142024e-01 -2.42930502e-01 3.33076984e-01 -2.76453584e-01
5.74034274e-01 3.03774416e-01 7.12410450e-01 -5.42412817e-01
9.33872640e-01 9.69749033e-01 4.46874239e-02 -7.92896807e-01
-2.55908221e-01 -8.31689779e-03 -7.71637619e-01 -2.17808157e-01
8.30839396e-01 -7.04573989e-01 -8.04217517e-01 7.88900256e-01
-1.40067077e+00 -3.98141861e-01 -3.10967922e-01 -9.06471014e-02
-7.16195881e-01 4.18555439e-01 -4.22250897e-01 -6.59794748e-01
-2.73349196e-01 -1.02899027e+00 1.41313517e+00 2.81614453e-01
-1.91050947e-01 -1.04584551e+00 2.21415192e-01 3.22971702e-01
3.96719337e-01 5.98760188e-01 5.01547575e-01 -5.53046577e-02
-5.10452867e-01 2.17703089e-01 -2.32363626e-01 4.27848727e-01
3.87925267e-01 3.30866814e-01 -1.24867368e+00 -7.14418769e-01
6.25034794e-03 -2.51741886e-01 7.73182452e-01 5.49645364e-01
1.14222813e+00 -5.14575005e-01 -4.34909314e-01 1.08146083e+00
1.24896646e+00 2.20237523e-01 8.99469197e-01 -3.90408859e-02
8.81803513e-01 8.37886512e-01 6.67194426e-02 2.60935843e-01
3.10126364e-01 7.05793202e-01 5.03045440e-01 -5.21394372e-01
-2.93978721e-01 -2.98520833e-01 2.22259313e-01 1.64048746e-01
-2.68393103e-02 -2.85355777e-01 -4.06908035e-01 5.18804431e-01
-1.64825010e+00 -6.65393293e-01 2.92460442e-01 2.26283145e+00
6.11466348e-01 -4.89131242e-01 1.35496616e-01 -2.13945404e-01
9.14508045e-01 5.29821776e-02 -4.86353993e-01 -5.08293033e-01
-5.36044277e-02 5.04828513e-01 3.60760659e-01 9.54244316e-01
-1.00385916e+00 9.52899873e-01 6.95129395e+00 6.32868946e-01
-1.03030574e+00 2.73271620e-01 1.00160384e+00 -2.49613792e-01
-5.81555367e-01 -8.06845278e-02 -3.67606938e-01 9.31790024e-02
3.16756576e-01 2.77118832e-01 7.42804348e-01 4.73215580e-01
-3.76082137e-02 -3.34378034e-02 -1.24039257e+00 1.13072813e+00
5.86164594e-01 -1.28986585e+00 -2.17548888e-02 1.05942199e-02
1.00816250e+00 -4.24457490e-01 3.70805472e-01 -1.70885235e-01
2.52481788e-01 -1.18843687e+00 6.57524407e-01 4.67855930e-01
1.11906445e+00 -9.01244283e-01 3.09888899e-01 -6.01584435e-01
-1.15607786e+00 2.06894010e-01 -4.71870089e-03 3.03954273e-01
2.06625313e-01 2.85442531e-01 -6.64301991e-01 4.43298399e-01
7.75165975e-01 5.21952271e-01 -2.58747846e-01 7.37236798e-01
-7.03691542e-02 -6.68337420e-02 -3.59609187e-01 6.38092995e-01
-1.98594615e-01 -2.86719292e-01 6.69456422e-01 8.23368967e-01
3.87743831e-01 6.85880184e-02 -1.57683920e-02 8.22438002e-01
-5.09051859e-01 5.40747344e-02 -5.28786302e-01 5.14694631e-01
3.90243977e-01 1.52357888e+00 -6.93594515e-01 -4.85639572e-02
-2.09634542e-01 1.59162724e+00 2.78487384e-01 3.95138651e-01
-6.84935570e-01 -1.87641695e-01 9.50304747e-01 3.77941102e-01
4.03221160e-01 9.66982171e-02 -5.74591979e-02 -9.54715669e-01
1.85648516e-01 -7.64772594e-01 1.52307421e-01 -8.53297949e-01
-1.35540664e+00 7.17594028e-01 1.51996374e-01 -7.79282272e-01
-1.85781434e-01 -4.11404848e-01 -8.13974500e-01 8.64581406e-01
-1.35271168e+00 -1.55190134e+00 -3.20708960e-01 6.42975688e-01
3.61602306e-01 -1.35624349e-01 6.14520848e-01 2.65222162e-01
-3.89148891e-01 8.43745768e-01 -1.41670629e-01 -1.11213334e-01
9.87085998e-01 -1.02870727e+00 6.64474666e-01 7.14373767e-01
-2.28723198e-01 5.92584312e-01 6.15055680e-01 -6.37288690e-01
-1.14485192e+00 -1.26838756e+00 6.04404449e-01 -3.22687954e-01
7.01086968e-02 -6.67147338e-01 -7.72764921e-01 1.06999171e+00
6.87198341e-01 6.51097894e-02 4.52032059e-01 -2.02015504e-01
-3.85637820e-01 -3.39146823e-01 -1.80125153e+00 8.25721502e-01
1.13390267e+00 -3.72648060e-01 -1.80036705e-02 2.77798325e-01
6.05862617e-01 -6.79526567e-01 -4.53908741e-01 3.43947083e-01
8.34436893e-01 -1.18655694e+00 8.28517675e-01 -1.36505485e-01
1.71028495e-01 -4.61446017e-01 -9.22063440e-02 -1.32346082e+00
-1.28661782e-01 -1.28612769e+00 3.59606624e-01 1.45462906e+00
-3.70804332e-02 -7.36365378e-01 8.05383921e-01 7.68252969e-01
5.67915924e-02 -6.57530725e-01 -1.16525018e+00 -5.04079700e-01
3.62354144e-02 1.46207828e-02 7.17057884e-01 8.98777246e-01
-3.86847496e-01 -1.50654972e-01 -4.93438393e-01 1.28509495e-02
8.52591753e-01 -2.58466095e-01 8.89337003e-01 -8.02280366e-01
-8.15540403e-02 -1.90318167e-01 -3.50589395e-01 -5.95693469e-01
3.96935999e-01 -6.41978323e-01 -3.65209579e-02 -1.17722023e+00
2.08147109e-01 -2.98436970e-01 1.86137512e-01 7.29207933e-01
2.12536484e-01 9.46119964e-01 2.69366115e-01 -1.12193912e-01
6.74661398e-02 6.06226206e-01 1.09066343e+00 1.37078583e-01
-2.36354336e-01 -3.38718981e-01 -7.23615885e-01 7.10459054e-01
5.92043459e-01 -3.01905364e-01 -3.86678547e-01 -4.01809692e-01
-6.55869320e-02 -8.39108229e-02 5.99534214e-01 -8.68883669e-01
1.96305960e-02 -1.73733104e-02 6.10832989e-01 -1.71948418e-01
6.95957899e-01 -8.97238374e-01 6.55016541e-01 3.45470905e-01
-1.06348433e-01 1.67678535e-01 4.53761399e-01 3.01307708e-01
2.32686758e-01 3.04770563e-02 1.25544262e+00 -7.01866150e-02
-1.48338571e-01 5.95143080e-01 -1.04797915e-01 -2.29883805e-01
1.28277028e+00 -3.04685682e-01 6.35449961e-02 -6.29541755e-01
-5.83817899e-01 7.16923773e-02 1.04780495e+00 3.80782068e-01
7.84917057e-01 -1.46230102e+00 -8.80903721e-01 6.78021431e-01
-4.08502936e-01 -7.50662535e-02 3.71557325e-01 6.82687759e-01
-9.10485148e-01 -2.25504905e-01 -4.78101701e-01 -5.54174900e-01
-1.69347763e+00 3.18159401e-01 6.99145257e-01 2.82452554e-01
-4.22105372e-01 1.05370820e+00 5.46492279e-01 -6.38936162e-01
4.63492349e-02 7.12064728e-02 2.24484891e-01 -8.71563628e-02
4.69259501e-01 3.48463565e-01 1.28016889e-01 -7.13030636e-01
-5.38786471e-01 8.29194307e-01 -2.07677379e-01 -4.69357401e-01
1.25832248e+00 -2.35553622e-01 -5.15799880e-01 3.40915844e-02
1.39918494e+00 2.01812163e-01 -1.67725027e+00 4.04496938e-02
-8.99886012e-01 -9.00886118e-01 -1.35279328e-01 -5.86722493e-01
-1.50756133e+00 5.07592559e-01 7.36961782e-01 -2.76829213e-01
1.43557417e+00 -1.09857805e-01 8.18288922e-01 -2.61839151e-01
4.46055792e-02 -8.56588483e-01 9.51490626e-02 1.91029862e-01
1.13977361e+00 -9.47075784e-01 3.12032811e-02 -6.94284976e-01
-3.91449034e-01 1.06762612e+00 6.87483072e-01 -2.46719569e-01
6.99141443e-01 5.41442871e-01 1.39273092e-01 3.70914117e-02
-3.73938352e-01 3.52074236e-01 2.43083283e-01 8.62795532e-01
3.73319685e-01 -2.82915920e-01 -3.76558751e-02 -1.70894936e-01
-2.38337010e-01 -7.09804893e-02 4.69005436e-01 9.11182284e-01
1.08427994e-01 -1.28954029e+00 -5.83767951e-01 1.69759825e-01
-3.07396919e-01 -1.11997284e-01 -4.70216721e-01 6.26894116e-01
4.34981287e-01 6.00809872e-01 4.63055521e-01 -1.11088336e-01
1.50031909e-01 -6.66091517e-02 1.07615483e+00 -4.99366522e-01
-5.24069190e-01 2.00224862e-01 -1.53167695e-01 -6.61848783e-01
-6.25319958e-01 -7.65982211e-01 -9.10956979e-01 -6.50394976e-01
-3.50182280e-02 -3.10609549e-01 5.06886601e-01 7.02082157e-01
4.52437550e-01 1.94104999e-01 8.78574073e-01 -1.37499380e+00
1.24631219e-01 -4.55521643e-01 -6.71570182e-01 2.32694402e-01
6.22224927e-01 -5.22199929e-01 -3.39848548e-01 3.23406041e-01] | [12.750121116638184, -0.1631173938512802] |
68834f2f-f0e5-416c-93ba-ccbdb339d35c | one-pass-incomplete-multi-view-clustering | 1903.00637 | null | http://arxiv.org/abs/1903.00637v1 | http://arxiv.org/pdf/1903.00637v1.pdf | One-Pass Incomplete Multi-view Clustering | Real data are often with multiple modalities or from multiple heterogeneous
sources, thus forming so-called multi-view data, which receives more and more
attentions in machine learning. Multi-view clustering (MVC) becomes its
important paradigm. In real-world applications, some views often suffer from
instances missing. Clustering on such multi-view datasets is called incomplete
multi-view clustering (IMC) and quite challenging. To date, though many
approaches have been developed, most of them are offline and have high
computational and memory costs especially for large scale datasets. To address
this problem, in this paper, we propose an One-Pass Incomplete Multi-view
Clustering framework (OPIMC). With the help of regularized matrix factorization
and weighted matrix factorization, OPIMC can relatively easily deal with such
problem. Different from the existing and sole online IMC method, OPIMC can
directly get clustering results and effectively determine the termination of
iteration process by introducing two global statistics. Finally, extensive
experiments conducted on four real datasets demonstrate the efficiency and
effectiveness of the proposed OPIMC method. | ['Songcan Chen', 'Menglei Hu'] | 2019-03-02 | null | null | null | null | ['incomplete-multi-view-clustering'] | ['computer-vision'] | [-1.10206731e-01 -6.05133533e-01 -1.93502620e-01 -1.61501259e-01
-9.26826596e-01 -5.39898753e-01 3.81952733e-01 -1.36067988e-02
-1.41205579e-01 3.43697131e-01 2.45546594e-01 8.63005891e-02
-3.06866527e-01 -4.59373772e-01 -2.43899256e-01 -8.27463150e-01
2.46133909e-01 5.65659404e-01 1.28783777e-01 2.49086782e-01
1.89564973e-01 -5.43003790e-02 -1.47643018e+00 4.52986479e-01
1.05385172e+00 6.76072359e-01 3.51998746e-01 3.92910779e-01
-1.97598800e-01 5.35370111e-01 -1.42852098e-01 -3.88675392e-01
7.47459605e-02 -4.35515672e-01 -6.76591575e-01 5.91163337e-01
-3.43094856e-01 -4.82550897e-02 1.29389152e-01 1.09104276e+00
5.36484361e-01 2.20823064e-01 3.82736027e-01 -1.43658578e+00
-3.18310440e-01 5.36505282e-01 -1.12856102e+00 -8.38808790e-02
2.75741547e-01 -3.77166897e-01 7.58204997e-01 -1.14638865e+00
5.37169456e-01 1.24902368e+00 3.08525681e-01 2.61002570e-01
-9.45398808e-01 -3.56621116e-01 3.71773720e-01 5.02307475e-01
-1.44663918e+00 -2.98166662e-01 8.98942411e-01 -3.76761943e-01
6.54456764e-02 3.57381642e-01 3.61590356e-01 6.73774302e-01
-4.37844634e-01 1.10653532e+00 1.47727799e+00 -1.16953544e-01
2.54825056e-01 1.71822235e-01 -1.21969603e-01 2.14473993e-01
2.01141953e-01 -5.47899663e-01 -2.20965132e-01 -3.34640384e-01
4.32606876e-01 5.83863199e-01 -4.67896014e-01 -5.16745090e-01
-1.40196645e+00 6.12886429e-01 -2.54158508e-02 3.28572273e-01
-3.59840930e-01 -3.69186252e-01 5.60537159e-01 6.26956820e-02
3.46076161e-01 -3.22647661e-01 -2.54638374e-01 -2.58824378e-01
-6.97581589e-01 -7.96463042e-02 6.05611980e-01 9.96925950e-01
7.49182999e-01 -2.97703505e-01 4.37941611e-01 1.11331594e+00
2.92263359e-01 5.01547873e-01 5.38071275e-01 -8.87512207e-01
7.56962121e-01 1.04470801e+00 5.82387447e-02 -1.30749595e+00
-3.01441312e-01 -2.76470065e-01 -1.24869120e+00 -2.75816679e-01
2.14289784e-01 -1.27340451e-01 -4.48708534e-01 1.36986136e+00
8.60485554e-01 2.38899589e-01 7.15234131e-02 1.11048913e+00
8.18714082e-01 7.98089385e-01 -1.81713834e-01 -7.90995419e-01
1.22390330e+00 -8.35193574e-01 -7.75950730e-01 1.02963999e-01
3.81436646e-01 -9.68701720e-01 8.03622484e-01 6.47464633e-01
-8.79229188e-01 -4.22710270e-01 -9.26136911e-01 2.71815628e-01
-2.21460536e-01 1.88843250e-01 5.61336935e-01 3.95579129e-01
-5.52844465e-01 1.24916948e-01 -8.84306729e-01 -2.86151946e-01
1.69542104e-01 2.22104415e-01 -6.15576744e-01 -7.11877704e-01
-7.84619808e-01 3.62378865e-01 5.16739249e-01 3.64668101e-01
-6.07280970e-01 -3.20839375e-01 -4.36411440e-01 -1.40846103e-01
9.51534808e-01 -3.17555189e-01 7.73819208e-01 -9.07521367e-01
-1.11963868e+00 4.98498529e-01 -3.39674264e-01 2.24746600e-01
3.72584492e-01 -3.75527553e-02 -7.50300765e-01 2.46926486e-01
2.60608435e-01 4.62942757e-02 7.60232985e-01 -1.83328021e+00
-6.58551514e-01 -8.46366048e-01 -7.52956942e-02 5.36497474e-01
-2.98150152e-01 3.68700027e-02 -1.06567430e+00 -3.84334594e-01
6.32120252e-01 -9.48873281e-01 -2.74871439e-01 -4.26548511e-01
-3.56623948e-01 -1.43874303e-01 1.11424422e+00 -6.58960521e-01
1.36486661e+00 -2.17411113e+00 5.90716183e-01 2.35126376e-01
3.69223505e-01 2.09481210e-01 9.92013812e-02 6.80411816e-01
-2.91269161e-02 2.68201083e-02 -3.27434957e-01 -3.13468188e-01
-2.40953416e-01 2.95367956e-01 2.02552333e-01 6.31778300e-01
-3.99793833e-01 4.30165201e-01 -7.43429184e-01 -8.20225000e-01
2.47714654e-01 2.89316326e-01 -4.26918209e-01 1.70835063e-01
7.49896979e-03 6.60063505e-01 -6.49207115e-01 8.68303597e-01
1.10875475e+00 -6.46429658e-01 4.69922930e-01 -2.76222199e-01
-5.45043945e-02 -7.90346622e-01 -1.71840632e+00 1.76476288e+00
-3.85623485e-01 -5.25522754e-02 4.82464314e-01 -1.15787244e+00
3.73545885e-01 5.64986527e-01 7.64334679e-01 -2.86926210e-01
1.85195670e-01 2.83850342e-01 -1.30377039e-01 -6.46621466e-01
3.28350842e-01 -2.08905920e-01 2.79259011e-02 6.87866271e-01
-1.69151127e-01 4.44346309e-01 2.43504807e-01 4.84330416e-01
5.83715796e-01 -2.44056433e-02 3.67454916e-01 -6.68543056e-02
1.01701760e+00 -2.21249997e-03 9.39328015e-01 7.01574013e-02
-1.90726668e-01 8.57828319e-01 3.42567384e-01 -2.60465562e-01
-6.24576032e-01 -7.16150165e-01 1.23972550e-01 7.30327249e-01
4.14613873e-01 -7.35871851e-01 -6.91325366e-01 -6.13698840e-01
-2.22255364e-01 -7.56425187e-02 -3.71025801e-01 2.50046402e-01
-2.70053208e-01 -8.68253529e-01 -6.87953904e-02 2.78066158e-01
5.34631908e-01 -5.60980797e-01 -1.32148892e-01 2.67695785e-01
-7.52011180e-01 -1.30065799e+00 -5.63041151e-01 -3.26338083e-01
-9.21345294e-01 -1.52527571e+00 -7.97598898e-01 -5.51582873e-01
8.21095943e-01 9.62994099e-01 7.52101600e-01 1.76567852e-01
-1.75969288e-01 5.44046283e-01 -7.82687426e-01 -5.64426966e-02
-7.21020922e-02 -6.51781727e-03 1.92615956e-01 6.55701101e-01
2.83198833e-01 -6.38194680e-01 -7.22036302e-01 5.75285614e-01
-1.35132015e+00 2.55854309e-01 5.74821770e-01 6.68107986e-01
8.95197868e-01 5.41765451e-01 7.50664413e-01 -1.23164773e+00
3.84296447e-01 -6.98949456e-01 -5.36649704e-01 5.23616374e-01
-5.68020999e-01 -2.97315240e-01 8.76097381e-01 -1.50905311e-01
-1.26215363e+00 1.00345291e-01 1.89300552e-01 -8.47534418e-01
-1.16463043e-01 8.69454801e-01 -7.63817072e-01 2.70646870e-01
7.97748286e-03 4.17391896e-01 5.90448678e-02 -6.15177512e-01
4.63796139e-01 7.82882452e-01 3.05860519e-01 -5.31492472e-01
8.28560650e-01 8.00673783e-01 -1.56765953e-01 -6.04502559e-01
-7.24792600e-01 -8.73956800e-01 -6.76707983e-01 -4.77952659e-01
9.40094292e-01 -1.11377347e+00 -7.85614133e-01 4.73607689e-01
-7.21859038e-01 4.10982221e-01 3.84556264e-01 6.22205019e-01
-1.30579919e-01 6.71122372e-01 -3.32756639e-01 -8.46857429e-01
-9.83477384e-02 -1.18359149e+00 8.18339527e-01 2.46469289e-01
4.85596240e-01 -1.07626593e+00 -5.01758754e-02 8.13657105e-01
6.00902326e-02 3.97077471e-01 8.64463866e-01 -3.03587317e-01
-5.35232067e-01 -1.73455849e-01 -2.66157746e-01 1.17034361e-01
3.37995648e-01 -6.66996837e-02 -7.40603209e-01 -4.78645444e-01
2.24068329e-01 -1.53895974e-01 5.42893589e-01 2.00609401e-01
1.26987815e+00 -2.59787552e-02 -3.58308524e-01 4.47739452e-01
1.81404662e+00 2.37079814e-01 2.45201677e-01 7.56596923e-02
9.97473657e-01 6.34885073e-01 8.84146869e-01 9.11168039e-01
5.97910881e-01 4.18384612e-01 5.20348072e-01 8.72174203e-02
5.36355138e-01 -1.64790422e-01 8.75274464e-02 1.69077778e+00
-2.76642978e-01 -1.32913798e-01 -8.11798692e-01 4.97634113e-01
-2.24202633e+00 -1.06109297e+00 -4.75106537e-01 2.20673823e+00
4.25071120e-01 -2.66935706e-01 2.00276151e-01 3.10870856e-01
1.04236734e+00 2.02771887e-01 -6.61843717e-01 2.02209041e-01
-8.89683515e-02 -4.79292482e-01 1.30271912e-01 4.81575802e-02
-1.06223071e+00 3.57753068e-01 4.72058344e+00 9.53921735e-01
-6.38736725e-01 3.84618610e-01 6.09371006e-01 2.79077585e-03
-4.28048968e-01 2.52237111e-01 -4.39243108e-01 6.33690536e-01
4.72142041e-01 -3.33725102e-02 6.16452515e-01 7.81402469e-01
3.61899763e-01 -3.64919752e-01 -7.74045587e-01 1.56370628e+00
2.30858967e-01 -8.61335635e-01 -3.24580558e-02 2.26112217e-01
9.97435927e-01 -3.36884260e-01 -1.74815655e-01 2.31361002e-01
5.58430552e-02 -5.29954076e-01 1.56710804e-01 2.62234032e-01
7.21539259e-01 -1.15569687e+00 6.39313936e-01 6.89419210e-01
-1.66863835e+00 -7.14413300e-02 -4.47570175e-01 2.21302837e-01
2.55260408e-01 8.21500063e-01 -5.76559380e-02 1.30647743e+00
6.31540000e-01 8.99464011e-01 -6.35680735e-01 8.79933774e-01
2.61531562e-01 3.95033896e-01 -1.97264597e-01 3.44417334e-01
1.13442622e-01 -6.39022470e-01 3.40920806e-01 5.47565520e-01
3.23987514e-01 5.18014848e-01 5.70497155e-01 2.91345865e-01
3.79381701e-02 4.21545327e-01 -4.58168834e-01 5.14135398e-02
4.27623838e-01 1.54596889e+00 -9.51202691e-01 -3.71914387e-01
-8.37258697e-01 1.00398362e+00 2.13038594e-01 3.00722063e-01
-8.77671480e-01 -2.52567865e-02 1.51548833e-01 -1.90313101e-01
1.59267306e-01 -1.40315980e-01 -1.17439618e-02 -1.69516289e+00
3.03864390e-01 -1.09446919e+00 6.65740073e-01 -4.78724808e-01
-1.35563421e+00 3.18853587e-01 -6.79269359e-02 -1.82262671e+00
1.61159918e-01 -1.90888941e-01 -4.22062755e-01 4.63532567e-01
-1.24800289e+00 -1.08647072e+00 -4.83159453e-01 1.07175136e+00
6.61827624e-01 -2.84188557e-02 5.41663051e-01 6.09746575e-01
-8.73413444e-01 2.17569917e-01 5.78832567e-01 -7.29454905e-02
7.72560775e-01 -1.09972692e+00 -4.25041258e-01 8.81742775e-01
-2.71717533e-02 6.23363674e-01 2.59383440e-01 -4.85902518e-01
-1.99034679e+00 -1.00468898e+00 3.40069860e-01 -1.35430902e-01
3.78945887e-01 -1.23770975e-01 -8.73976469e-01 4.65971351e-01
2.57634193e-01 7.80652091e-02 1.08659279e+00 1.63308382e-01
-1.22336596e-01 -2.14025766e-01 -9.48624551e-01 3.64423037e-01
8.67281199e-01 -3.57352227e-01 -1.92672893e-01 4.12288636e-01
3.45036358e-01 -2.84101576e-01 -1.08261013e+00 4.06742126e-01
3.13938737e-01 -1.20997608e+00 7.98186958e-01 -3.30924988e-01
3.87845993e-01 -7.54548550e-01 -3.72662097e-01 -1.25813901e+00
-2.16483578e-01 -2.20304161e-01 -1.60836905e-01 1.71319461e+00
2.92501729e-02 -4.72085834e-01 6.90613806e-01 5.15757799e-01
1.22749873e-01 -6.40262604e-01 -7.26162672e-01 -4.84737247e-01
-3.07465732e-01 -3.59287351e-01 5.76363683e-01 1.15543914e+00
7.81823620e-02 2.81820863e-01 -5.96641481e-01 3.00665081e-01
8.94363642e-01 7.10341811e-01 7.93229342e-01 -1.29961252e+00
-3.60262990e-01 2.63676792e-02 -1.74573958e-01 -6.69727504e-01
-1.13995373e-01 -7.62887836e-01 -2.42944092e-01 -1.77606058e+00
6.98267639e-01 -4.23556209e-01 -1.84923217e-01 5.70308864e-02
-5.31132817e-01 -5.69908172e-02 2.60212660e-01 7.05281556e-01
-1.10239553e+00 5.68711400e-01 1.22448266e+00 7.91165531e-02
-3.09164394e-02 4.01969217e-02 -6.36478066e-01 7.41233349e-01
5.86857796e-01 -3.07156742e-01 -7.51330972e-01 -4.55102712e-01
1.25871152e-01 6.36734724e-01 -6.10198714e-02 -8.13955605e-01
3.96077007e-01 -3.19923848e-01 2.76074111e-01 -1.00715637e+00
2.10127920e-01 -1.17722178e+00 5.12169361e-01 1.70871377e-01
2.64792114e-01 2.99633533e-01 -3.00723821e-01 1.06863999e+00
-6.93203449e-01 6.17961474e-02 7.02990413e-01 -4.95813191e-01
-8.40618074e-01 5.34786165e-01 -2.08608657e-01 2.51565605e-01
1.28173578e+00 4.88593020e-02 -1.45661339e-01 -3.34941328e-01
-9.29230154e-01 7.68767059e-01 5.87883472e-01 3.46000791e-01
7.20191717e-01 -1.70431304e+00 -5.66541016e-01 -7.76029825e-02
3.17612201e-01 2.40535453e-01 9.06997740e-01 1.16651523e+00
-2.34492630e-01 1.37878194e-01 2.32558489e-01 -6.85197413e-01
-1.34936523e+00 1.00013638e+00 -2.43202031e-01 -2.33006895e-01
-3.81136566e-01 1.75578326e-01 1.31525382e-01 -5.39846241e-01
5.23552671e-03 3.93989533e-01 -2.96570659e-01 3.85481089e-01
4.80954617e-01 5.89520454e-01 -1.59989089e-01 -8.12677741e-01
-3.39683145e-01 6.68439925e-01 1.88528430e-02 -5.54639325e-02
1.34191072e+00 -5.69431543e-01 -4.90387589e-01 6.36060655e-01
1.39565372e+00 -1.23738863e-01 -8.81930470e-01 -4.01062518e-01
-1.34815738e-01 -5.54225862e-01 -2.78829634e-01 -3.54877561e-01
-1.45125735e+00 8.82445037e-01 4.09402460e-01 1.56089604e-01
1.26912832e+00 -1.98054329e-01 8.47277045e-01 1.21772148e-01
5.51667690e-01 -1.38557720e+00 -1.10368110e-01 -3.74413989e-02
4.91778940e-01 -1.55242181e+00 2.22615033e-01 -6.02561653e-01
-9.55588758e-01 8.53543878e-01 7.66058922e-01 1.66368037e-01
8.12429726e-01 -1.03971526e-01 3.09064910e-02 -2.43301004e-01
-7.58170784e-01 -1.21435113e-01 -4.33528423e-02 2.46652082e-01
1.29777610e-01 1.49519265e-01 -4.56321329e-01 7.18854845e-01
4.55541641e-01 -1.51597247e-01 4.88705933e-01 1.07316244e+00
-2.75762022e-01 -1.33958423e+00 -6.99248433e-01 4.98504698e-01
-4.93961275e-01 2.98303485e-01 -1.33714110e-01 6.43887877e-01
7.55650625e-02 1.30281699e+00 -4.42729443e-01 -4.94424760e-01
8.29412118e-02 -9.51437429e-02 5.97748235e-02 -3.72303128e-01
-2.80502617e-01 5.03948450e-01 -3.59929651e-01 -4.27186906e-01
-1.03473747e+00 -9.87850428e-01 -1.15527487e+00 -2.61767894e-01
-5.28957784e-01 4.68483955e-01 4.87388730e-01 8.72399271e-01
3.98609638e-01 2.16192946e-01 1.10111046e+00 -6.48636341e-01
-3.22124869e-01 -6.23346925e-01 -8.73620749e-01 7.39006698e-01
-9.34206620e-02 -5.75447917e-01 -2.38606885e-01 1.46937594e-01] | [8.270694732666016, 4.635233402252197] |
0c5fac6b-b2ae-4e3e-ab63-fa054797804a | semantic-image-segmentation-with-deep-1 | 2210.13296 | null | https://arxiv.org/abs/2210.13296v1 | https://arxiv.org/pdf/2210.13296v1.pdf | Semantic Image Segmentation with Deep Learning for Vine Leaf Phenotyping | Plant phenotyping refers to a quantitative description of the plants properties, however in image-based phenotyping analysis, our focus is primarily on the plants anatomical, ontogenetical and physiological properties.This technique reinforced by the success of Deep Learning in the field of image based analysis is applicable to a wide range of research areas making high-throughput screens of plants possible, reducing the time and effort needed for phenotypic characterization.In this study, we use Deep Learning methods (supervised and unsupervised learning based approaches) to semantically segment grapevine leaves images in order to develop an automated object detection (through segmentation) system for leaf phenotyping which will yield information regarding their structure and function.In these directions we studied several deep learning approaches with promising results as well as we reported some future challenging tasks in the area of precision agriculture.Our work contributes to plant lifecycle monitoring through which dynamic traits such as growth and development can be captured and quantified, targeted intervention and selective application of agrochemicals and grapevine variety identification which are key prerequisites in sustainable agriculture. | ['Nestoras C. Tsirliganis', 'George Ioannakis', 'Alexandra D. Solomou', 'Chairi Kiourt', 'Petros N. Tamvakis'] | 2022-10-24 | null | null | null | null | ['plant-phenotyping'] | ['computer-vision'] | [ 4.68479276e-01 -1.12715609e-01 -1.08798377e-01 -1.26799671e-02
-8.64375606e-02 -1.12980056e+00 8.70088488e-02 8.37332249e-01
1.85498714e-01 3.57901841e-01 -6.52326107e-01 -5.50521612e-01
-4.45374817e-01 -1.14715481e+00 -1.99707106e-01 -9.49840784e-01
-2.20529199e-01 5.06463885e-01 4.37405407e-02 -2.41421863e-01
8.78578424e-02 1.04617882e+00 -1.61993134e+00 1.44642919e-01
8.48509133e-01 9.39081967e-01 1.00141978e+00 7.99623191e-01
-1.71985283e-01 2.30410561e-01 -3.57979953e-01 2.62147576e-01
1.46392006e-02 -2.62932807e-01 -7.33962059e-01 3.55215251e-01
2.42514595e-01 -2.21214160e-01 4.60358322e-01 1.05175495e+00
4.63561296e-01 -4.17309105e-01 6.00288570e-01 -8.32617939e-01
-7.57496178e-01 6.42699897e-01 -8.02972674e-01 -1.24885380e-01
-1.71304762e-01 2.90276855e-01 7.26457834e-01 -3.79207104e-01
5.99222600e-01 1.00837171e+00 4.71587867e-01 -7.70012587e-02
-1.52142155e+00 -1.73209086e-01 -6.07026964e-02 3.48546505e-01
-9.17856157e-01 -1.09834462e-01 3.39253902e-01 -6.00477993e-01
3.24441075e-01 2.83080041e-01 7.84551740e-01 3.85996550e-01
2.30625585e-01 7.89609969e-01 8.54024351e-01 -4.44369525e-01
2.46798977e-01 -8.47739652e-02 5.75149581e-02 6.46668434e-01
3.94880295e-01 2.94815361e-01 2.62773365e-01 3.26707095e-01
7.71372378e-01 -3.85081992e-02 -9.43246409e-02 -7.00973451e-01
-8.12480927e-01 7.10996032e-01 5.53450346e-01 6.10150695e-01
-7.18895972e-01 -3.56272571e-02 5.00997066e-01 3.06725707e-02
1.01468302e-02 6.56333506e-01 -1.06860113e+00 2.84947693e-01
-9.37356770e-01 -2.99946475e-03 6.63799465e-01 6.97055042e-01
5.90575933e-01 2.61979163e-01 -4.77766991e-01 7.93668151e-01
4.28634107e-01 6.13868535e-01 1.45427957e-01 -1.22125661e+00
-6.81190431e-01 9.56572175e-01 8.36878493e-02 -8.14335644e-01
-5.75736403e-01 -1.72673792e-01 -7.35803246e-01 6.48179770e-01
5.80518425e-01 -3.89247872e-02 -9.83078420e-01 1.26190126e+00
2.17105523e-01 -4.18640494e-01 -1.48157835e-01 6.30465686e-01
8.94769251e-01 5.02327442e-01 5.39410889e-01 -5.42788394e-02
1.74330163e+00 -3.01309347e-01 -4.34200019e-01 6.45980984e-02
9.25493479e-01 -7.92758346e-01 9.79076028e-01 5.99325359e-01
-8.79878819e-01 -5.25852859e-01 -1.06763911e+00 3.06410223e-01
-8.61459374e-01 8.53103697e-01 9.48468804e-01 9.20184731e-01
-8.07565928e-01 8.88211429e-01 -8.72997165e-01 -9.56346929e-01
9.45338190e-01 5.37182868e-01 -5.41680932e-01 1.17686149e-02
-3.42320114e-01 9.60012197e-01 6.06769025e-01 4.97934878e-01
-1.34973264e+00 -6.81815565e-01 -5.62644243e-01 4.95480329e-01
3.14923078e-01 -1.30304068e-01 1.00128460e+00 -7.04974234e-01
-1.81865168e+00 1.33733737e+00 2.07528949e-01 -4.09289390e-01
-1.37489006e-01 9.00310725e-02 8.71964768e-02 5.66324554e-02
4.66341386e-03 8.61367762e-01 3.65100622e-01 -1.24017191e+00
-6.09444737e-01 -8.91817391e-01 -1.72709689e-01 -3.93640190e-01
-2.77679771e-01 -3.80411297e-02 3.20100605e-01 -7.85560086e-02
3.33774924e-01 -9.06837404e-01 -2.50951916e-01 4.28983927e-01
-2.90784359e-01 3.90276909e-01 1.44876504e+00 -7.59968281e-01
1.79593593e-01 -1.68733561e+00 1.15019657e-01 -1.64813533e-01
1.42396748e-01 1.11907399e+00 -3.63215625e-01 3.86900902e-01
-1.20714679e-02 3.39896500e-01 -3.29634666e-01 4.18887198e-01
-3.96356195e-01 6.64438531e-02 2.07729608e-01 4.09460783e-01
6.30736887e-01 9.54769611e-01 -8.24473679e-01 -1.72597051e-01
8.14967275e-01 3.18206936e-01 2.29805753e-01 1.76166445e-01
-3.01601976e-01 5.34312069e-01 -4.07757550e-01 1.23744369e+00
1.13618481e+00 1.93311602e-01 3.87043923e-01 -1.86390638e-01
-6.25172794e-01 -4.70094442e-01 -5.74451506e-01 1.22169340e+00
-4.47121322e-01 6.45123124e-01 5.03599524e-01 -1.43385231e+00
1.23854423e+00 2.29329810e-01 6.50557697e-01 -2.61671811e-01
2.96269804e-01 -1.50143588e-02 1.74114823e-01 -4.92877930e-01
-1.94165111e-01 3.63971949e-01 6.15918875e-01 -1.54828578e-01
3.37568074e-01 -3.67816091e-01 5.52752674e-01 -1.94496229e-01
8.51396263e-01 5.12265205e-01 2.46819243e-01 -7.29812086e-01
5.98537087e-01 5.47662735e-01 4.25016791e-01 2.64036775e-01
-6.78315163e-01 2.49798119e-01 7.67336786e-01 -3.08928043e-01
-1.07240438e+00 -7.84370661e-01 -3.30081165e-01 9.45625842e-01
-1.38754696e-01 4.95293438e-02 -7.10162103e-01 -3.18269730e-01
7.79764578e-02 5.13229609e-01 -6.05039656e-01 -2.76925415e-01
-7.83273056e-02 -1.12478328e+00 4.82155055e-01 5.41728199e-01
6.47613585e-01 -1.55739331e+00 -8.01380813e-01 3.25215310e-01
3.71711850e-01 -1.05666280e+00 7.84096777e-01 7.60543764e-01
-1.03427696e+00 -1.29035342e+00 -8.40671539e-01 -9.65241671e-01
2.81209469e-01 2.41867691e-01 1.02434802e+00 -1.02934740e-01
-1.04500771e+00 -2.64094863e-02 -5.42056024e-01 -8.67071629e-01
-5.46306789e-01 1.63857907e-01 -6.54326618e-01 -3.85522425e-01
4.56855446e-01 -6.92767262e-01 -6.23706579e-01 4.46190611e-02
-6.24445438e-01 -3.41643691e-01 9.38471079e-01 7.29594111e-01
7.75814354e-01 9.13546886e-03 3.64390820e-01 -1.17251003e+00
2.17426896e-01 -1.35365307e-01 -1.25067568e+00 6.83125556e-01
-4.34765249e-01 -3.99878383e-01 6.37478113e-01 -9.62458998e-02
-8.88415515e-01 3.20131332e-01 2.24135891e-01 3.43408525e-01
-8.93879116e-01 8.94142091e-01 -6.33633137e-01 -3.79634321e-01
6.40786767e-01 4.61255088e-02 1.73365965e-01 -3.55610013e-01
2.92469651e-01 3.21972787e-01 4.64801520e-01 -3.05191338e-01
6.47649527e-01 3.19589943e-01 5.99873543e-01 -1.47756505e+00
-5.06648242e-01 -5.72294950e-01 -1.22398615e+00 -2.37237558e-01
9.36167717e-01 -4.87213850e-01 -1.40048087e+00 6.35265172e-01
-8.79862964e-01 -4.58850682e-01 -2.46666789e-01 3.52597475e-01
-4.74921674e-01 4.32831496e-01 -4.75370258e-01 -8.43764901e-01
-5.72308064e-01 -1.03867865e+00 1.08985817e+00 5.90151787e-01
6.22894280e-02 -9.83432233e-01 2.15912033e-02 4.81864884e-02
3.67103696e-01 5.23678780e-01 1.09734082e+00 -7.65478536e-02
-4.25660789e-01 -4.61688966e-01 -8.21330905e-01 3.85854781e-01
3.03996027e-01 8.08416426e-01 -1.04128718e+00 3.02044060e-02
-3.84354085e-01 -4.72941846e-01 7.05654740e-01 1.10047054e+00
1.28642416e+00 5.28030276e-01 -3.88151109e-01 6.31368160e-01
1.72206271e+00 4.44117576e-01 7.82762468e-01 3.39617193e-01
7.18750894e-01 9.91244018e-01 9.05144453e-01 4.97236758e-01
-3.47940117e-01 3.62163723e-01 9.60623324e-01 -4.57867920e-01
-1.09876640e-01 3.51521939e-01 -1.12733999e-02 -1.10143013e-01
-2.42817655e-01 -4.65456665e-01 -9.59869742e-01 5.58219552e-01
-1.58252561e+00 -9.94674087e-01 -7.37312555e-01 2.14011693e+00
4.66986328e-01 -2.70658672e-01 -1.21610221e-02 3.16469610e-01
6.98575616e-01 -2.84776628e-01 -6.15726113e-01 -6.25483036e-01
-5.33045292e-01 4.68516469e-01 8.97635877e-01 8.99545848e-02
-1.32063127e+00 1.33449221e+00 6.60306549e+00 2.25954831e-01
-1.32572734e+00 -3.60629201e-01 7.30356336e-01 7.23994732e-01
2.71319240e-01 3.30550343e-01 -6.14599049e-01 -1.95664942e-01
7.52320945e-01 3.89623970e-01 3.10492396e-01 7.36182272e-01
6.39303386e-01 -4.25272673e-01 -6.16910636e-01 1.66105285e-01
-5.84730506e-01 -1.17597854e+00 -2.91703165e-01 1.01643719e-01
4.00107712e-01 -1.79917142e-01 -1.57105953e-01 -1.15559669e-02
4.17409807e-01 -8.99122000e-01 3.01735792e-02 1.48990273e-01
6.69621229e-01 -4.14460391e-01 7.27011800e-01 5.89358881e-02
-1.25244784e+00 -1.99928105e-01 -7.26554155e-01 1.61638290e-01
-2.25967169e-01 7.25022972e-01 -1.00353920e+00 5.77055275e-01
7.17355371e-01 7.09032297e-01 -8.71106744e-01 1.22779953e+00
-2.03075245e-01 9.22642589e-01 -2.26567388e-01 1.59235373e-01
1.56292871e-01 -4.59164411e-01 1.28376469e-01 9.80417907e-01
3.21757436e-01 -4.34611738e-01 3.16289663e-01 1.22637880e+00
3.03479224e-01 1.90249816e-01 -4.39930558e-01 -7.00636625e-01
2.08654940e-01 1.79151797e+00 -1.54361629e+00 2.69913375e-02
-1.67238098e-02 8.29939187e-01 -1.25794023e-01 -2.32529473e-02
-3.03251237e-01 -4.19167548e-01 2.34546423e-01 3.52965556e-02
5.16592085e-01 -3.25816095e-01 -4.88502979e-01 -5.22705197e-01
-5.68678677e-01 -5.24735451e-01 4.93007265e-02 -7.52255499e-01
-7.48547614e-01 -1.13013163e-01 -1.01386912e-01 -6.12624705e-01
3.22896808e-01 -1.23987603e+00 -7.92104840e-01 9.71169293e-01
-1.34604895e+00 -1.61646652e+00 -8.87854040e-01 -1.61471665e-01
4.34882820e-01 -1.98382989e-01 1.51345289e+00 7.55085126e-02
-7.57606447e-01 -1.26334906e-01 5.37695765e-01 -2.08678246e-01
4.50948089e-01 -1.32553363e+00 9.10071880e-02 7.14759767e-01
-2.48784702e-02 -1.56639665e-01 4.81884241e-01 -6.09044552e-01
-1.29252136e+00 -9.75361347e-01 5.04204869e-01 6.92183524e-02
4.18693364e-01 1.12987027e-01 -6.10308945e-01 4.77411360e-01
6.82654828e-02 -1.45598665e-01 6.83195293e-01 1.87500626e-01
4.23998892e-01 -2.16573045e-01 -1.26828694e+00 3.69555980e-01
3.22062612e-01 3.97180617e-02 5.02109230e-01 5.61542153e-01
1.19250752e-02 3.60707566e-02 -9.64454472e-01 7.93479323e-01
7.45355904e-01 -8.00113380e-01 9.21236813e-01 -5.28897762e-01
3.64259630e-01 -3.29067141e-01 -6.03600368e-02 -1.03475082e+00
-8.58852983e-01 -1.67741507e-01 8.81768018e-02 1.61558020e+00
2.24195033e-01 -6.94014877e-02 9.33313251e-01 3.78913954e-02
-7.83745944e-02 -3.44971150e-01 1.66082397e-01 -6.83866978e-01
3.68646793e-02 2.16880858e-01 4.61856484e-01 6.98909938e-01
-6.16606951e-01 -5.22338226e-02 1.52946740e-01 3.67771447e-01
3.34476978e-01 3.86710733e-01 4.35135663e-01 -1.79857957e+00
1.22378498e-01 -6.21470273e-01 -8.05646300e-01 -2.94280984e-02
2.36858726e-02 -6.97472990e-01 1.98535365e-03 -1.54745555e+00
7.48109519e-02 -2.42134854e-01 5.02233021e-02 4.04057413e-01
-8.09499547e-02 -3.70272622e-02 -2.72907480e-03 -2.14846849e-01
2.33905151e-01 7.92599991e-02 1.28220904e+00 -3.79427910e-01
-3.09304714e-01 4.21200037e-01 -6.65669799e-01 5.95851064e-01
1.20059144e+00 -1.96636766e-01 -3.84132266e-01 -3.59061748e-01
-2.34657019e-01 -1.59837604e-01 6.26745343e-01 -1.01272380e+00
-4.77404386e-01 -3.09182405e-01 7.73250163e-01 -7.80176044e-01
1.23486280e-01 -8.48937571e-01 -1.46250397e-01 8.12130809e-01
-2.74969280e-01 -3.33508044e-01 5.15371561e-01 1.45346060e-01
8.01753998e-02 -7.06756711e-01 1.09743822e+00 -3.48386556e-01
-1.05203485e+00 3.48584771e-01 -8.64667952e-01 -5.71583331e-01
1.36556435e+00 -3.61014396e-01 -1.41163602e-01 -6.79145679e-02
-1.02788210e+00 1.13446303e-01 2.37094074e-01 2.88204402e-02
2.04976737e-01 -8.20704401e-01 -8.55863631e-01 1.39560783e-02
2.66181797e-01 -2.41517872e-01 1.37166262e-01 7.93243825e-01
-1.53711116e+00 6.08996153e-01 -1.15384793e+00 -8.71671855e-01
-1.37493396e+00 6.16031706e-01 2.63387233e-01 -1.68722689e-01
-1.24573916e-01 6.97607577e-01 2.74276644e-01 -4.62481260e-01
1.53003797e-01 -3.15192997e-01 -9.42506552e-01 1.83093160e-01
8.59945193e-02 4.26270813e-01 2.16503799e-01 -3.91989589e-01
-2.28938490e-01 3.74135613e-01 2.26446196e-01 5.16889215e-01
1.54820585e+00 5.92894442e-02 -2.08673865e-01 5.41940272e-01
6.68601871e-01 -5.26414275e-01 -1.10518622e+00 5.03417730e-01
4.62964058e-01 -3.58611912e-01 4.06065464e-01 -1.08108008e+00
-1.16767454e+00 1.07553446e+00 1.38855445e+00 6.79872513e-01
1.26073134e+00 -3.09194058e-01 3.13962400e-01 4.60167319e-01
-1.22682817e-01 -9.37172711e-01 -5.92756927e-01 2.39324063e-01
5.75606287e-01 -1.72805810e+00 -1.63622331e-02 -7.71001220e-01
-1.86990634e-01 1.37207830e+00 5.42774737e-01 -1.04731001e-01
7.18863249e-01 4.00096357e-01 2.42624834e-01 -3.54117066e-01
-4.87158209e-01 -8.10489774e-01 -1.26035571e-01 1.32074094e+00
1.15407109e+00 3.60920250e-01 -4.07833129e-01 8.98445174e-02
1.64757371e-01 1.11236461e-01 3.56201887e-01 9.76938546e-01
-6.00494385e-01 -1.45390832e+00 -3.61827433e-01 4.49188948e-01
-3.83748770e-01 1.75358523e-02 -8.32172513e-01 7.29961693e-01
1.51984185e-01 7.14598298e-01 -4.15524483e-01 9.68664885e-02
1.55999720e-01 -1.73831210e-01 8.37598562e-01 -9.27724183e-01
-4.95117337e-01 3.95065285e-02 -2.34932527e-01 -1.17480338e-01
-3.42262536e-01 -8.50718677e-01 -7.20448077e-01 -3.33985984e-01
-6.73377156e-01 -4.93238747e-01 1.07279050e+00 8.27943504e-01
2.72886097e-01 9.09436822e-01 5.51250696e-01 -8.10294986e-01
3.87155376e-02 -9.51999784e-01 -1.01380539e+00 -8.31657425e-02
-2.60063171e-01 -5.12499094e-01 3.90831530e-01 4.27552223e-01] | [9.155440330505371, -1.5614023208618164] |
1a50111e-f70f-47a3-9faf-6824f11806dd | learning-deep-neural-network-representations | 1708.06850 | null | http://arxiv.org/abs/1708.06850v2 | http://arxiv.org/pdf/1708.06850v2.pdf | Learning Deep Neural Network Representations for Koopman Operators of Nonlinear Dynamical Systems | The Koopman operator has recently garnered much attention for its value in
dynamical systems analysis and data-driven model discovery. However, its
application has been hindered by the computational complexity of extended
dynamic mode decomposition; this requires a combinatorially large basis set to
adequately describe many nonlinear systems of interest, e.g. cyber-physical
infrastructure systems, biological networks, social systems, and fluid
dynamics. Often the dictionaries generated for these problems are manually
curated, requiring domain-specific knowledge and painstaking tuning. In this
paper we introduce a deep learning framework for learning Koopman operators of
nonlinear dynamical systems. We show that this novel method automatically
selects efficient deep dictionaries, outperforming state-of-the-art methods. We
benchmark this method on partially observed nonlinear systems, including the
glycolytic oscillator and show it is able to predict quantitatively 100 steps
into the future, using only a single timepoint, and qualitative oscillatory
behavior 400 steps into the future. | ['Soumya Kundu', 'Enoch Yeung', 'Nathan Hodas'] | 2017-08-22 | null | null | null | null | ['model-discovery'] | ['miscellaneous'] | [-1.23554617e-01 -1.72946095e-01 7.52477422e-02 2.76847035e-01
-1.77289903e-01 -7.75996089e-01 7.39843011e-01 2.07296833e-01
-3.64282459e-01 8.48993599e-01 -2.33103812e-01 -2.57507056e-01
-5.38447917e-01 -5.57963371e-01 -3.16206247e-01 -1.11650777e+00
-6.22108161e-01 9.41623628e-01 1.51356116e-01 -5.94916105e-01
1.39514595e-01 5.77153206e-01 -1.15036368e+00 -2.44471118e-01
7.23493934e-01 8.53284121e-01 1.59257203e-01 9.30715442e-01
3.84610742e-01 4.16874558e-01 -3.33021909e-01 1.12571679e-01
2.12005377e-01 -5.94519317e-01 -7.14332402e-01 -3.13820317e-02
-1.95064396e-01 2.90319204e-01 -6.05713725e-01 7.04736054e-01
6.21491551e-01 2.55648822e-01 6.93606317e-01 -1.01135314e+00
-4.68032748e-01 3.08083385e-01 -7.82800168e-02 5.27826011e-01
-1.30870277e-02 5.44376433e-01 1.01333380e+00 -6.01929545e-01
7.99772441e-01 8.02100539e-01 8.01926255e-01 3.98417115e-01
-1.76679993e+00 -4.58878100e-01 -1.46628961e-01 2.07193747e-01
-1.31140888e+00 -3.82593900e-01 9.06977832e-01 -8.05658758e-01
1.09197319e+00 -2.89201289e-02 9.57134187e-01 9.35217559e-01
5.40183008e-01 1.57960981e-01 1.05444694e+00 -1.87138557e-01
4.29043591e-01 -2.42129579e-01 -4.86004949e-02 1.02155602e+00
8.99688378e-02 3.07177722e-01 -3.50340903e-01 -2.35583112e-01
7.92356551e-01 -2.04148218e-01 -3.99387002e-01 -3.50090504e-01
-1.24122703e+00 6.74113393e-01 1.75295025e-01 5.74738979e-01
-3.45023185e-01 1.92552164e-01 3.31762284e-01 4.97626960e-01
4.29424912e-01 9.45107102e-01 -7.93078542e-01 -3.84966940e-01
-8.68110240e-01 5.42142451e-01 9.79440272e-01 3.58453780e-01
7.24647105e-01 5.76445311e-02 5.28881490e-01 5.05001843e-01
2.14670636e-02 5.40687263e-01 4.61061567e-01 -1.06497085e+00
-2.04051569e-01 6.68438494e-01 1.58670381e-01 -9.87733483e-01
-9.70270872e-01 -5.15345454e-01 -1.38363802e+00 1.07929230e-01
6.01664186e-01 -2.59932667e-01 -7.29137003e-01 1.68394887e+00
1.80313185e-01 1.56039909e-01 -5.85295670e-02 9.06082451e-01
3.68654996e-01 8.90211761e-01 -4.25096810e-01 -5.63704371e-01
9.82551992e-01 -4.52156663e-01 -5.51058412e-01 1.04865961e-01
6.61213458e-01 -3.29295218e-01 7.46104181e-01 5.32608807e-01
-1.09225106e+00 -2.42467061e-01 -7.62456536e-01 1.37209624e-01
-3.57759982e-01 -1.51198015e-01 6.61019325e-01 -1.96772274e-02
-9.25895929e-01 1.04733932e+00 -1.30926633e+00 -4.69593257e-01
1.48070484e-01 6.10640168e-01 -3.95399451e-01 4.82784510e-01
-1.21092105e+00 1.03414595e+00 2.42676884e-01 7.43002743e-02
-1.14200044e+00 -8.48047376e-01 -5.25096655e-01 3.45398158e-01
1.85815424e-01 -7.26318359e-01 8.39417040e-01 -5.54188251e-01
-1.74705827e+00 8.12322140e-01 -4.56837155e-02 -5.77392757e-01
3.51587921e-01 3.95234674e-01 -2.84582198e-01 5.47379032e-02
-2.92963773e-01 1.23444378e-01 8.60662878e-01 -9.23868358e-01
1.05929501e-01 -1.77963257e-01 -1.44977169e-02 -1.97232008e-01
-4.25268263e-01 -2.86658853e-01 8.88223872e-02 -5.20656586e-01
1.22007295e-01 -1.37973380e+00 -5.72357357e-01 -7.09023103e-02
-2.09438324e-01 -2.06005737e-01 7.25564659e-01 -5.22253215e-01
1.34202170e+00 -1.71063197e+00 1.08566976e+00 2.14305773e-01
6.58879757e-01 2.96208173e-01 1.31373808e-01 8.30611706e-01
-1.16971120e-01 7.92495832e-02 -4.90724683e-01 -1.86533749e-01
4.66102585e-02 2.68119276e-01 -2.94501245e-01 6.14271760e-01
2.53233880e-01 8.31802607e-01 -8.27724040e-01 -2.31660187e-01
2.11872682e-01 3.93251777e-01 -6.45333052e-01 2.29744270e-01
-3.23647022e-01 7.39270449e-01 -1.34722769e-01 3.72698545e-01
2.75616705e-01 -7.58886039e-01 1.75252795e-01 -1.61108136e-01
-1.95261434e-01 5.50947003e-02 -9.47275221e-01 1.45515645e+00
-5.15420198e-01 8.24287415e-01 1.46895215e-01 -1.63626623e+00
6.90926790e-01 3.65828991e-01 1.10759985e+00 -4.26348865e-01
1.48386210e-01 2.32979387e-01 5.28187931e-01 -5.25889933e-01
2.10002493e-02 -3.07835042e-01 -1.64206877e-01 5.71646452e-01
2.79737979e-01 -3.73080552e-01 4.23093438e-01 1.71482667e-01
1.33752787e+00 -3.25877756e-01 3.87118429e-01 -7.22141504e-01
7.40149677e-01 1.15479052e-01 4.31173682e-01 4.84076142e-01
-2.03072339e-01 2.03969374e-01 8.41634870e-01 -7.60652721e-01
-1.23902380e+00 -8.43524992e-01 -2.16950402e-01 7.85022020e-01
-1.48537066e-02 -5.56668758e-01 -5.75447142e-01 -9.97044221e-02
5.85712865e-02 -9.06985030e-02 -7.61048138e-01 -4.82845783e-01
-8.14750373e-01 -1.00560129e+00 4.37267929e-01 -9.09005664e-03
7.78806955e-03 -1.06690574e+00 -5.33583820e-01 3.88180643e-01
6.25968575e-02 -1.02705371e+00 -1.78027779e-01 2.45122194e-01
-7.48833954e-01 -1.09004605e+00 -7.19754457e-01 -4.99440163e-01
6.60992205e-01 -2.89142698e-01 1.14344418e+00 1.24280795e-01
-7.02419221e-01 1.87201008e-01 -2.63903327e-02 4.91882674e-02
-8.44748735e-01 3.71865720e-01 6.58611536e-01 -2.43201002e-01
-2.33239740e-01 -9.13613856e-01 -4.70034927e-01 2.75663674e-01
-8.97513926e-01 -1.18257320e-02 3.34418982e-01 1.11906314e+00
3.94881487e-01 2.95180619e-01 5.47050416e-01 -3.79538894e-01
7.95515716e-01 -4.67543453e-01 -1.02434647e+00 -2.57056057e-02
-6.86999440e-01 4.02263939e-01 1.15909946e+00 -6.91788793e-01
-3.74997109e-01 -5.85923018e-03 1.01379298e-01 -3.06162119e-01
1.30270213e-01 8.03742230e-01 3.82318258e-01 -3.64687115e-01
7.22443044e-01 5.09343684e-01 3.12886536e-01 -2.44722635e-01
9.14888456e-02 1.41943738e-01 3.10691476e-01 -7.15373158e-01
7.95869946e-01 4.57875907e-01 6.18521392e-01 -1.17402983e+00
-5.08753181e-01 -1.32291064e-01 -7.12934375e-01 -4.42953914e-01
5.57148218e-01 -4.56341952e-01 -1.32094061e+00 6.05114937e-01
-9.96312499e-01 -5.93738735e-01 -2.39511028e-01 1.54276520e-01
-6.85831547e-01 8.68202150e-02 -7.09465802e-01 -7.52022088e-01
-1.14907458e-01 -1.09024715e+00 8.63642395e-01 -2.11233757e-02
-1.37065619e-01 -1.34268188e+00 7.60703743e-01 -2.39362717e-02
6.28672779e-01 5.31218708e-01 9.60848272e-01 -2.80753732e-01
-3.46686542e-01 -2.12029845e-01 2.22671822e-01 1.14760369e-01
9.57683846e-03 2.64000386e-01 -5.23883283e-01 -4.86577958e-01
-2.49626100e-01 -2.12887779e-01 9.34417725e-01 4.48477924e-01
1.07394540e+00 -3.67783487e-01 -2.97343493e-01 6.57484949e-01
1.11060703e+00 -2.20584329e-02 2.19832342e-02 -2.51285851e-01
5.89369357e-01 4.98485565e-01 -2.14853361e-01 6.19222701e-01
3.42482120e-01 5.63001513e-01 4.04734045e-01 1.11251630e-01
4.14480209e-01 3.15836310e-01 2.65044540e-01 1.19738865e+00
-3.94173920e-01 -1.51938036e-01 -1.32343745e+00 5.17660201e-01
-1.87875938e+00 -1.08495629e+00 6.51270300e-02 1.92687023e+00
8.59346688e-01 2.14072868e-01 2.62203574e-01 2.93757975e-01
2.84845650e-01 8.69612992e-02 -8.94578755e-01 -1.50845736e-01
-1.52743563e-01 3.25467050e-01 2.05191106e-01 7.55186677e-01
-8.70244563e-01 7.47077346e-01 6.49031448e+00 3.94163281e-01
-1.35059071e+00 2.09219608e-04 4.65188980e-01 7.79137984e-02
2.35430405e-01 4.57462817e-02 -4.89633024e-01 5.09774804e-01
1.10402369e+00 -4.43708122e-01 9.96163011e-01 3.06244195e-01
5.42360783e-01 2.99837720e-02 -1.06433392e+00 1.07324910e+00
-4.93295938e-01 -1.63496900e+00 -4.98004824e-01 3.93661350e-01
8.41451705e-01 1.81220859e-01 1.63646194e-03 1.89551681e-01
2.01439723e-01 -1.06607914e+00 2.40492806e-01 9.09772277e-01
4.28092867e-01 -4.84030426e-01 4.92614329e-01 5.90530396e-01
-1.02373898e+00 -1.95300877e-01 -2.26695865e-01 -4.54370916e-01
2.03722492e-01 9.66661394e-01 -4.17660117e-01 -2.63137445e-02
4.08967882e-01 9.52943444e-01 -2.11468369e-01 6.16294026e-01
4.46250707e-01 7.50739634e-01 -6.17147386e-01 -9.58812162e-02
1.09525740e-01 -3.84715110e-01 7.93116808e-01 9.13991511e-01
3.74376416e-01 2.24683106e-01 2.45715573e-01 1.06524086e+00
-4.34541330e-02 -3.49661291e-01 -6.48105681e-01 -6.96821272e-01
1.20795175e-01 1.30452812e+00 -8.04630041e-01 -2.35085085e-01
7.90190026e-02 6.40921533e-01 3.42018723e-01 3.38869512e-01
-7.51778603e-01 -1.46167189e-01 1.02999079e+00 3.46189201e-01
2.24304974e-01 -9.07660723e-01 2.20279358e-02 -1.52121162e+00
-3.38945121e-01 -7.36067176e-01 7.55556673e-02 -3.48667026e-01
-1.27785683e+00 4.34849083e-01 -8.34710822e-02 -9.70974505e-01
-4.25166219e-01 -8.41010094e-01 -4.95574355e-01 6.34264290e-01
-9.87522185e-01 -6.06430769e-01 -2.05374151e-01 4.32280689e-01
1.15261294e-01 -2.97273517e-01 9.59772229e-01 2.49935612e-01
-9.18528676e-01 -4.45546620e-02 4.94087875e-01 -9.67945531e-03
3.62037122e-01 -1.35328150e+00 2.97057301e-01 5.81844091e-01
-4.37381119e-02 4.95813668e-01 1.15345526e+00 -1.56990856e-01
-1.77846217e+00 -7.21419990e-01 6.39848113e-01 -3.53202343e-01
1.10828030e+00 -5.61568379e-01 -9.21251416e-01 3.19448262e-01
-5.60730770e-02 4.13142413e-01 5.02111793e-01 -1.80288963e-02
5.87908961e-02 -2.55511701e-01 -8.04182112e-01 4.50306565e-01
8.74988616e-01 -5.69846034e-01 -8.49313140e-02 5.09193778e-01
3.84146243e-01 -4.43356961e-01 -1.15256906e+00 2.94254065e-01
5.90608180e-01 -7.44256496e-01 9.48224723e-01 -8.77373993e-01
4.69004571e-01 -2.18062997e-01 3.35997611e-01 -1.47649026e+00
-5.43548584e-01 -9.95290458e-01 -7.65393555e-01 4.56626594e-01
3.50621045e-01 -7.34179616e-01 4.75967020e-01 2.37898767e-01
3.58372331e-02 -9.16751027e-01 -9.82614636e-01 -8.05879116e-01
2.71395713e-01 -9.76593122e-02 1.43271402e-01 1.15082157e+00
2.94080079e-01 3.94094110e-01 -2.96562105e-01 2.04240054e-01
5.54958403e-01 2.59529024e-01 6.28384352e-01 -1.52242696e+00
-5.89019835e-01 -9.92778063e-01 -5.64617515e-01 -7.83363342e-01
3.73935997e-01 -8.51279020e-01 -7.88125694e-02 -1.01667631e+00
-7.51644075e-02 -3.19462866e-01 -2.34950960e-01 1.99620932e-01
8.83107856e-02 2.49607161e-01 6.78636804e-02 2.31300622e-01
-4.94029701e-01 5.97184420e-01 1.16575515e+00 -1.40370354e-01
-3.46065283e-01 -1.85614392e-01 -2.88957119e-01 5.40977716e-01
8.15480053e-01 -4.57624018e-01 -1.69554472e-01 6.30472749e-02
6.95397675e-01 2.79918373e-01 5.52117288e-01 -1.19290519e+00
2.27006614e-01 -3.42439502e-01 1.26538277e-02 -8.59649405e-02
3.32006037e-01 -6.37884617e-01 2.14284465e-01 8.33864570e-01
-4.22750056e-01 2.69699961e-01 1.50396749e-01 5.93785882e-01
-8.77794847e-02 2.09182039e-01 9.98009503e-01 -1.28637433e-01
-1.31567180e-01 6.16535485e-01 -7.66055822e-01 3.47077340e-01
8.29780102e-01 3.34636152e-01 -1.39232531e-01 -4.28567410e-01
-1.14708877e+00 3.01730990e-01 4.80213255e-01 1.60770029e-01
2.51046985e-01 -1.00731814e+00 -6.55433953e-01 2.51871407e-01
-1.73998773e-01 -6.66272342e-02 1.57683477e-01 1.14050865e+00
-6.90214336e-01 4.64647561e-01 -2.08309218e-01 -7.94350505e-01
-9.01214302e-01 4.81979102e-01 6.86565220e-01 -6.99456990e-01
-5.02310634e-01 4.84253645e-01 -6.21883422e-02 -3.95735145e-01
-3.29950064e-01 -6.70741081e-01 8.04828554e-02 -7.27277026e-02
1.02441020e-01 5.22955418e-01 -7.29699880e-02 -6.23346925e-01
-3.83313268e-01 6.23327851e-01 2.93212414e-01 -1.94448680e-02
1.44971967e+00 4.56436798e-02 -5.21234989e-01 6.28283978e-01
1.11990237e+00 -4.11743402e-01 -1.26451695e+00 -2.79633999e-01
1.21404648e-01 2.17776030e-01 5.18628620e-02 -4.81064826e-01
-9.27361608e-01 1.07713592e+00 2.27466270e-01 9.45919871e-01
1.11336994e+00 5.31233735e-02 8.69764149e-01 8.98688853e-01
3.09558839e-01 -8.44475091e-01 1.32515281e-01 9.03952122e-01
8.60345185e-01 -1.26962388e+00 -1.27426069e-02 2.39010096e-01
-7.82161653e-02 1.38354433e+00 1.90950885e-01 -5.75391829e-01
1.24692166e+00 3.85769874e-01 -4.42743301e-01 -2.33206615e-01
-1.19207847e+00 -5.58015220e-02 3.14871192e-01 2.70496786e-01
3.58667523e-01 9.27744880e-02 -1.34760708e-01 3.74576569e-01
-1.58691585e-01 -2.23966494e-01 6.20152593e-01 5.81753850e-01
-3.53563339e-01 -9.41517055e-01 2.32597217e-02 4.49043036e-01
-2.75973678e-01 1.28394097e-01 -4.01699543e-01 6.10286236e-01
-1.48103870e-02 3.85273129e-01 -6.68478534e-02 -2.09822088e-01
-3.95082235e-02 1.02947563e-01 5.38484991e-01 -3.02633673e-01
-4.52202708e-01 -1.47117540e-01 -3.36695641e-01 -4.30653393e-01
-4.00734991e-01 -6.27237976e-01 -1.20779026e+00 -6.14379585e-01
-2.12283775e-01 1.85913786e-01 3.63038987e-01 1.20165730e+00
6.66105926e-01 5.62637269e-01 5.81815124e-01 -1.28383994e+00
-4.48195100e-01 -9.32779431e-01 -4.50076252e-01 3.01367551e-01
9.08913851e-01 -7.35863686e-01 -6.54682159e-01 1.79122597e-01] | [6.565204620361328, 3.58657169342041] |
4e177ef1-5178-4943-91dc-6f152358bf0f | multimodal-emotion-cause-pair-extraction-in | 2110.08020 | null | https://arxiv.org/abs/2110.08020v1 | https://arxiv.org/pdf/2110.08020v1.pdf | Multimodal Emotion-Cause Pair Extraction in Conversations | Emotion cause analysis has received considerable attention in recent years. Previous studies primarily focused on emotion cause extraction from texts in news articles or microblogs. It is also interesting to discover emotions and their causes in conversations. As conversation in its natural form is multimodal, a large number of studies have been carried out on multimodal emotion recognition in conversations, but there is still a lack of work on multimodal emotion cause analysis. In this work, we introduce a new task named Multimodal Emotion-Cause Pair Extraction in Conversations, aiming to jointly extract emotions and their associated causes from conversations reflected in multiple modalities (text, audio and video). We accordingly construct a multimodal conversational emotion cause dataset, Emotion-Cause-in-Friends, which contains 9,272 multimodal emotion-cause pairs annotated on 13,509 utterances in the sitcom Friends. We finally benchmark the task by establishing a baseline system that incorporates multimodal features for emotion-cause pair extraction. Preliminary experimental results demonstrate the potential of multimodal information fusion for discovering both emotions and causes in conversations. | ['Jianfei Yu', 'Zhaoyu Li', 'Rui Xia', 'Zixiang Ding', 'Fanfan Wang'] | 2021-10-15 | null | null | null | null | ['multimodal-emotion-recognition', 'emotion-cause-pair-extraction', 'emotion-cause-extraction', 'multimodal-emotion-recognition'] | ['computer-vision', 'natural-language-processing', 'natural-language-processing', 'speech'] | [ 3.06511343e-01 1.15338378e-01 1.23718590e-01 -7.06947684e-01
-1.06596160e+00 -7.14480340e-01 8.92835736e-01 1.59652844e-01
1.28711313e-01 8.51832330e-01 1.01377952e+00 2.84814149e-01
-1.43449128e-01 -1.78466797e-01 -2.30816141e-01 -6.08500540e-01
-1.80227503e-01 -3.50168301e-03 -5.50928175e-01 -3.32462430e-01
-2.09677382e-03 1.10328466e-01 -2.08943009e+00 9.53210711e-01
5.22954345e-01 9.43336248e-01 -2.70076722e-01 1.09995735e+00
-6.38804436e-01 1.57414627e+00 -8.15641522e-01 -7.58868575e-01
-6.69758916e-01 -8.34150612e-01 -1.15033937e+00 3.19955885e-01
2.48177699e-03 -7.75452778e-02 1.77804790e-02 6.39111161e-01
6.92726374e-01 -2.24156175e-02 7.50474393e-01 -2.02496839e+00
7.77996406e-02 8.98087263e-01 -3.32429826e-01 -9.95723158e-02
1.32759118e+00 -4.74388421e-01 1.16893530e+00 -6.24713480e-01
6.30239785e-01 1.58543336e+00 4.25833762e-01 5.15712321e-01
-7.54081309e-01 -3.82697403e-01 -6.19527064e-02 3.69952083e-01
-1.01813293e+00 -5.14151275e-01 1.01403499e+00 -2.06520572e-01
1.06527388e+00 6.96734965e-01 5.04067421e-01 1.42862344e+00
-1.31657317e-01 1.26581872e+00 1.05878961e+00 -4.63647395e-01
-1.82373840e-02 2.38962725e-01 2.14835271e-01 3.31759930e-01
-7.85401583e-01 -5.57190359e-01 -1.04924214e+00 -6.75272703e-01
1.72972292e-01 -4.02687728e-01 -4.13219303e-01 2.10885152e-01
-1.43653667e+00 7.48952627e-01 -1.97028480e-02 5.67568064e-01
-7.52881706e-01 5.89997955e-02 7.71143556e-01 6.61058307e-01
4.98199731e-01 2.72125900e-01 -5.89163125e-01 -8.62053216e-01
-3.82153958e-01 3.71903270e-01 1.34160674e+00 7.37388909e-01
6.59974694e-01 -5.31881571e-01 1.76284820e-01 1.13240957e+00
2.01143965e-01 4.25574511e-01 1.53328255e-01 -7.24255502e-01
3.58499736e-01 8.57232928e-01 7.48630166e-02 -1.48460305e+00
-8.81567299e-01 5.63373446e-01 -7.03281939e-01 -8.87541115e-01
1.13540560e-01 -7.68330336e-01 -1.72804460e-01 1.83106065e+00
5.66139579e-01 1.60819829e-01 4.85354751e-01 9.72555161e-01
1.71212840e+00 8.28658164e-01 2.95380224e-02 -4.59615886e-01
1.68892491e+00 -4.02280360e-01 -1.40696168e+00 3.24813962e-01
6.31082892e-01 -1.31432295e+00 6.15054250e-01 4.84892845e-01
-7.93990076e-01 -1.13654453e-02 -3.01577151e-01 7.01745301e-02
-5.05253077e-01 1.37528509e-01 9.37047541e-01 5.23652732e-01
-6.49188697e-01 -2.10536852e-01 6.89642923e-03 -7.74535179e-01
-3.15738738e-01 3.35785389e-01 -7.32641459e-01 1.32372126e-01
-1.56635749e+00 7.15943992e-01 4.17168215e-02 -2.53206678e-02
-5.68349242e-01 -4.06914115e-01 -1.03925312e+00 -2.27333143e-01
2.75396526e-01 -4.27764595e-01 1.44368494e+00 -1.24924421e+00
-1.55215716e+00 7.70284235e-01 -6.20066702e-01 1.70893118e-01
-2.83763885e-01 -2.80349433e-01 -1.04060686e+00 2.88630128e-01
-1.90675735e-01 7.64793992e-01 7.36259520e-01 -1.45464778e+00
-8.54873776e-01 -2.20551312e-01 -1.00891896e-01 3.69324535e-01
-5.99447608e-01 7.76706994e-01 -5.15501678e-01 -2.81953812e-01
-1.13274761e-01 -9.01166618e-01 1.26154095e-01 -8.25952530e-01
-7.67395377e-01 -6.53568149e-01 9.57484901e-01 -4.15644586e-01
1.28171277e+00 -1.93839502e+00 3.80622655e-01 7.56854713e-02
2.99393713e-01 -5.74269235e-01 -2.61311084e-01 8.88228178e-01
-4.33457345e-01 4.83262166e-02 6.97824210e-02 -5.82393527e-01
3.90545487e-01 4.08842772e-01 -5.43174624e-01 1.37135208e-01
3.27947021e-01 7.11005926e-01 -7.43890882e-01 -7.18826950e-01
2.51928151e-01 7.27959573e-01 -3.00316542e-01 6.46731615e-01
6.86854357e-03 4.71943110e-01 -5.10145247e-01 6.58655584e-01
4.61733341e-01 5.29962592e-02 2.08487496e-01 -4.05799925e-01
-1.01339228e-01 2.52793044e-01 -9.86157179e-01 1.45394075e+00
-4.88662064e-01 9.58108544e-01 3.59400302e-01 -4.03193980e-01
8.51818085e-01 1.10376704e+00 8.88906837e-01 -3.89603227e-02
7.25350380e-01 -6.98232353e-02 -8.85679051e-02 -1.22037542e+00
8.87275040e-01 -3.36777091e-01 -8.22668970e-01 4.89471138e-01
3.03178728e-02 -2.64617264e-01 1.91289529e-01 3.33297580e-01
1.29874980e+00 -5.12302518e-01 1.75498694e-01 4.56627011e-01
6.79162383e-01 1.06188886e-01 2.19830781e-01 1.63665801e-01
-3.79877836e-01 4.26274329e-01 1.00561643e+00 -2.17588529e-01
-3.58191431e-01 -5.64561784e-01 -1.31806388e-01 1.35717905e+00
-6.35771677e-02 -7.34533012e-01 -6.12734079e-01 -5.75039983e-01
-2.25573063e-01 5.08514702e-01 -6.55509651e-01 2.74284154e-01
-2.16423780e-01 -8.30807090e-01 8.66006434e-01 -4.49583456e-02
3.53511214e-01 -1.08500433e+00 -1.07138269e-01 1.75764188e-01
-1.04312599e+00 -1.56607163e+00 1.37933195e-01 8.01687762e-02
-2.05889374e-01 -1.20932209e+00 -5.90825140e-01 -5.89086950e-01
2.31378719e-01 7.72264525e-02 1.31089914e+00 -4.02747989e-01
-2.13436857e-01 1.18925548e+00 -7.71035254e-01 -6.12458825e-01
-5.58364451e-01 4.68582734e-02 1.90829728e-02 5.91140747e-01
7.55661249e-01 -2.94389904e-01 3.17876264e-02 4.08688754e-01
-8.23881745e-01 -1.45948324e-02 2.70606786e-01 4.13464040e-01
-1.58268526e-01 -3.20196301e-02 8.65694940e-01 -3.02615583e-01
1.05685723e+00 -8.59472692e-01 3.41680586e-01 -5.80072636e-03
3.78119260e-01 -4.25233692e-01 1.38473749e-01 -2.72105634e-01
-1.41102409e+00 1.74043193e-01 -3.18690091e-01 -1.10806555e-01
-1.03784049e+00 7.52509356e-01 -4.58168983e-01 3.19982827e-01
8.72240439e-02 4.76394556e-02 4.05476689e-02 -2.58597493e-01
6.36432350e-01 1.08489692e+00 3.99333358e-01 -4.88599777e-01
5.21464311e-02 4.11401153e-01 -2.57849365e-01 -1.28388584e+00
-6.76272929e-01 -9.41589475e-01 -3.38397682e-01 -9.12741303e-01
1.11598849e+00 -9.73158419e-01 -1.24989235e+00 3.88421655e-01
-1.57048213e+00 3.93094510e-01 3.79930347e-01 5.03357053e-01
-4.98477578e-01 2.73978680e-01 -6.73951089e-01 -1.14528012e+00
-3.17054354e-02 -8.22480261e-01 1.38827574e+00 3.35905924e-02
-1.00497925e+00 -9.46262717e-01 1.20630942e-01 4.83989865e-01
2.90486123e-02 4.43814784e-01 6.38059914e-01 -7.37840950e-01
1.03276357e-01 -4.48341966e-01 -2.81672269e-01 -2.14422971e-01
4.31065679e-01 1.96415916e-01 -1.31665802e+00 5.67188203e-01
-2.08941862e-01 -7.74407327e-01 5.42264700e-01 2.21665040e-01
5.04290998e-01 -2.52303094e-01 -1.93789646e-01 -2.69550145e-01
6.51486337e-01 1.22691862e-01 6.02608740e-01 -2.21001938e-01
7.41727710e-01 1.48730528e+00 5.29595077e-01 8.20144534e-01
8.27040911e-01 4.64899868e-01 6.52800977e-01 -4.29851979e-01
2.78020024e-01 2.25929007e-01 5.06857812e-01 1.18550885e+00
5.92180900e-02 -4.88882422e-01 -7.14582860e-01 6.43261254e-01
-1.97269559e+00 -1.21826482e+00 -8.95807028e-01 1.20467877e+00
8.38052034e-01 -8.09594810e-01 3.60445529e-01 3.18758667e-01
8.49857748e-01 1.45261392e-01 1.26822650e-01 -5.75365543e-01
-4.82069820e-01 -8.03375125e-01 -5.17150104e-01 1.96583137e-01
-1.46978140e+00 5.18455744e-01 5.98884630e+00 6.25353873e-01
-7.45526254e-01 -4.53393646e-02 6.23989642e-01 -9.60092172e-02
-4.04919744e-01 -2.42942080e-01 -4.75348532e-01 1.85760811e-01
7.88086534e-01 2.89666384e-01 1.73416391e-01 6.09696329e-01
3.93634617e-01 -3.19799840e-01 -1.14261091e+00 1.50950360e+00
3.91638696e-01 -9.11178589e-01 -2.29022205e-01 -1.19430490e-01
7.04899371e-01 -2.81021923e-01 -3.39108050e-01 2.36542478e-01
-1.56839371e-01 -7.60854006e-01 2.54748732e-01 6.02235734e-01
2.24864900e-01 -1.16302633e+00 1.03505337e+00 5.06416112e-02
-1.16148257e+00 9.17236134e-02 8.20558369e-02 -3.41902584e-01
5.60360193e-01 9.50745940e-01 -9.86325026e-01 4.93694752e-01
6.92274213e-01 8.70792985e-01 -1.85944557e-01 5.86638689e-01
-2.24843130e-01 6.83814347e-01 -3.08183819e-01 -6.19042099e-01
3.91129628e-02 2.64324933e-01 7.71310091e-01 1.71908951e+00
1.72000825e-01 2.34904736e-01 1.37920320e-01 4.48147833e-01
-2.02307194e-01 4.17604893e-01 -7.83702254e-01 -2.36166224e-01
1.76589191e-01 1.71431303e+00 -5.30951738e-01 -3.60718578e-01
-3.49723965e-01 1.03130400e+00 -1.72714502e-01 1.83070734e-01
-9.69744027e-01 -3.29026818e-01 9.59732771e-01 -8.70174706e-01
-1.77447960e-01 1.23897702e-01 1.27702817e-01 -1.19078660e+00
-2.35625967e-01 -9.63039935e-01 5.31571209e-01 -9.54629064e-01
-1.75880873e+00 8.04243326e-01 1.63552091e-02 -9.65811372e-01
-7.32093275e-01 -3.74790698e-01 -6.24123812e-01 5.41186869e-01
-1.23054826e+00 -1.04582608e+00 -3.81466925e-01 8.38023424e-01
4.14574414e-01 -7.34138787e-02 1.15824270e+00 3.98268044e-01
-7.00222373e-01 2.03091115e-01 -6.60486519e-01 1.09110244e-01
1.00957835e+00 -1.09569240e+00 -6.03024244e-01 1.63460359e-01
-9.49062631e-02 4.38440621e-01 1.01051629e+00 -5.45101047e-01
-1.93890679e+00 -5.19967735e-01 1.57011151e+00 -6.73159063e-01
8.64963293e-01 -3.10237676e-01 -5.33290327e-01 1.91225052e-01
1.05380917e+00 -6.11852288e-01 1.24227405e+00 6.32168233e-01
-2.21793488e-01 2.06075087e-01 -1.12997961e+00 5.43653786e-01
5.20444036e-01 -6.41952515e-01 -6.48950517e-01 2.57606447e-01
5.62377393e-01 2.35065520e-02 -1.08342063e+00 3.87855828e-01
4.86363858e-01 -1.04922247e+00 6.32122040e-01 -5.83828986e-01
8.04251909e-01 -5.67669012e-02 -3.41689557e-01 -1.24222386e+00
4.20495570e-01 -1.12389994e+00 1.22818105e-01 2.14958334e+00
4.28088635e-01 -2.34637707e-01 3.72092307e-01 9.89247084e-01
-1.49731934e-02 -2.10001141e-01 -9.75092530e-01 -2.60161255e-02
-4.81485724e-01 -1.04881573e+00 7.81053543e-01 1.29563725e+00
7.88025200e-01 9.02494848e-01 -6.83187366e-01 1.17602043e-01
2.05903634e-01 3.55307221e-01 1.07954085e+00 -9.30721700e-01
1.10412784e-01 -4.87071723e-01 -3.25387955e-01 -6.39196515e-01
4.65541214e-01 -5.01412928e-01 3.59336078e-01 -1.39886391e+00
4.71388221e-01 1.16879076e-01 2.26101071e-01 7.39270926e-01
-3.83803517e-01 3.19257639e-02 1.59273311e-01 -7.72635406e-03
-9.64296222e-01 8.22514117e-01 9.42531347e-01 2.45829299e-02
-2.21802235e-01 -2.96895385e-01 -3.74281526e-01 1.05555308e+00
5.68305612e-01 -1.98078781e-01 -1.39961571e-01 2.33640999e-01
6.95103407e-01 3.70698184e-01 4.15741861e-01 -3.71287137e-01
6.01770505e-02 -5.22624748e-03 -4.36238237e-02 -8.84414434e-01
5.90155244e-01 -8.85438323e-01 -1.03003487e-01 -3.74705017e-01
-4.30661798e-01 7.27499425e-02 3.12880367e-01 3.25053990e-01
-8.21476579e-01 5.18136052e-03 -1.28957639e-02 2.65343279e-01
-8.14594030e-01 -4.38770503e-01 -1.16169298e+00 -2.24246517e-01
7.87381113e-01 2.24712104e-01 -3.49833190e-01 -1.13134873e+00
-7.07996786e-01 2.81384319e-01 -1.63931921e-01 7.07378805e-01
7.84811437e-01 -1.50232232e+00 -8.26451182e-01 -4.33806568e-01
2.99331963e-01 -7.83883929e-01 5.45406461e-01 1.26047254e+00
4.75526661e-01 3.85478765e-01 2.55686074e-01 -4.95333076e-01
-1.86303139e+00 2.62620270e-01 -1.10756725e-01 -4.68382761e-02
8.23629051e-02 9.46724951e-01 3.29095349e-02 -7.56956816e-01
3.76610935e-01 -5.73941395e-02 -7.20518827e-01 9.17288661e-01
7.50682294e-01 3.02557617e-01 -2.16844231e-01 -1.02226543e+00
-4.96846408e-01 3.60479146e-01 4.49124992e-01 -2.44679600e-01
1.28398776e+00 -6.41497970e-01 -8.36440384e-01 8.29548240e-01
1.56225455e+00 -7.11466698e-03 -3.28462899e-01 1.73239261e-01
1.28646329e-01 5.13683036e-02 -3.32910828e-02 -8.22799385e-01
-7.15564430e-01 6.22568190e-01 2.83840150e-01 9.47334826e-01
1.38418281e+00 4.86212283e-01 8.53118658e-01 3.86824757e-01
6.48088306e-02 -8.91108155e-01 1.71057627e-01 8.80176842e-01
1.44205773e+00 -1.46508777e+00 -4.33049828e-01 -8.02172303e-01
-9.79951739e-01 1.27444279e+00 2.09740967e-01 4.86846775e-01
5.25587618e-01 5.56871831e-01 5.02793610e-01 -6.76776290e-01
-1.00780416e+00 -5.76030850e-01 5.06514192e-01 2.56035894e-01
8.57841909e-01 1.19303040e-01 -1.27262920e-01 1.08689415e+00
-2.51905739e-01 -2.69266486e-01 4.90044892e-01 9.15505409e-01
-1.28172547e-01 -1.05833220e+00 -1.03063703e+00 2.02865601e-01
-7.14352608e-01 -1.34523228e-01 -1.37755394e+00 5.82150578e-01
2.52085458e-02 1.90229058e+00 -2.94971336e-02 -7.30552733e-01
2.45375022e-01 4.23537463e-01 1.24588147e-01 -9.73035172e-02
-9.47804391e-01 4.82858390e-01 1.01336122e+00 -6.51929975e-01
-1.25552285e+00 -9.61447060e-01 -1.37438405e+00 -3.40668350e-01
-1.29638851e-01 2.18018755e-01 8.13948572e-01 1.11752963e+00
4.65302974e-01 6.84573829e-01 8.91952276e-01 -1.01697099e+00
4.76394296e-01 -1.02599001e+00 -4.56451446e-01 5.60920596e-01
3.84008199e-01 -3.99449050e-01 -5.29175639e-01 3.03264648e-01] | [13.079998016357422, 5.650579452514648] |
f1e5f167-d626-4f6d-89aa-14fa97e5ed94 | compositional-scalable-object-slam | 2011.02658 | null | https://arxiv.org/abs/2011.02658v1 | https://arxiv.org/pdf/2011.02658v1.pdf | Compositional Scalable Object SLAM | We present a fast, scalable, and accurate Simultaneous Localization and Mapping (SLAM) system that represents indoor scenes as a graph of objects. Leveraging the observation that artificial environments are structured and occupied by recognizable objects, we show that a compositional scalable object mapping formulation is amenable to a robust SLAM solution for drift-free large scale indoor reconstruction. To achieve this, we propose a novel semantically assisted data association strategy that obtains unambiguous persistent object landmarks, and a 2.5D compositional rendering method that enables reliable frame-to-model RGB-D tracking. Consequently, we deliver an optimized online implementation that can run at near frame rate with a single graphics card, and provide a comprehensive evaluation against state of the art baselines. An open source implementation will be provided at https://placeholder. | ['Michael Kaess', 'Wei Dong', 'Akash Sharma'] | 2020-11-05 | null | null | null | null | ['object-slam'] | ['computer-vision'] | [ 1.63926393e-01 -2.48811215e-01 1.18082963e-01 -5.16284525e-01
-9.19283807e-01 -7.38317966e-01 7.29139090e-01 1.83279589e-01
-3.44556153e-01 5.98056793e-01 -3.85261467e-03 -2.12604314e-01
-6.48303926e-02 -7.63379633e-01 -1.03842151e+00 -1.83730841e-01
-2.19411001e-01 7.74614096e-01 5.29731452e-01 -9.02139172e-02
2.14782342e-01 8.83868992e-01 -1.69897234e+00 -2.47253135e-01
5.66152632e-01 1.17305934e+00 4.70849901e-01 9.02955055e-01
4.11086529e-02 5.75054526e-01 -2.15819627e-01 9.73934755e-02
4.00251478e-01 4.26119976e-02 -6.16505206e-01 5.73562570e-02
9.83091295e-01 -3.85383546e-01 -4.24805909e-01 7.88152039e-01
5.51454723e-01 2.20632777e-01 7.06112385e-02 -1.27594292e+00
-3.90715957e-01 -2.54817575e-01 -2.13922307e-01 -2.31461957e-01
9.80268538e-01 -2.39435434e-01 5.79785585e-01 -1.18145633e+00
8.30417097e-01 1.12501228e+00 9.63709474e-01 2.39900932e-01
-1.26800311e+00 -4.99476254e-01 3.03602427e-01 -1.15620054e-01
-1.88153303e+00 -7.96070099e-01 6.86957896e-01 -1.22897327e-01
8.90462101e-01 6.07804537e-01 7.96567202e-01 9.92480397e-01
1.91585153e-01 3.65968913e-01 8.54822993e-01 -3.22768539e-01
4.67828065e-01 -5.76236583e-02 -1.15612172e-01 8.20100307e-01
4.17944223e-01 9.29100662e-02 -9.13031638e-01 -2.94562608e-01
1.01537371e+00 2.16402352e-01 -1.60063356e-01 -1.21306038e+00
-1.37099898e+00 3.04305106e-01 8.81294012e-01 -2.77582467e-01
-2.56895602e-01 8.10924053e-01 -1.77437082e-01 -5.18973656e-02
2.63722330e-01 1.28747717e-01 -2.78829604e-01 -2.05619723e-01
-7.23246813e-01 2.21646577e-01 5.15677452e-01 1.60515678e+00
1.05530941e+00 -3.29118490e-01 5.13373733e-01 1.77496195e-01
5.74260175e-01 9.67599571e-01 -7.31014907e-02 -1.44095480e+00
2.29927093e-01 3.57812166e-01 5.92346072e-01 -1.13995123e+00
-4.28782731e-01 -1.60742044e-01 -4.23313946e-01 2.01090828e-01
-5.88103794e-02 4.44193929e-01 -7.01826453e-01 1.29800081e+00
8.28595042e-01 4.74382043e-01 -4.18619961e-02 8.73686433e-01
5.94517827e-01 2.70092458e-01 -2.93990701e-01 2.40968242e-01
1.18409979e+00 -9.26790476e-01 -5.74825823e-01 -5.97578287e-01
3.73472899e-01 -7.05821395e-01 8.81052673e-01 8.16200897e-02
-8.45336795e-01 -4.36873645e-01 -1.13714898e+00 -5.28324902e-01
-2.24262640e-01 -1.79723933e-01 9.68149006e-01 5.73356211e-01
-1.35102606e+00 2.60472983e-01 -1.49197030e+00 -7.42404163e-01
2.93040514e-01 4.40118790e-01 -5.83299518e-01 -2.54019529e-01
-3.26645821e-01 8.89259219e-01 2.48174801e-01 7.16793686e-02
-7.99280167e-01 -4.76600796e-01 -1.10997069e+00 -5.94025970e-01
2.27711409e-01 -9.39530194e-01 1.16196787e+00 -2.67332166e-01
-1.35888898e+00 8.75890255e-01 -7.42009163e-01 -4.43090707e-01
6.20507002e-01 -5.35664558e-01 -1.07613452e-01 4.43350375e-02
2.45233357e-01 5.84031522e-01 1.85265705e-01 -1.75576973e+00
-8.03341568e-01 -6.15153313e-01 -2.69937925e-02 4.05704230e-01
1.84936985e-01 -7.98718557e-02 -7.41939604e-01 -4.92073596e-02
1.10369587e+00 -1.03046286e+00 -4.94048804e-01 5.79642415e-01
-1.99030071e-01 5.36896110e-01 7.33173788e-01 -4.79765058e-01
5.96728206e-01 -2.16567159e+00 -4.60909680e-02 3.72379869e-01
1.06745154e-01 -5.12027085e-01 1.28336504e-01 3.58806342e-01
6.57152116e-01 -3.15311164e-01 -2.00514477e-02 -1.16244853e+00
2.25116685e-01 4.78198528e-01 -4.42436546e-01 9.28245962e-01
-4.73090976e-01 8.90549242e-01 -1.10522091e+00 -3.31934959e-01
5.64395249e-01 8.08058858e-01 -4.27133769e-01 1.36711091e-01
5.19884266e-02 6.20162666e-01 -3.69973779e-01 1.07301784e+00
8.79764676e-01 -2.69968897e-01 1.25269711e-01 4.07868177e-02
-4.03116792e-01 6.18345976e-01 -1.50780487e+00 2.66090369e+00
-5.26462138e-01 6.33227050e-01 2.83136427e-01 -3.63680497e-02
9.67520058e-01 -1.01474598e-01 4.30699080e-01 -5.71449339e-01
-2.20659181e-01 4.84960079e-01 -9.24172044e-01 3.87271017e-01
1.01129472e+00 4.85275835e-01 -1.16000593e-01 1.06400564e-01
-2.95433402e-01 -4.92401958e-01 -4.56467330e-01 2.47994125e-01
1.39310765e+00 6.67349994e-01 4.82471943e-01 -2.10624576e-01
1.76495701e-01 3.39570999e-01 4.99516875e-01 9.24338102e-01
-1.37590915e-01 5.97618580e-01 -6.52843058e-01 -7.10660517e-01
-8.73869359e-01 -1.51101017e+00 -2.75473177e-01 7.55184829e-01
1.00755286e+00 -5.77935219e-01 -5.37012368e-02 -1.66645557e-01
3.30585271e-01 2.41795868e-01 -1.87651023e-01 3.20161045e-01
-6.21592402e-01 -1.25027537e-01 1.62264138e-01 6.27571881e-01
5.24065375e-01 -4.96411085e-01 -1.02521729e+00 1.87607318e-01
-1.28977969e-01 -1.26756048e+00 -1.70876369e-01 3.12296450e-01
-7.34890819e-01 -8.82210672e-01 -5.11731058e-02 -8.54771316e-01
9.29754436e-01 8.36779237e-01 1.14726496e+00 1.19330443e-01
-2.72622347e-01 8.20144475e-01 -1.83368355e-01 -2.65413582e-01
6.98648952e-03 -1.70420691e-01 5.51306069e-01 -4.39944297e-01
4.78120521e-03 -8.05091619e-01 -6.90071344e-01 4.06383544e-01
-2.14361742e-01 4.13890630e-01 2.04613671e-01 1.57949626e-01
1.34738457e+00 -4.94997829e-01 -3.60092402e-01 -3.60153347e-01
-2.81553060e-01 -1.04787618e-01 -1.17602730e+00 1.08357564e-01
-4.94234711e-01 -2.52866268e-01 1.53634965e-01 -4.29412313e-02
-6.75983012e-01 8.23522270e-01 2.23539323e-02 -2.86063075e-01
-3.43084276e-01 -5.19141033e-02 -1.72464564e-01 -9.65385497e-01
5.12670159e-01 3.77507627e-01 -3.96086514e-01 -4.79275346e-01
6.12878621e-01 4.99331653e-01 1.02885330e+00 -8.97229791e-01
1.21592343e+00 1.08787119e+00 1.75805286e-01 -5.69892287e-01
-5.82898676e-01 -7.87482321e-01 -8.59557927e-01 -1.91605896e-01
4.53458458e-01 -1.33680320e+00 -7.23351717e-01 1.57233238e-01
-1.29374754e+00 -5.65174103e-01 -2.83979595e-01 4.59887445e-01
-8.96269441e-01 1.53559238e-01 -3.06486964e-01 -1.00103998e+00
2.41755508e-02 -8.42362285e-01 1.69104755e+00 3.19056422e-03
-2.29506016e-01 -7.27107346e-01 3.55424643e-01 1.50233328e-01
2.64850289e-01 5.83484590e-01 -1.63597569e-01 -1.45463973e-01
-1.42584777e+00 -1.92147687e-01 -2.71512449e-01 -4.80054736e-01
2.63322413e-01 -3.04876447e-01 -9.38577294e-01 -5.77554286e-01
-2.79369086e-01 -1.08590245e-01 3.02488327e-01 -1.31090051e-02
6.66446924e-01 -1.54002443e-01 -7.78248072e-01 1.25108910e+00
1.69213712e+00 3.38267331e-04 4.49279308e-01 8.11674356e-01
9.82651234e-01 9.79184955e-02 7.43933618e-01 4.40014511e-01
1.03005195e+00 9.86658931e-01 6.67983115e-01 -5.79742081e-02
-2.57226110e-01 -5.51497400e-01 -2.49404665e-02 6.83632433e-01
-1.83726139e-02 2.10662186e-02 -1.17461634e+00 4.47028130e-01
-1.94025445e+00 -5.51161110e-01 -3.03179890e-01 2.39137459e+00
3.82795990e-01 -1.46718279e-01 -3.14700603e-01 -2.08909392e-01
3.12753797e-01 1.97031081e-01 -3.27452034e-01 2.77605742e-01
-1.08204283e-01 2.76274215e-02 9.57736254e-01 9.66043830e-01
-1.09981883e+00 1.08183444e+00 6.41317225e+00 1.03440702e-01
-6.44152105e-01 1.96809635e-01 -2.46385112e-01 -1.79266999e-03
-2.87537754e-01 2.77351648e-01 -8.52215171e-01 1.84494972e-01
9.05946493e-01 -9.22034308e-02 6.02923989e-01 1.23413682e+00
3.86671871e-02 -2.25728840e-01 -1.09324777e+00 1.32164764e+00
-1.22427456e-01 -1.64880574e+00 -2.82677174e-01 3.71135592e-01
6.57416642e-01 4.67520773e-01 -2.43899241e-01 -2.50637680e-01
5.43164313e-01 -5.00101447e-01 1.28350759e+00 3.73173058e-01
7.67411232e-01 -4.71886247e-01 2.80289739e-01 3.87805879e-01
-1.76118588e+00 2.12630942e-01 -4.72881913e-01 -2.96007842e-01
3.84211212e-01 2.19015047e-01 -8.00198317e-01 8.67990017e-01
7.87617147e-01 6.24587178e-01 -5.98629653e-01 1.34918952e+00
-2.15050802e-01 -1.73750892e-01 -7.40525603e-01 1.23015285e-01
-1.73627883e-01 -1.35409497e-02 3.64877582e-01 1.00705421e+00
6.17303252e-01 -2.09981482e-02 6.26595616e-01 5.70923805e-01
-3.05583148e-04 -1.35545418e-01 -7.44475067e-01 7.11274683e-01
1.01073456e+00 9.68629241e-01 -1.02679121e+00 -1.14828967e-01
-2.10728094e-01 1.44065678e+00 4.45620626e-01 1.32346079e-01
-8.32988143e-01 -4.37789448e-02 9.16776061e-01 1.21474251e-01
5.47798946e-02 -1.01513231e+00 -4.15943176e-01 -1.11024535e+00
3.20514441e-01 -2.45144084e-01 -1.34262040e-01 -9.75675642e-01
-7.34104455e-01 5.63055992e-01 -2.93516278e-01 -1.20327306e+00
-3.81423458e-02 -2.55173355e-01 9.04130265e-02 5.69406688e-01
-1.62567294e+00 -1.39272368e+00 -1.07986343e+00 7.24388421e-01
1.37536287e-01 4.21671957e-01 1.18165636e+00 3.36726934e-01
7.84979668e-03 7.17496574e-02 1.78063005e-01 -4.85900417e-02
4.35062885e-01 -1.32093728e+00 9.09499466e-01 1.27472734e+00
5.25619984e-01 7.84818292e-01 7.84809113e-01 -8.24132740e-01
-2.05098629e+00 -1.21186411e+00 7.81062782e-01 -9.97998595e-01
4.43164676e-01 -9.42652047e-01 -3.68059158e-01 1.23337412e+00
-3.90310407e-01 6.16117597e-01 2.46427223e-01 2.26054490e-02
-2.32468680e-01 -1.47342667e-01 -1.14068246e+00 4.08497453e-01
1.73229754e+00 -6.73511147e-01 -2.47418031e-01 5.94369411e-01
9.06353831e-01 -1.32050669e+00 -6.50227189e-01 2.89733201e-01
6.29855454e-01 -1.00849903e+00 1.41892695e+00 1.14451893e-01
-5.95885038e-01 -8.20756018e-01 -9.57596123e-01 -6.68021977e-01
-2.91018605e-01 -7.53826976e-01 -5.46851277e-01 9.62233663e-01
-2.24713385e-01 -7.75712371e-01 9.22825634e-01 8.49950492e-01
-3.38201702e-01 -1.64533466e-01 -1.34087443e+00 -1.11033404e+00
-9.29034472e-01 -6.42208338e-01 7.91058004e-01 5.86388230e-01
-5.42882979e-01 -3.03693652e-01 -3.64525467e-01 9.88158822e-01
1.05828798e+00 2.98952311e-01 1.32178199e+00 -1.13351512e+00
1.01668447e-01 9.22112465e-02 -9.56758082e-01 -1.50666356e+00
4.03948948e-02 -5.18718123e-01 4.76764828e-01 -1.85960925e+00
-1.86510429e-01 -9.59280014e-01 -9.59725380e-02 2.73826331e-01
3.00708383e-01 6.29729033e-01 -4.17336123e-03 3.52149308e-01
-1.18442142e+00 5.38644195e-01 5.85096776e-01 1.95322007e-01
-1.34104401e-01 -9.11562666e-02 -2.93612957e-01 7.04540610e-01
4.71812516e-01 -4.47219431e-01 -1.58808097e-01 -8.15812111e-01
7.27876276e-02 -1.81214109e-01 7.00418472e-01 -1.52933621e+00
4.96314198e-01 -2.42477775e-01 4.63461995e-01 -7.85705745e-01
1.02754712e+00 -1.17728615e+00 6.64820969e-01 4.88183051e-01
1.95742890e-01 3.00329328e-01 1.17380157e-01 6.91169202e-01
2.41494954e-01 3.95294130e-01 3.13869268e-01 -1.18933603e-01
-1.21730232e+00 6.07851624e-01 3.34007621e-01 -4.56782699e-01
1.00626564e+00 -3.82667154e-01 -3.14072728e-01 -2.72007018e-01
-2.73182154e-01 8.47374126e-02 1.43135703e+00 3.67661297e-01
8.29584301e-01 -1.68702149e+00 -3.72163177e-01 3.13507408e-01
3.45810980e-01 3.98314387e-01 -4.00753208e-02 5.31884313e-01
-1.11935747e+00 3.39353800e-01 -1.89934263e-03 -9.21638429e-01
-1.14599073e+00 3.44886214e-01 2.22464785e-01 4.60821331e-01
-9.15882111e-01 9.11577284e-01 -1.11919813e-01 -7.18745232e-01
4.07261193e-01 -3.44352216e-01 7.94479489e-01 -7.24268615e-01
6.51735663e-01 5.11570454e-01 1.54815957e-01 -9.31049109e-01
-1.13559794e+00 7.59505212e-01 6.35275722e-01 -2.32603878e-01
1.21065545e+00 -7.12088287e-01 -3.04225922e-01 5.83512366e-01
9.04275358e-01 5.06754577e-01 -1.59689653e+00 -2.41741210e-01
6.78045154e-02 -1.09321117e+00 -1.82509143e-02 -4.30263579e-01
-2.42129505e-01 1.51147574e-01 7.94952035e-01 -1.21048041e-01
7.66320109e-01 1.83717266e-01 6.11541569e-01 6.57447875e-01
1.54398131e+00 -5.71781754e-01 -2.88419425e-01 5.05442798e-01
6.75980151e-01 -1.25106418e+00 2.91523606e-01 -5.67948520e-01
3.91519219e-02 7.71625280e-01 4.91209567e-01 -9.76688713e-02
2.10610002e-01 6.48842096e-01 1.32475957e-01 -9.91688222e-02
-2.37070978e-01 -1.64988533e-01 9.76185948e-02 1.03792620e+00
-1.14876546e-01 8.01853016e-02 4.86077905e-01 -6.25612736e-02
-3.56803507e-01 -1.55972198e-01 1.65564626e-01 1.38480449e+00
-5.62664270e-01 -9.79920208e-01 -7.10812390e-01 -2.63740271e-01
3.01595122e-01 4.43697460e-02 -1.00173622e-01 6.67494953e-01
-4.04859334e-02 8.04802179e-01 1.30806044e-01 -3.17927867e-01
3.09110731e-01 -2.09903359e-01 7.99147129e-01 -5.74308395e-01
-1.15599763e-02 -1.57131463e-01 -2.72710454e-02 -1.47195017e+00
-4.31463778e-01 -6.97429419e-01 -1.43604243e+00 -4.68708098e-01
-1.27837777e-01 -1.17294863e-01 1.16286123e+00 5.68402529e-01
8.27485919e-01 1.62573591e-01 4.13741499e-01 -1.45279121e+00
1.30259186e-01 -3.17583650e-01 -4.42492813e-01 7.53019750e-02
6.91193581e-01 -6.90845013e-01 -1.38070863e-02 -1.16442651e-01] | [7.4131317138671875, -2.257479429244995] |
4b517bb2-85e2-4edb-973f-010c8976b993 | low-discrepancy-points-via-energetic | 2111.10722 | null | https://arxiv.org/abs/2111.10722v2 | https://arxiv.org/pdf/2111.10722v2.pdf | A Deterministic Sampling Method via Maximum Mean Discrepancy Flow with Adaptive Kernel | We propose a novel deterministic sampling method to approximate a target distribution $\rho^*$ by minimizing the kernel discrepancy, also known as the Maximum Mean Discrepancy (MMD). By employing the general \emph{energetic variational inference} framework (Wang et al., 2021), we convert the problem of minimizing MMD to solving a dynamic ODE system of the particles. We adopt the implicit Euler numerical scheme to solve the ODE systems. This leads to a proximal minimization problem in each iteration of updating the particles, which can be solved by optimization algorithms such as L-BFGS. The proposed method is named EVI-MMD. To overcome the long-existing issue of bandwidth selection of the Gaussian kernel, we propose a novel way to specify the bandwidth dynamically. Through comprehensive numerical studies, we have shown the proposed adaptive bandwidth significantly improves the EVI-MMD. We use the EVI-MMD algorithm to solve two types of sampling problems. In the first type, the target distribution is given by a fully specified density function. The second type is a "two-sample problem", where only training data are available. The EVI-MMD method is used as a generative learning model to generate new samples that follow the same distribution as the training data. With the recommended settings of the tuning parameters, we show that the proposed EVI-MMD method outperforms some existing methods for both types of problems. | ['Chun Liu', 'Lulu Kang', 'Yiwei Wang', 'Yindong Chen'] | 2021-11-21 | null | null | null | null | ['numerical-integration'] | ['miscellaneous'] | [-5.46016097e-02 -1.27220415e-02 -6.80602193e-02 8.06205440e-04
-7.58006215e-01 -1.47279426e-01 5.05237222e-01 -3.75271916e-01
-5.45278609e-01 1.23564029e+00 -2.74959058e-01 -2.37343013e-01
-7.05702826e-02 -9.81483340e-01 -7.14216590e-01 -1.12822962e+00
5.44292629e-01 4.99912471e-01 2.43414015e-01 2.27060914e-02
5.34115911e-01 3.33971202e-01 -1.54215682e+00 -4.05600041e-01
1.39080620e+00 1.02176321e+00 5.08915603e-01 3.25707734e-01
-2.74508148e-01 1.51922449e-01 -5.49212575e-01 -1.19679384e-01
1.65160149e-01 -7.47927845e-01 -5.10554790e-01 -6.05220608e-02
-3.80938023e-01 -1.85586050e-01 1.05086498e-01 1.23289025e+00
7.40915537e-01 6.50758266e-01 1.14825666e+00 -1.12124133e+00
-5.40623009e-01 1.52833223e-01 -7.81302750e-01 3.93645801e-02
-3.79426554e-02 -2.14951336e-02 5.64952552e-01 -9.52164471e-01
5.34578860e-01 1.14179313e+00 6.36338115e-01 7.80258894e-01
-1.37710774e+00 -5.65312803e-01 -7.64039680e-02 5.84038254e-03
-1.88613713e+00 -1.03566758e-01 9.02724087e-01 -5.30058324e-01
4.61047262e-01 2.06792831e-01 6.50146365e-01 8.63620639e-01
3.40293288e-01 5.54958642e-01 1.06938314e+00 -5.42570412e-01
7.38435328e-01 4.15869325e-01 9.38500185e-03 5.50454438e-01
5.11126935e-01 4.87779863e-02 -1.98645964e-01 -5.32476783e-01
7.67795563e-01 -3.33821982e-01 -2.51590282e-01 -3.34323674e-01
-7.53003418e-01 1.15571570e+00 -9.78811830e-02 -2.01298743e-02
-3.60559762e-01 1.30642712e-01 1.79110885e-01 -2.91643322e-01
7.44344115e-01 2.55342126e-01 -1.54267013e-01 -7.76012093e-02
-1.03762233e+00 7.26520181e-01 8.01518321e-01 6.75363064e-01
7.70933270e-01 1.34793893e-01 -2.56136775e-01 1.04225457e+00
8.29810202e-01 8.49429667e-01 4.84761596e-01 -9.65932250e-01
2.73579508e-01 2.21859038e-01 7.53616154e-01 -7.12613940e-01
-4.86401143e-03 -4.64339107e-01 -8.50065172e-01 3.59103829e-01
4.76473331e-01 -4.66205955e-01 -7.01328039e-01 1.89327335e+00
8.16006243e-01 2.24796280e-01 -3.47055234e-02 8.62031639e-01
5.30268908e-01 1.10692525e+00 -1.49059519e-01 -7.83881545e-01
8.10327649e-01 -7.08907783e-01 -7.42055953e-01 2.37830088e-01
2.93434024e-01 -6.61968410e-01 1.09163177e+00 4.35415834e-01
-1.14118123e+00 -3.71942222e-01 -1.02992797e+00 3.97936493e-01
-6.35568276e-02 1.73624039e-01 2.74638772e-01 7.03047037e-01
-7.76891351e-01 6.03598893e-01 -8.14519465e-01 3.96311395e-02
1.79626435e-01 8.52447823e-02 3.70863914e-01 3.08379024e-01
-1.09478128e+00 7.22743511e-01 3.97620559e-01 1.25834212e-01
-8.30929577e-01 -8.35168660e-01 -5.28519571e-01 -1.49872333e-01
2.54507631e-01 -6.75010264e-01 1.15836084e+00 -6.10509634e-01
-2.06458735e+00 5.16497254e-01 -4.76719946e-01 -5.90334423e-02
6.93849683e-01 -3.57379243e-02 -5.73622482e-03 -2.38960847e-01
1.78173721e-01 2.75010049e-01 8.53367448e-01 -1.47250736e+00
-2.03859061e-01 1.33529492e-02 -3.04438293e-01 2.13827193e-01
-3.19792219e-02 -3.79826277e-01 -3.13343614e-01 -6.16504073e-01
5.55520505e-03 -9.25578713e-01 -3.22129965e-01 -2.13862434e-01
-5.40806949e-01 -3.66414487e-01 7.75804937e-01 -4.22089159e-01
1.34139836e+00 -1.88900054e+00 3.04523200e-01 4.69359726e-01
-1.31641790e-01 3.58618885e-01 4.10819888e-01 4.16267276e-01
2.55397499e-01 8.64979252e-02 -5.57976305e-01 -4.38221276e-01
4.19809893e-02 2.05212235e-01 -2.50897944e-01 3.67510796e-01
-2.65940994e-01 5.10839641e-01 -7.56033957e-01 -5.37277877e-01
1.41399920e-01 6.66198194e-01 -5.99380136e-01 4.15112734e-01
-4.60096985e-01 5.99960268e-01 -6.61922753e-01 1.10029131e-01
1.02144825e+00 -3.51744145e-01 2.19637342e-02 -2.37049222e-01
-3.81350517e-01 -1.66977882e-01 -1.59458506e+00 1.33241701e+00
-2.93404043e-01 1.39648229e-01 1.69635996e-01 -1.09625733e+00
1.08789861e+00 2.29976512e-02 3.89571667e-01 -2.39639342e-01
2.13171616e-01 3.83109540e-01 -2.67824441e-01 -3.53957087e-01
3.33282351e-01 -4.86973494e-01 1.03904486e-01 3.48684371e-01
-1.63879663e-01 -2.70414889e-01 2.54870117e-01 6.51983768e-02
3.82218748e-01 2.85272062e-01 2.93623239e-01 -7.22931445e-01
6.89254403e-01 4.96803783e-03 6.78061664e-01 7.72592664e-01
1.81657732e-01 4.39549565e-01 4.85892326e-01 -4.61115837e-02
-1.07173157e+00 -1.15873933e+00 -5.33877611e-01 4.11028951e-01
3.68088424e-01 -2.32703134e-01 -1.07975054e+00 -2.93928266e-01
1.17990896e-01 9.52255309e-01 -5.06703854e-01 -1.45993397e-01
-3.71050715e-01 -1.29175043e+00 8.75157565e-02 5.86626902e-02
6.94576740e-01 -8.01761389e-01 -4.37490880e-01 1.92639664e-01
-1.28566623e-01 -4.31143999e-01 -5.32340109e-01 -1.35073632e-01
-6.74154758e-01 -7.14350820e-01 -1.00192702e+00 -4.34125543e-01
6.36622965e-01 -1.29531547e-01 6.90864205e-01 -1.64993331e-01
-1.73027590e-01 1.18523531e-01 -1.18583061e-01 -2.33669847e-01
-4.73903447e-01 -5.34827225e-02 1.47584379e-01 1.27081290e-01
-9.61093903e-02 -5.73174477e-01 -8.63725305e-01 4.08470601e-01
-6.82988226e-01 1.91057011e-01 3.18438560e-01 9.21522737e-01
1.14272916e+00 2.00461671e-01 7.80756652e-01 -7.91452825e-01
9.11614120e-01 -6.40252054e-01 -9.73645389e-01 1.06645152e-01
-9.31817412e-01 3.91278952e-01 7.30256736e-01 -7.28370547e-01
-1.36121893e+00 -2.14987338e-01 -4.05230969e-01 -4.53695744e-01
2.80327201e-01 5.15148818e-01 -3.48838449e-01 -2.16243416e-01
3.76475722e-01 4.69146103e-01 1.54551178e-01 -6.77257538e-01
2.55423039e-01 6.57727718e-01 2.92849690e-01 -9.76534307e-01
5.72902918e-01 2.52482802e-01 2.15839624e-01 -8.61839890e-01
-7.78649449e-01 4.60649207e-02 -1.08364090e-01 -2.47062474e-01
9.16955531e-01 -4.00504678e-01 -9.28444624e-01 6.88708425e-01
-9.58010018e-01 -5.18002808e-01 -4.13617611e-01 6.59011662e-01
-7.47093856e-01 4.18206930e-01 -4.10533696e-01 -1.20014107e+00
-4.20708001e-01 -1.24932265e+00 7.79068172e-01 5.15476346e-01
6.95012510e-02 -1.07384050e+00 4.43449199e-01 -4.92967106e-03
3.99273753e-01 3.82908612e-01 8.48418474e-01 -2.47312203e-01
-4.05068398e-01 1.94078073e-01 2.67779268e-02 3.65325034e-01
1.17663119e-03 8.94218460e-02 -5.05402446e-01 -3.47456217e-01
3.51428241e-01 -1.97753478e-02 5.88830888e-01 8.32517445e-01
1.31698000e+00 -3.31680924e-01 -4.95756775e-01 7.40741432e-01
1.57473063e+00 5.19004524e-01 5.52478731e-01 9.75919217e-02
5.01279116e-01 1.10208035e-01 7.39500880e-01 8.31967592e-01
1.19285718e-01 7.79803455e-01 1.40620768e-01 2.42946848e-01
2.32320637e-01 -4.24636185e-01 1.64181337e-01 9.04913127e-01
-1.36723444e-01 -3.08105648e-01 -6.74989998e-01 2.29650840e-01
-2.04578304e+00 -9.61148083e-01 -1.55505329e-01 2.55137253e+00
1.02637494e+00 -9.18899477e-03 1.42733485e-01 -2.20673487e-01
8.37814689e-01 -9.87256467e-02 -7.87778437e-01 -2.81312108e-01
2.29698256e-01 2.92796880e-01 3.37628186e-01 7.99057186e-01
-6.48586333e-01 5.79104841e-01 6.49587727e+00 1.33677828e+00
-1.10021722e+00 1.79581419e-01 5.77793717e-01 -2.73553692e-02
-5.17933667e-01 6.71796426e-02 -1.27136171e+00 1.13934290e+00
8.23479950e-01 -3.69121194e-01 4.33493078e-01 7.64086425e-01
6.92278743e-01 -4.21573043e-01 -6.14999235e-01 1.08037174e+00
-1.55190259e-01 -1.39816689e+00 -1.73828881e-02 1.72695816e-01
7.48471677e-01 -5.39100528e-01 -3.18991393e-02 2.37874940e-01
1.39169663e-01 -9.12022352e-01 6.94681227e-01 8.30142975e-01
8.07701468e-01 -8.24388742e-01 5.03729403e-01 6.25299215e-01
-1.15754759e+00 3.55970860e-01 -5.67953706e-01 1.94577038e-01
6.05396390e-01 1.00809491e+00 -3.38600069e-01 4.21367973e-01
4.44695324e-01 3.09025586e-01 8.22405666e-02 1.11725760e+00
-3.56450956e-03 6.80489182e-01 -5.53371131e-01 -3.24823558e-01
-1.66865848e-02 -9.99249756e-01 6.83438838e-01 7.60663450e-01
7.59979546e-01 1.38318360e-01 1.64242927e-02 1.50140989e+00
2.13819876e-01 1.70643970e-01 -3.01082611e-01 1.31107971e-01
6.93266153e-01 9.78656709e-01 -5.58222413e-01 -3.40765268e-01
4.18815687e-02 7.12881207e-01 2.25987390e-01 5.45165241e-01
-1.07973790e+00 -4.36728030e-01 4.98805523e-01 6.96886703e-02
4.00881231e-01 -2.11619452e-01 -7.25152642e-02 -1.08192289e+00
-9.04282928e-02 -4.80071127e-01 8.77266154e-02 -6.95039988e-01
-1.22672319e+00 2.10468277e-01 5.04128575e-01 -8.67841423e-01
-3.95540118e-01 -4.86319423e-01 -7.99046993e-01 1.17930901e+00
-1.08788073e+00 -5.52148521e-01 -1.35814399e-01 5.28627932e-01
3.18746030e-01 -2.46083774e-02 5.71644306e-01 3.20101619e-01
-7.75572598e-01 5.35601735e-01 7.17542410e-01 -4.87788945e-01
3.59631836e-01 -1.06202483e+00 -2.18974441e-01 5.27332485e-01
-5.35828352e-01 5.09814858e-01 9.79660749e-01 -8.62156570e-01
-1.16756260e+00 -9.04109538e-01 4.34604943e-01 9.35689360e-02
2.73214638e-01 -4.08133805e-01 -9.44049954e-01 2.70350248e-01
-1.01475440e-01 -4.32183817e-02 4.77441639e-01 -4.21783596e-01
3.92005533e-01 -5.99254966e-02 -1.49006152e+00 3.88327211e-01
8.05141985e-01 -3.13530900e-02 -2.98050612e-01 3.59510124e-01
6.93567634e-01 -3.89727116e-01 -9.32003677e-01 4.16961670e-01
3.24211270e-01 -8.37067604e-01 8.89947891e-01 -2.76184022e-01
9.03223455e-02 -4.92947012e-01 -2.02604041e-01 -1.47623968e+00
-1.11024536e-01 -8.43921900e-01 -2.02861041e-01 1.31778169e+00
3.66196483e-01 -9.31548119e-01 7.55537450e-01 6.15114868e-01
-7.34392926e-03 -1.18201196e+00 -1.11574924e+00 -7.52534568e-01
2.78486937e-01 -2.36299828e-01 5.28114736e-01 5.22051096e-01
-3.84142429e-01 9.99636278e-02 -5.20914257e-01 -9.69437184e-04
1.00420046e+00 1.75623313e-01 6.26808167e-01 -9.77124572e-01
-6.58322334e-01 -2.13429481e-01 2.53257662e-01 -1.37852919e+00
1.36371329e-01 -6.18622303e-01 1.40624493e-01 -1.47128201e+00
2.44099066e-01 -7.41232276e-01 3.06964405e-02 -2.56968945e-01
-2.61666954e-01 -2.21142903e-01 -1.72140315e-01 1.80472121e-01
-2.17614830e-01 1.08479524e+00 1.37554908e+00 2.83780068e-01
-4.60070342e-01 2.43517503e-01 -3.52220148e-01 5.96400082e-01
7.22116888e-01 -4.23134208e-01 -7.04462826e-01 9.63832662e-02
1.44957229e-01 2.19135046e-01 1.59090966e-01 -8.18700552e-01
-7.35695139e-02 -5.75242460e-01 2.15129793e-01 -7.45115340e-01
4.85424101e-01 -2.98741549e-01 5.23025990e-01 4.03225750e-01
-5.02472222e-02 -2.18647152e-01 3.15370853e-03 6.21668637e-01
9.21380296e-02 -7.59257495e-01 1.12457025e+00 -6.26110733e-02
-1.81244761e-01 2.54836768e-01 -5.55467248e-01 3.67670447e-01
1.08587229e+00 -3.18697169e-02 -1.51806161e-01 -3.17420512e-01
-8.13610852e-01 1.48109449e-02 3.99645895e-01 -3.05341631e-01
5.11577487e-01 -1.58405817e+00 -3.49979818e-01 1.35629177e-01
-3.75884295e-01 2.53733903e-01 2.51007557e-01 9.03046191e-01
-4.21062738e-01 -1.39773926e-02 2.19862252e-01 -5.55987716e-01
-6.99223101e-01 4.59417164e-01 5.84886968e-01 -2.33586192e-01
-4.35565203e-01 7.23685980e-01 1.86613038e-01 -3.42593521e-01
-8.64555091e-02 4.05467190e-02 -1.41150812e-02 -1.75646320e-01
2.48897940e-01 6.04343951e-01 -3.14444423e-01 -4.30090368e-01
-1.78232729e-01 6.46416068e-01 1.95688561e-01 -3.82891834e-01
1.18943429e+00 -2.16458172e-01 -8.32342803e-02 4.33736891e-01
1.13520348e+00 1.37148604e-01 -1.38273537e+00 1.04913741e-01
-4.80973810e-01 -4.42511529e-01 3.11896503e-02 -5.02394080e-01
-9.05529082e-01 6.12533748e-01 5.55802524e-01 1.28076345e-01
7.32810676e-01 -1.92236185e-01 7.87353992e-01 -2.45925458e-03
3.51433665e-01 -1.21417367e+00 -5.78903556e-02 4.10938770e-01
6.64710760e-01 -8.83437693e-01 2.93284971e-02 -4.50549960e-01
-4.42999721e-01 8.78120780e-01 7.78377771e-01 -2.57880300e-01
9.33808088e-01 1.98502302e-01 -3.44900727e-01 1.08165555e-01
-3.94376487e-01 8.78399536e-02 1.91700444e-01 3.41773599e-01
1.90073013e-01 4.77692252e-03 -9.87266660e-01 7.22802281e-01
-8.43346417e-02 1.68335959e-01 2.80315846e-01 6.15894854e-01
-5.36454856e-01 -1.32563841e+00 -4.17063832e-01 5.30413866e-01
-1.13977320e-01 1.86507940e-01 1.19516589e-01 7.80830383e-01
1.55715227e-01 7.49205828e-01 4.24900558e-03 -1.39380500e-01
1.54091194e-02 1.71395093e-01 5.15179932e-01 -3.32005471e-01
1.10062428e-01 1.42974094e-01 -1.37175873e-01 -2.06850648e-01
-2.98424214e-01 -6.51554525e-01 -1.35252595e+00 -4.23850089e-01
-5.06700635e-01 7.27320552e-01 5.82448065e-01 9.52311695e-01
2.90152162e-01 3.02047193e-01 5.46666563e-01 -9.43357468e-01
-9.17789638e-01 -7.86075115e-01 -8.36924076e-01 2.14927778e-01
-1.06984340e-02 -1.22240245e+00 -6.31005704e-01 -3.78624320e-01] | [6.945245265960693, 3.9291462898254395] |
03894c0e-2d5c-4f49-a9ba-c989a3bbbd5c | strong-transcenter-improved-multi-object | 2210.13570 | null | https://arxiv.org/abs/2210.13570v1 | https://arxiv.org/pdf/2210.13570v1.pdf | Strong-TransCenter: Improved Multi-Object Tracking based on Transformers with Dense Representations | Transformer networks have been a focus of research in many fields in recent years, being able to surpass the state-of-the-art performance in different computer vision tasks. A few attempts have been made to apply this method to the task of Multiple Object Tracking (MOT), among those the state-of-the-art was TransCenter, a transformer-based MOT architecture with dense object queries for accurately tracking all the objects while keeping reasonable runtime. TransCenter is the first center-based transformer framework for MOT, and is also among the first to show the benefits of using transformer-based architectures for MOT. In this paper we show an improvement to this tracker using post processing mechanism based in the Track-by-Detection paradigm: motion model estimation using Kalman filter and target Re-identification using an embedding network. Our new tracker shows significant improvements in the IDF1 and HOTA metrics and comparable results on the MOTA metric (70.9%, 59.8% and 75.8% respectively) on the MOTChallenge MOT17 test dataset and improvement on all 3 metrics (67.5%, 56.3% and 73.0%) on the MOT20 test dataset. Our tracker is currently ranked first among transformer-based trackers in these datasets. The code is publicly available at: https://github.com/amitgalor18/STC_Tracker | ['Ben-Zion Bobrovsky', 'Roy Orfaig', 'Amit Galor'] | 2022-10-24 | null | null | null | null | ['multiple-object-tracking-with-transformer'] | ['computer-vision'] | [-1.53025359e-01 -4.12877858e-01 9.45987180e-02 1.17522873e-01
-6.90818727e-01 -6.80548072e-01 8.68438840e-01 -8.28576609e-02
-4.21037853e-01 4.91371274e-01 -1.41904965e-01 4.44836356e-02
-2.14041010e-01 -4.48117286e-01 -7.03396380e-01 -8.00771236e-01
-2.79559493e-01 8.37217331e-01 1.02474201e+00 -1.78826034e-01
-2.27974337e-02 3.11061949e-01 -1.68813014e+00 -9.98585299e-02
3.27494949e-01 1.11950779e+00 1.01524591e-01 9.04404342e-01
2.73933679e-01 5.92487574e-01 -5.63729465e-01 -3.37150693e-01
4.83388036e-01 6.22293577e-02 -3.46979409e-01 -4.98496741e-01
1.11399019e+00 5.21459244e-02 -4.48055655e-01 9.19423163e-01
7.36167133e-01 -2.16744214e-01 4.65321094e-01 -1.56170845e+00
-3.92547995e-01 5.11076808e-01 -6.22354984e-01 5.20398438e-01
1.52032629e-01 2.78491586e-01 7.33641744e-01 -6.64251983e-01
4.95551378e-01 1.34906673e+00 1.20045948e+00 4.04029459e-01
-1.20198822e+00 -1.00437796e+00 -2.73802280e-01 4.44319665e-01
-1.36442149e+00 -4.25208181e-01 1.76412836e-01 -5.66976488e-01
9.51103985e-01 3.00006688e-01 6.39021635e-01 9.87035990e-01
5.93753576e-01 6.50367975e-01 1.16006577e+00 -1.03191644e-01
-1.66621581e-01 2.05912188e-01 2.11361125e-01 7.11125970e-01
5.01422226e-01 6.30655646e-01 -5.38934052e-01 1.03361020e-02
4.97594178e-01 -5.83346449e-02 3.19098420e-02 -4.52938974e-01
-1.60275805e+00 6.76451981e-01 5.80733061e-01 4.02032346e-01
-3.63332152e-01 6.77334607e-01 5.27112007e-01 3.45878750e-01
3.36349070e-01 4.93388660e-02 -2.35452458e-01 -4.01529163e-01
-1.09903586e+00 1.85065448e-01 7.20946610e-01 8.83932352e-01
3.54219615e-01 2.31061101e-01 -3.88909012e-01 2.36540407e-01
6.67863309e-01 1.04839516e+00 2.66711414e-01 -7.70525098e-01
7.18656257e-02 4.75789964e-01 1.68574780e-01 -8.60275865e-01
-3.15459579e-01 -6.83596909e-01 -4.83759999e-01 4.68564004e-01
5.10855556e-01 9.46008712e-02 -8.62942457e-01 1.51677907e+00
6.41993880e-01 6.88378930e-01 5.54495640e-02 6.35558665e-01
9.62660372e-01 5.43945611e-01 -1.05482005e-01 2.89832354e-01
1.61290073e+00 -9.45442021e-01 -5.69579422e-01 5.52348830e-02
4.16063815e-01 -1.08935142e+00 3.13961536e-01 3.37875098e-01
-6.54220939e-01 -6.08704269e-01 -1.09808195e+00 3.78667563e-01
-4.91479307e-01 1.23586178e-01 3.75164092e-01 8.34500432e-01
-1.49828291e+00 5.12906134e-01 -1.24918699e+00 -1.10240662e+00
3.03560913e-01 7.78324246e-01 -1.74352422e-01 8.17340836e-02
-6.49891555e-01 1.00982964e+00 2.69656420e-01 -2.00859711e-01
-1.47966778e+00 -7.49626637e-01 -3.89445901e-01 -1.93627134e-01
3.65733892e-01 -7.46042609e-01 1.15547574e+00 -2.33988762e-01
-1.16569030e+00 6.43039703e-01 9.89166945e-02 -1.03427768e+00
6.89604104e-01 -3.62401783e-01 -5.72702885e-01 -1.82390317e-01
2.05192864e-01 8.70387018e-01 8.45660090e-01 -8.83788347e-01
-9.27924514e-01 -1.62948042e-01 -1.19586587e-02 -2.18442127e-01
-3.42073292e-01 1.16292462e-01 -3.33582848e-01 -3.46448243e-01
-2.82458395e-01 -1.32334793e+00 2.12722689e-01 2.32857183e-01
-1.69144928e-01 -5.03235400e-01 1.59266758e+00 -2.46717364e-01
8.56431425e-01 -1.94022775e+00 -1.21294864e-01 -4.01795983e-01
4.56881434e-01 5.76922834e-01 -1.25439137e-01 5.31241596e-01
2.52255857e-01 -3.92816067e-01 2.76601255e-01 -4.87785965e-01
1.66736513e-01 -1.64620727e-02 -4.99799252e-02 8.60271037e-01
-2.01060131e-01 8.51453483e-01 -9.29722190e-01 -4.49408352e-01
7.11727142e-01 8.53085518e-01 -1.42241254e-01 -1.42312258e-01
-2.89691743e-02 3.87150824e-01 -3.21760386e-01 7.67428935e-01
6.44051611e-01 -2.76181579e-01 -2.82139421e-01 -5.82462430e-01
-4.68918920e-01 -1.03274144e-01 -1.27826464e+00 1.35789156e+00
-7.51714483e-02 1.19487882e+00 -1.27330478e-02 -6.82600737e-01
7.82640219e-01 4.53580618e-01 1.01189208e+00 -4.61854726e-01
3.46260816e-01 3.08905333e-01 1.57135978e-01 -5.63369282e-02
6.45755470e-01 2.08817810e-01 -4.31682430e-02 4.58960831e-02
1.59620360e-01 4.08080310e-01 3.57309341e-01 3.47619742e-01
1.47712851e+00 2.19808057e-01 -1.99042466e-02 -5.88718951e-01
5.04085541e-01 3.47524256e-01 3.69435668e-01 8.76889527e-01
-5.05504847e-01 2.72071004e-01 -3.61279994e-01 -4.75132763e-01
-7.73170233e-01 -1.20654237e+00 -1.46846205e-01 9.08047974e-01
2.87333190e-01 -6.67885423e-01 -4.45898592e-01 -4.80940610e-01
2.82068491e-01 2.48269245e-01 -6.54363275e-01 8.84923190e-02
-7.16327667e-01 -5.23198485e-01 8.95187795e-01 4.60840732e-01
6.54730082e-01 -8.05849075e-01 -1.01234078e+00 3.49869877e-01
-2.02724095e-02 -1.22014964e+00 -3.63695115e-01 8.58123302e-02
-8.02725613e-01 -1.40826142e+00 -5.53978920e-01 -4.85683680e-01
1.35372728e-01 5.05291820e-01 9.95581985e-01 2.74608210e-02
-3.95126224e-01 6.74673796e-01 -2.79112369e-01 -6.70369864e-01
-2.16245368e-01 -6.97606727e-02 4.12709564e-01 -1.33518994e-01
3.12978357e-01 -3.72471094e-01 -5.62629640e-01 6.80427790e-01
-5.76246738e-01 -3.18917930e-01 4.50781971e-01 5.79497933e-01
4.36947167e-01 -1.62230685e-01 -2.63494421e-02 -1.69925928e-01
1.62681509e-02 -3.77501130e-01 -1.12140810e+00 -3.09712309e-02
-6.26556993e-01 -3.38052139e-02 3.53811949e-01 -6.21650875e-01
-3.81016880e-01 2.34250844e-01 7.28468224e-02 -8.29681158e-01
8.49656165e-02 -2.12076172e-01 4.91852194e-01 -6.92800164e-01
6.24963820e-01 2.61867285e-01 -6.21354096e-02 -5.46990812e-01
4.37115356e-02 4.04684037e-01 5.61574042e-01 -2.52798527e-01
1.29384565e+00 8.48825216e-01 3.12576205e-01 -6.86578631e-01
-5.34971893e-01 -9.11037743e-01 -3.89545262e-01 -5.46674967e-01
9.15373147e-01 -1.06546450e+00 -1.05568266e+00 6.12823546e-01
-8.51864874e-01 -1.25400349e-01 -2.52397805e-01 6.95275605e-01
-2.90521353e-01 2.35151738e-01 -2.80828923e-01 -7.25326777e-01
-6.07186139e-01 -1.29290199e+00 1.17652512e+00 4.66100961e-01
1.62710145e-01 -9.25134301e-01 4.52972293e-01 1.54917449e-01
9.72383082e-01 5.47785640e-01 -1.89144135e-01 -5.69018185e-01
-1.08990669e+00 -1.21941216e-01 -1.99672356e-01 -1.55987993e-01
6.36748299e-02 -4.52714339e-02 -9.70135689e-01 -9.04999614e-01
-3.87112051e-01 -5.70215192e-03 1.01210964e+00 5.37927508e-01
1.87603623e-01 6.68679774e-02 -1.07638299e+00 5.53669572e-01
1.74136615e+00 1.13096781e-01 4.18134004e-01 6.29209340e-01
7.04587400e-01 -1.20109826e-01 7.67486632e-01 2.47353718e-01
6.41424835e-01 1.24323273e+00 8.76236856e-01 3.01997513e-01
-6.09157205e-01 1.20738056e-02 7.53790259e-01 5.39062023e-01
1.31610170e-01 -4.88833457e-01 -9.70635712e-01 9.75854635e-01
-1.94111264e+00 -1.03640842e+00 -5.96873045e-01 2.19221091e+00
3.13061886e-02 2.10296497e-01 5.23261428e-01 -9.36699212e-02
6.93567753e-01 5.36567830e-02 -4.20520723e-01 2.11068481e-01
1.79074556e-01 1.00437172e-01 1.00645518e+00 2.69271940e-01
-1.48437893e+00 9.72638309e-01 5.64702559e+00 8.06191564e-01
-1.14989054e+00 4.05389756e-01 -4.55901027e-01 -1.30579144e-01
6.14851177e-01 1.44972339e-01 -1.40586352e+00 5.62030375e-01
1.39963722e+00 -1.12291165e-01 2.22969159e-01 6.43488884e-01
1.09449297e-03 -8.31756964e-02 -9.66769576e-01 1.02939534e+00
-5.28088734e-02 -1.31206191e+00 -3.99896413e-01 2.97250748e-01
3.65635544e-01 8.25036526e-01 8.47722068e-02 4.62614954e-01
6.37634635e-01 -7.11126983e-01 8.46060097e-01 4.20134068e-01
4.52235818e-01 -3.20266008e-01 8.87666702e-01 1.90431908e-01
-2.05344248e+00 8.30607638e-02 -3.99513096e-01 2.49609724e-01
1.69439867e-01 3.34645391e-01 -9.46247458e-01 6.87871695e-01
1.33014858e+00 9.74763453e-01 -8.68771255e-01 1.70919657e+00
4.22920227e-01 7.68672287e-01 -9.46049809e-01 -1.97603837e-01
3.14337403e-01 2.87672013e-01 1.19416821e+00 1.23428261e+00
4.36578244e-01 -5.68988919e-01 3.30648184e-01 3.69838893e-01
1.94206774e-01 -3.62176508e-01 -6.58662260e-01 1.65289238e-01
5.31316400e-01 1.54341388e+00 -6.92178786e-01 -4.09335196e-01
-2.29140759e-01 4.90949035e-01 -1.54914752e-01 -3.01563561e-01
-1.29224730e+00 -1.93610236e-01 7.94330001e-01 1.30134225e-01
9.01192188e-01 -1.38706341e-01 3.24663013e-01 -8.20183694e-01
-1.18274406e-01 -6.78087056e-01 5.04695415e-01 -5.87491632e-01
-9.66638327e-01 7.19526887e-01 2.04131067e-01 -1.62707913e+00
-1.21057309e-01 -6.58108771e-01 -4.95921850e-01 3.82948905e-01
-1.42421865e+00 -1.32289183e+00 -5.41495502e-01 6.01303756e-01
3.99627268e-01 -3.27745229e-01 4.92189854e-01 6.98170722e-01
-3.11841697e-01 7.78742552e-01 5.28892994e-01 -5.52222393e-02
8.22365582e-01 -1.19890082e+00 5.07088959e-01 9.53520238e-01
2.86056787e-01 3.37542385e-01 9.86731768e-01 -5.40272415e-01
-1.88887715e+00 -1.34009290e+00 6.24289870e-01 -9.87179637e-01
9.02435064e-01 -3.66589785e-01 -4.34574246e-01 8.61326754e-01
6.65655911e-01 2.02525482e-01 1.05121396e-01 -2.28030965e-01
-2.23098114e-01 -4.24639404e-01 -1.26591444e+00 2.24576294e-01
9.28620458e-01 2.95608453e-02 -3.18322361e-01 3.71305138e-01
4.53892857e-01 -5.65773189e-01 -1.19680464e+00 4.50929761e-01
7.87477136e-01 -9.93085980e-01 1.10783863e+00 2.87660241e-01
-5.51443994e-01 -1.02486658e+00 -3.41751099e-01 -1.07335770e+00
-4.50852305e-01 -6.61654413e-01 -4.11987692e-01 1.33416700e+00
5.16754836e-02 -8.26563358e-01 8.82206917e-01 -3.62708449e-01
-2.32022926e-01 -3.86630982e-01 -1.25318468e+00 -1.31261957e+00
-2.90232897e-01 -2.38073766e-01 3.43726099e-01 5.14688849e-01
-7.06886947e-01 2.57985085e-01 -5.61762869e-01 3.00439239e-01
1.23761094e+00 -7.46540949e-02 1.14651954e+00 -1.58143270e+00
-1.56637877e-01 -1.50519162e-01 -9.84916627e-01 -8.82187128e-01
-5.08672118e-01 -8.65981102e-01 -1.73664317e-02 -1.59141183e+00
2.36821905e-01 -4.14850295e-01 -4.10405248e-01 5.60509324e-01
3.16840172e-01 6.07030511e-01 7.05542207e-01 4.46380675e-01
-1.06612825e+00 3.92863214e-01 7.71803677e-01 -4.09771979e-01
2.55775392e-01 -5.57928756e-02 -2.40690246e-01 2.47363433e-01
8.02208960e-01 -1.09920919e+00 -3.04274186e-02 -4.30351764e-01
-3.26593637e-01 -3.11910957e-01 8.34990382e-01 -1.83740282e+00
6.81958795e-01 5.05742967e-01 2.19768256e-01 -1.11210239e+00
6.88487113e-01 -1.19408882e+00 8.42540622e-01 1.01069701e+00
3.71263295e-01 6.07188761e-01 7.02535450e-01 6.29773676e-01
1.38675451e-01 2.73645729e-01 8.68020177e-01 1.07273199e-01
-1.03370476e+00 2.99133837e-01 -2.77029723e-01 -7.34365657e-02
1.24470067e+00 -5.01709461e-01 -6.61223114e-01 -1.86900422e-02
-2.92899072e-01 3.78161013e-01 4.04974282e-01 7.23925054e-01
3.01018775e-01 -1.45586193e+00 -9.30399835e-01 -2.43200108e-01
1.84810117e-01 -6.92028463e-01 -8.37205425e-02 1.36708832e+00
-3.49720150e-01 8.11693072e-01 -3.22576135e-01 -1.36850941e+00
-1.86095464e+00 2.61381984e-01 4.80206728e-01 -6.13061428e-01
-7.40855157e-01 6.46462739e-01 -3.43342498e-02 -2.86586195e-01
4.06658351e-01 -2.65132427e-01 -3.07561606e-02 -2.68247455e-01
4.21429068e-01 7.18727946e-01 1.56058297e-02 -8.89015973e-01
-9.74744141e-01 9.37774420e-01 -9.07114968e-02 6.28093928e-02
1.28428602e+00 5.97224869e-02 2.56854117e-01 1.62922218e-01
8.97811890e-01 -1.17233358e-01 -1.19563794e+00 -1.39572337e-01
-2.06157696e-02 -4.15475816e-01 1.54451936e-01 -9.22988772e-01
-1.10247993e+00 6.38713479e-01 1.68024683e+00 2.62030303e-01
8.14227343e-01 8.23607817e-02 7.42246747e-01 2.78593868e-01
7.46518195e-01 -4.87391442e-01 2.38397103e-02 6.08197451e-01
4.53836083e-01 -1.23659945e+00 4.61736359e-02 -5.45375654e-03
-1.01591930e-01 7.74807274e-01 6.86079443e-01 -3.79171222e-01
5.71133256e-01 5.44137001e-01 1.58066377e-01 -5.43163836e-01
-8.06378365e-01 -3.84355724e-01 2.58911937e-01 7.96256065e-01
2.95883507e-01 6.92287162e-02 -1.58399437e-02 -5.08367538e-01
-8.58725384e-02 -2.79992074e-02 1.42287940e-01 9.11935449e-01
-4.82562184e-01 -1.09051132e+00 -7.48332441e-01 3.78264427e-01
-5.10807037e-01 4.41804640e-02 -3.19710344e-01 1.16855121e+00
4.54866961e-02 1.15338945e+00 -2.48649150e-01 -7.97151566e-01
4.15415823e-01 -4.60708082e-01 6.07065916e-01 -1.53053090e-01
-1.10365033e+00 -5.37652299e-02 6.27948418e-02 -6.27906144e-01
-6.75697565e-01 -9.27587986e-01 -9.17406201e-01 -6.59793019e-01
-5.86127639e-01 2.01720491e-01 7.41218567e-01 6.35497808e-01
5.13627827e-01 7.30774343e-01 5.28490171e-02 -9.30427790e-01
-4.32135105e-01 -1.04396558e+00 -2.11858645e-01 -5.46669355e-03
5.31292140e-01 -1.23046768e+00 -1.67499512e-01 -1.84417978e-01] | [6.3498759269714355, -2.048076629638672] |
067c44a9-50bd-4c2b-9a58-5a1a5beea990 | multibodysync-multi-body-segmentation-and | 2101.06605 | null | https://arxiv.org/abs/2101.06605v3 | https://arxiv.org/pdf/2101.06605v3.pdf | MultiBodySync: Multi-Body Segmentation and Motion Estimation via 3D Scan Synchronization | We present MultiBodySync, a novel, end-to-end trainable multi-body motion segmentation and rigid registration framework for multiple input 3D point clouds. The two non-trivial challenges posed by this multi-scan multibody setting that we investigate are: (i) guaranteeing correspondence and segmentation consistency across multiple input point clouds capturing different spatial arrangements of bodies or body parts; and (ii) obtaining robust motion-based rigid body segmentation applicable to novel object categories. We propose an approach to address these issues that incorporates spectral synchronization into an iterative deep declarative network, so as to simultaneously recover consistent correspondences as well as motion segmentation. At the same time, by explicitly disentangling the correspondence and motion segmentation estimation modules, we achieve strong generalizability across different object categories. Our extensive evaluations demonstrate that our method is effective on various datasets ranging from rigid parts in articulated objects to individually moving objects in a 3D scene, be it single-view or full point clouds. | ['Leonidas Guibas', 'Shi-Min Hu', 'Federica Arrigoni', 'Minhyuk Sung', 'Tolga Birdal', 'He Wang', 'Jiahui Huang'] | 2021-01-17 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Huang_MultiBodySync_Multi-Body_Segmentation_and_Motion_Estimation_via_3D_Scan_Synchronization_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Huang_MultiBodySync_Multi-Body_Segmentation_and_Motion_Estimation_via_3D_Scan_Synchronization_CVPR_2021_paper.pdf | cvpr-2021-1 | ['motion-segmentation'] | ['computer-vision'] | [-5.56237735e-02 -1.23788372e-01 4.77311090e-02 -2.51091242e-01
-1.02861452e+00 -9.85151827e-01 4.21644449e-01 -3.17996979e-01
-2.22480759e-01 2.52517057e-03 -1.23320937e-01 1.80017784e-01
-1.75430179e-01 -3.59426022e-01 -9.94657576e-01 -5.22426009e-01
1.38015732e-01 1.27952683e+00 5.10296345e-01 -8.95402208e-02
-1.07149415e-01 8.66950989e-01 -1.33131516e+00 -2.76982784e-01
6.84981585e-01 5.43283105e-01 5.15150912e-02 7.62542903e-01
1.59387454e-01 4.97985110e-02 7.89079815e-03 -2.48285010e-01
7.65400887e-01 2.28390750e-02 -1.22126567e+00 6.09147787e-01
1.07515526e+00 -2.78626382e-01 -2.30105426e-02 9.63548005e-01
3.37951660e-01 3.71309996e-01 4.64863718e-01 -1.07601237e+00
-3.37203443e-01 2.70754218e-01 -1.04022408e+00 -3.05911656e-02
3.79076838e-01 4.42810357e-01 7.89408267e-01 -9.60599780e-01
6.30646825e-01 1.50566673e+00 9.19315398e-01 5.53220987e-01
-1.36091471e+00 -3.48570764e-01 2.98928857e-01 -4.14882958e-01
-1.38205016e+00 -4.50249195e-01 8.16814840e-01 -7.41079032e-01
7.59138703e-01 1.83859900e-01 7.48499990e-01 6.20248795e-01
1.67529404e-01 5.76738715e-01 6.42061234e-01 1.15155376e-01
-1.38862163e-01 -6.66084468e-01 1.43883944e-01 7.35507250e-01
1.73336074e-01 -1.98203892e-01 -3.15973848e-01 -3.21233928e-01
1.13097739e+00 8.67294148e-02 -1.33557871e-01 -8.91841233e-01
-1.67264783e+00 4.26215023e-01 2.04559654e-01 8.32409598e-04
-3.38661671e-01 5.86586297e-01 2.16617450e-01 -7.50542432e-02
4.04700339e-01 3.09994556e-02 -6.31262660e-01 1.77835137e-01
-9.61446524e-01 6.40261948e-01 6.94159508e-01 1.25877678e+00
7.41015375e-01 -3.69474152e-03 3.30992877e-01 4.03236479e-01
6.33140087e-01 8.60953510e-01 9.26119015e-02 -1.26757240e+00
5.69589853e-01 4.04066592e-01 5.39977327e-02 -8.61758173e-01
-7.61372447e-01 -1.60722643e-01 -8.09168816e-01 1.32657155e-01
5.42753518e-01 -2.25215312e-02 -1.16631699e+00 1.84418094e+00
1.03779280e+00 2.95361847e-01 -2.19490170e-01 1.30076718e+00
7.87974477e-01 3.56961727e-01 -1.06015772e-01 9.47353840e-02
1.54641688e+00 -1.01694667e+00 -2.23938331e-01 -4.13313538e-01
3.02843869e-01 -7.35074878e-01 8.37946951e-01 -2.89903767e-02
-1.54979372e+00 -5.42267442e-01 -6.30789101e-01 -4.50906843e-01
1.69692710e-01 -3.34617764e-01 4.40596759e-01 2.13352203e-01
-1.01667917e+00 7.14950800e-01 -1.40543246e+00 -2.47261062e-01
2.55274594e-01 7.51477003e-01 -6.04539871e-01 1.55376419e-01
-4.84954566e-01 7.41286874e-01 1.67706609e-01 1.98853776e-01
-7.80077994e-01 -7.78863132e-01 -1.04531085e+00 -2.53193319e-01
4.29792255e-01 -1.46037102e+00 1.33779800e+00 -8.21563005e-01
-1.44979978e+00 1.27715695e+00 -1.75460890e-01 -1.22031301e-01
8.83217037e-01 -4.21731144e-01 1.80579439e-01 7.42416605e-02
1.47441342e-01 6.66986048e-01 7.20311701e-01 -1.38176107e+00
-3.03250104e-01 -7.65349030e-01 -4.03213203e-01 5.74022412e-01
4.91653919e-01 8.47033113e-02 -8.81352425e-01 -5.32098651e-01
7.80018866e-01 -1.45357788e+00 -3.29157680e-01 1.96721822e-01
-5.81622779e-01 -1.39496595e-01 9.36849415e-01 -6.96246266e-01
2.50117362e-01 -1.79404855e+00 9.10727739e-01 2.97356546e-01
2.14431956e-01 -2.18384758e-01 -1.32207230e-01 -1.71612099e-01
-1.96026154e-02 -7.03010336e-02 -5.99062920e-01 -6.95124924e-01
-2.07425822e-02 4.01700079e-01 -9.91051570e-02 9.83331382e-01
1.65025383e-01 1.21839833e+00 -7.42581427e-01 -6.67118490e-01
3.06078732e-01 4.22790766e-01 -4.74241078e-01 1.13053709e-01
-2.83421397e-01 1.03647590e+00 -5.25050640e-01 9.01346207e-01
8.25871706e-01 -4.74622935e-01 -1.64934903e-01 -3.25278074e-01
-4.46773507e-02 -2.99835969e-02 -1.64531481e+00 2.30055833e+00
-8.48534778e-02 9.53644738e-02 4.68186319e-01 -5.57304740e-01
4.35538620e-01 3.12020540e-01 1.08471584e+00 -2.02674475e-02
4.65931185e-02 1.41966134e-01 -1.71493068e-01 -2.42548048e-01
5.31906545e-01 -2.84550130e-01 -2.10051417e-01 4.09885824e-01
1.73584580e-01 -5.58632970e-01 -2.33936921e-01 -1.49634713e-02
6.12432480e-01 6.91354215e-01 9.15802643e-02 -4.31882262e-01
2.81200528e-01 2.12924361e-01 7.27244616e-01 3.47352505e-01
-3.70823704e-02 1.11611128e+00 -2.76650637e-01 -5.99861503e-01
-9.98884559e-01 -1.26661503e+00 -5.74337989e-02 8.19206953e-01
6.50377572e-01 -1.02473898e-02 -4.44664240e-01 -3.21321994e-01
1.65181205e-01 2.79787206e-03 -1.82222977e-01 2.57876456e-01
-1.11793268e+00 -5.47573209e-01 2.69816279e-01 5.69620013e-01
2.41211742e-01 -5.64476192e-01 -8.49846661e-01 2.32596770e-01
-4.61096108e-01 -1.41671479e+00 -1.04005480e+00 8.90099071e-03
-1.18864274e+00 -1.12793291e+00 -7.64222980e-01 -7.09560692e-01
5.34575284e-01 5.76937616e-01 1.27021194e+00 -1.67730823e-02
-8.43048692e-02 7.52508700e-01 1.45631552e-01 6.30994216e-02
-3.45515102e-01 6.16509393e-02 2.43505269e-01 -2.48457611e-01
-2.40626901e-01 -8.09637845e-01 -5.55996895e-01 6.40450001e-01
-8.09896886e-01 1.90359324e-01 1.99642703e-01 2.46193185e-01
1.19152415e+00 -6.19163036e-01 -2.04048693e-01 -5.42981982e-01
-1.40004918e-01 -3.30328137e-01 -7.28197396e-01 3.63971561e-01
-1.17440693e-01 -2.04275288e-02 1.25901759e-01 -5.76291263e-01
-6.78918719e-01 6.63881063e-01 -1.90976351e-01 -9.20276344e-01
-2.53142715e-01 1.32796198e-01 -2.25474104e-01 -4.25153106e-01
3.23704123e-01 1.08839296e-01 1.41039178e-01 -4.95503664e-01
7.12529182e-01 -1.21552676e-01 1.16974521e+00 -1.01121747e+00
1.16733968e+00 9.06036556e-01 2.14743838e-01 -5.65696597e-01
-5.49159646e-01 -8.82553875e-01 -1.27669144e+00 -1.99850917e-01
1.31847620e+00 -1.17389953e+00 -7.16382027e-01 6.54427290e-01
-1.40477872e+00 -4.07700926e-01 -2.62179673e-01 3.87312680e-01
-9.18547869e-01 7.62115657e-01 -7.83009946e-01 -2.80587703e-01
-5.87630093e-01 -1.39982188e+00 1.77663779e+00 3.25344168e-02
-3.69312704e-01 -9.07313943e-01 3.14864844e-01 4.94334131e-01
-2.52826542e-01 6.93997025e-01 4.73128319e-01 -3.43585461e-01
-9.33650851e-01 -3.58484243e-03 2.07905352e-01 -2.76203394e-01
2.01774448e-01 2.18205407e-01 -6.00723743e-01 -5.69226444e-01
3.86832170e-02 -2.44696796e-01 4.56336200e-01 5.46964109e-01
8.40131760e-01 -2.24427670e-01 -3.36848378e-01 1.19342637e+00
1.24051440e+00 -3.99476945e-01 2.23050892e-01 1.69511080e-01
1.38335407e+00 5.47288358e-01 3.75715196e-01 1.44241273e-01
8.32741082e-01 1.01810205e+00 6.49961352e-01 -2.59967744e-01
-1.15302905e-01 -1.78697873e-02 1.36469811e-01 1.14198685e+00
-3.08373451e-01 2.11050157e-02 -1.11997032e+00 6.40549958e-01
-1.92674851e+00 -7.83654749e-01 -4.36294675e-01 2.10858274e+00
6.81113899e-01 -2.90915281e-01 4.58006322e-01 -5.98505795e-01
6.79326355e-01 4.66254465e-02 -8.53298783e-01 2.14582846e-01
-9.79456082e-02 -3.68296131e-02 5.30338883e-01 4.29263949e-01
-1.17908394e+00 9.93447185e-01 5.94409943e+00 1.70370638e-01
-9.36630845e-01 2.27904305e-01 9.46607515e-02 -1.97786540e-01
-3.77316594e-01 1.13782592e-01 -7.18523264e-01 1.24176972e-01
4.12150413e-01 6.61009774e-02 3.37263614e-01 6.50386870e-01
2.61190683e-02 3.28689098e-01 -1.30267489e+00 1.08996940e+00
-3.46249267e-02 -1.13918662e+00 1.47016451e-01 1.32352442e-01
7.69309521e-01 5.33491492e-01 -2.21932963e-01 -2.97314703e-01
6.83447778e-01 -8.22906971e-01 1.27753675e+00 4.71364528e-01
4.26370859e-01 -4.42497820e-01 1.26141295e-01 5.31605363e-01
-1.42680192e+00 5.82173467e-01 -1.57420233e-01 3.82652819e-01
6.76289439e-01 3.34554315e-01 -1.73108459e-01 9.68889177e-01
7.65920401e-01 6.92062914e-01 -1.99406981e-01 9.92583752e-01
1.02511033e-01 8.51757526e-02 -7.49541461e-01 8.98263156e-01
6.44620955e-02 -4.25409943e-01 9.00895715e-01 9.04778659e-01
3.23204398e-01 3.72360706e-01 5.73144495e-01 1.08391774e+00
-5.90974512e-03 -2.56911993e-01 -4.18953955e-01 4.25017416e-01
2.83217520e-01 1.36327195e+00 -1.05329621e+00 -1.23657018e-01
-3.69448811e-01 1.03420091e+00 2.73690701e-01 2.07693622e-01
-8.86135459e-01 4.94381428e-01 8.77158582e-01 7.87235498e-02
2.42882803e-01 -8.20092380e-01 -3.19359988e-01 -1.47383523e+00
1.46008089e-01 -6.85725331e-01 4.75262195e-01 -8.08981240e-01
-1.20994163e+00 2.67970949e-01 2.58910388e-01 -1.24326158e+00
-2.13088751e-01 -2.55331218e-01 -4.41099674e-01 8.10728669e-01
-1.17223334e+00 -1.62844360e+00 -4.03795570e-01 8.13821554e-01
5.90877593e-01 4.89874840e-01 3.68823975e-01 2.08240137e-01
-4.05584961e-01 1.73959136e-01 -2.68366277e-01 2.20546704e-02
5.92110753e-01 -1.21679854e+00 7.42675602e-01 1.01715612e+00
2.99417138e-01 6.79361343e-01 5.97148418e-01 -6.46998107e-01
-1.87663043e+00 -1.13203442e+00 2.15736985e-01 -9.67803478e-01
3.32520992e-01 -2.34215662e-01 -1.18923855e+00 1.12491047e+00
-1.28694326e-01 2.61111736e-01 2.86166608e-01 -1.04939006e-01
-2.33970031e-01 3.21881175e-01 -1.09771168e+00 5.17851233e-01
1.40371406e+00 -2.67893523e-01 -7.08014607e-01 4.08445388e-01
9.08170283e-01 -1.39374077e+00 -1.10602844e+00 5.86605132e-01
4.27545875e-01 -5.71672738e-01 1.50939572e+00 -5.74171841e-01
2.67450839e-01 -6.85791731e-01 -1.38661668e-01 -1.05204868e+00
-2.74505347e-01 -7.58933663e-01 -2.48396471e-01 9.48582947e-01
-1.12672001e-02 -3.94376010e-01 7.47942269e-01 9.29650664e-01
-4.95873302e-01 -4.12925333e-01 -1.20271754e+00 -7.62199163e-01
3.37669134e-01 -4.28780645e-01 5.69893718e-01 1.09593868e+00
-6.77549720e-01 8.82434174e-02 -3.95324022e-01 7.68745363e-01
1.00542855e+00 6.60387576e-01 1.24477708e+00 -1.40313315e+00
-3.79077494e-01 -4.16303456e-01 -3.42897832e-01 -1.34494960e+00
2.73653120e-01 -9.58598614e-01 4.13535297e-01 -1.42847908e+00
3.15319896e-01 -3.66161883e-01 3.58159065e-01 5.73887527e-01
-8.39297771e-02 1.69861570e-01 2.82486051e-01 7.85670877e-01
-6.04708493e-01 4.02706385e-01 1.49373198e+00 -1.79110602e-01
-2.78510213e-01 9.49478745e-02 -3.94434869e-01 1.00282812e+00
2.36208871e-01 -5.20055413e-01 -2.18166605e-01 -1.00799835e+00
1.98709276e-02 4.01806772e-01 1.01902449e+00 -8.27817261e-01
4.37347442e-01 -5.13952076e-01 2.04782501e-01 -9.08121109e-01
4.00427610e-01 -9.13971722e-01 7.30612695e-01 2.95246512e-01
1.63269788e-01 4.41016406e-01 3.79789025e-01 5.97013175e-01
7.54615441e-02 1.76072940e-01 9.63899791e-01 -4.16260779e-01
-7.15468347e-01 7.83056378e-01 2.07246721e-01 2.48248398e-01
9.88072574e-01 -5.15096724e-01 -3.16804312e-02 1.59899350e-02
-9.34741318e-01 6.56936109e-01 9.74164188e-01 5.47874331e-01
5.37974238e-01 -1.26100016e+00 -6.45854354e-01 7.28449151e-02
5.90744754e-03 1.07593024e+00 3.66742373e-01 1.14383352e+00
-5.79031587e-01 -1.96825732e-02 -6.79960754e-03 -1.38438678e+00
-1.46631324e+00 1.51568964e-01 7.88369000e-01 -6.96821883e-03
-1.00593555e+00 8.43672216e-01 2.76721448e-01 -1.00089192e+00
-1.17743716e-01 -7.13910282e-01 4.94949877e-01 -5.41033149e-01
-6.74566478e-02 3.37390333e-01 6.13316819e-02 -1.17233551e+00
-6.19388878e-01 1.48204613e+00 2.42418826e-01 -1.39079943e-01
1.33209741e+00 -3.52453589e-01 -2.51167297e-01 4.49368864e-01
1.05042756e+00 -1.34007975e-01 -1.56659603e+00 -2.77018845e-01
-1.74194485e-01 -2.92097241e-01 -3.36558938e-01 -2.38906547e-01
-1.14240885e+00 6.20940387e-01 3.66508245e-01 -2.20319301e-01
6.23323560e-01 5.37290335e-01 9.62511837e-01 2.47325301e-01
5.82526505e-01 -7.39045143e-01 -7.08913803e-02 6.53136134e-01
8.38464737e-01 -1.12642598e+00 1.90898478e-01 -5.89253426e-01
-4.38967168e-01 1.06905818e+00 6.36584282e-01 -1.53444141e-01
3.73515517e-01 1.67332634e-01 3.59102413e-02 -5.63186169e-01
-3.05880964e-01 -7.48852417e-02 7.32344627e-01 3.72534782e-01
1.15238227e-01 -1.36372015e-01 1.79690704e-01 1.51982471e-01
-2.03943729e-01 -2.17063665e-01 2.21705884e-01 9.75071251e-01
-9.40035060e-02 -8.95092189e-01 -6.97582543e-01 1.86814507e-03
-3.38125020e-01 4.67896193e-01 -3.73178184e-01 9.00274932e-01
-9.06766430e-02 4.53488827e-01 2.08113432e-01 -1.76143005e-01
5.29666185e-01 -6.08666800e-02 7.60491550e-01 -6.55956626e-01
-5.84420800e-01 3.78776550e-01 -3.26398760e-01 -8.23356926e-01
-9.46366310e-01 -1.03507400e+00 -1.61688662e+00 -2.59016544e-01
-2.59832084e-01 -2.43189305e-01 4.18009371e-01 1.20489705e+00
3.90755713e-01 2.97037274e-01 2.09030107e-01 -1.49571371e+00
-6.57480896e-01 -5.50783575e-01 -3.13889652e-01 7.75628269e-01
5.35698056e-01 -6.09573305e-01 -1.05326161e-01 3.49537522e-01] | [8.076740264892578, -2.5034186840057373] |
bd90f8ce-520c-42e8-9c64-3ab6af7a1cf5 | bc-vad-a-robust-bone-conduction-voice | 2212.02996 | null | https://arxiv.org/abs/2212.02996v1 | https://arxiv.org/pdf/2212.02996v1.pdf | BC-VAD: A Robust Bone Conduction Voice Activity Detection | Voice Activity Detection (VAD) is a fundamental module in many audio applications. Recent state-of-the-art VAD systems are often based on neural networks, but they require a computational budget that usually exceeds the capabilities of a small battery-operated device when preserving the performance of larger models. In this work, we rely on the input from a bone conduction microphone (BCM) to design an efficient VAD (BC-VAD) robust against residual non-stationary noises originating from the environment or speakers not wearing the BCM.We first show that a larger VAD system (58k parameters) achieves state-of-the-art results on a publicly available benchmark but fails when running on bone conduction signals. We then compare its variant BC-VAD (5k parameters and trained on BC data) with a baseline especially designed for a BCM and show that the proposed method achieves better performances under various metrics while keeping the realtime processing requirement for a microcontroller. | ['Milos Cernak', 'Damien Ronssin', "Niccolo' Polvani"] | 2022-12-06 | null | null | null | null | ['activity-detection'] | ['computer-vision'] | [ 9.49356258e-02 -2.34751999e-01 2.33800426e-01 1.57539278e-01
-1.09835601e+00 -2.63561279e-01 2.20510155e-01 -3.01774949e-01
-3.80498022e-01 3.58971268e-01 2.39545748e-01 -4.07451272e-01
1.60127595e-01 -1.85301885e-01 -6.45062864e-01 -9.11791921e-01
3.15660425e-02 5.89434393e-02 5.45498908e-01 3.65061462e-02
-3.23678166e-01 4.89387482e-01 -1.99084997e+00 1.67047054e-01
4.13347334e-01 1.32995045e+00 3.00677150e-01 1.07247508e+00
2.22658336e-01 5.38595259e-01 -1.30115736e+00 2.63261765e-01
5.05749928e-03 -5.12943387e-01 -1.64118916e-01 -4.81504381e-01
4.42254126e-01 -1.04457974e-01 -3.93786609e-01 6.29083514e-01
1.35856581e+00 1.27481483e-02 4.51308042e-01 -1.18317473e+00
9.37725827e-02 6.06037974e-01 8.72868462e-04 2.56376952e-01
1.88905984e-01 9.58535373e-02 6.62759364e-01 -8.68250072e-01
8.13996792e-02 9.23484445e-01 9.55806553e-01 9.56776023e-01
-1.17294800e+00 -7.67556369e-01 -2.56366193e-01 4.74916309e-01
-1.54695344e+00 -1.00185931e+00 1.13136411e+00 -7.97076002e-02
1.40263784e+00 6.36815488e-01 7.02623785e-01 1.54766273e+00
-6.81534782e-02 6.68368578e-01 8.43985498e-01 -4.78084981e-01
8.92378688e-01 4.68304753e-02 -7.08097592e-02 3.07038572e-04
-1.80529669e-01 -1.29819095e-01 -1.01935542e+00 -3.18895310e-01
3.71619433e-01 -7.43211091e-01 -6.31510675e-01 -5.06691970e-02
-9.58189070e-01 4.35282201e-01 -1.11068465e-01 5.03228366e-01
-5.02125204e-01 3.60569268e-01 4.14565802e-01 3.34551573e-01
2.26128176e-01 6.70540035e-02 -5.18531084e-01 -7.48984456e-01
-1.23509228e+00 -1.24105290e-01 1.01145267e+00 6.16686165e-01
1.68085527e-02 7.38473475e-01 -1.51721865e-01 1.12583375e+00
3.43912572e-01 7.82518029e-01 8.41584206e-01 -7.82452047e-01
2.48617381e-01 -1.09377347e-01 -2.49589875e-01 -4.59792793e-01
-5.40977836e-01 -6.55666232e-01 -8.14043760e-01 2.84229577e-01
3.00720572e-01 -1.98301181e-01 -9.65615153e-01 1.70076311e+00
2.62710482e-01 4.99358833e-01 6.29974976e-02 7.90378153e-01
9.76949453e-01 7.50651360e-01 -3.69752735e-01 -6.02978706e-01
9.47722435e-01 -1.02557838e+00 -1.14570177e+00 -2.67873585e-01
1.34972394e-01 -6.29337430e-01 1.28011394e+00 1.08696711e+00
-1.09149909e+00 -6.92681611e-01 -1.37924528e+00 1.61899939e-01
7.03025460e-02 -1.72258631e-04 -8.33690539e-03 1.38063121e+00
-1.40262365e+00 2.71709114e-01 -9.69083846e-01 -8.73903483e-02
3.75970267e-02 5.51988423e-01 -5.81376329e-02 5.53744256e-01
-1.08842874e+00 6.95102811e-01 -4.27495152e-01 1.31336734e-01
-1.44722772e+00 -6.35493338e-01 -4.80683178e-01 -5.08232899e-02
1.86067984e-01 -3.24251771e-01 1.49921978e+00 -7.37352610e-01
-2.36657643e+00 2.68354654e-01 -2.26281285e-01 -6.55125916e-01
5.90376675e-01 -4.45339352e-01 -8.39132965e-01 1.37610003e-01
-6.27348185e-01 1.82615727e-01 1.38617945e+00 -9.26912725e-01
-4.15260315e-01 -3.63634303e-02 -4.81559098e-01 -1.87204629e-01
-6.25281155e-01 -1.06836639e-01 -3.87871742e-01 -6.37686610e-01
2.72070151e-02 -9.08915222e-01 2.16815025e-01 -1.37855589e-01
-4.11322445e-01 -1.21832490e-01 1.02595711e+00 -8.78768444e-01
1.52405894e+00 -2.33515573e+00 5.22756279e-02 1.53799355e-01
-1.26194537e-01 6.92439079e-01 -6.29547313e-02 3.54970619e-02
-8.72691870e-02 -3.02619666e-01 -2.35193625e-01 -7.23126888e-01
1.46912159e-02 7.32783135e-03 -2.32186429e-02 5.55388093e-01
-1.43556476e-01 3.45050395e-01 -6.09006993e-02 -1.68231159e-01
2.50241876e-01 9.34618473e-01 -5.18396497e-01 3.23820591e-01
2.19147861e-01 3.06388617e-01 2.92999029e-01 5.66796601e-01
5.09483337e-01 4.27130431e-01 -4.52100672e-02 -7.41405264e-02
2.34811883e-02 7.41878867e-01 -1.36402249e+00 1.76482475e+00
-8.89768720e-01 1.00717866e+00 7.36696124e-01 -8.53382528e-01
9.27920997e-01 8.87801290e-01 2.77802318e-01 -5.84297836e-01
3.74577522e-01 6.31298721e-01 2.06124589e-01 -4.22868311e-01
8.72972831e-02 1.57458261e-01 2.01452628e-01 3.64039212e-01
2.44634360e-01 -2.43965104e-01 -6.10902965e-01 -3.16989630e-01
1.38176477e+00 -2.92063415e-01 -3.55224386e-02 -2.70501316e-01
6.09585166e-01 -5.61773539e-01 4.13578004e-01 7.48008370e-01
-4.92520869e-01 6.71320200e-01 6.11156747e-02 1.66319221e-01
-7.36025393e-01 -1.10540128e+00 -1.44676879e-01 9.33921814e-01
-3.30803841e-01 -3.62968922e-01 -1.19247770e+00 1.10619115e-02
-2.73200750e-01 8.15429509e-01 -2.49601752e-01 -1.26102358e-01
-6.52168989e-01 -5.34438431e-01 1.16994083e+00 5.18649399e-01
2.48178661e-01 -9.53788936e-01 -7.96740592e-01 5.74157655e-01
-1.79722279e-01 -1.10903990e+00 -1.58105671e-01 7.59436488e-01
-6.18807793e-01 -3.71538103e-01 -7.58237302e-01 -8.19542766e-01
-3.33616853e-01 -2.52628382e-02 8.45966935e-01 -3.50340039e-01
-3.31347920e-02 5.68042696e-01 -9.66808274e-02 -9.34450626e-01
-7.64139473e-01 5.55608906e-02 6.69995725e-01 3.12492531e-02
2.89087474e-01 -1.12239766e+00 -3.93362552e-01 2.43995294e-01
-5.43699443e-01 -3.44960362e-01 4.13331151e-01 6.23395920e-01
4.96820390e-01 2.57385433e-01 8.16275179e-01 -1.61529351e-02
6.92374825e-01 -1.64081648e-01 -3.23161662e-01 -1.97712526e-01
-4.52034265e-01 -2.26249546e-01 5.04147470e-01 -1.03426027e+00
-7.27272391e-01 2.26652101e-01 -7.51853347e-01 -5.87528884e-01
-2.04418391e-01 -6.19041314e-03 -8.06352079e-01 5.79554029e-02
6.88589394e-01 3.19979459e-01 -1.44913509e-01 -8.62256527e-01
3.15207914e-02 1.47187364e+00 9.27624464e-01 -1.71650171e-01
5.49092591e-01 2.70809680e-01 -1.61390483e-01 -1.38400280e+00
3.18367518e-02 -3.02274466e-01 -3.28547925e-01 -3.34488392e-01
5.49600005e-01 -1.17325985e+00 -6.84602499e-01 1.00809908e+00
-9.90562618e-01 -5.86566865e-01 -2.01299712e-01 6.83076024e-01
-5.73689342e-01 6.53520897e-02 -5.83092570e-01 -1.19607997e+00
-7.52931833e-01 -1.07191610e+00 8.02702785e-01 -5.65481000e-02
-2.92557865e-01 -6.09693527e-01 2.44624391e-01 3.70011032e-01
6.64786518e-01 -1.48638487e-01 7.38887608e-01 -5.83304703e-01
8.92539471e-02 -9.81049091e-02 8.12273920e-01 8.40261579e-01
1.36542335e-01 -1.76278815e-01 -1.98027456e+00 -1.30900085e-01
5.99332035e-01 -2.58656032e-02 7.44786143e-01 7.43466139e-01
9.73363340e-01 -1.01001605e-01 3.04608047e-02 3.62141430e-01
9.12904501e-01 2.72860944e-01 6.93858445e-01 1.50866345e-01
5.28854191e-01 2.78812766e-01 1.63825259e-01 2.92879969e-01
-8.08913559e-02 7.65237212e-01 5.50795913e-01 -2.24188138e-02
-5.10275900e-01 3.91345732e-02 8.65449369e-01 1.44117141e+00
1.04518898e-01 -5.61420500e-01 -7.47231543e-01 7.38351047e-01
-1.38474631e+00 -5.36322474e-01 -3.29190046e-01 2.31213403e+00
9.85011160e-01 3.30560237e-01 3.62036735e-01 1.22940278e+00
6.02446795e-01 5.85905500e-02 -7.16162145e-01 -7.82940924e-01
-2.30894953e-01 7.40305901e-01 1.69079363e-01 3.94328535e-01
-8.27005088e-01 3.87492210e-01 6.67864084e+00 1.13324738e+00
-1.51220977e+00 5.68571568e-01 4.71928390e-03 -6.94455981e-01
5.12708984e-02 -8.74812245e-01 -5.55791616e-01 3.10301572e-01
1.86514711e+00 3.37879360e-01 5.59095204e-01 1.01948845e+00
4.44312364e-01 -3.49692367e-02 -1.20629692e+00 1.39168322e+00
1.42425463e-01 -7.54465878e-01 -4.38500851e-01 2.43108682e-02
2.88269609e-01 -1.84034556e-02 1.91138953e-01 2.57930994e-01
-4.68514979e-01 -1.12182832e+00 1.13747692e+00 7.64478743e-02
9.00545776e-01 -7.32811987e-01 6.65330827e-01 4.47679609e-01
-1.17295873e+00 -8.88531432e-02 -1.82814315e-01 -1.41363814e-01
9.53567550e-02 7.18227446e-01 -9.94904160e-01 1.05664402e-03
8.62723231e-01 1.06586359e-01 -4.22718555e-01 1.03778005e+00
-1.38518423e-01 1.39635110e+00 -5.97208023e-01 -2.47117087e-01
-1.84324324e-01 6.09080553e-01 8.03911924e-01 1.20835137e+00
5.04648983e-01 -3.49662393e-01 -5.83654046e-01 3.43723804e-01
-2.04613637e-02 -7.69020692e-02 -9.58432108e-02 2.37972513e-01
6.44297838e-01 9.48130608e-01 -1.35649279e-01 -1.22031480e-01
-3.51662785e-01 9.23123777e-01 -2.71791667e-01 1.13602221e-01
-9.54534531e-01 -4.08299446e-01 1.11445677e+00 8.76221731e-02
4.00687695e-01 -1.56411737e-01 -1.82336494e-01 -6.07667923e-01
1.92250460e-01 -1.09644008e+00 -2.61530522e-02 -8.37807715e-01
-7.94724226e-01 8.10208857e-01 -4.72747952e-01 -1.34684551e+00
-3.09794962e-01 -6.87248290e-01 -4.56737161e-01 7.43703365e-01
-1.13756955e+00 -6.97795749e-01 -2.79841542e-01 8.80858123e-01
6.75399840e-01 -1.48707241e-01 1.12122416e+00 7.23943233e-01
-5.45721889e-01 8.87824833e-01 2.19525114e-01 -1.25981674e-01
7.33946800e-01 -1.06931996e+00 3.75696421e-01 8.96489382e-01
3.02863598e-01 2.50594199e-01 1.03028238e+00 -1.05544560e-01
-1.58204758e+00 -9.03257430e-01 4.90144134e-01 -4.05646265e-01
3.70263070e-01 -8.90285552e-01 -9.23136413e-01 1.07848093e-01
3.09413552e-01 -7.98678771e-02 6.56419218e-01 -2.21778870e-01
-1.38948202e-01 -4.36490297e-01 -1.04697788e+00 3.95527780e-01
1.00232816e+00 -9.02368426e-01 -5.45583546e-01 -9.27156135e-02
9.10229027e-01 -4.13043678e-01 -6.23830974e-01 2.05177858e-01
7.72174776e-01 -9.35297310e-01 8.30334902e-01 -8.55416358e-02
-4.29521114e-01 -3.76061171e-01 -3.67529631e-01 -1.41564357e+00
1.56478211e-01 -1.02611446e+00 -6.69168055e-01 1.41692650e+00
3.60322267e-01 -4.72500205e-01 6.29601121e-01 1.47044659e-01
-2.81021953e-01 -4.83110219e-01 -1.68241668e+00 -1.18973780e+00
-8.09188485e-02 -1.29152942e+00 4.78279203e-01 2.88294256e-01
-4.59570857e-03 3.51546466e-01 -5.78274608e-01 4.34822768e-01
3.44583869e-01 -6.19189858e-01 7.10220039e-01 -1.13912845e+00
-3.89128715e-01 -1.63045898e-01 -5.85362434e-01 -9.45821464e-01
-1.11190397e-02 -2.58028090e-01 4.27817464e-01 -1.06122434e+00
-4.58990455e-01 -1.13459088e-01 -5.02586484e-01 1.92295745e-01
2.53255337e-01 4.19043601e-01 -5.83343506e-02 -8.61262307e-02
-1.42934158e-01 5.14830649e-01 5.97987771e-01 -3.45374435e-01
-7.41433799e-01 1.83322862e-01 -1.30585328e-01 7.63947308e-01
7.03871310e-01 -6.71381176e-01 -4.06967938e-01 -2.51939088e-01
-2.10615367e-01 1.07187234e-01 2.70696551e-01 -1.83167589e+00
4.62085187e-01 4.67242151e-01 7.70704597e-02 -6.19328916e-01
1.02345753e+00 -1.00197339e+00 5.22379041e-01 5.26690423e-01
-8.36708918e-02 -2.74979711e-01 4.16157335e-01 4.08673435e-01
-1.72306836e-01 2.43792590e-02 6.89993501e-01 4.90246624e-01
-2.73977041e-01 -3.05289447e-01 -1.02512598e+00 -1.64005280e-01
6.21994436e-01 -1.45334288e-01 -2.83319056e-01 -7.10040748e-01
-7.32904136e-01 -5.39372683e-01 -1.05584078e-01 4.70584780e-01
4.75297570e-01 -1.19657969e+00 -5.35556376e-01 4.08989727e-01
-1.52049735e-01 -3.81299779e-02 2.43453249e-01 9.41837490e-01
-4.34771746e-01 5.04994154e-01 -1.51575888e-02 -8.55547190e-01
-1.65812731e+00 2.95137435e-01 7.28212237e-01 3.26216578e-01
-5.39504468e-01 1.00502741e+00 -2.75880218e-01 -2.23240610e-02
9.89406705e-01 -6.14985287e-01 -6.60263523e-02 5.56774139e-02
7.36851156e-01 6.79576218e-01 7.92106390e-01 -5.88393033e-01
-6.66282058e-01 3.98647279e-01 7.44072735e-01 -6.01790428e-01
1.32469535e+00 -1.07405670e-01 3.40243667e-01 8.89852762e-01
1.06034780e+00 5.55649519e-01 -9.81653869e-01 -4.82359789e-02
-4.29077983e-01 -1.83020439e-02 4.49493706e-01 -7.24064708e-01
-1.09520018e+00 1.27430868e+00 1.39894700e+00 1.32036805e-01
1.54907286e+00 -2.09011868e-01 1.09216404e+00 3.34344357e-01
4.60212708e-01 -1.16666281e+00 1.43424451e-01 1.73046231e-01
8.17579150e-01 -5.16161680e-01 -3.97667408e-01 -1.82980612e-01
-4.25105929e-01 1.01468170e+00 2.44166777e-01 1.96089849e-01
1.01160347e+00 6.20909393e-01 6.02310777e-01 3.78390610e-01
-8.01705718e-01 9.22712460e-02 2.07472846e-01 9.05183911e-01
1.18419351e-02 -7.38859829e-03 3.24923657e-02 1.14575636e+00
-5.47212720e-01 -2.57857710e-01 5.83214760e-01 7.54698396e-01
-4.48499501e-01 -7.64168918e-01 -6.77402377e-01 1.60783291e-01
-6.96723104e-01 -1.14228964e-01 -4.20966327e-01 4.71465319e-01
1.44305006e-01 1.64263022e+00 5.37045263e-02 -7.92396188e-01
4.41992491e-01 3.27434540e-01 2.06784174e-01 -1.57011181e-01
-9.51232910e-01 5.74107766e-01 1.89598322e-01 -5.81082761e-01
-4.51370776e-01 -6.12880111e-01 -1.03114772e+00 -2.23022774e-01
-5.18299997e-01 -1.32075874e-02 1.17733693e+00 5.64935386e-01
3.66071761e-01 9.98134792e-01 5.18611968e-01 -8.61816704e-01
-3.64685297e-01 -1.40988469e+00 -8.77998292e-01 -1.74646571e-01
7.61672676e-01 -5.20616412e-01 -8.36434722e-01 -5.14285825e-02] | [14.834747314453125, 5.8069844245910645] |
db0d4160-ff23-4e81-a696-053f91b0e6c0 | cost-effective-task-offloading-scheduling-for | 2306.14588 | null | https://arxiv.org/abs/2306.14588v1 | https://arxiv.org/pdf/2306.14588v1.pdf | Cost-Effective Task Offloading Scheduling for Hybrid Mobile Edge-Quantum Computing | In this paper, we aim to address the challenge of hybrid mobile edge-quantum computing (MEQC) for sustainable task offloading scheduling in mobile networks. We develop cost-effective designs for both task offloading mode selection and resource allocation, subject to the individual link latency constraint guarantees for mobile devices, while satisfying the required success ratio for their computation tasks. Specifically, this is a time-coupled offloading scheduling optimization problem in need of a computationally affordable and effective solution. To this end, we propose a deep reinforcement learning (DRL)-based Lyapunov approach. More precisely, we reformulate the original time-coupled challenge into a mixed-integer optimization problem by introducing a penalty part in terms of virtual queues constructed by time-coupled constraints to the objective function. Subsequently, a Deep Q-Network (DQN) is adopted for task offloading mode selection. In addition, we design the Deep Deterministic Policy Gradient (DDPG)-based algorithm for partial-task offloading decision-making. Finally, tested in a realistic network setting, extensive experiment results demonstrate that our proposed approach is significantly more cost-effective and sustainable compared to existing methods. | ['Dusit Niyato', 'Han Yu', 'Minrui Xu', 'Yue Xiao', 'Yulan Gao', 'Ziqiang Ye'] | 2023-06-26 | null | null | null | null | ['decision-making'] | ['reasoning'] | [ 1.71637595e-01 -2.05280602e-01 -5.20174861e-01 9.65175033e-02
-6.37484074e-01 -3.17598224e-01 -2.89948046e-01 -5.09260774e-01
-4.03402716e-01 9.89689469e-01 -2.22106948e-01 -7.12625861e-01
-4.53122377e-01 -6.79377854e-01 -4.73431915e-01 -9.04457331e-01
-1.71860531e-01 4.22965407e-01 -2.94031590e-01 -2.39771828e-01
2.38859444e-03 2.84764707e-01 -9.50706899e-01 -4.86449271e-01
1.21544993e+00 1.29332709e+00 5.15865803e-01 5.06552875e-01
5.50921619e-01 5.09367466e-01 3.39861983e-03 -3.59851390e-01
2.41586819e-01 -3.90534371e-01 -7.74003446e-01 1.80199414e-01
-5.44381559e-01 -2.96056211e-01 -3.97250682e-01 1.03500545e+00
1.09667861e+00 2.08710238e-01 2.01673359e-01 -1.54219127e+00
-2.56004244e-01 2.10525930e-01 -4.93560672e-01 4.60470438e-01
-1.44745603e-01 1.71937510e-01 1.10325289e+00 -5.77869177e-01
5.55653274e-01 7.32703269e-01 2.08445027e-01 6.75218761e-01
-1.06849468e+00 -4.74334329e-01 4.27501462e-02 4.63463694e-01
-1.12790465e+00 -4.59474981e-01 6.49881959e-01 -8.84816423e-02
7.50348389e-01 3.39823782e-01 7.95846820e-01 5.91075480e-01
3.92809004e-01 7.45134950e-01 7.62541831e-01 -2.61516243e-01
6.31141841e-01 -2.45683193e-01 -3.95174742e-01 4.64390457e-01
2.89963990e-01 1.75076589e-01 -4.01089102e-01 -1.66994944e-01
5.70220172e-01 1.50931049e-02 -2.71659762e-01 -3.42166483e-01
-9.87180531e-01 4.61208880e-01 4.50025439e-01 -8.42293352e-02
-8.32461119e-01 6.98591352e-01 4.45179433e-01 3.29063147e-01
6.94579542e-01 1.05791524e-01 -5.37392974e-01 -4.49978113e-01
-8.57348323e-01 2.31392264e-01 6.51176929e-01 1.08795547e+00
4.66928214e-01 8.24097916e-02 -5.14570951e-01 2.76201159e-01
3.61123145e-01 6.54712021e-01 -1.47076219e-01 -1.27864003e+00
5.35829067e-01 -7.09473491e-02 4.99614626e-01 -7.68273711e-01
-5.88025928e-01 -1.19336915e+00 -7.86312997e-01 -3.45794320e-01
-4.16386537e-02 -7.36875594e-01 -2.55426556e-01 1.87328732e+00
3.10083330e-01 5.74733317e-01 -2.24179160e-02 1.43200052e+00
-1.50171980e-01 6.74774230e-01 4.28159535e-02 -9.39341724e-01
1.35883641e+00 -9.38394904e-01 -9.46121991e-01 -5.15744425e-02
8.13354492e-01 -4.69319969e-01 7.46085227e-01 -1.12701487e-02
-1.34620416e+00 8.59570801e-02 -9.93291140e-01 3.77782464e-01
3.28485519e-01 1.34510621e-01 7.34650970e-01 8.47315848e-01
-1.04871893e+00 5.29301941e-01 -7.35953927e-01 -8.63585472e-02
4.44940627e-01 8.94147158e-01 5.35463929e-01 -2.99779624e-02
-1.32534552e+00 3.70343626e-01 4.87369299e-02 5.15913427e-01
-9.46750224e-01 -4.58957404e-01 -3.05164009e-01 4.27084684e-01
1.05272579e+00 -1.16266584e+00 1.41252828e+00 -8.28730881e-01
-1.90422070e+00 3.66318882e-01 -1.96781784e-01 -3.78891498e-01
4.41463143e-01 1.34708837e-01 -2.75864542e-01 1.91045329e-01
2.61311978e-01 2.96627562e-02 8.96574378e-01 -9.80385363e-01
-7.19604075e-01 -3.41558069e-01 4.20868158e-01 6.29697621e-01
-4.39931691e-01 -5.51060215e-02 -4.66607630e-01 -2.61665642e-01
-2.76882231e-01 -1.39044690e+00 -4.85143036e-01 -4.94042367e-01
-3.80207002e-01 -5.84399030e-02 5.23059726e-01 -5.28250456e-01
1.66017306e+00 -1.97788346e+00 4.84301388e-01 1.81572616e-01
4.77998465e-01 2.91878209e-02 4.96741291e-03 3.10202420e-01
6.32683277e-01 1.84333965e-01 3.74999531e-02 -4.51571226e-01
1.08517826e-01 1.66911483e-01 6.39540702e-02 3.85967642e-01
-7.66134188e-02 1.05893922e+00 -1.17656624e+00 -4.70552623e-01
-1.74124971e-01 2.39677038e-02 -8.64915013e-01 7.10549131e-02
-1.42062128e-01 4.86341268e-01 -1.02442181e+00 7.96226740e-01
6.99490130e-01 -5.11475921e-01 5.69406152e-01 -4.78887223e-02
-5.83504215e-02 7.67861530e-02 -7.77866304e-01 1.47462082e+00
-9.23714280e-01 2.40682274e-01 5.56832373e-01 -1.08270621e+00
2.05890968e-01 3.18229377e-01 8.41170788e-01 -9.00581479e-01
4.05366689e-01 6.53718770e-01 9.89942104e-02 -5.56569338e-01
5.96001744e-01 -2.61052012e-01 -1.53112113e-01 5.56840658e-01
-3.35173666e-01 3.41600001e-01 -8.09040591e-02 1.93909287e-01
1.00879681e+00 7.99057931e-02 -2.00855374e-01 -4.91529733e-01
4.78836775e-01 -2.42263466e-01 9.48076248e-01 5.30162752e-01
-7.82466352e-01 -5.30661233e-02 7.33907342e-01 -5.40725850e-02
-9.04859722e-01 -5.33234835e-01 2.11469382e-01 1.36258805e+00
6.33035183e-01 -1.52152359e-01 -9.01389897e-01 -4.79995430e-01
-2.24351719e-01 3.04677546e-01 -2.01958969e-01 -2.21066996e-01
-5.60343385e-01 -9.84985352e-01 -7.80088753e-02 1.66952699e-01
3.92054260e-01 -7.28240907e-01 -6.71977341e-01 4.99918342e-01
-4.54777986e-01 -1.37329888e+00 -1.02162910e+00 3.17969322e-01
-7.98911989e-01 -3.60850900e-01 -9.82106030e-01 -8.66571903e-01
5.19844651e-01 6.39507413e-01 9.12076116e-01 3.96382250e-02
1.71762943e-01 2.03340590e-01 -2.11113662e-01 1.46864533e-01
2.95967817e-01 5.43365836e-01 3.26849848e-01 4.08149093e-01
-1.88496366e-01 -6.03604913e-01 -1.00593841e+00 3.75623047e-01
-5.88657379e-01 5.60842641e-02 5.03895223e-01 9.84166563e-01
8.04676950e-01 1.43974602e-01 1.12449527e+00 -5.51618457e-01
8.22059453e-01 -6.67939663e-01 -6.32786810e-01 3.20809811e-01
-7.59606719e-01 -3.03608160e-02 9.28488553e-01 -1.05751336e-01
-7.56584823e-01 -6.45904765e-02 1.74906448e-01 -4.25678492e-01
8.72786820e-01 5.35174489e-01 -2.51681507e-01 -2.59460062e-01
3.31216939e-02 2.06710249e-01 -2.11946025e-01 1.54548869e-01
8.54144022e-02 9.06139851e-01 -9.23345685e-02 -6.99053824e-01
5.66348433e-01 4.71226066e-01 4.82276380e-01 -3.55690747e-01
-9.47729588e-01 -1.80023357e-01 -4.25658263e-02 -5.40629327e-01
7.11934388e-01 -9.70369160e-01 -1.42025518e+00 1.88631434e-02
-9.16435003e-01 -4.73565608e-01 1.39164664e-02 5.76284587e-01
-1.01830554e+00 1.39478296e-01 -4.84443933e-01 -1.15857875e+00
-3.77653509e-01 -1.40321445e+00 9.97243464e-01 3.79793048e-01
7.05908954e-01 -8.10680807e-01 -3.43322337e-01 5.00967443e-01
7.23159075e-01 -6.84503838e-02 5.45676768e-01 2.85888702e-01
-8.64941299e-01 1.07964501e-01 -5.61146915e-01 -5.28578050e-02
-2.82645106e-01 -5.26357532e-01 -4.64528590e-01 -6.63695633e-01
-7.16968104e-02 -2.06340790e-01 4.81030822e-01 7.52687454e-01
1.29519367e+00 -1.10964105e-01 -4.48916227e-01 7.96774507e-01
1.47493231e+00 1.95832133e-01 2.77323484e-01 2.52675056e-01
6.27567053e-01 3.31562668e-01 9.53311503e-01 8.78447771e-01
7.25861847e-01 8.94840777e-01 8.95319521e-01 1.64744735e-01
4.17784423e-01 1.75726071e-01 4.22507912e-01 1.07763028e+00
-2.95795530e-01 -6.65500104e-01 -5.30355394e-01 3.03147942e-01
-2.31377077e+00 -6.07861757e-01 2.30508596e-02 2.27063465e+00
3.81219178e-01 2.54348636e-01 2.13850915e-01 -1.28399178e-01
9.06990886e-01 -5.28278574e-02 -1.03802037e+00 -4.12052602e-01
1.37351230e-01 -1.79066688e-01 8.25873017e-01 2.67065942e-01
-8.14076006e-01 9.19859350e-01 4.94653940e+00 1.16872454e+00
-1.14306903e+00 5.70906043e-01 7.99649179e-01 -3.52118254e-01
-4.11288977e-01 1.04763016e-01 -3.59224349e-01 7.26934671e-01
1.14644885e+00 -6.59380555e-01 1.18716764e+00 5.51064551e-01
9.52133775e-01 3.93294804e-02 -5.64921439e-01 1.14077866e+00
-5.48137188e-01 -1.24867475e+00 -7.55724549e-01 5.33159435e-01
7.49968708e-01 1.18648901e-03 2.11415991e-01 4.11500901e-01
-4.07098770e-01 -5.63636005e-01 8.32252085e-01 4.13307190e-01
1.05112481e+00 -1.08844471e+00 7.77998567e-01 2.40525380e-01
-1.29627466e+00 -4.87082303e-01 -2.96398431e-01 -3.39007676e-01
5.69393098e-01 7.66438961e-01 -6.39124662e-02 8.33369672e-01
2.34304816e-01 6.97954893e-01 3.02025497e-01 1.03564298e+00
1.76782995e-01 5.68815649e-01 -5.52282259e-02 -4.82224762e-01
3.17913949e-01 -4.12399173e-01 6.42228067e-01 6.33391142e-01
2.88798839e-01 1.48202136e-01 4.30966079e-01 7.60557652e-01
-4.75629449e-01 -2.89717782e-02 -1.87841982e-01 -2.37265676e-01
6.81997418e-01 1.56827760e+00 -9.81667578e-01 -9.77252126e-02
-9.69272256e-02 1.14958084e+00 1.36272341e-01 5.41765809e-01
-1.19396460e+00 -5.27863145e-01 7.91653872e-01 -1.42732888e-01
6.13477603e-02 -3.62562567e-01 -2.96246439e-01 -1.27727878e+00
1.93099752e-01 -2.92020023e-01 -9.01102871e-02 -5.42377830e-01
-9.94319916e-01 4.76763964e-01 -6.82273328e-01 -1.55428195e+00
1.09443448e-01 -2.82301813e-01 -5.37626028e-01 8.67794693e-01
-2.02654338e+00 -7.11348295e-01 3.87233905e-02 4.32980537e-01
2.00733498e-01 1.33270994e-01 3.09834421e-01 1.02619433e+00
-1.04826212e+00 5.33986807e-01 4.82115924e-01 -5.60036063e-01
2.46713996e-01 -1.08795321e+00 -2.20466964e-02 6.55057192e-01
-7.12254882e-01 1.27487630e-01 6.07075512e-01 -5.41297853e-01
-2.20488763e+00 -1.11455750e+00 6.70254946e-01 2.71182865e-01
7.43639410e-01 -4.47360605e-01 -8.13139752e-02 1.17295429e-01
-1.22521900e-01 3.72786880e-01 4.70727473e-01 -1.40114039e-01
6.04465246e-01 -2.20326245e-01 -9.89949107e-01 6.18916452e-01
1.41014647e+00 -4.59838867e-01 6.80967450e-01 8.21226120e-01
1.17488313e+00 -4.36393172e-01 -6.28007293e-01 2.30413139e-01
4.83908743e-01 -3.15784156e-01 6.51804328e-01 -6.34462237e-01
2.12094933e-01 -1.27851605e-01 -2.02340811e-01 -1.19015837e+00
-4.70364958e-01 -1.45633852e+00 -4.09919620e-01 6.60718381e-01
3.72086704e-01 -4.84551609e-01 1.09642184e+00 3.15366238e-01
-3.52077812e-01 -1.35614800e+00 -1.23870039e+00 -9.31802094e-01
-3.37310910e-01 -3.89481574e-01 3.87091398e-01 6.72154725e-01
2.12056637e-02 4.50704843e-01 -8.59948635e-01 1.83112636e-01
7.67137051e-01 2.18148872e-01 1.51736677e-01 -7.13145971e-01
-5.28637469e-01 -2.81520844e-01 2.46727746e-02 -1.18396115e+00
2.32426748e-01 -9.38974857e-01 1.31804973e-01 -1.46695399e+00
2.82964438e-01 -8.30254853e-01 -5.23213565e-01 -3.89250107e-02
-2.10080996e-01 4.42960151e-02 4.00193334e-01 3.90865952e-01
-1.30235755e+00 1.01678753e+00 1.66176522e+00 2.22748503e-01
-2.74818152e-01 4.55808878e-01 -6.88640714e-01 5.18719368e-02
8.42207551e-01 -6.26748681e-01 -4.30992126e-01 -4.66558695e-01
7.71138132e-01 8.82434964e-01 3.06660645e-02 -7.06745148e-01
1.43548220e-01 -4.64886636e-01 -5.75109959e-01 -3.14610779e-01
2.61542529e-01 -8.41634393e-01 -2.75933057e-01 8.88130069e-01
-1.81209911e-02 3.51747647e-02 -3.42710048e-01 9.15797412e-01
2.94272602e-01 -2.28711031e-02 4.39341307e-01 3.25427234e-01
-3.95218283e-01 7.98544288e-01 -4.74330366e-01 2.00719729e-01
1.14905655e+00 8.29876363e-02 -1.06648825e-01 -4.71029460e-01
-9.04086411e-01 9.08182502e-01 6.76050410e-02 2.02915128e-02
2.28108257e-01 -1.41359997e+00 -4.28722262e-01 -4.53924507e-01
-2.16247946e-01 -5.50760686e-01 6.74588323e-01 1.43994176e+00
-3.89013678e-01 5.67734182e-01 -7.74940774e-02 -3.14208269e-01
-5.91181219e-01 4.02370334e-01 6.02464199e-01 -6.33871734e-01
7.24467561e-02 6.90011382e-01 -3.21258932e-01 -2.83601433e-01
5.78958425e-04 -1.77036952e-02 1.79465786e-01 -7.21980259e-02
-1.10259056e-01 7.70960212e-01 6.11866675e-02 -3.63978416e-01
-6.21060789e-01 2.67762214e-01 3.68693113e-01 -2.01388448e-01
1.14155912e+00 -6.42328620e-01 -1.13007888e-01 -3.48351151e-02
1.12663877e+00 -1.73297942e-01 -1.20055103e+00 -2.00325385e-01
-1.05896965e-01 -1.97411433e-01 6.86517656e-01 -5.95088959e-01
-1.20628715e+00 6.81482553e-01 6.94835663e-01 5.98215103e-01
1.40093303e+00 -4.25668061e-01 1.44687510e+00 1.51536524e-01
7.31716752e-01 -1.36681390e+00 -1.36689823e-02 5.38609445e-01
9.63025540e-02 -1.11763084e+00 -2.56010950e-01 -4.28741068e-01
-4.79904950e-01 9.33276534e-01 5.32731771e-01 3.13765518e-02
7.01834202e-01 -1.11765943e-01 -5.32828629e-01 -1.03426194e-02
-9.80674863e-01 -5.43411076e-01 -9.81676430e-02 1.69416234e-01
2.72382796e-01 7.56864071e-01 -8.99382710e-01 7.44715095e-01
3.57582420e-01 4.31915671e-01 6.46411657e-01 1.05039525e+00
-3.30073774e-01 -1.16013670e+00 1.03907637e-01 6.40050530e-01
-5.90359747e-01 -1.69911876e-01 3.19916964e-01 2.40522753e-02
1.56441867e-01 9.10247922e-01 -2.01099753e-01 -5.68696141e-01
1.98091399e-02 -4.99080926e-01 2.18179092e-01 -4.87920761e-01
-4.21709627e-01 2.15907320e-01 7.54548758e-02 -6.28895998e-01
-3.44429910e-01 -3.77919108e-01 -1.26875031e+00 -3.87736499e-01
-6.99048400e-01 2.70298511e-01 7.58453786e-01 1.07021022e+00
1.05189347e+00 1.03393292e+00 1.14176071e+00 -1.02734065e+00
-7.00599432e-01 -3.02513778e-01 -8.06489050e-01 -2.96285033e-01
2.42074326e-01 -6.94916129e-01 -2.07570523e-01 -7.15571046e-01] | [5.869668960571289, 1.742750644683838] |
039985d3-0806-4ad7-b729-72470f5fb778 | pillarnext-rethinking-network-designs-for-3d | 2305.04925 | null | https://arxiv.org/abs/2305.04925v1 | https://arxiv.org/pdf/2305.04925v1.pdf | PillarNeXt: Rethinking Network Designs for 3D Object Detection in LiDAR Point Clouds | In order to deal with the sparse and unstructured raw point clouds, LiDAR based 3D object detection research mostly focuses on designing dedicated local point aggregators for fine-grained geometrical modeling. In this paper, we revisit the local point aggregators from the perspective of allocating computational resources. We find that the simplest pillar based models perform surprisingly well considering both accuracy and latency. Additionally, we show that minimal adaptions from the success of 2D object detection, such as enlarging receptive field, significantly boost the performance. Extensive experiments reveal that our pillar based networks with modernized designs in terms of architecture and training render the state-of-the-art performance on the two popular benchmarks: Waymo Open Dataset and nuScenes. Our results challenge the common intuition that the detailed geometry modeling is essential to achieve high performance for 3D object detection. | ['Xiaodong Yang', 'Chenxu Luo', 'Jinyu Li'] | 2023-05-08 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Li_PillarNeXt_Rethinking_Network_Designs_for_3D_Object_Detection_in_LiDAR_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_PillarNeXt_Rethinking_Network_Designs_for_3D_Object_Detection_in_LiDAR_CVPR_2023_paper.pdf | cvpr-2023-1 | ['2d-object-detection'] | ['computer-vision'] | [-2.17226014e-01 -3.07327569e-01 4.10420373e-02 -1.98142916e-01
-6.06335342e-01 -6.30882502e-01 6.19651794e-01 2.38638669e-02
-2.89981037e-01 1.62584215e-01 -3.23021770e-01 -4.12815630e-01
-6.38833269e-02 -1.00098300e+00 -1.05658567e+00 -3.73209208e-01
-3.06357205e-01 6.04785323e-01 6.77810550e-01 -2.15874732e-01
3.48430753e-01 1.30439079e+00 -1.82938635e+00 -7.38682672e-02
7.27831364e-01 1.27027404e+00 1.99661866e-01 4.31865484e-01
-3.77222657e-01 -5.92590757e-02 -3.67954820e-01 -1.59172997e-01
8.03278625e-01 7.70707846e-01 -3.33727509e-01 -2.58823127e-01
9.72599924e-01 -6.59432352e-01 -1.87649995e-01 9.52247858e-01
6.37295783e-01 1.66958332e-01 6.56234145e-01 -1.12737036e+00
-4.62816089e-01 2.01564968e-01 -8.00290704e-01 3.35073113e-01
-9.75451320e-02 2.87355453e-01 1.03609550e+00 -1.46110010e+00
3.88949990e-01 1.57248306e+00 9.62478101e-01 3.45434874e-01
-9.88841116e-01 -6.81618333e-01 5.66913843e-01 -1.80630103e-01
-1.80474627e+00 -3.57492417e-01 6.55177414e-01 -2.62400210e-01
1.37627029e+00 2.96050698e-01 4.93951648e-01 5.41000426e-01
3.48249897e-02 5.49024999e-01 4.56558675e-01 -1.33243680e-01
2.31601581e-01 -2.72610188e-01 1.37087420e-01 5.90179145e-01
7.95360804e-01 1.02573834e-01 -3.66457880e-01 -2.72887439e-01
1.13747489e+00 1.35221675e-01 1.98000759e-01 -6.98928833e-01
-1.04077959e+00 7.56931961e-01 9.71192181e-01 -9.15461630e-02
-3.73506099e-01 3.63753647e-01 2.39813551e-01 -6.40066713e-02
6.87192082e-01 3.45812500e-01 -4.09687608e-01 2.51176119e-01
-7.50906229e-01 5.13670444e-01 4.41349000e-01 1.31174088e+00
9.96685147e-01 8.61910284e-02 1.02679636e-02 5.83245397e-01
3.73498410e-01 7.94066608e-01 -4.57674921e-01 -1.10311902e+00
5.90543389e-01 9.56064343e-01 2.27492273e-01 -9.37951207e-01
-5.08582711e-01 -7.55544662e-01 -7.41328120e-01 5.48754394e-01
1.23870127e-01 1.56360686e-01 -1.15690029e+00 1.25947618e+00
6.66220307e-01 2.39753261e-01 -4.17006820e-01 9.97309029e-01
1.15822911e+00 5.38405895e-01 -4.71430384e-02 4.61924106e-01
1.25163555e+00 -5.53113639e-01 2.41102159e-01 -3.79584342e-01
5.81136465e-01 -6.45121634e-01 1.01937747e+00 -2.58331634e-02
-1.17323685e+00 -6.32906914e-01 -1.06035388e+00 -2.95092791e-01
-4.26476002e-01 -8.84630382e-02 8.49237919e-01 4.78120118e-01
-1.02004588e+00 4.55385089e-01 -8.67490053e-01 -4.63046640e-01
9.29595053e-01 6.35549307e-01 -9.16461740e-03 -1.25807911e-01
-4.19185519e-01 6.62136078e-01 5.11343665e-02 -8.81261379e-02
-6.57933891e-01 -1.12622702e+00 -4.34542090e-01 2.64133900e-01
4.35176730e-01 -1.14986360e+00 1.24329793e+00 2.11054325e-01
-9.58458543e-01 1.02523172e+00 -1.62467062e-01 -6.23874605e-01
3.15203518e-01 -3.08306873e-01 2.43570805e-01 6.27081990e-02
1.89560816e-01 1.07258976e+00 7.11107254e-01 -1.25258422e+00
-9.75599766e-01 -7.57471025e-01 2.69547224e-01 2.63449669e-01
-1.66220367e-01 -1.68888241e-01 -6.30298257e-01 -2.16439605e-01
3.69302511e-01 -7.86331594e-01 -4.23026174e-01 3.15796494e-01
-3.17338556e-01 -7.00203896e-01 1.01516008e+00 3.49247843e-01
6.19333208e-01 -2.07868838e+00 -3.35236907e-01 1.53183416e-01
6.84721649e-01 1.80254087e-01 -7.26488605e-02 2.11531445e-01
2.34933913e-01 3.40973288e-01 1.52922079e-01 -5.99521697e-01
1.52229637e-01 1.32492721e-01 -4.67423022e-01 4.67270643e-01
3.00231844e-01 1.07375622e+00 -5.54497421e-01 -4.72279340e-01
5.89595854e-01 5.58195055e-01 -7.86272585e-01 -8.80261660e-02
-2.54474282e-01 9.87826847e-03 -8.70734274e-01 1.06857908e+00
1.15075755e+00 -5.35449743e-01 -5.55195332e-01 -2.24027634e-01
-4.22131866e-01 3.88414055e-01 -1.13466191e+00 1.60773492e+00
-2.67123044e-01 2.84979582e-01 3.74444246e-01 -4.59811807e-01
8.86936247e-01 -1.86771825e-01 5.64711988e-01 -4.51610535e-01
1.15471847e-01 3.00051123e-01 -1.89722866e-01 3.28105569e-01
6.32561624e-01 1.60049796e-01 1.17909402e-01 1.37786105e-01
-1.32306680e-01 -5.62193274e-01 -1.69413447e-01 1.45233706e-01
1.04866076e+00 -4.44719382e-02 1.32808834e-01 -3.94922644e-01
8.37298557e-02 2.40598395e-01 2.78552145e-01 1.22677231e+00
-2.39729937e-02 6.15816236e-01 -1.54765062e-02 -8.01570952e-01
-1.01825702e+00 -1.21860492e+00 -3.78834963e-01 9.96538699e-01
3.54433089e-01 -4.41696048e-01 -4.31956202e-01 -4.83988017e-01
5.16801238e-01 3.14688236e-01 -2.48773947e-01 2.23313913e-01
-6.25708401e-01 -6.43150866e-01 3.96432370e-01 6.71854198e-01
4.16083425e-01 -5.78451753e-01 -8.50153506e-01 1.32724166e-01
3.55504602e-01 -1.27067673e+00 -8.16159397e-02 2.66899407e-01
-1.25841641e+00 -7.54308939e-01 -4.53056604e-01 -4.45962757e-01
4.62771714e-01 1.08805180e+00 1.31770253e+00 2.17921510e-01
-3.50092471e-01 2.63667464e-01 -2.00565383e-01 -9.36411023e-01
3.96278501e-01 5.19630492e-01 1.90514311e-01 -6.83347583e-01
6.41363621e-01 -9.08900976e-01 -7.70627618e-01 3.20245385e-01
-5.90328515e-01 -1.55481562e-01 6.25197887e-01 1.90465912e-01
9.00807500e-01 -6.29287362e-02 6.30184785e-02 -3.64527702e-01
2.48641580e-01 -3.08161676e-01 -8.73009861e-01 -3.09416771e-01
-4.37881112e-01 -2.63495773e-01 1.85072839e-01 -1.77824482e-01
-5.42052984e-01 1.45412967e-01 -2.02569157e-01 -8.99918497e-01
-3.19065183e-01 -8.39053243e-02 1.31048843e-01 -7.38651872e-01
7.75986493e-01 -2.79312357e-02 -5.21353185e-01 -7.57200599e-01
3.51945281e-01 4.05052930e-01 2.94279635e-01 -8.03413749e-01
1.05037344e+00 1.04686952e+00 4.37178850e-01 -8.72414231e-01
-6.64717197e-01 -5.68590760e-01 -6.88823462e-01 -5.22329658e-02
4.98102307e-01 -1.16179729e+00 -1.03989136e+00 2.03499645e-01
-1.51131082e+00 -8.73214379e-03 -4.90277201e-01 1.86737537e-01
-4.21899468e-01 -9.35393944e-03 -3.03518444e-01 -9.72266734e-01
-4.89481270e-01 -1.09947658e+00 1.70354342e+00 8.62638727e-02
2.66236544e-01 -3.68414432e-01 -1.02589048e-01 -1.16669491e-01
3.54868919e-01 1.83129370e-01 9.09036934e-01 -3.51989597e-01
-1.38423550e+00 -2.90693790e-01 -7.66568422e-01 -6.95433691e-02
-1.25717729e-01 -1.64910555e-01 -1.09751117e+00 -1.27483040e-01
-1.36371225e-01 6.04286306e-02 9.63872731e-01 5.29264927e-01
1.28462863e+00 1.97127864e-01 -6.66637838e-01 9.14196134e-01
1.43348062e+00 -1.37108549e-01 3.28844279e-01 7.32330158e-02
9.81829584e-01 3.52819681e-01 4.63692069e-01 3.79804313e-01
4.34346616e-01 6.38746560e-01 1.06386030e+00 -1.02314278e-01
-2.09282279e-01 -2.04090208e-01 -2.65142411e-01 3.33880663e-01
-5.00443578e-01 -2.47907951e-01 -1.19081426e+00 4.99530166e-01
-1.80631042e+00 -7.51617193e-01 -2.54679114e-01 2.05258465e+00
1.26357958e-01 4.91739303e-01 1.66000202e-01 -3.30096155e-01
6.36502981e-01 3.46607745e-01 -6.56586468e-01 8.70138258e-02
-1.34263307e-01 2.91979462e-01 8.24475229e-01 2.66229928e-01
-1.12647164e+00 1.23532915e+00 6.84160757e+00 1.04547119e+00
-1.00544059e+00 1.67414621e-01 3.88548940e-01 -5.34502208e-01
-8.77859518e-02 -2.60774881e-01 -1.46804619e+00 1.23443089e-01
5.16530812e-01 2.15810180e-01 1.73221394e-01 1.25117338e+00
1.25537604e-01 1.90812036e-01 -1.05267107e+00 1.17961073e+00
-3.39310974e-01 -1.80601752e+00 2.36285597e-01 4.47182149e-01
6.60283267e-01 1.03909707e+00 1.61798060e-01 2.27369666e-01
3.77087474e-01 -9.67093885e-01 8.41716886e-01 2.50272870e-01
6.51949942e-01 -6.60898209e-01 5.00752091e-01 4.68714714e-01
-1.51762128e+00 1.33255571e-01 -8.40258896e-01 -2.98296928e-01
8.45437497e-02 7.67059684e-01 -9.58021641e-01 2.88189590e-01
1.13325548e+00 4.66301918e-01 -6.05157435e-01 1.29036534e+00
2.34398022e-01 1.91241890e-01 -1.17181170e+00 -1.97664827e-01
5.31546175e-01 8.21398124e-02 8.67184341e-01 9.14331734e-01
5.07543802e-01 5.35789095e-02 3.50676298e-01 1.14989555e+00
-1.72729447e-01 -5.06375022e-02 -1.03166676e+00 3.18453908e-01
7.87673056e-01 1.11473095e+00 -8.63985896e-01 -1.44393504e-01
-5.79633236e-01 1.65367857e-01 4.63477373e-01 1.24718264e-01
-4.43800300e-01 -7.17752147e-03 1.13318169e+00 5.80400229e-01
7.05953300e-01 -7.40234733e-01 -8.06164980e-01 -8.92231226e-01
7.32951090e-02 -2.95917064e-01 -7.34528992e-03 -7.60905981e-01
-1.29041588e+00 4.88984674e-01 7.24529941e-03 -1.16984439e+00
1.96708500e-01 -8.49215806e-01 -6.44713163e-01 8.10224891e-01
-1.70153213e+00 -1.30338383e+00 -5.00613570e-01 5.00459254e-01
4.98305321e-01 2.44996667e-01 4.31552500e-01 4.07509238e-01
-2.26571247e-01 2.52831995e-01 -3.41212809e-01 -1.36734456e-01
2.10824177e-01 -9.81257319e-01 1.05232573e+00 7.01556683e-01
1.96756169e-01 8.07751536e-01 4.72940594e-01 -7.75998890e-01
-1.72021592e+00 -1.11591244e+00 3.55290383e-01 -8.65812778e-01
3.74531895e-01 -6.66242421e-01 -7.66722143e-01 5.24499714e-01
-4.19676811e-01 4.32142049e-01 1.11403845e-01 4.67203200e-01
-3.90213728e-01 -1.01282634e-01 -1.10359192e+00 5.78731179e-01
1.81607544e+00 -3.58943164e-01 -3.79150659e-01 4.56041873e-01
1.02664864e+00 -6.39780879e-01 -6.90607429e-01 8.83309841e-01
4.75101590e-01 -1.02094686e+00 1.64620125e+00 -5.73765039e-01
6.88182637e-02 -4.08362627e-01 -5.37959933e-01 -7.71983325e-01
-6.48622751e-01 -3.73264611e-01 -3.55161965e-01 7.06052423e-01
1.14277914e-01 -5.69259107e-01 1.30383289e+00 2.50181615e-01
-5.08242488e-01 -7.84664810e-01 -1.16984677e+00 -9.27556336e-01
7.03070611e-02 -8.06156456e-01 1.21907091e+00 5.59056103e-01
-8.60709012e-01 3.72214556e-01 9.72950906e-02 6.41077101e-01
8.87613237e-01 3.98045301e-01 1.14653981e+00 -1.85603452e+00
3.40006322e-01 -6.32646263e-01 -5.60304463e-01 -1.52760303e+00
-2.82101184e-01 -7.36421764e-01 -1.24829389e-01 -1.41621768e+00
-1.38355240e-01 -8.89895141e-01 -1.45067498e-01 3.10653239e-01
5.82817234e-02 5.84704101e-01 3.01810354e-01 4.72445786e-01
-7.74416685e-01 4.51534241e-01 1.13278019e+00 3.59401200e-03
-3.17016065e-01 9.77930874e-02 -6.76869988e-01 9.73282278e-01
5.34297287e-01 -4.35371041e-01 -1.43676654e-01 -9.94688213e-01
4.19374138e-01 -4.93118107e-01 7.82433510e-01 -1.22695625e+00
3.62884909e-01 -1.41840622e-01 4.18036491e-01 -1.42024016e+00
8.31639230e-01 -9.41321254e-01 -2.68696368e-01 2.36217692e-01
3.39699149e-01 1.90177597e-02 4.03401941e-01 5.01680315e-01
2.19747916e-01 7.48412237e-02 6.93455279e-01 -3.18493813e-01
-7.55271137e-01 8.76029372e-01 2.45653212e-01 -2.55496502e-01
8.25804234e-01 -5.52903652e-01 -4.93361712e-01 -2.49198023e-02
-4.12477523e-01 3.19990605e-01 6.28771007e-01 3.33453298e-01
6.57083154e-01 -1.35160756e+00 -5.63630581e-01 2.07446635e-01
1.99147806e-01 7.12903321e-01 2.09076807e-01 4.73951042e-01
-5.96390843e-01 7.45898545e-01 4.97543961e-02 -1.16974723e+00
-9.77467358e-01 4.26700234e-01 2.44786367e-01 2.42093839e-02
-8.51111829e-01 1.06870627e+00 6.35351658e-01 -5.56399524e-01
2.23651931e-01 -6.18868053e-01 3.42922419e-01 -2.23462552e-01
2.80560464e-01 4.83884960e-01 5.08534670e-01 -3.56234670e-01
-6.31962478e-01 9.68670666e-01 -1.26142144e-01 4.47333366e-01
1.53304708e+00 -1.21020824e-01 1.33911490e-01 1.00235052e-01
9.24094558e-01 -8.91960487e-02 -1.23213911e+00 -3.69504064e-01
-3.33113670e-01 -6.81838572e-01 4.64200169e-01 -3.38730514e-01
-7.43112743e-01 1.03264415e+00 6.38277411e-01 4.32455957e-01
5.83448410e-01 4.28231031e-01 6.53449833e-01 9.12135661e-01
8.29573810e-01 -7.19644725e-01 -2.81384438e-01 8.43960404e-01
7.66729951e-01 -1.28238142e+00 2.17710480e-01 -8.59102070e-01
2.36638293e-01 8.02559674e-01 7.91220844e-01 -5.90609133e-01
6.88787282e-01 3.45127225e-01 -3.94289404e-01 -7.64824808e-01
-5.41391909e-01 -4.82574761e-01 3.12574893e-01 6.67848706e-01
-1.24073409e-01 4.00740877e-02 3.59306484e-01 3.08080107e-01
-2.74397224e-01 -2.69611984e-01 -2.27404445e-01 8.21360528e-01
-9.40162897e-01 -6.94547057e-01 -6.06271923e-01 7.05131173e-01
-1.06075332e-01 -1.82406813e-01 -4.19548869e-01 1.00585496e+00
1.80362403e-01 4.39703226e-01 5.20041645e-01 -1.70596898e-01
5.76084793e-01 -2.26488963e-01 4.88276601e-01 -1.04479527e+00
-3.93430769e-01 -3.40376571e-02 -1.67526796e-01 -8.42411101e-01
-1.79596305e-01 -3.81733388e-01 -1.08167875e+00 -6.70640945e-01
-3.90081644e-01 -2.29586154e-01 8.86307478e-01 7.65540302e-01
9.04836297e-01 3.41957927e-01 3.04965258e-01 -1.52129376e+00
-5.36098719e-01 -7.46667445e-01 -3.94373566e-01 -2.73039371e-01
3.54863316e-01 -9.59242463e-01 -2.48449981e-01 -6.18157923e-01] | [7.688499927520752, -2.9190080165863037] |
61474bd1-f630-46b2-940b-6066c7cfc205 | reward-balancing-for-statistical-spoken | 1707.06299 | null | http://arxiv.org/abs/1707.06299v1 | http://arxiv.org/pdf/1707.06299v1.pdf | Reward-Balancing for Statistical Spoken Dialogue Systems using Multi-objective Reinforcement Learning | Reinforcement learning is widely used for dialogue policy optimization where
the reward function often consists of more than one component, e.g., the
dialogue success and the dialogue length. In this work, we propose a structured
method for finding a good balance between these components by searching for the
optimal reward component weighting. To render this search feasible, we use
multi-objective reinforcement learning to significantly reduce the number of
training dialogues required. We apply our proposed method to find optimized
component weights for six domains and compare them to a default baseline. | ['Milica Gašić', 'Tsung-Hsien Wen', 'Pei-Hao Su', 'Lina Rojas-Barahona', 'Nikola Mrkšić', 'Iñigo Casanueva', 'Paweł Budzianowski', 'Stefan Ultes', 'Steve Young'] | 2017-07-19 | reward-balancing-for-statistical-spoken-1 | https://aclanthology.org/W17-5509 | https://aclanthology.org/W17-5509.pdf | ws-2017-8 | ['multi-objective-reinforcement-learning'] | ['methodology'] | [-8.58626664e-02 1.00586936e-01 -4.58260626e-01 -3.49069744e-01
-8.26779246e-01 -6.13913655e-01 5.92632651e-01 2.86143392e-01
-8.48635733e-01 1.10249484e+00 3.17229658e-01 -2.84774899e-01
-1.07083451e-02 -6.71192944e-01 -5.14963940e-02 -5.39077759e-01
1.47051975e-01 4.99130636e-01 3.23329955e-01 -5.57804167e-01
8.75325263e-01 1.19017556e-01 -1.05813897e+00 -1.05512887e-01
1.16618001e+00 7.97399104e-01 4.33418304e-01 7.00383186e-01
-3.55463684e-01 7.86754012e-01 -9.57524180e-01 -2.48417079e-01
1.81104299e-02 -7.14810431e-01 -1.07547915e+00 2.29127258e-01
-3.35478604e-01 -6.61969125e-01 3.11935823e-02 9.21764553e-01
6.46484256e-01 5.28105021e-01 6.56334698e-01 -1.18494487e+00
3.68769057e-02 6.21210873e-01 -3.97340387e-01 2.34736606e-01
4.88214463e-01 2.47399807e-01 1.24114454e+00 -3.51695091e-01
2.43053436e-01 1.46036065e+00 1.09089568e-01 6.09618008e-01
-1.10908353e+00 -2.54166812e-01 2.88923204e-01 1.98403388e-01
-7.64325678e-01 -4.51833338e-01 7.82896936e-01 -1.49027705e-01
1.06627345e+00 5.60292788e-03 5.79739988e-01 7.41438508e-01
4.31907736e-02 6.30731642e-01 1.15108883e+00 -7.17369556e-01
5.14404535e-01 3.56186211e-01 -6.10197820e-02 5.76634467e-01
-3.02098304e-01 -1.72851384e-01 -2.35777333e-01 -5.85564435e-01
7.15303719e-01 -4.51388627e-01 -2.17003971e-02 -1.80713028e-01
-7.89391756e-01 1.29160392e+00 -5.98794483e-02 2.30267912e-01
-3.67692888e-01 -2.45802030e-02 4.84887093e-01 5.03543794e-01
1.21638075e-01 9.50706542e-01 -5.93750298e-01 -7.07962871e-01
-3.13853681e-01 5.65373123e-01 1.07779181e+00 4.35255826e-01
7.52406061e-01 -6.10288754e-02 -4.02066857e-01 1.45785916e+00
3.78112614e-01 1.46420419e-01 6.14377320e-01 -1.19110692e+00
6.10211670e-01 5.84585845e-01 6.99803948e-01 -6.69526994e-01
-3.10754448e-01 4.51084733e-01 -1.97381496e-01 2.04683065e-01
6.33818030e-01 -7.45475411e-01 -3.80187333e-01 1.61868155e+00
4.58774328e-01 -2.88084149e-01 2.75178492e-01 8.65389466e-01
5.70609808e-01 7.51331568e-01 7.95863122e-02 -3.74008149e-01
1.20877361e+00 -1.09070110e+00 -9.36039805e-01 -3.25343490e-01
6.50545657e-01 -8.83895874e-01 1.22779024e+00 1.68878213e-01
-1.24079823e+00 -1.99498475e-01 -9.32671666e-01 3.98663044e-01
5.24737202e-02 1.47321701e-01 3.98280621e-01 6.47111535e-01
-7.07670629e-01 8.39817643e-01 -4.45844382e-01 -1.36429027e-01
-3.62476915e-01 5.44319689e-01 2.69603759e-01 2.77353019e-01
-1.39343786e+00 1.19764900e+00 6.92806065e-01 -2.43974954e-01
-6.66739047e-01 9.72098410e-02 -5.90692103e-01 -7.69080967e-03
7.06910968e-01 -2.77730614e-01 1.77550304e+00 -7.27233291e-01
-2.43242979e+00 2.53984869e-01 1.14634201e-01 -2.56867886e-01
3.08010668e-01 -2.43861347e-01 8.67248252e-02 2.10134402e-01
-2.75103927e-01 4.70450372e-01 8.43009293e-01 -9.78936076e-01
-7.81050801e-01 2.96467971e-02 6.25321507e-01 8.82902145e-01
-5.28790116e-01 1.89303815e-01 -4.54531014e-01 -3.52265686e-01
-4.25467312e-01 -8.87431264e-01 -5.95352829e-01 -5.31430244e-01
-1.06428728e-01 -7.00814366e-01 3.50151122e-01 -5.38903832e-01
1.31838906e+00 -1.83761883e+00 2.11888283e-01 2.41421849e-01
-9.78642777e-02 2.49774516e-01 -2.75427461e-01 5.52726865e-01
4.59988385e-01 -1.44604743e-01 -1.51753023e-01 -1.15715072e-01
-7.91721698e-03 3.90329450e-01 1.33682176e-01 2.18482584e-01
1.99491844e-01 4.33082223e-01 -1.00233221e+00 -6.22639179e-01
4.46824506e-02 -6.21218197e-02 -6.28180325e-01 9.26310480e-01
-4.41784650e-01 3.35640728e-01 -6.41302526e-01 2.44436115e-01
1.95679352e-01 -7.08704814e-03 5.26457906e-01 1.97914109e-01
-1.99173093e-01 6.71774864e-01 -1.39608908e+00 1.22789419e+00
-5.42049885e-01 6.37575835e-02 1.45330489e-01 -9.62040424e-01
1.22797287e+00 3.52337599e-01 6.65158689e-01 -6.79865897e-01
3.67459029e-01 6.74006296e-03 2.32875362e-01 -4.84856158e-01
8.68573785e-01 -1.50505096e-01 -2.65670707e-03 8.09579849e-01
2.62672491e-02 -1.98688164e-01 5.48387110e-01 -9.93097872e-02
9.14557040e-01 -6.61882088e-02 5.57826877e-01 -5.30011952e-02
7.82568514e-01 2.74399389e-03 4.91031945e-01 6.70307279e-01
-4.54502344e-01 1.19539797e-02 1.08628273e+00 1.83216017e-02
-1.12704253e+00 -3.70970100e-01 2.50518531e-01 1.43136990e+00
1.10036172e-01 -4.56547499e-01 -8.42593014e-01 -8.43295991e-01
-1.14002727e-01 6.34463310e-01 -2.74302185e-01 -2.16071814e-01
-7.46978462e-01 -7.04534352e-01 4.66530710e-01 1.52306452e-01
4.22150224e-01 -1.26310122e+00 -7.08490431e-01 5.05497873e-01
-3.01549464e-01 -8.45104754e-01 -7.21951902e-01 1.79981902e-01
-7.71754444e-01 -1.05740893e+00 -8.46932948e-01 -5.58191895e-01
4.70455676e-01 1.72399104e-01 9.67742145e-01 2.22229034e-01
1.77440792e-01 2.12455422e-01 -5.17834902e-01 -2.50595540e-01
-6.86726809e-01 2.70991296e-01 -1.62135884e-01 -3.78284067e-01
-4.37350497e-02 -1.21857964e-01 -6.21226668e-01 5.63928783e-01
-6.90811396e-01 -2.99919605e-01 4.24663574e-01 1.08406770e+00
1.16805390e-01 -3.33326757e-02 8.05711269e-01 -7.51654625e-01
1.72608709e+00 -3.41213733e-01 -7.72568583e-01 4.06671971e-01
-9.55902874e-01 4.79687601e-01 7.60606587e-01 -7.20662773e-01
-1.01485133e+00 -1.51462033e-01 -2.86965370e-01 1.33611947e-01
7.55805448e-02 5.45558989e-01 9.71280560e-02 -2.12301806e-01
5.30784607e-01 -7.66048059e-02 1.56392500e-01 -1.65855557e-01
4.02394801e-01 8.17322195e-01 -1.25652671e-01 -8.50555718e-01
2.56861657e-01 -4.13454652e-01 -4.03031707e-01 -7.29601681e-01
-3.69552553e-01 -4.70103145e-01 -2.05280766e-01 -3.68149787e-01
4.49643791e-01 -3.21422249e-01 -1.04769301e+00 8.66591856e-02
-1.08677590e+00 -5.69897294e-01 -8.23685154e-02 5.11399925e-01
-8.36371362e-01 5.31770706e-01 -5.44858754e-01 -1.22657418e+00
-5.73471844e-01 -1.31126559e+00 5.87723911e-01 3.99191737e-01
-2.92685390e-01 -8.52629483e-01 4.24110651e-01 1.16735190e-01
3.48551810e-01 -2.17817366e-01 9.83979583e-01 -8.06529939e-01
7.28595480e-02 1.13249250e-01 -1.12125434e-01 2.01548994e-01
1.55290842e-01 -5.34048080e-02 -3.02788466e-01 -3.36711049e-01
-5.73555864e-02 -7.44540513e-01 2.98368603e-01 2.28240982e-01
7.39018261e-01 -6.43164992e-01 2.04664350e-01 -1.33290617e-02
1.18869865e+00 7.93192804e-01 5.08284450e-01 6.13307178e-01
2.28925288e-01 6.93790197e-01 1.10064089e+00 1.04712510e+00
3.87092650e-01 9.56853986e-01 2.66740352e-01 1.90755367e-01
5.32905042e-01 -5.56763075e-02 4.01715934e-01 7.35493124e-01
-1.48303390e-01 -1.45807505e-01 -8.80242348e-01 3.26219648e-01
-2.03404331e+00 -7.84468114e-01 5.85866451e-01 2.19880557e+00
1.23932099e+00 3.04561406e-01 7.17891097e-01 -2.90847267e-03
5.89134991e-01 2.65446603e-01 -5.66703320e-01 -9.08036113e-01
3.82728606e-01 -6.80505671e-03 4.69989806e-01 7.66573131e-01
-8.16481948e-01 1.12933338e+00 7.11501312e+00 6.55774295e-01
-9.69656408e-01 -3.22413534e-01 4.73477364e-01 1.26039729e-01
-2.22252131e-01 1.28053933e-01 -7.87449181e-01 2.72186637e-01
1.03227806e+00 -2.69030690e-01 8.50261569e-01 7.65093207e-01
4.15232867e-01 -4.01666731e-01 -5.89717388e-01 6.23608291e-01
-2.82743335e-01 -9.17765081e-01 -2.49041334e-01 -9.56512988e-02
4.81446177e-01 -3.55084777e-01 -3.76962006e-01 3.43966037e-01
6.13487482e-01 -7.96004236e-01 2.42315024e-01 6.97735250e-02
3.15191358e-01 -9.50852513e-01 5.82000673e-01 4.31791931e-01
-8.37533116e-01 -2.91106969e-01 -3.23083103e-01 1.92544177e-01
3.36874500e-02 -6.14777431e-02 -1.29713452e+00 1.40554890e-01
2.23494142e-01 7.56745860e-02 -6.23749793e-02 1.11229432e+00
-2.59880394e-01 4.40368176e-01 -1.35777533e-01 -6.69048011e-01
5.52007496e-01 -5.06344140e-01 4.96559292e-01 1.09682941e+00
3.68640088e-02 2.75569111e-01 6.09649062e-01 4.40190315e-01
3.07730865e-02 6.02213323e-01 -3.89966309e-01 -3.03165764e-01
5.30836999e-01 1.26414478e+00 -6.44701719e-01 -1.56812891e-02
-2.67408431e-01 7.37370133e-01 6.00549459e-01 2.17415035e-01
-8.22796524e-01 -5.79208255e-01 5.89580834e-01 -4.20380563e-01
3.42537463e-01 -4.21990782e-01 9.32852086e-03 -7.85738647e-01
-3.33879262e-01 -1.21183586e+00 2.65938312e-01 -2.09763005e-01
-9.60015535e-01 7.15017855e-01 1.39327735e-01 -1.01106095e+00
-7.32852876e-01 -3.50623786e-01 -7.23338485e-01 9.45923269e-01
-1.62585390e+00 -3.68785739e-01 6.58487082e-02 6.24683857e-01
7.50437081e-01 -3.67000759e-01 9.00172651e-01 -4.18809466e-02
-6.72506392e-01 6.86994970e-01 1.43831879e-01 -1.20481573e-01
8.63330543e-01 -1.46588039e+00 8.48675147e-02 1.96899444e-01
-3.18649709e-01 4.44209009e-01 9.35356796e-01 -4.25908953e-01
-1.14964211e+00 -2.61437207e-01 6.38488173e-01 4.43953186e-01
7.16229737e-01 2.75754742e-02 -7.28381991e-01 -9.62177143e-02
5.25078893e-01 -8.87817085e-01 6.28289938e-01 2.46569753e-01
1.33660793e-01 5.92450202e-02 -1.12772572e+00 8.08264911e-01
2.48290837e-01 -1.58817112e-01 -4.46813196e-01 2.36902073e-01
5.73259354e-01 -5.88176847e-01 -1.04403853e+00 1.04225799e-01
3.70240420e-01 -5.84310472e-01 6.59956157e-01 -5.40271521e-01
4.37668771e-01 1.05490565e-01 6.85204715e-02 -1.77637541e+00
-1.04652457e-01 -1.00273621e+00 -2.31211588e-01 1.06407845e+00
4.98311132e-01 -5.01904070e-01 7.84381211e-01 8.85746598e-01
2.32248932e-01 -1.10761297e+00 -8.06654334e-01 -4.44000393e-01
-4.38013114e-02 2.19752982e-01 5.23255229e-01 6.33669853e-01
3.11636120e-01 8.33061576e-01 -6.73189938e-01 -3.03918421e-01
2.47911647e-01 7.11908638e-02 6.86705470e-01 -9.61531222e-01
-4.68727469e-01 -5.23583889e-01 2.97145873e-01 -1.25170445e+00
2.40146533e-01 -1.30518988e-01 4.12793964e-01 -1.40715742e+00
-1.29886806e-01 -5.79986870e-01 -1.31470695e-01 3.79154593e-01
-4.26568002e-01 -6.57655001e-01 2.77100861e-01 1.16959192e-01
-6.50813282e-01 8.80334139e-01 1.39368594e+00 7.78293908e-02
-7.98277974e-01 3.25595289e-01 -7.10794747e-01 5.22826910e-01
1.34231853e+00 -4.36694503e-01 -5.30612350e-01 -2.46196911e-02
3.25310417e-02 6.67543769e-01 -4.67115968e-01 -4.45123285e-01
4.92484160e-02 -8.69438231e-01 -1.08107448e-01 -3.36166024e-01
3.73226553e-01 -5.84046066e-01 -6.35690272e-01 3.26897085e-01
-7.78406739e-01 3.15204263e-01 2.99144909e-02 4.58654791e-01
-2.72416264e-01 -1.01680183e+00 8.79079819e-01 -2.77505666e-01
-4.26870674e-01 -1.38285741e-01 -5.48312008e-01 4.02645245e-02
9.25404131e-01 5.99468080e-03 -2.27761313e-01 -6.91575825e-01
-2.72839665e-01 6.54371321e-01 2.70171613e-02 4.03523743e-01
7.00997174e-01 -9.67072666e-01 -4.84405011e-01 -3.08218867e-01
-1.01387516e-01 -4.79549021e-01 -2.32244968e-01 4.56340760e-01
-3.32023233e-01 4.61777419e-01 -4.48330820e-01 -6.58943579e-02
-1.50659955e+00 1.05089560e-01 4.29740608e-01 -7.92096019e-01
-2.75782585e-01 3.97770762e-01 -4.74031538e-01 -5.56319356e-01
7.25000739e-01 -4.56492566e-02 -9.89550054e-01 6.01502396e-02
5.65900803e-01 5.46639562e-01 -3.32664847e-01 -3.25544327e-01
-7.79470205e-02 2.05706909e-01 -2.05344588e-01 -6.40253961e-01
1.14495254e+00 -2.39905059e-01 -3.90122905e-02 3.89453232e-01
9.44450259e-01 -1.74770817e-01 -1.34091532e+00 -1.99684143e-01
1.60246953e-01 -5.58034658e-01 3.45334387e-03 -7.41013706e-01
-6.86282635e-01 3.17790806e-01 5.23062170e-01 5.63274145e-01
9.69745755e-01 -3.30981165e-01 7.96667278e-01 7.32130289e-01
2.03343004e-01 -1.64078474e+00 4.70900267e-01 9.36664999e-01
6.85835779e-01 -1.33757353e+00 1.15919828e-01 -4.99932393e-02
-1.18409157e+00 1.34137499e+00 9.62775767e-01 -8.27343762e-02
2.75917560e-01 -1.87771786e-02 1.06279857e-01 5.95229194e-02
-9.18957770e-01 -3.61995101e-01 1.55066416e-01 3.09935093e-01
7.31388986e-01 1.37264039e-02 -1.02457154e+00 1.15954399e-01
7.21178651e-02 -2.41139784e-01 4.42903817e-01 1.01173842e+00
-7.60510862e-01 -1.82464337e+00 -4.11956668e-01 4.44345027e-01
-4.93593156e-01 1.65953159e-01 -5.12011349e-01 3.14001292e-01
-4.97380733e-01 1.35362983e+00 -3.34682792e-01 -3.20932597e-01
7.18149483e-01 -1.01776943e-01 4.88118321e-01 -4.64801788e-01
-9.37820554e-01 2.93186277e-01 5.31706274e-01 -3.41727167e-01
-4.46810842e-01 -4.34230834e-01 -1.27389383e+00 -3.11169088e-01
-6.31832659e-01 4.61520672e-01 6.74564540e-01 1.05248368e+00
-1.40248731e-01 3.84783924e-01 1.20339656e+00 -4.88244176e-01
-1.15989196e+00 -1.17544067e+00 -3.81811023e-01 1.39713272e-01
1.55068412e-01 -7.85255492e-01 1.25448704e-02 -5.89787900e-01] | [13.06534194946289, 8.02713680267334] |
0bb4cea1-4755-4585-af3a-ab640f65c8f8 | kalman-filter-is-all-you-need-optimization | null | null | https://openreview.net/forum?id=cMBKc-0OTY5 | https://openreview.net/pdf?id=cMBKc-0OTY5 | Kalman Filter Is All You Need: Optimization Works When Noise Estimation Fails | Determining the noise parameters of a Kalman Filter (KF) has been studied for decades. A huge body of research focuses on the task of noise estimation under various conditions, since precise noise estimation is considered equivalent to minimization of the filtering errors. However, we show that even a small violation of the KF assumptions can significantly modify the effective noise, breaking the equivalence between the tasks and making noise estimation an inferior strategy. We show that such violations are common, and are often not trivial to handle or even notice. Consequentially, we argue that a robust solution is needed - rather than choosing a dedicated model per problem.
To that end, we apply gradient-based optimization to the filtering errors directly, with relation to an efficient parameterization of the symmetric and positive-definite parameters of the KF. In a variety of state-estimation and tracking problems, we show that the optimization improves both the accuracy of the KF and its robustness to design decisions.
In addition, we demonstrate how an optimized neural network model can seem to reduce the errors significantly compared to a KF - and how this reduction vanishes once the KF is optimized similarly. This indicates how complicated models can be wrongly identified as superior to the KF, while in fact they were merely more optimized. | ['Netanel Yannay', 'Shie Mannor', 'Ido Greenberg'] | 2021-09-29 | null | null | null | null | ['noise-estimation'] | ['medical'] | [ 2.10056156e-01 1.46799490e-01 2.41585061e-01 1.04153547e-02
-4.81383115e-01 -7.29032516e-01 4.82234776e-01 -1.24960104e-02
-5.90424359e-01 8.26746702e-01 -1.25618803e-03 -5.19128025e-01
-6.16850495e-01 -3.60553175e-01 -7.02211082e-01 -1.08331716e+00
1.64486632e-01 4.01604436e-02 1.74541533e-01 -2.97786862e-01
2.58180112e-01 5.71209073e-01 -1.50615299e+00 -6.62737787e-01
6.66676819e-01 9.76705194e-01 -3.88018228e-02 8.36481512e-01
3.82460564e-01 3.02846760e-01 -6.34356260e-01 -2.31496066e-01
5.49692929e-01 -3.47559035e-01 -6.00195348e-01 -7.14160725e-02
4.06672657e-01 -2.40025911e-02 -9.37647894e-02 1.31618834e+00
6.27811074e-01 3.11601758e-01 5.02326310e-01 -8.97439778e-01
1.52719095e-02 3.84071946e-01 1.18847407e-01 5.67792542e-02
-7.80369788e-02 9.05728564e-02 8.12418282e-01 -3.70376974e-01
3.90314817e-01 9.43567395e-01 1.07919240e+00 3.65325212e-01
-1.30808175e+00 -3.34747732e-01 -6.89856783e-02 -2.64252394e-01
-1.44753385e+00 -6.38453305e-01 4.67644781e-01 -6.09261811e-01
5.36569417e-01 4.05045480e-01 6.79950714e-01 7.58524477e-01
2.57575274e-01 1.86428696e-01 1.11659825e+00 -5.73626339e-01
2.44786933e-01 2.25416243e-01 3.05195153e-01 4.83677208e-01
7.35519648e-01 4.43360627e-01 -3.35324593e-02 -1.11193404e-01
8.65868449e-01 -4.62657005e-01 -7.30922341e-01 -5.78961909e-01
-1.22316098e+00 6.23773158e-01 2.96440590e-02 4.85032171e-01
-2.03876376e-01 2.45650202e-01 3.93577129e-01 5.93975902e-01
2.26345763e-01 9.66150522e-01 -6.90524697e-01 -2.98692018e-01
-6.26886666e-01 4.31682348e-01 1.06469011e+00 4.85471487e-01
4.92762893e-01 3.42602372e-01 1.80114269e-01 4.16741401e-01
1.75469756e-01 5.59765339e-01 2.19657093e-01 -1.02985024e+00
2.88061559e-01 1.86908469e-01 5.85864604e-01 -1.14470744e+00
-7.34939098e-01 -8.85175288e-01 -9.10423160e-01 4.13439661e-01
1.20877564e+00 -5.49427390e-01 -4.95898753e-01 1.83506548e+00
2.22819880e-01 4.92378063e-02 1.79923163e-03 9.84612644e-01
-1.31970480e-01 4.19361770e-01 -2.82469481e-01 -5.96940577e-01
9.19700146e-01 -3.74922425e-01 -1.05969620e+00 -2.71499425e-01
7.60262251e-01 -6.48511052e-01 8.69473040e-01 7.10662544e-01
-9.21553850e-01 -3.74775499e-01 -1.20626128e+00 4.09893125e-01
-3.18986684e-01 2.26605833e-01 3.37523431e-01 9.46843028e-01
-8.98574710e-01 1.15583086e+00 -6.63499534e-01 -2.78551936e-01
-3.51058573e-01 4.81410056e-01 -2.36487761e-01 6.59863353e-01
-1.20622480e+00 1.30343878e+00 4.94858533e-01 6.78893566e-01
-1.64249510e-01 -6.06792331e-01 -5.18054247e-01 1.81594826e-02
7.05048740e-01 -6.48568571e-01 1.27748251e+00 -9.77452099e-01
-1.75495911e+00 2.03566521e-01 1.47305802e-02 -5.08590698e-01
6.57723486e-01 -4.97063786e-01 -3.78816605e-01 -2.23408267e-01
-3.59173566e-01 -1.19020045e-01 1.19078183e+00 -1.03606582e+00
-3.63790870e-01 -1.71084866e-01 1.28721178e-01 -7.60407448e-02
-4.81323004e-02 -1.86800584e-01 1.65468790e-02 -6.40735984e-01
2.82542080e-01 -1.05572855e+00 -5.65569103e-01 -1.97238252e-01
-4.43576902e-01 1.26406148e-01 3.55514407e-01 -6.19787216e-01
1.50497341e+00 -2.02597690e+00 8.43278244e-02 5.92734814e-01
-4.49485257e-02 4.75471377e-01 2.98009157e-01 3.50696683e-01
-3.49972486e-01 1.21862993e-01 -1.49581909e-01 9.96362716e-02
2.06623018e-01 1.01280935e-01 -3.93053681e-01 9.23976421e-01
1.69234231e-01 5.53976893e-01 -7.00618982e-01 1.13895155e-01
2.31658503e-01 4.61698472e-01 -4.17225987e-01 -2.71521181e-01
3.68688479e-02 5.00850916e-01 -4.39534724e-01 -9.79504883e-02
4.30176467e-01 -1.97019652e-01 3.22352618e-01 -4.32802051e-01
-3.02188694e-01 2.33406350e-01 -1.83552623e+00 1.11755061e+00
-3.98350447e-01 6.86336339e-01 3.94327879e-01 -1.26924992e+00
6.12745464e-01 3.06822062e-01 2.12962404e-01 -4.47022766e-01
4.71592128e-01 3.61399740e-01 1.34669021e-01 -4.27308381e-01
3.62819374e-01 -3.64145100e-01 6.96519241e-02 1.56555086e-01
-1.95316091e-01 -1.69622526e-01 1.48879305e-01 -1.89686403e-01
7.98483670e-01 6.62130788e-02 4.93644893e-01 -7.30893433e-01
6.28599465e-01 -1.97526529e-01 5.42530477e-01 9.64666665e-01
-6.18547574e-02 2.30365872e-01 5.53146660e-01 -1.29355028e-01
-9.40756500e-01 -6.40740514e-01 -2.74562448e-01 5.00808060e-01
-3.40340361e-02 -4.49688315e-01 -8.21583390e-01 -3.08890134e-01
4.46189977e-02 4.53378052e-01 -4.76646990e-01 -2.74608999e-01
-6.07652009e-01 -7.89751828e-01 3.95995200e-01 2.93324977e-01
2.41454154e-01 -2.80371040e-01 -7.44197845e-01 2.44372755e-01
1.86499059e-01 -8.85702550e-01 -1.66691959e-01 5.47995985e-01
-8.30035150e-01 -1.04819286e+00 -5.36342442e-01 -2.41393194e-01
5.75466871e-01 -1.28664553e-01 8.22947323e-01 3.57858129e-02
1.27822876e-01 5.44475198e-01 -9.11727771e-02 -4.50887710e-01
-5.33399701e-01 -7.26337582e-02 4.76100862e-01 8.52249470e-03
-1.76466793e-01 -5.11355281e-01 -2.70539612e-01 4.29909438e-01
-7.34846771e-01 -4.02044564e-01 3.84938836e-01 7.94415116e-01
1.33491635e-01 3.40472877e-01 5.29448688e-01 -6.91812634e-01
7.65038550e-01 1.42915115e-01 -1.19226003e+00 6.44791648e-02
-7.54252017e-01 4.29522902e-01 8.57728183e-01 -5.05360842e-01
-7.82612443e-01 2.95857042e-01 -1.00076348e-01 -1.83769748e-01
9.32343677e-03 4.75757360e-01 -1.47073179e-01 -4.71615463e-01
7.26723909e-01 -6.51082247e-02 1.90960303e-01 -5.03773272e-01
3.77444983e-01 4.01490182e-01 4.97828811e-01 -4.45532084e-01
1.06230807e+00 2.03867495e-01 3.53753000e-01 -1.08897126e+00
-8.28085065e-01 -4.93941516e-01 -4.90819991e-01 -2.31163561e-01
5.18513978e-01 -6.33331895e-01 -9.91268396e-01 3.26408476e-01
-9.15385008e-01 -2.64262468e-01 -4.44642901e-01 7.24197149e-01
-6.10678911e-01 5.40179193e-01 -4.21696395e-01 -1.26134217e+00
7.34346211e-02 -1.16735446e+00 6.01726174e-01 1.00987449e-01
-2.07304522e-01 -1.14740241e+00 2.20672898e-02 -2.61781245e-01
6.00095332e-01 1.79714784e-02 4.94696856e-01 -4.68097448e-01
-3.78035754e-01 -2.88979173e-01 -2.69849114e-02 6.15350544e-01
-1.01426817e-01 1.40145496e-01 -9.53396857e-01 -3.76045585e-01
3.84994805e-01 2.83225387e-01 7.55374253e-01 5.74156106e-01
5.90883255e-01 -3.64584893e-01 -1.51036054e-01 4.18121725e-01
1.66566694e+00 1.19220167e-01 4.34490770e-01 5.03553748e-01
4.74357486e-01 5.68895638e-01 3.53860646e-01 3.47876042e-01
-3.44785929e-01 8.28614533e-01 2.40998387e-01 1.32232741e-01
3.16518277e-01 1.14549451e-01 3.13390672e-01 8.63888085e-01
-2.76626498e-01 6.73379898e-02 -6.60651386e-01 1.89102113e-01
-1.75665629e+00 -7.89065778e-01 -4.64773536e-01 2.62260389e+00
6.14102781e-01 3.18686217e-01 2.02323183e-01 4.97524828e-01
5.83385825e-01 -1.53454110e-01 -2.29264349e-01 -3.39870155e-01
-1.16793782e-01 1.09973298e-02 1.04984975e+00 7.64576793e-01
-1.21072710e+00 3.57986450e-01 7.40478706e+00 7.82429218e-01
-1.13020384e+00 -1.78507656e-01 1.39103895e-02 1.38000265e-01
-1.38586834e-02 2.49243334e-01 -8.67747843e-01 5.44401169e-01
1.11032689e+00 -2.47418299e-01 4.51726317e-01 7.53744423e-01
5.70283353e-01 -2.84940362e-01 -7.59880781e-01 8.51937354e-01
-2.45956585e-01 -1.06967664e+00 -3.57252419e-01 1.63011953e-01
4.63439792e-01 -5.37207127e-01 -2.19829932e-01 1.21009156e-01
-5.51226661e-02 -7.32302189e-01 8.25522244e-01 9.46949184e-01
2.44474150e-02 -7.14469135e-01 8.23182523e-01 4.53819364e-01
-9.45574522e-01 -1.85394779e-01 -4.39985156e-01 -4.46239114e-01
2.22730950e-01 9.66009736e-01 -3.00962478e-01 6.56268597e-01
2.54376382e-01 3.13891441e-01 -4.55795646e-01 1.41078901e+00
-1.20834880e-01 6.55356526e-01 -5.94812989e-01 -1.10997483e-01
8.65388960e-02 -5.43764770e-01 7.03603804e-01 1.10685432e+00
3.68476421e-01 -1.21408522e-01 -3.67116705e-02 6.06069624e-01
4.72627014e-01 2.83063464e-02 -4.60367084e-01 -9.30969492e-02
6.73878286e-03 9.87235606e-01 -5.92839122e-01 -1.73134387e-01
-2.68242866e-01 4.46262449e-01 -6.04756027e-02 4.81878072e-01
-6.95673168e-01 -5.89606583e-01 7.22798288e-01 -9.91853178e-02
3.23066324e-01 -3.50248426e-01 -2.05245435e-01 -1.12513483e+00
1.96139857e-01 -7.80919850e-01 -3.46165635e-02 -4.62566882e-01
-1.01836908e+00 2.25881174e-01 -1.24463039e-02 -1.32617056e+00
-4.61083621e-01 -8.06511402e-01 -2.77925611e-01 8.38732064e-01
-1.21874213e+00 -2.95481324e-01 1.52740166e-01 1.96296945e-01
-5.68218380e-02 3.04273427e-01 5.15444756e-01 3.53052199e-01
-6.99111819e-01 2.51712829e-01 4.84094709e-01 -1.35929659e-01
5.23062289e-01 -1.29223657e+00 9.74282771e-02 1.18722391e+00
-1.65812433e-01 9.36503828e-01 1.37339950e+00 -5.20010054e-01
-1.47379649e+00 -6.79863453e-01 8.65292907e-01 -3.59103799e-01
1.07379293e+00 -2.18164071e-01 -8.88831615e-01 4.08654064e-01
-5.19538634e-02 -2.04061538e-01 2.31474750e-02 4.13741246e-02
1.56960949e-01 -8.53540078e-02 -5.85765362e-01 6.65329814e-01
8.31745148e-01 -5.13054192e-01 -4.41037059e-01 2.90517867e-01
3.90550703e-01 -3.54931444e-01 -9.15308893e-01 3.97491336e-01
6.84088647e-01 -9.59495842e-01 9.54988658e-01 -6.49810195e-01
-2.69136190e-01 -6.27512634e-01 -5.52289523e-02 -1.40101373e+00
-2.41159588e-01 -1.04185462e+00 -2.47937869e-02 9.97048140e-01
4.14608419e-01 -8.32614481e-01 6.10673010e-01 6.75237715e-01
-1.54281873e-02 -5.47539353e-01 -8.89086485e-01 -1.25955093e+00
4.29681279e-02 -6.37383163e-01 7.10702091e-02 8.13174486e-01
-2.62764208e-02 3.27315122e-01 -4.41426307e-01 4.95924860e-01
3.60233307e-01 -2.23453581e-01 8.34535301e-01 -1.37660015e+00
-5.30160606e-01 -5.95793307e-01 -3.54859561e-01 -1.24633098e+00
-3.18558747e-03 -2.80138642e-01 2.69800633e-01 -1.09275651e+00
-6.61285162e-01 -4.24106389e-01 -2.08606888e-02 7.90719502e-03
-5.60756251e-02 -1.43940911e-01 2.06834316e-01 -1.64540447e-02
-2.08937362e-01 1.64352074e-01 1.11349905e+00 1.01559736e-01
-2.01651588e-01 6.28862143e-01 -5.33251405e-01 9.19979811e-01
8.25017750e-01 -2.41209894e-01 -3.22064847e-01 -5.97309042e-03
8.68004441e-01 8.89403746e-02 5.76231301e-01 -1.22708380e+00
2.98438519e-01 8.47871080e-02 1.57835469e-01 -1.89024091e-01
1.96306676e-01 -1.02920008e+00 4.36221361e-01 4.84970182e-01
-2.23880157e-01 -8.24204832e-02 2.06936315e-01 7.41980910e-01
-2.76032370e-03 -7.69448340e-01 6.57996953e-01 -6.80885240e-02
-5.59498489e-01 -2.62088090e-01 -6.32082164e-01 -8.29218253e-02
5.56230724e-01 -3.18026245e-01 -1.35171741e-01 -6.58588409e-01
-1.08161366e+00 -1.50629595e-01 3.53010118e-01 -3.29415016e-02
-4.90529872e-02 -9.73324656e-01 -1.78550407e-01 3.00946832e-01
-3.77883971e-01 -3.44630271e-01 1.82504103e-01 1.34188557e+00
-2.35612512e-01 6.04617357e-01 2.50452429e-01 -5.05191326e-01
-1.03877974e+00 4.69673961e-01 7.50127792e-01 -9.95482504e-02
-5.29013395e-01 5.64450383e-01 -8.11342895e-02 -2.92170614e-01
4.33325529e-01 -7.74679124e-01 6.80894852e-02 7.47314245e-02
4.43879873e-01 4.60938454e-01 3.25372994e-01 -3.89070481e-01
-1.32940248e-01 8.73824894e-01 4.27531093e-01 1.24801164e-02
8.77530217e-01 -3.92711639e-01 1.30167350e-01 5.71114063e-01
9.39661443e-01 2.24882156e-01 -1.22887671e+00 1.05688497e-01
1.12885348e-01 -3.13061506e-01 1.43908188e-01 -5.68501890e-01
-8.52978289e-01 6.62314534e-01 5.74487150e-01 6.10937476e-01
1.03891551e+00 -5.54718852e-01 2.64609575e-01 8.12583327e-01
3.12536180e-01 -1.19555402e+00 -5.98049045e-01 6.65208757e-01
6.48230493e-01 -8.20308089e-01 3.94236259e-02 -5.13809860e-01
-1.65577367e-01 1.29528129e+00 1.02916040e-01 -2.77576566e-01
7.34279037e-01 3.10295373e-01 5.07919118e-02 2.14771658e-01
-3.76397967e-01 -3.80632609e-01 3.32325906e-01 4.33543354e-01
1.89502746e-01 -1.74583003e-01 -5.30315101e-01 3.66151243e-01
-3.02141756e-01 -1.59167070e-02 7.07584739e-01 7.96385944e-01
-5.84415674e-01 -9.88749266e-01 -6.02647781e-01 4.73276258e-01
-4.97711211e-01 1.33154213e-01 -1.08575486e-01 1.08281672e+00
-6.37895241e-02 9.89844263e-01 -2.89135784e-01 -2.27511004e-01
6.06597304e-01 2.29896158e-01 4.84072298e-01 -1.94897931e-02
-6.33747041e-01 2.27837503e-01 3.10191065e-01 -5.40885210e-01
-5.01594543e-01 -5.10694981e-01 -7.33004510e-01 -2.86115974e-01
-9.32780623e-01 4.24724996e-01 7.68405795e-01 1.21943378e+00
4.24647890e-02 5.17305791e-01 2.18060106e-01 -7.37670541e-01
-1.11495221e+00 -6.80483758e-01 -7.97312856e-01 6.84683248e-02
5.72108507e-01 -8.60533118e-01 -9.22375500e-01 -2.24013299e-01] | [6.6008405685424805, 3.559725761413574] |
3d36cdd3-cbdc-49ba-9291-3f4c81816658 | context-aware-cnns-for-person-head-detection | 1511.07917 | null | http://arxiv.org/abs/1511.07917v1 | http://arxiv.org/pdf/1511.07917v1.pdf | Context-aware CNNs for person head detection | Person detection is a key problem for many computer vision tasks. While face
detection has reached maturity, detecting people under a full variation of
camera view-points, human poses, lighting conditions and occlusions is still a
difficult challenge. In this work we focus on detecting human heads in natural
scenes. Starting from the recent local R-CNN object detector, we extend it with
two types of contextual cues. First, we leverage person-scene relations and
propose a Global CNN model trained to predict positions and scales of heads
directly from the full image. Second, we explicitly model pairwise relations
among objects and train a Pairwise CNN model using a structured-output
surrogate loss. The Local, Global and Pairwise models are combined into a joint
CNN framework. To train and test our full model, we introduce a large dataset
composed of 369,846 human heads annotated in 224,740 movie frames. We evaluate
our method and demonstrate improvements of person head detection against
several recent baselines in three datasets. We also show improvements of the
detection speed provided by our model. | ['Tuan-Hung Vu', 'Ivan Laptev', 'Anton Osokin'] | 2015-11-24 | context-aware-cnns-for-person-head-detection-1 | http://openaccess.thecvf.com/content_iccv_2015/html/Vu_Context-Aware_CNNs_for_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Vu_Context-Aware_CNNs_for_ICCV_2015_paper.pdf | iccv-2015-12 | ['head-detection'] | ['computer-vision'] | [-1.57979921e-01 -9.79808643e-02 1.75911993e-01 -7.87095428e-01
-6.09879673e-01 -4.04896677e-01 5.52200377e-01 -2.61497736e-01
-7.74806440e-01 4.48567718e-01 3.77295583e-01 3.93472672e-01
3.75306338e-01 -3.56027186e-01 -7.61665463e-01 -4.83117133e-01
-1.92548335e-02 5.26004314e-01 3.31183076e-01 1.02562971e-01
-1.51725248e-01 7.01942027e-01 -1.60301983e+00 -1.19674064e-01
2.44154185e-01 8.55694234e-01 -1.71034381e-01 8.35036635e-01
6.67767882e-01 6.22085333e-01 -6.18497968e-01 -6.86411083e-01
3.27545792e-01 -8.72951224e-02 -5.62140167e-01 4.05160397e-01
1.02549040e+00 -6.43513381e-01 -5.59682727e-01 9.74810243e-01
9.39152718e-01 1.37691125e-01 4.30087447e-01 -1.23410487e+00
-5.14173150e-01 2.78444648e-01 -1.00406158e+00 4.44352031e-02
6.07414067e-01 3.53006363e-01 8.88542295e-01 -1.07624352e+00
4.59782720e-01 1.70623958e+00 8.42801929e-01 7.60106266e-01
-1.07174301e+00 -7.92546630e-01 2.26406008e-01 2.02406719e-01
-1.67068577e+00 -8.12612653e-01 4.42543924e-01 -2.93027192e-01
8.15842152e-01 -2.71063261e-02 5.27829587e-01 1.05917847e+00
-2.30674252e-01 9.60681617e-01 8.52600098e-01 -2.88024575e-01
-1.65079162e-01 1.96866784e-02 2.47735620e-01 9.40639138e-01
4.19748247e-01 -1.19639486e-01 -6.28931940e-01 -6.55341707e-03
7.37276196e-01 8.46453682e-02 -1.71537206e-01 -3.23860049e-01
-8.27261329e-01 8.51602137e-01 6.81116045e-01 -1.56557128e-01
-1.49903074e-01 1.07617833e-01 2.54999012e-01 -4.49564964e-01
3.04237753e-01 -6.17048182e-02 -2.20114559e-01 4.75696415e-01
-9.90337253e-01 6.01511121e-01 8.34564686e-01 1.13885307e+00
5.36087334e-01 -2.86173731e-01 -5.25466561e-01 7.67098904e-01
4.42101061e-01 8.32178473e-01 -2.60385703e-02 -9.00042176e-01
2.44461238e-01 6.04352236e-01 5.32130562e-02 -9.72088039e-01
-8.41726243e-01 -2.57498562e-01 -6.21901751e-01 -2.12491155e-02
6.43939972e-01 -2.10067242e-01 -8.99891257e-01 1.85615027e+00
6.33936882e-01 -2.80704815e-02 -5.30846238e-01 1.09502125e+00
1.06711566e+00 2.61880517e-01 6.89039826e-02 6.88592717e-02
1.60058248e+00 -1.11063778e+00 -5.38032234e-01 -5.55191278e-01
9.99281630e-02 -4.55909610e-01 4.56185520e-01 1.51596546e-01
-1.36057734e+00 -6.70985639e-01 -6.63076520e-01 -5.41012168e-01
-2.85490066e-01 4.84152406e-01 2.09324494e-01 5.82560480e-01
-1.39418852e+00 1.34673119e-01 -7.33473241e-01 -9.00380492e-01
5.06511807e-01 8.49886477e-01 -4.91349965e-01 -2.15482593e-01
-7.16385126e-01 8.20975661e-01 -1.01220623e-01 3.39050800e-01
-9.17106390e-01 -2.12848857e-01 -1.01159465e+00 5.69443516e-02
3.62067848e-01 -9.22397077e-01 1.40185833e+00 -4.00626004e-01
-1.11824965e+00 1.25935411e+00 -5.61329246e-01 -3.94847065e-01
6.86385989e-01 -4.98182476e-01 -1.52740432e-02 4.79684845e-02
2.62460858e-01 1.04897726e+00 7.29064047e-01 -1.05983222e+00
-7.69933045e-01 -8.42840552e-01 -2.36072868e-01 7.33316690e-02
-3.14683855e-01 7.42116213e-01 -1.19093478e+00 -2.33616635e-01
-3.25967938e-01 -1.08346975e+00 -1.09253056e-01 2.63003349e-01
-6.81853473e-01 -4.78351623e-01 7.92135954e-01 -7.41248488e-01
7.54174531e-01 -1.74465215e+00 1.97576404e-01 -1.19347461e-02
4.79805708e-01 2.10623816e-01 7.12196678e-02 -2.30299562e-01
1.39062345e-01 -3.25945377e-01 -1.77070703e-02 -1.26182497e+00
2.06696317e-01 -4.72655371e-02 1.97589681e-01 8.17159593e-01
1.44784093e-01 1.14105105e+00 -5.44724762e-01 -7.64905035e-01
2.76318908e-01 9.25671279e-01 -6.11280859e-01 3.57035190e-01
3.46302271e-01 4.05291617e-01 -1.12123355e-01 9.52823758e-01
8.05082798e-01 -2.85145849e-01 1.37923965e-02 -2.19173938e-01
7.59454668e-02 -1.08856343e-01 -1.01504529e+00 1.39144194e+00
3.88976187e-02 6.24774933e-01 4.04736280e-01 -5.67923307e-01
7.50442564e-01 1.63489625e-01 1.32879198e-01 -5.36926270e-01
3.81576747e-01 -1.93327352e-01 -1.81218445e-01 -4.04136091e-01
4.59797442e-01 3.17779779e-02 -6.70868084e-02 -1.39626283e-02
3.33097100e-01 3.74171853e-01 1.42070785e-01 9.65026319e-02
1.08007896e+00 5.85023239e-02 3.28735203e-01 -1.26436710e-01
5.48783600e-01 -6.69042647e-01 6.23094738e-01 8.24298799e-01
-7.24990368e-01 1.00509584e+00 3.85768443e-01 -6.42631054e-01
-9.10091281e-01 -8.77594113e-01 -7.40201175e-02 1.47602379e+00
-9.89115387e-02 -2.10551441e-01 -1.00004137e+00 -7.32767045e-01
1.96092933e-01 6.07788451e-02 -9.08744097e-01 1.91661000e-01
-9.40340281e-01 -7.71755099e-01 6.16666555e-01 8.81108701e-01
4.88586724e-01 -1.12956595e+00 -5.36420465e-01 -1.33733675e-01
3.83390416e-03 -1.64980602e+00 -8.05628240e-01 -9.97643322e-02
-2.41730422e-01 -1.01209092e+00 -9.55090761e-01 -6.74814463e-01
4.43498939e-01 4.41929281e-01 1.32064462e+00 7.50507638e-02
-6.56710505e-01 4.19582516e-01 7.23783374e-02 -4.75608319e-01
3.70923132e-01 3.05342395e-02 4.07595098e-01 9.23926085e-02
8.23606312e-01 -1.55989334e-01 -8.99173200e-01 3.20571750e-01
-2.66852796e-01 -2.75399804e-01 4.58287537e-01 4.61004764e-01
1.54468566e-01 -4.98964638e-01 2.37271458e-01 -6.16345167e-01
1.24531657e-01 -1.47719845e-01 -6.16899669e-01 1.50769114e-01
-5.23918599e-04 -4.41699356e-01 1.77274048e-01 -1.46981299e-01
-1.03325915e+00 5.72515905e-01 -5.71183115e-02 -5.37859201e-01
-4.32366073e-01 -2.09967524e-01 -5.53017795e-01 -4.96757478e-02
5.14133275e-01 3.58682461e-02 -1.42545268e-01 -4.43054438e-01
3.48470688e-01 4.89550203e-01 1.01999199e+00 -3.92369211e-01
8.43277335e-01 7.39880741e-01 -1.40868694e-01 -1.03662777e+00
-1.03243327e+00 -7.79505789e-01 -1.02951658e+00 -2.93214023e-01
1.18499935e+00 -1.24211681e+00 -1.15225434e+00 6.49966538e-01
-1.42878425e+00 -2.03081861e-01 1.96836829e-01 1.12670682e-01
-7.46652931e-02 3.70403767e-01 -7.76934147e-01 -1.08921516e+00
-3.99138242e-01 -1.15590608e+00 1.72198546e+00 4.19500619e-01
-8.99694189e-02 -8.21669459e-01 -7.94051439e-02 4.27780956e-01
2.08777472e-01 2.15437591e-01 -1.35065719e-01 -5.28483510e-01
-5.50676346e-01 -6.09486520e-01 -6.05773449e-01 -3.25848572e-02
-1.50715545e-01 -6.61400259e-02 -1.31534028e+00 -5.84005833e-01
-2.68304378e-01 -4.39753503e-01 1.06860030e+00 7.68302321e-01
1.06452549e+00 -2.05683112e-01 -6.31477833e-01 8.04683089e-01
9.36640084e-01 -2.14215919e-01 3.50917578e-01 1.38040751e-01
1.08040071e+00 8.98413062e-01 4.16601636e-02 5.66679597e-01
6.91376030e-01 8.81011963e-01 2.36179247e-01 -2.44123355e-01
-1.81261137e-01 -1.73971951e-01 3.78169954e-01 1.69657812e-01
-2.31526539e-01 -2.00892866e-01 -8.93203497e-01 4.69531119e-01
-1.82723379e+00 -8.90546679e-01 5.15782647e-02 2.06056094e+00
4.50712562e-01 1.24136887e-01 8.98219585e-01 -2.79010475e-01
1.01919568e+00 -1.64043549e-02 -5.93354225e-01 3.61982912e-01
-9.02936459e-02 4.54358906e-02 6.48187518e-01 3.77734870e-01
-1.52546096e+00 1.04004049e+00 6.55510616e+00 1.56055510e-01
-8.34766269e-01 1.07438520e-01 6.42605245e-01 -4.27716583e-01
6.96975648e-01 -4.90304798e-01 -1.66214037e+00 2.31291249e-01
6.05238795e-01 3.45759273e-01 1.04272150e-01 9.10320699e-01
1.02622695e-01 9.42818820e-02 -1.42480314e+00 1.25801265e+00
6.36435390e-01 -6.51502550e-01 -3.58018786e-01 2.92430639e-01
5.75138807e-01 8.77536237e-02 1.14382386e-01 3.58161211e-01
3.52958798e-01 -1.30184734e+00 8.58024001e-01 4.09765899e-01
5.23859262e-01 -7.32495666e-01 8.10689509e-01 1.74923643e-01
-1.42821777e+00 -1.86610684e-01 -3.80781740e-01 -3.23602445e-02
2.59469539e-01 2.37801284e-01 -6.39845312e-01 1.24039603e-02
1.13930118e+00 7.21607685e-01 -1.08998525e+00 1.14938164e+00
-2.07741603e-01 2.63406307e-01 -4.95343804e-01 1.61883861e-01
-2.26983160e-01 1.75330326e-01 3.37060183e-01 1.61679935e+00
-8.34164489e-03 2.00933427e-01 3.46071661e-01 8.47971022e-01
-4.02153999e-01 -2.61528164e-01 -5.56114554e-01 5.27323484e-01
3.59008640e-01 1.62408578e+00 -6.20621085e-01 -2.28806823e-01
-6.33680284e-01 9.75888789e-01 5.61452150e-01 3.03889930e-01
-9.09017205e-01 -7.53163546e-02 5.95115304e-01 1.20164685e-01
3.21206480e-01 -9.34211761e-02 2.99314428e-02 -1.16515350e+00
8.12164992e-02 -6.77799582e-01 3.26479107e-01 -5.78003049e-01
-1.42681158e+00 4.01597530e-01 8.17809632e-05 -6.32837355e-01
-1.95214406e-01 -8.40852976e-01 -7.15405047e-01 7.79359221e-01
-1.39390874e+00 -1.71828878e+00 -6.65171325e-01 7.55013704e-01
6.00454926e-01 -1.07879207e-01 4.17388737e-01 3.71628046e-01
-1.02519917e+00 9.81245160e-01 -4.65888441e-01 7.73870111e-01
1.02197039e+00 -1.32849705e+00 6.30227745e-01 8.57946038e-01
-2.10313842e-01 9.13630068e-01 7.67454445e-01 -4.79893714e-01
-1.36820400e+00 -1.21204960e+00 8.97874951e-01 -8.81916523e-01
2.19959870e-01 -8.82880628e-01 -7.75656939e-01 1.07084131e+00
1.33569539e-01 5.16152382e-01 3.95098656e-01 3.03333253e-01
-4.97448087e-01 -2.15268694e-02 -1.05502105e+00 3.21839631e-01
1.14122164e+00 -4.62892979e-01 -3.42197299e-01 3.53099108e-01
5.14356196e-01 -4.53071922e-01 -3.55276346e-01 2.74366975e-01
8.56931865e-01 -1.08252227e+00 1.32156706e+00 -4.04078126e-01
1.31257817e-01 -2.44238600e-01 -1.28410444e-01 -6.46465421e-01
-4.63393003e-01 -4.61336792e-01 -3.35106432e-01 1.30293489e+00
1.34165481e-01 -2.18344256e-01 9.76289451e-01 9.99511957e-01
1.54714525e-01 -3.86588365e-01 -7.72261679e-01 -5.05279481e-01
6.93854168e-02 -1.91655401e-02 3.53376448e-01 4.17465448e-01
-2.44332865e-01 6.55697286e-01 -7.34566629e-01 2.73179919e-01
1.16895604e+00 -2.15293959e-01 1.18974614e+00 -1.29416180e+00
-2.82562107e-01 -3.16479594e-01 -4.89258260e-01 -1.24243510e+00
2.69413292e-01 -5.13193727e-01 3.16422671e-01 -1.32100809e+00
1.01084995e+00 1.94655851e-01 -7.64437318e-02 4.67181414e-01
-4.37999845e-01 6.11757398e-01 4.16565090e-01 2.13264048e-01
-1.09541774e+00 3.93493205e-01 7.77267516e-01 -6.66667670e-02
-2.00410131e-02 -5.67757487e-02 -5.56325257e-01 9.92874563e-01
3.97108972e-01 -1.42178714e-01 2.61512995e-01 -5.10472894e-01
-1.72082260e-01 -8.26333091e-02 8.33576918e-01 -1.19917381e+00
7.42583334e-01 2.93828249e-01 1.08044505e+00 -7.85110056e-01
6.36008799e-01 -3.91707242e-01 -4.40185487e-01 2.60304958e-01
-2.24087894e-01 1.07700109e-01 6.55421764e-02 6.25138819e-01
1.75184771e-01 2.28232309e-01 1.06481910e+00 -9.34721753e-02
-4.51604098e-01 5.16432345e-01 5.70901036e-02 -3.82439233e-02
8.24826062e-01 -5.99519052e-02 -1.80724993e-01 -3.75199318e-01
-7.74520516e-01 4.47508574e-01 5.12844503e-01 4.16413695e-01
6.24368668e-01 -1.09740925e+00 -8.18746507e-01 1.29204661e-01
6.31787926e-02 7.41756931e-02 3.14952172e-02 7.67966092e-01
-1.14421085e-01 5.72720468e-01 3.07681393e-02 -7.88519681e-01
-1.68597937e+00 5.87391138e-01 4.81057107e-01 1.05580106e-01
-5.36249936e-01 1.19031107e+00 6.62081480e-01 -5.67882121e-01
6.33969724e-01 -2.20028707e-03 -4.00055557e-01 -7.90797621e-02
8.92342389e-01 2.98562467e-01 -1.20611563e-01 -1.27117026e+00
-5.72010934e-01 7.89600551e-01 -2.09301159e-01 -1.10080831e-01
1.27556360e+00 -1.99983865e-01 -5.15432991e-02 2.69905403e-02
1.27872372e+00 -1.77689686e-01 -1.45542145e+00 -4.16954279e-01
-6.92575946e-02 -4.54956770e-01 -2.13126600e-01 -3.12362283e-01
-1.13120437e+00 1.03073037e+00 9.00662243e-01 -2.14314833e-01
8.17833066e-01 3.86083603e-01 6.86408401e-01 3.55357051e-01
2.58504242e-01 -9.20740545e-01 1.97619945e-01 4.60255891e-01
6.74173772e-01 -1.73925233e+00 2.77577396e-02 -3.01697075e-01
-3.83115143e-01 8.61130834e-01 8.72136474e-01 -5.64199165e-02
3.91755015e-01 2.01656342e-01 -1.63503855e-01 -3.31881136e-01
-3.08915734e-01 -5.70974469e-01 3.95023227e-01 5.76086342e-01
6.13694906e-01 -1.30887504e-03 2.44696230e-01 6.20531619e-01
-2.48482436e-01 3.07230242e-02 1.03573933e-01 8.41942370e-01
-5.02255201e-01 -7.29473233e-01 -6.85346007e-01 1.95218831e-01
-6.38769984e-01 2.00133905e-01 -5.49037576e-01 7.38771677e-01
2.24381864e-01 9.28538084e-01 6.95911497e-02 -1.93345100e-01
5.00367284e-01 -6.01370772e-03 5.00048280e-01 -6.53934121e-01
-3.41354251e-01 2.59670913e-01 -2.98133671e-01 -5.25765717e-01
-4.84109014e-01 -1.03020227e+00 -7.91045845e-01 -4.19965059e-01
-3.98284972e-01 -3.64514172e-01 4.57784504e-01 9.90511775e-01
7.27271810e-02 2.53459573e-01 2.97122926e-01 -1.56911731e+00
-4.69093770e-01 -1.07518661e+00 -5.50306976e-01 4.65468913e-01
5.28053880e-01 -7.77613699e-01 -2.87496984e-01 2.40816563e-01] | [13.788528442382812, 0.4282793700695038] |
94f5c5a6-cf30-469f-96e3-0a1a014182f3 | megane-morphable-eyeglass-and-avatar-network | 2302.04868 | null | https://arxiv.org/abs/2302.04868v1 | https://arxiv.org/pdf/2302.04868v1.pdf | MEGANE: Morphable Eyeglass and Avatar Network | Eyeglasses play an important role in the perception of identity. Authentic virtual representations of faces can benefit greatly from their inclusion. However, modeling the geometric and appearance interactions of glasses and the face of virtual representations of humans is challenging. Glasses and faces affect each other's geometry at their contact points, and also induce appearance changes due to light transport. Most existing approaches do not capture these physical interactions since they model eyeglasses and faces independently. Others attempt to resolve interactions as a 2D image synthesis problem and suffer from view and temporal inconsistencies. In this work, we propose a 3D compositional morphable model of eyeglasses that accurately incorporates high-fidelity geometric and photometric interaction effects. To support the large variation in eyeglass topology efficiently, we employ a hybrid representation that combines surface geometry and a volumetric representation. Unlike volumetric approaches, our model naturally retains correspondences across glasses, and hence explicit modification of geometry, such as lens insertion and frame deformation, is greatly simplified. In addition, our model is relightable under point lights and natural illumination, supporting high-fidelity rendering of various frame materials, including translucent plastic and metal within a single morphable model. Importantly, our approach models global light transport effects, such as casting shadows between faces and glasses. Our morphable model for eyeglasses can also be fit to novel glasses via inverse rendering. We compare our approach to state-of-the-art methods and demonstrate significant quality improvements. | ['Jason Saragih', 'Hongdong Li', 'Stephen Lombardi', 'Tomas Simon', 'Shunsuke Saito', 'Junxuan Li'] | 2023-02-09 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Li_MEGANE_Morphable_Eyeglass_and_Avatar_Network_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_MEGANE_Morphable_Eyeglass_and_Avatar_Network_CVPR_2023_paper.pdf | cvpr-2023-1 | ['inverse-rendering'] | ['computer-vision'] | [ 1.47916645e-01 2.90755540e-01 4.13264692e-01 -1.69462636e-01
-4.83892970e-02 -6.21298313e-01 5.01029909e-01 -4.09007728e-01
3.81866664e-01 5.15628040e-01 1.64691154e-02 2.06852645e-01
5.41462779e-01 -9.75931704e-01 -9.31995749e-01 -3.60790640e-01
3.67118955e-01 4.09381479e-01 4.03053552e-01 -3.07337165e-01
-1.45953270e-02 8.37873578e-01 -1.86370146e+00 1.69445530e-01
7.71883786e-01 8.97328079e-01 -2.34269440e-01 5.42106271e-01
-1.38720602e-01 4.22572136e-01 -3.96187842e-01 -6.94651544e-01
4.07862723e-01 -3.54552716e-01 -3.90863836e-01 6.60524011e-01
1.06674743e+00 -5.28293192e-01 -2.88204521e-01 6.75567627e-01
1.82114601e-01 2.05885097e-02 6.54642642e-01 -1.15103984e+00
-8.90021205e-01 -3.67045820e-01 -8.25829446e-01 -7.58128762e-01
7.26613462e-01 1.69714823e-01 5.40018082e-01 -1.03996480e+00
8.50284874e-01 1.60765433e+00 7.19555199e-01 7.30143785e-01
-1.56707156e+00 -6.01806939e-01 3.69250804e-01 -3.56933028e-01
-1.49303758e+00 -7.24000633e-01 9.56837773e-01 -5.86661875e-01
6.20982766e-01 5.57820737e-01 1.12301826e+00 7.14074075e-01
3.04582596e-01 2.19281539e-01 1.14957511e+00 -4.28432971e-01
-7.80138969e-02 3.03254664e-01 -2.79134482e-01 1.08076918e+00
1.15543187e-01 2.26893783e-01 -6.99512303e-01 -3.81363302e-01
1.36193871e+00 7.54843950e-02 -4.89935368e-01 -5.88888705e-01
-8.65995824e-01 3.21523398e-01 2.41332456e-01 -4.05906379e-01
3.85337286e-02 4.50185776e-01 -3.10156524e-01 1.00294746e-01
8.13155174e-01 1.48074642e-01 -1.26934201e-01 2.00991705e-01
-6.78343594e-01 4.69909161e-01 6.78641796e-01 1.24325693e+00
1.17811620e+00 1.43567072e-02 2.40175679e-01 8.03291798e-01
7.02045858e-01 7.29844391e-01 -1.87041566e-01 -1.24047387e+00
-7.59598613e-02 7.65205324e-01 2.59795368e-01 -1.16147053e+00
-8.45337287e-02 8.02536309e-02 -4.17953581e-01 8.17825615e-01
1.64262086e-01 2.15658337e-01 -1.22057533e+00 1.43713665e+00
8.36783707e-01 4.14006144e-01 -3.71812314e-01 7.93537080e-01
1.12872422e+00 2.44733855e-01 -4.35453415e-01 -1.67586133e-01
1.32432115e+00 -6.02887928e-01 -8.23219955e-01 -1.30170032e-01
1.49508879e-01 -1.17911422e+00 1.12197423e+00 1.13885656e-01
-1.77087116e+00 -2.10587040e-01 -9.00554955e-01 -6.02841616e-01
1.56047821e-01 -1.29056334e-01 4.63475436e-01 6.65044785e-01
-1.33944225e+00 4.63902175e-01 -7.38472223e-01 -7.58156776e-02
2.68059611e-01 6.22429371e-01 -2.73569703e-01 1.15274517e-02
-6.40794039e-01 7.19903052e-01 -7.44556785e-01 9.06364545e-02
-4.39304948e-01 -1.01369727e+00 -9.97344434e-01 -2.86485165e-01
2.15164453e-01 -1.15358269e+00 1.04123604e+00 -9.32316601e-01
-2.08260989e+00 1.23657656e+00 -6.19950533e-01 3.05407107e-01
6.79500997e-01 2.18411498e-02 -6.62164986e-02 5.42651266e-02
-4.63145554e-01 4.27837610e-01 1.03521562e+00 -1.96210301e+00
1.95697382e-01 -5.00163257e-01 3.45158935e-01 2.63023496e-01
-2.21346188e-02 -1.11031048e-01 -6.26189888e-01 -5.74180007e-01
3.81427169e-01 -1.13950753e+00 8.34411159e-02 8.22441578e-01
-4.15883422e-01 5.79008281e-01 1.11104512e+00 -5.30725002e-01
8.34713101e-01 -2.05271649e+00 1.46561965e-01 2.91434944e-01
5.15199721e-01 -3.59063782e-02 1.10142194e-01 2.72109121e-01
2.44688436e-01 1.93853989e-01 -2.17072010e-01 -1.11756480e+00
-3.46475579e-02 2.35456303e-01 -2.21819669e-01 5.89800596e-01
-3.60804722e-02 8.59052956e-01 -4.71755058e-01 -4.32496041e-01
3.31494689e-01 1.13236272e+00 -7.08440363e-01 7.07452372e-02
-2.89612800e-01 6.23692155e-01 -1.72707349e-01 8.26645136e-01
1.08252156e+00 -1.33458838e-01 1.91133097e-01 -3.82298052e-01
-5.34612201e-02 1.02598235e-01 -1.30075073e+00 1.42682385e+00
-6.50407970e-01 4.96195734e-01 4.50329423e-01 2.25304142e-01
9.82863426e-01 2.66121596e-01 3.16475868e-01 -4.45682228e-01
1.67478397e-01 1.16032824e-01 -4.50202167e-01 -1.04190163e-01
6.82925761e-01 -3.98913920e-01 5.22073328e-01 3.47281635e-01
-5.36268830e-01 -8.51329267e-01 -6.00053370e-01 8.41659382e-02
5.36253333e-01 4.48873341e-01 -1.84082657e-01 -1.44526809e-01
2.05170810e-01 -4.32278067e-01 5.49855888e-01 -7.81103075e-02
3.67226750e-01 1.17138505e+00 1.80986747e-01 -4.23410386e-01
-9.35065866e-01 -1.13041973e+00 -3.64406228e-01 3.11058760e-01
6.43260539e-01 -3.97007883e-01 -8.42195332e-01 -6.48859888e-02
2.31037989e-01 3.73299062e-01 -6.92690372e-01 -5.59227988e-02
-5.86812437e-01 -4.30432111e-01 7.52454326e-02 1.97798684e-01
2.96802074e-01 -4.59973872e-01 -4.26162213e-01 -3.89229208e-02
4.47727442e-02 -1.19264436e+00 -7.26954341e-01 -9.71872866e-01
-9.09981549e-01 -1.17225039e+00 -5.64042568e-01 -4.54109669e-01
9.88870621e-01 4.21444535e-01 1.36177576e+00 5.31067491e-01
-4.94639248e-01 7.92708814e-01 2.20218524e-02 -3.37645859e-01
-5.12299538e-01 -6.66480541e-01 8.67754593e-02 4.98282492e-01
-3.93974364e-01 -8.73066366e-01 -8.38974476e-01 8.23175132e-01
-8.31138909e-01 4.36089545e-01 -3.90232980e-01 3.59656811e-01
7.21386969e-01 -5.39275944e-01 -3.06890041e-01 -8.09278548e-01
9.51997489e-02 1.37973398e-01 -7.67801583e-01 1.00412823e-01
-1.66020438e-01 -3.03126425e-01 1.18582614e-01 -4.33239281e-01
-1.32619250e+00 1.27151217e-02 3.19902837e-01 -8.02087128e-01
1.69985175e-01 -3.27171385e-01 -3.62905502e-01 -7.70197213e-01
4.33664024e-01 -2.65232891e-01 1.64915115e-01 -5.64565778e-01
3.38249892e-01 3.09128076e-01 2.43295997e-01 -7.02616990e-01
1.07255411e+00 1.22771847e+00 3.60179275e-01 -1.17402434e+00
-1.83861226e-01 1.54868364e-01 -7.11113811e-01 -4.41727728e-01
5.17899930e-01 -8.52226913e-01 -1.05087721e+00 7.41418660e-01
-1.46756172e+00 -5.11028469e-01 -4.43353206e-01 2.56388456e-01
-4.64390963e-01 3.55440646e-01 -5.54484427e-01 -9.70987558e-01
1.33087695e-01 -1.27162290e+00 1.58874786e+00 1.32179216e-01
-2.03173131e-01 -1.05482924e+00 -9.52166542e-02 3.75611603e-01
1.75711259e-01 7.88112640e-01 6.94026351e-01 9.72311616e-01
-1.19859684e+00 8.72270837e-02 4.28384990e-02 3.11093815e-02
4.74909008e-01 6.76585793e-01 -1.29737294e+00 -2.23885313e-01
-6.31493703e-02 9.05835629e-03 4.86909926e-01 4.50822592e-01
7.64570475e-01 -2.16338262e-01 -4.82040256e-01 1.08974588e+00
1.34525204e+00 4.16178778e-02 8.87525201e-01 -1.96747959e-01
1.24163139e+00 9.31199074e-01 2.33656332e-01 3.48534167e-01
5.60224116e-01 1.14364970e+00 7.07197666e-01 -5.23677707e-01
-6.03667080e-01 -8.94297808e-02 2.83402741e-01 6.32923126e-01
-5.54229379e-01 -1.98680922e-01 -6.14656746e-01 1.30459577e-01
-1.46681273e+00 -5.63085675e-01 -6.74892008e-01 2.48862910e+00
6.83230877e-01 -4.03896302e-01 -7.23904744e-02 -1.90022305e-01
4.57958251e-01 -6.80879354e-02 -4.56005961e-01 -5.25160611e-01
-2.64251620e-01 1.13730527e-01 3.55975598e-01 1.05518723e+00
-4.76458192e-01 9.95861113e-01 6.77118349e+00 2.36026719e-01
-1.10735810e+00 9.32875052e-02 5.77249765e-01 -5.38302720e-01
-1.03631568e+00 1.33691072e-01 -7.31391311e-01 1.77381232e-01
3.06228638e-01 8.89045671e-02 4.58115757e-01 2.63503671e-01
3.25216085e-01 -8.18473697e-02 -1.01980078e+00 1.00978255e+00
2.93927133e-01 -1.42425907e+00 1.84285671e-01 5.19302130e-01
9.89187360e-01 -5.60230076e-01 3.00220996e-01 -6.73351645e-01
1.77399695e-01 -9.91467118e-01 1.01595426e+00 7.65792310e-01
1.27304041e+00 -4.48366135e-01 -1.03390120e-01 -2.12198704e-01
-1.27661002e+00 5.66099107e-01 -1.74172640e-01 -4.86861132e-02
3.84433001e-01 5.65045953e-01 -4.16077346e-01 3.67715061e-01
4.91571039e-01 4.92602438e-01 -2.40446880e-01 7.28302002e-01
-1.62808661e-04 3.93538438e-02 -3.58432472e-01 5.57528317e-01
-4.03402001e-01 -6.26241028e-01 7.05992758e-01 5.37055314e-01
2.83343941e-01 3.05493385e-01 4.77086157e-02 1.23621166e+00
-1.83742285e-01 7.86562040e-02 -7.26777971e-01 4.35580194e-01
3.64788383e-01 8.66722107e-01 -5.82019389e-01 -9.39263776e-02
-5.42095959e-01 1.11954725e+00 1.51141033e-01 6.35234296e-01
-8.29220593e-01 1.77852273e-01 1.27339482e+00 9.70492303e-01
-7.31033310e-02 -4.51847404e-01 -4.30154204e-01 -1.33142149e+00
4.38295484e-01 -5.19632876e-01 -5.19484222e-01 -9.22769368e-01
-9.04065907e-01 6.25419378e-01 -9.26492363e-02 -1.22466373e+00
2.55700409e-01 -4.19904172e-01 -3.97191823e-01 1.02617908e+00
-1.62883806e+00 -1.70623755e+00 -5.24779022e-01 6.44567609e-01
2.41608158e-01 4.28086162e-01 7.86078393e-01 1.62213519e-01
-2.86644876e-01 6.14832997e-01 -7.34003484e-02 -4.21247303e-01
7.14340329e-01 -9.00442302e-01 6.71923399e-01 6.22430325e-01
-1.38073623e-01 7.84960568e-01 5.66604018e-01 -7.38480628e-01
-1.68002832e+00 -1.03978646e+00 4.60641146e-01 -7.03725994e-01
9.36975032e-02 -6.91481531e-01 -1.03104258e+00 9.26112771e-01
-2.43333708e-02 2.17430070e-01 5.54281950e-01 -1.57303989e-01
-4.49395210e-01 6.43504933e-02 -1.29821670e+00 8.85272384e-01
1.47344649e+00 -6.59360468e-01 -5.86966192e-03 2.41272524e-01
6.49506569e-01 -9.76957381e-01 -8.23796093e-01 2.32414663e-01
9.17885005e-01 -1.23992074e+00 1.16041839e+00 -1.13319591e-01
1.92840114e-01 -5.04651904e-01 -1.86969549e-03 -1.22122467e+00
-9.72259268e-02 -1.22056055e+00 -2.72716641e-01 1.18273497e+00
6.80676922e-02 -8.77184510e-01 7.89400220e-01 1.30647933e+00
-3.41770202e-01 -6.61652744e-01 -6.25823855e-01 -6.58773899e-01
-2.23037973e-01 -1.32885724e-01 9.96131122e-01 1.03954792e+00
-3.67420584e-01 -2.34315574e-01 -3.70018274e-01 3.14039379e-01
7.77766526e-01 2.45481715e-01 1.11840296e+00 -1.45420098e+00
-2.77643263e-01 -1.96575999e-01 -4.53750819e-01 -1.00209033e+00
1.37144163e-01 -5.70517182e-01 -1.97591051e-01 -1.33600664e+00
-1.13864273e-01 -7.02655733e-01 7.23129570e-01 2.34622881e-01
-1.09701045e-02 7.75575817e-01 2.44839773e-01 2.83221871e-01
3.76762331e-01 6.92978561e-01 1.78517330e+00 1.13991112e-01
-5.22052348e-01 -2.08643749e-01 -4.34320658e-01 1.14222538e+00
3.48249048e-01 -1.00018591e-01 -5.37037134e-01 -7.33713627e-01
3.11176717e-01 -4.39213552e-02 6.38657153e-01 -5.61465561e-01
-1.59842402e-01 -3.23574215e-01 3.16204160e-01 -1.27422854e-01
1.07021689e+00 -7.98070371e-01 7.65084982e-01 -5.86186498e-02
2.90475965e-01 -7.66290799e-02 3.94009411e-01 5.08656442e-01
1.47356153e-01 2.42245674e-01 9.32370007e-01 -2.95272321e-01
8.98568928e-02 7.13577867e-01 -1.17339775e-01 -2.21418887e-01
9.57793176e-01 -7.84036040e-01 -2.70926863e-01 -4.29879874e-01
-7.55805790e-01 -2.99445868e-01 1.59013176e+00 1.73779786e-01
8.54563653e-01 -1.31238472e+00 -5.42280614e-01 5.34307182e-01
-1.07482411e-01 3.03612441e-01 2.45142251e-01 5.93051612e-01
-8.51586044e-01 -2.72522241e-01 9.55578014e-02 -6.08473122e-01
-1.75154710e+00 6.70224428e-02 7.30935872e-01 4.71433640e-01
-7.50521064e-01 9.49900329e-01 8.78930926e-01 -3.29720855e-01
-1.05132639e-01 -5.36180079e-01 3.25711161e-01 -2.45663702e-01
4.46444660e-01 4.30011094e-01 1.45395264e-01 -1.05613387e+00
-1.83923632e-01 1.39191031e+00 1.84008464e-01 -1.31177202e-01
9.54553306e-01 -3.29080582e-01 -3.16728204e-01 4.23398584e-01
9.94719923e-01 5.16223550e-01 -1.42870069e+00 -1.21037878e-01
-9.61457312e-01 -1.11594605e+00 9.44752246e-03 -2.84827292e-01
-1.24990237e+00 8.73187244e-01 1.18571274e-01 -6.75324872e-02
8.53864312e-01 -1.23599045e-01 1.00066829e+00 -2.95532823e-01
5.06831527e-01 -7.12381661e-01 -2.59834956e-02 1.17854789e-01
1.13112009e+00 -8.05555046e-01 1.87742576e-01 -1.40819335e+00
-2.58091152e-01 9.61352587e-01 6.08603299e-01 -1.55166052e-02
7.22160518e-01 3.69086266e-01 1.44675627e-01 -4.34207350e-01
-3.95814389e-01 1.16395563e-01 5.85449636e-01 6.84218228e-01
5.37755728e-01 6.93887621e-02 3.16793770e-02 -1.83495864e-01
-2.76628047e-01 -2.31601462e-01 7.47699201e-01 8.95955741e-01
5.13193384e-02 -1.27109075e+00 -6.95654988e-01 1.08149692e-01
-6.01242594e-02 -1.19895837e-03 -5.86780250e-01 7.02360809e-01
1.37629598e-01 8.08283865e-01 3.00499111e-01 -1.56424418e-01
5.65256059e-01 -2.03779668e-01 9.55239534e-01 -7.57847846e-01
-3.80206078e-01 2.86930323e-01 3.40754003e-03 -8.08690310e-01
-5.10244668e-01 -6.53376758e-01 -1.28429806e+00 -6.00017607e-01
-3.91485572e-01 -4.89040554e-01 6.61529660e-01 5.36352396e-01
5.90200007e-01 3.35543782e-01 6.57488465e-01 -1.37740672e+00
2.69237787e-01 -2.58747041e-01 -8.55443478e-01 5.15153944e-01
4.97824520e-01 -1.06635332e+00 -4.05307889e-01 3.55033517e-01] | [12.738277435302734, -0.33872440457344055] |
54b751e9-d015-44eb-add1-44e9b9c66d4a | joint-visual-grounding-and-tracking-with | 2303.12027 | null | https://arxiv.org/abs/2303.12027v1 | https://arxiv.org/pdf/2303.12027v1.pdf | Joint Visual Grounding and Tracking with Natural Language Specification | Tracking by natural language specification aims to locate the referred target in a sequence based on the natural language description. Existing algorithms solve this issue in two steps, visual grounding and tracking, and accordingly deploy the separated grounding model and tracking model to implement these two steps, respectively. Such a separated framework overlooks the link between visual grounding and tracking, which is that the natural language descriptions provide global semantic cues for localizing the target for both two steps. Besides, the separated framework can hardly be trained end-to-end. To handle these issues, we propose a joint visual grounding and tracking framework, which reformulates grounding and tracking as a unified task: localizing the referred target based on the given visual-language references. Specifically, we propose a multi-source relation modeling module to effectively build the relation between the visual-language references and the test image. In addition, we design a temporal modeling module to provide a temporal clue with the guidance of the global semantic information for our model, which effectively improves the adaptability to the appearance variations of the target. Extensive experimental results on TNL2K, LaSOT, OTB99, and RefCOCOg demonstrate that our method performs favorably against state-of-the-art algorithms for both tracking and grounding. Code is available at https://github.com/lizhou-cs/JointNLT. | ['Zhenyu He', 'Kaige Mao', 'Zikun Zhou', 'Li Zhou'] | 2023-03-21 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhou_Joint_Visual_Grounding_and_Tracking_With_Natural_Language_Specification_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhou_Joint_Visual_Grounding_and_Tracking_With_Natural_Language_Specification_CVPR_2023_paper.pdf | cvpr-2023-1 | ['visual-grounding', 'visual-tracking'] | ['computer-vision', 'computer-vision'] | [-4.69489932e-01 -3.96336526e-01 -5.32665253e-01 -1.80308685e-01
-6.51019573e-01 -7.49070168e-01 6.67002916e-01 -1.38818398e-01
-1.94483444e-01 3.47589672e-01 -1.61820520e-02 -1.60960346e-01
2.80446321e-01 -4.52871829e-01 -5.58595657e-01 -5.98376632e-01
1.84004530e-01 2.53946513e-01 9.39933002e-01 -6.71417266e-02
-2.39513465e-03 3.45100999e-01 -1.25670862e+00 1.63546026e-01
5.83493114e-01 1.04746306e+00 5.53246140e-01 3.50988090e-01
-4.33612019e-01 7.88342893e-01 -3.07717234e-01 -2.71611065e-01
2.36068994e-01 -4.02179033e-01 -4.47269380e-01 3.85935694e-01
7.49915242e-01 -3.42988878e-01 -6.29803538e-01 1.36331153e+00
2.88574129e-01 -4.10219207e-02 1.90723687e-01 -1.72461700e+00
-6.13822639e-01 3.58926475e-01 -7.78718531e-01 2.14489594e-01
3.98003131e-01 4.24511373e-01 8.42694938e-01 -9.57376599e-01
6.52948380e-01 1.38749516e+00 6.25341654e-01 7.48440683e-01
-8.48670125e-01 -9.21496093e-01 7.36549497e-01 2.04818651e-01
-1.57940876e+00 -4.96908605e-01 6.32888496e-01 -6.11375749e-01
2.33748049e-01 -5.51409200e-02 7.25183308e-01 8.89278889e-01
7.00352341e-02 9.56612170e-01 7.30497301e-01 -2.78464884e-01
-1.07964844e-01 -2.44314298e-02 3.09355706e-01 1.12187493e+00
4.21171218e-01 5.89693427e-01 -6.75443649e-01 -2.73639690e-02
6.85618579e-01 1.90882906e-01 -3.68964791e-01 -7.57015288e-01
-1.41222715e+00 4.48982865e-01 6.33064568e-01 2.36365527e-01
-1.21608868e-01 4.22980398e-01 2.67438740e-01 -4.50976864e-02
2.47444123e-01 -3.34871292e-01 -2.06499040e-01 2.92885900e-01
-1.04846883e+00 1.84795037e-01 3.15319479e-01 1.57342577e+00
6.92102253e-01 1.77668640e-03 -5.39516389e-01 2.64793515e-01
1.05832982e+00 7.95417488e-01 2.08725587e-01 -5.67387283e-01
5.50440013e-01 6.41920626e-01 3.84409189e-01 -1.05087388e+00
-2.57727295e-01 -5.54295421e-01 -4.02516127e-01 1.28895521e-01
6.30269766e-01 -3.77949737e-02 -1.09662247e+00 1.87936246e+00
7.13630915e-01 5.59947133e-01 -4.86729033e-02 1.13598716e+00
1.05950105e+00 4.62328732e-01 3.57175589e-01 -1.10914983e-01
1.64516509e+00 -1.21040499e+00 -8.33508015e-01 -3.97176921e-01
6.77636683e-01 -8.89883220e-01 8.44956219e-01 -2.26484895e-01
-8.25727224e-01 -8.13400507e-01 -9.65836227e-01 1.28649607e-01
-3.42072636e-01 4.17796135e-01 2.99001396e-01 2.80774117e-01
-1.02959454e+00 -1.27937868e-02 -8.68536651e-01 -6.58724368e-01
1.83124527e-01 5.49282171e-02 -1.46296144e-01 -7.07773045e-02
-1.23906755e+00 6.90213621e-01 6.86841428e-01 4.61491019e-01
-1.00973451e+00 -4.93010551e-01 -8.89552534e-01 -2.38308594e-01
6.96983457e-01 -7.95376122e-01 1.22627127e+00 -5.82069695e-01
-1.05202901e+00 9.76168871e-01 -3.56194675e-01 -4.05035794e-01
6.56288683e-01 -1.75928846e-01 -6.03932261e-01 2.25216895e-02
4.52146620e-01 7.63990879e-01 7.14512825e-01 -1.39781702e+00
-1.11587656e+00 -1.77227363e-01 -8.51362757e-03 4.08639312e-02
9.76664349e-02 1.59020901e-01 -1.39427078e+00 -7.02482998e-01
2.87132889e-01 -1.04414916e+00 -3.70379686e-02 5.04417181e-01
-5.23848772e-01 -3.05630058e-01 1.16559708e+00 -4.87357765e-01
1.32560170e+00 -2.36769056e+00 -1.89450920e-01 -4.47067730e-02
3.74504775e-01 4.05990750e-01 -2.27665812e-01 3.25857520e-01
7.51632527e-02 -2.25555077e-01 3.90917599e-01 -3.11851442e-01
8.70240033e-02 -4.93120961e-02 -4.54826027e-01 6.99293852e-01
-7.09272549e-02 1.18110740e+00 -1.17576456e+00 -9.39306855e-01
2.65425771e-01 3.06143433e-01 -1.22715078e-01 2.15475753e-01
-4.70663697e-01 4.36350852e-01 -7.24074006e-01 8.62093985e-01
6.42326355e-01 -4.85003084e-01 6.07580282e-02 -5.03059626e-01
-2.88675994e-01 -2.28966400e-01 -1.26001799e+00 1.76337564e+00
6.38671294e-02 4.96352464e-01 -1.17837060e-02 -3.73890460e-01
9.54375446e-01 2.45165005e-01 3.78043711e-01 -6.68550313e-01
5.48093729e-02 7.76995672e-04 -1.15020230e-01 -2.84888834e-01
3.85752708e-01 7.28759319e-02 -5.17560132e-02 1.00960903e-01
7.36452313e-03 4.30682153e-01 2.20119491e-01 4.85873699e-01
6.27011418e-01 5.85359514e-01 2.58997172e-01 -1.40036657e-01
6.35040402e-01 3.58149290e-01 8.85047734e-01 7.97457278e-01
-6.93293869e-01 4.09833699e-01 2.47588083e-02 -3.56848240e-01
-6.41197085e-01 -1.01430154e+00 2.78081983e-01 1.10715103e+00
1.04303730e+00 -7.17008889e-01 -5.10621488e-01 -8.41350377e-01
-1.69372708e-02 4.67860162e-01 -6.52742028e-01 -1.67005882e-01
-4.25783008e-01 -3.86972837e-02 4.43029255e-01 6.67348802e-01
7.03448772e-01 -8.76568854e-01 -5.39515078e-01 3.96493226e-02
-3.28065425e-01 -1.40510511e+00 -9.98262882e-01 -2.74024576e-01
-5.88235021e-01 -1.24684620e+00 -6.28177643e-01 -8.59643936e-01
6.60270989e-01 8.34190369e-01 9.01703119e-01 6.01206958e-01
7.64324144e-02 5.91325104e-01 -2.86299378e-01 -1.31334811e-01
-2.09087938e-01 -1.79612949e-01 -4.54627685e-02 1.06198393e-01
3.42328787e-01 9.78360027e-02 -5.72090447e-01 7.47011840e-01
-6.63003564e-01 4.21060979e-01 3.80352020e-01 5.23799956e-01
7.80374646e-01 -1.73687756e-01 1.22299865e-01 -4.80607212e-01
1.06113717e-01 -3.43540132e-01 -1.03866935e+00 6.83120728e-01
-3.92203957e-01 -7.52212480e-02 1.99849635e-01 -5.18978894e-01
-9.27179039e-01 3.61164898e-01 3.02747190e-01 -8.96963477e-01
-5.50841317e-02 2.78418332e-01 -3.68951112e-01 -2.52088964e-01
2.66075253e-01 4.62441951e-01 -1.72737822e-01 -3.44835192e-01
5.95145762e-01 1.73947275e-01 8.07683051e-01 -5.36847949e-01
1.36420608e+00 5.35286546e-01 -6.66900873e-02 -3.68455946e-01
-9.67562497e-01 -8.27750623e-01 -6.22265577e-01 -4.72576410e-01
9.02018368e-01 -1.10334468e+00 -7.39056110e-01 4.01333898e-01
-1.26233840e+00 -3.42701882e-01 -8.75179321e-02 4.20197368e-01
-4.86582220e-01 5.87689698e-01 -3.07375461e-01 -6.25173628e-01
-2.27920234e-01 -1.12374818e+00 1.35584426e+00 5.13673723e-01
1.65520877e-01 -9.60118353e-01 -7.46088699e-02 3.85918841e-03
1.61798686e-01 1.25418186e-01 3.17035019e-01 -4.31087017e-01
-1.09731817e+00 -1.61950052e-01 -5.32005489e-01 -2.65768409e-01
3.71444449e-02 -2.51405463e-02 -6.07907057e-01 -4.90168035e-01
-3.28896880e-01 -7.22834915e-02 7.14235425e-01 3.28879416e-01
6.45265102e-01 3.87191623e-02 -8.44146311e-01 7.66162813e-01
1.42212892e+00 3.62103403e-01 4.23721641e-01 3.86812627e-01
9.14250791e-01 3.26809734e-01 1.25241351e+00 1.27000406e-01
4.40280616e-01 1.23552179e+00 4.37738895e-01 -2.38654450e-01
-5.38738966e-01 -6.06895566e-01 4.97111678e-01 4.19500262e-01
1.98971525e-01 -2.72130519e-01 -9.82244134e-01 4.78049636e-01
-2.21947265e+00 -1.03080320e+00 -1.49355128e-01 2.03032923e+00
4.93354946e-01 5.39644212e-02 2.66648144e-01 -5.58839679e-01
1.18613422e+00 3.70756239e-01 -4.16703403e-01 6.85877323e-01
6.43173084e-02 -7.30120778e-01 4.78350639e-01 4.74905849e-01
-1.14411438e+00 1.39953780e+00 5.37259483e+00 1.02416062e+00
-1.25268316e+00 2.40569100e-01 -6.17263541e-02 1.50731534e-01
-6.14241464e-03 2.12011993e-01 -1.42882252e+00 6.58812344e-01
3.69928122e-01 -4.12968069e-01 5.82178086e-02 7.95497954e-01
2.73569971e-01 1.34399012e-01 -1.02605343e+00 1.01166654e+00
-1.42335117e-01 -1.26811051e+00 5.17587848e-02 -1.17244415e-01
3.71961623e-01 6.56498522e-02 -4.73469645e-02 2.79722661e-01
3.30282122e-01 -4.06616062e-01 1.26269591e+00 5.81956744e-01
6.32702351e-01 -1.82573348e-01 4.39522594e-01 3.44122618e-01
-2.08790731e+00 2.25153580e-01 -2.09849715e-01 5.34571767e-01
2.92947948e-01 -1.71592589e-02 -6.30020142e-01 8.02922726e-01
6.49770856e-01 8.88517976e-01 -7.62439430e-01 1.31470585e+00
-4.95887905e-01 3.50994110e-01 -8.26223940e-02 1.56298250e-01
3.51566911e-01 1.00687876e-01 6.76370859e-01 1.07170320e+00
1.93578511e-01 -6.60870746e-02 8.81186068e-01 8.54017258e-01
2.77084649e-01 -1.00970462e-01 -4.67824161e-01 9.56857204e-02
6.05691135e-01 1.15547752e+00 -9.69223559e-01 -3.63446504e-01
-5.94236016e-01 7.24948525e-01 2.00081810e-01 6.42048597e-01
-1.35664117e+00 -1.10331774e-01 4.09736097e-01 2.55624205e-01
5.34726560e-01 -3.55406374e-01 3.50593835e-01 -1.38503861e+00
3.62006091e-02 -7.48564839e-01 5.84227145e-01 -1.08831978e+00
-9.95488644e-01 7.32219577e-01 4.62875627e-02 -1.82113373e+00
-6.83828518e-02 -3.99599493e-01 -5.10526597e-01 8.64713967e-01
-1.53271139e+00 -1.53237140e+00 -7.64535308e-01 8.45815897e-01
4.56992239e-01 -9.59106442e-03 2.41503790e-01 3.88480246e-01
-4.95682836e-01 4.97064948e-01 -3.66084993e-01 4.61001337e-01
7.65462041e-01 -9.07521427e-01 4.52638537e-01 1.32196879e+00
3.28388274e-01 7.79110968e-01 5.89251816e-01 -9.16663587e-01
-1.30087614e+00 -1.45932138e+00 7.38167763e-01 -4.78490144e-01
8.26302171e-01 -3.22462589e-01 -8.35670352e-01 9.18399870e-01
1.18368603e-01 5.10125816e-01 1.77846894e-01 -3.87640856e-02
-4.45074320e-01 -6.20390847e-02 -6.10757113e-01 7.77754068e-01
1.32301819e+00 -5.17095745e-01 -5.26788414e-01 3.32541496e-01
9.05143082e-01 -7.48747945e-01 -4.00777429e-01 2.72333920e-01
4.25938040e-01 -6.42436981e-01 1.07121336e+00 -4.80180651e-01
-2.49160230e-01 -1.18112159e+00 -1.42890289e-01 -9.28702772e-01
-3.80060762e-01 -6.29529536e-01 -2.21508816e-01 1.55334198e+00
6.96581453e-02 -4.73297238e-01 7.91247189e-01 2.37586245e-01
-8.10598210e-02 -5.67940533e-01 -6.75127387e-01 -1.00420928e+00
-4.82951730e-01 -3.65752101e-01 6.24692023e-01 8.10781002e-01
-2.56169587e-01 2.13401169e-01 -5.48429906e-01 5.13429046e-01
8.73279810e-01 5.19047558e-01 1.03145194e+00 -1.11995828e+00
-1.11750066e-01 -2.06081152e-01 -5.81193507e-01 -1.50705659e+00
1.83607712e-01 -8.36054623e-01 3.47910970e-01 -1.53512907e+00
2.91415989e-01 -4.92177516e-01 -3.26227099e-01 6.79434121e-01
-3.22628796e-01 8.05801526e-02 6.29864097e-01 4.43228930e-01
-1.22898400e+00 5.13565838e-01 1.36196876e+00 -2.47631148e-01
-1.10847905e-01 1.84789263e-02 -5.74202657e-01 7.88003325e-01
4.83651698e-01 -6.84444070e-01 -4.45482075e-01 -4.77614701e-01
-2.77689546e-01 2.55795747e-01 7.54968226e-01 -1.08653665e+00
8.47998202e-01 -3.42262596e-01 2.58823246e-01 -1.02049327e+00
8.40670392e-02 -1.01813018e+00 2.56518304e-01 5.87538540e-01
-1.83414251e-01 4.69573326e-02 3.87780637e-01 8.46351385e-01
-2.16196820e-01 -5.54945087e-03 7.08560765e-01 8.86235610e-02
-1.48163760e+00 7.53334403e-01 1.01943187e-01 1.76172629e-01
1.28504109e+00 -2.35010415e-01 -6.59830868e-01 -2.56672055e-01
-7.62277782e-01 7.87503958e-01 6.11778677e-01 6.07297003e-01
5.47918499e-01 -1.69317937e+00 -4.74892527e-01 -5.94608160e-03
4.48000073e-01 -2.44689032e-01 2.33213902e-01 9.85946536e-01
-2.75878757e-01 5.30725718e-01 -2.58604698e-02 -9.99764264e-01
-1.50308394e+00 8.74280453e-01 5.46345472e-01 -2.06309453e-01
-7.06715941e-01 7.05400646e-01 7.63620317e-01 4.28096689e-02
4.77883250e-01 -2.69296616e-01 -7.19783306e-02 -1.64063245e-01
6.74106658e-01 -6.20564558e-02 -4.48661476e-01 -1.03830731e+00
-6.68653011e-01 9.57158148e-01 -8.65356624e-02 2.89370585e-02
5.60874045e-01 -4.92099077e-01 1.78110927e-01 1.49262011e-01
7.64909983e-01 4.16166633e-02 -1.53525686e+00 -6.97721481e-01
1.80658415e-01 -5.96753418e-01 -8.72135833e-02 -5.80194950e-01
-1.35646415e+00 7.79090285e-01 6.37374103e-01 5.33531830e-02
1.05339992e+00 2.12524176e-01 6.30433619e-01 1.30777448e-01
5.53293884e-01 -4.67165768e-01 1.27217069e-01 4.11609501e-01
6.72385216e-01 -1.06439745e+00 -1.78605244e-02 -7.00857103e-01
-4.72741455e-01 8.88571441e-01 1.00996053e+00 1.29708394e-01
4.81575876e-01 2.36094564e-01 4.18664068e-01 -3.27679932e-01
-5.86168289e-01 -7.39770234e-01 6.10755384e-01 6.90361559e-01
1.37384489e-01 -1.70599073e-01 -1.10786237e-01 5.32372355e-01
3.99387866e-01 1.54656053e-01 -1.57945737e-01 8.44473064e-01
-5.37762403e-01 -1.11459827e+00 -5.23301005e-01 -2.93303847e-01
-1.11246668e-01 -4.38596308e-02 -2.74256647e-01 1.10762453e+00
2.54197061e-01 8.51324797e-01 -1.67104691e-01 -4.88245308e-01
3.76456082e-01 -1.67053252e-01 2.70538330e-01 -6.38464749e-01
-2.83972204e-01 3.18452656e-01 -1.64143041e-01 -8.53888631e-01
-5.82777262e-01 -5.21198690e-01 -1.33912158e+00 -1.46634445e-01
-5.86738586e-01 3.02187204e-01 1.73197865e-01 9.36774254e-01
3.66164625e-01 5.31714320e-01 1.95905954e-01 -7.08223164e-01
-3.81144017e-01 -5.27525306e-01 -5.63789785e-01 4.15125310e-01
3.86246383e-01 -1.02365565e+00 -3.98414209e-02 2.15109810e-01] | [6.329048156738281, -2.091935634613037] |
f3ad7678-d966-424e-8f78-a533be1de737 | dense-teacher-dense-pseudo-labels-for-semi | 2207.02541 | null | https://arxiv.org/abs/2207.02541v2 | https://arxiv.org/pdf/2207.02541v2.pdf | Dense Teacher: Dense Pseudo-Labels for Semi-supervised Object Detection | To date, the most powerful semi-supervised object detectors (SS-OD) are based on pseudo-boxes, which need a sequence of post-processing with fine-tuned hyper-parameters. In this work, we propose replacing the sparse pseudo-boxes with the dense prediction as a united and straightforward form of pseudo-label. Compared to the pseudo-boxes, our Dense Pseudo-Label (DPL) does not involve any post-processing method, thus retaining richer information. We also introduce a region selection technique to highlight the key information while suppressing the noise carried by dense labels. We name our proposed SS-OD algorithm that leverages the DPL as Dense Teacher. On COCO and VOC, Dense Teacher shows superior performance under various settings compared with the pseudo-box-based methods. | ['Jian Sun', 'Haiyan Yu', 'Zeming Li', 'Weixin Mao', 'Songtao Liu', 'Zheng Ge', 'HongYu Zhou'] | 2022-07-06 | null | null | null | null | ['semi-supervised-object-detection'] | ['computer-vision'] | [ 1.76490508e-02 2.49303475e-01 -2.74636060e-01 -4.94920105e-01
-7.66081631e-01 -3.77254963e-01 6.03537917e-01 1.84463896e-02
-4.37050909e-01 5.01793325e-01 9.08320770e-02 2.80965455e-02
1.91492289e-01 -4.13781404e-01 -4.94181484e-01 -9.06371117e-01
3.12358528e-01 4.27134484e-01 1.01810026e+00 2.32920751e-01
7.16767013e-02 3.26742709e-01 -1.51644409e+00 2.89142549e-01
8.68959606e-01 1.02702725e+00 3.24986219e-01 1.83275044e-01
-1.57479629e-01 6.80961967e-01 -4.70792681e-01 -2.25100428e-01
5.07433116e-01 -1.27308771e-01 -3.41348290e-01 2.29155570e-01
7.02335954e-01 -3.43285501e-01 -3.49131554e-01 1.08227408e+00
5.90099335e-01 2.97345132e-01 8.42005074e-01 -1.06140757e+00
-3.42831016e-01 6.05253994e-01 -9.93694603e-01 1.49365917e-01
-1.00871637e-01 -1.03820637e-01 1.01846051e+00 -1.44457924e+00
6.06962740e-01 1.36608911e+00 9.60783780e-01 5.13419867e-01
-1.37963820e+00 -9.34745729e-01 3.60327780e-01 2.24908918e-01
-1.62686706e+00 -1.91358119e-01 8.42973590e-01 -5.38491368e-01
6.37410998e-01 -1.12851210e-01 4.84725654e-01 9.33129907e-01
-1.67851910e-01 1.01125860e+00 1.46436083e+00 -6.25806987e-01
2.85600275e-01 4.88139123e-01 4.71277088e-01 9.16016757e-01
2.87335634e-01 2.02238619e-01 -4.73995954e-01 -1.44062072e-01
5.45761168e-01 3.38453829e-01 -1.65621012e-01 -7.40464687e-01
-8.50754619e-01 8.58900368e-01 7.77967334e-01 -3.32667530e-02
-6.98633343e-02 -9.87298414e-02 1.89058185e-01 -2.42779940e-01
7.67867923e-01 2.99236357e-01 -3.58671665e-01 4.28752929e-01
-1.36554825e+00 -1.59391537e-02 3.80516350e-01 1.04381466e+00
9.85517979e-01 -9.86585319e-02 -5.97343028e-01 9.93090153e-01
6.28198087e-01 1.74207717e-01 4.53226596e-01 -5.17694235e-01
3.26153822e-02 7.75746703e-01 1.10146329e-01 -5.73113084e-01
-5.73594809e-01 -8.92599523e-01 -6.16919935e-01 2.52498209e-01
3.98989767e-01 1.13303155e-01 -1.59772885e+00 1.41125071e+00
8.19173276e-01 4.38457519e-01 -3.54373723e-01 8.86091232e-01
1.04958940e+00 5.57616353e-01 2.53010392e-01 1.08549334e-01
1.33253741e+00 -1.73189092e+00 -5.76235950e-01 -1.91024616e-01
6.89168692e-01 -5.91598213e-01 1.02664471e+00 3.02582532e-01
-5.74527860e-01 -8.31526279e-01 -1.09432983e+00 -2.46716470e-01
-4.52042848e-01 3.28173071e-01 5.11173725e-01 6.68998301e-01
-8.98270667e-01 3.23619664e-01 -6.83250010e-01 -3.23861182e-01
7.74947882e-01 3.73588800e-01 -2.54703701e-01 -2.44170129e-01
-7.42205381e-01 8.65504205e-01 5.06405354e-01 -2.37865284e-01
-1.04822695e+00 -7.21740246e-01 -8.11402798e-01 -8.79012980e-03
6.98306918e-01 -2.75941879e-01 1.21713042e+00 -7.93921471e-01
-1.42755246e+00 1.03376746e+00 -2.14314476e-01 -5.63298523e-01
3.99527788e-01 -3.82332742e-01 -5.68385385e-02 2.60006249e-01
2.11563021e-01 1.31651902e+00 1.21219838e+00 -1.38759530e+00
-7.74403811e-01 -1.65566519e-01 -1.26829565e-01 1.44697681e-01
-3.06949973e-01 1.11256011e-01 -8.05617034e-01 -9.34219658e-01
4.36103821e-01 -1.15529644e+00 -3.11917931e-01 2.85506755e-01
-6.80303574e-01 -4.04032111e-01 1.11261570e+00 -1.35224789e-01
1.22966254e+00 -2.35692692e+00 -2.68947184e-01 1.36860058e-01
6.26323521e-01 5.81146538e-01 -8.84916037e-02 -1.78853720e-02
-5.89296557e-02 -1.42519414e-01 -2.77692795e-01 -7.45837212e-01
5.68453036e-02 1.24044456e-01 -3.18559825e-01 7.22775161e-01
3.37372839e-01 7.36986935e-01 -8.58218551e-01 -9.22633767e-01
3.33314329e-01 4.59158987e-01 -5.07322252e-01 1.63012564e-01
-3.24589282e-01 2.42592752e-01 -4.68281537e-01 7.87088513e-01
9.31642830e-01 -4.27005410e-01 -3.13633680e-01 -9.28783789e-02
-1.47094935e-01 2.44223788e-01 -1.10885870e+00 1.58331048e+00
5.49778417e-02 5.17268419e-01 1.74162596e-01 -5.59076905e-01
1.07661879e+00 9.75201279e-02 4.43308383e-01 -3.64829153e-01
1.80449471e-01 1.68937072e-01 -2.28903845e-01 5.17760552e-02
4.72115070e-01 -9.20801908e-02 2.64381170e-01 3.52546722e-01
3.12069952e-01 4.51325886e-02 6.62155673e-02 3.72506082e-01
1.07840610e+00 3.58817905e-01 2.80894458e-01 -3.96984130e-01
3.53089720e-01 2.55281404e-02 8.21296751e-01 9.31074321e-01
-6.43988848e-01 8.86747181e-01 1.86806366e-01 -1.31641015e-01
-8.74255896e-01 -9.91697013e-01 -6.06064856e-01 1.21034956e+00
2.44877353e-01 -4.62029159e-01 -7.21242309e-01 -1.12224030e+00
2.51228333e-01 3.77555788e-01 -6.39079690e-01 4.94829081e-02
-2.65748590e-01 -6.95311904e-01 4.31569815e-01 5.86467266e-01
5.49719930e-01 -9.25217927e-01 -3.22166175e-01 1.51911736e-01
4.10978436e-01 -1.17718732e+00 -6.74362540e-01 8.21240544e-01
-8.28591466e-01 -8.94608021e-01 -9.48483944e-01 -9.12498355e-01
8.80432308e-01 7.09545255e-01 8.29737306e-01 -1.63231388e-01
-2.32897997e-01 -1.27795160e-01 -5.82476854e-01 -3.63164783e-01
3.42790149e-02 1.52344123e-01 9.04732794e-02 1.18593551e-01
4.12318587e-01 -2.72196174e-01 -6.58694923e-01 4.93624657e-01
-5.47982991e-01 2.83017755e-01 8.11491966e-01 9.80206013e-01
1.05318773e+00 -4.50657345e-02 4.48905081e-01 -1.35476613e+00
-3.64920162e-02 -3.23379129e-01 -6.85762465e-01 6.72202632e-02
-7.57284582e-01 1.18935280e-01 3.37015510e-01 -6.62547052e-01
-1.19517684e+00 5.84047377e-01 -2.39019170e-02 -7.13204682e-01
-2.29844347e-01 -5.51896766e-02 -5.79663515e-02 -3.36971164e-01
6.90439880e-01 -2.44091868e-01 -4.96741831e-01 -9.11712348e-01
5.49094141e-01 7.13513494e-01 4.90921497e-01 -3.22291464e-01
9.51233685e-01 7.90813923e-01 -1.65797338e-01 -4.54614669e-01
-1.59677732e+00 -1.07729661e+00 -9.03812587e-01 -4.28143144e-02
8.12424004e-01 -1.11188138e+00 -1.13231242e-01 4.62313175e-01
-8.35677624e-01 -3.27137172e-01 -3.97679597e-01 4.79732245e-01
-1.98034525e-01 3.03908557e-01 -5.90693057e-01 -8.05776834e-01
-3.16651016e-01 -1.02489531e+00 1.30178237e+00 4.10363436e-01
2.31371433e-01 -5.53692937e-01 1.09733097e-01 3.15945476e-01
2.32810736e-01 -3.36121209e-02 3.60850096e-01 -8.16497803e-01
-5.91599584e-01 -2.36788020e-01 -9.16077614e-01 1.81417838e-01
-1.29773065e-01 -2.03591481e-01 -1.42196405e+00 -3.55575502e-01
-2.54721254e-01 -5.70605338e-01 1.44956708e+00 4.59875911e-01
1.13903940e+00 1.87383965e-01 -6.41997516e-01 9.00071800e-01
1.28324366e+00 -7.00730979e-02 2.74011701e-01 3.22958231e-01
9.44062889e-01 4.45995361e-01 8.67901444e-01 3.62996012e-01
1.59179166e-01 8.41566086e-01 2.96167254e-01 -5.78368306e-01
-4.82130170e-01 -4.95033354e-01 2.65593410e-01 5.49289942e-01
4.28942859e-01 2.26811562e-02 -7.78503418e-01 4.19162005e-01
-1.87014580e+00 -6.01302445e-01 -2.32283458e-01 1.90774691e+00
9.07027245e-01 4.12572742e-01 2.05228895e-01 2.91085094e-02
8.34935427e-01 2.70071536e-01 -7.79156089e-01 1.94589972e-01
-1.06480874e-01 3.72031063e-01 7.31827259e-01 2.62299716e-01
-1.71415436e+00 1.29960561e+00 6.24411917e+00 1.17911601e+00
-8.27319503e-01 5.42294502e-01 4.28173363e-01 -1.39029562e-01
2.34606564e-01 8.72437805e-02 -1.52488315e+00 3.04633617e-01
5.85833907e-01 4.74818438e-01 -2.70778835e-01 1.33349812e+00
7.35859275e-02 -4.18338001e-01 -1.02568781e+00 8.12802315e-01
1.37336895e-01 -9.32475805e-01 -3.46963197e-01 -1.49264961e-01
1.07354712e+00 3.05940002e-01 3.84926796e-02 5.49021840e-01
5.02689958e-01 -7.19090164e-01 1.01156390e+00 1.50026292e-01
7.71429181e-01 -2.56927431e-01 5.60087383e-01 2.58108407e-01
-1.20087111e+00 -1.10945210e-01 -5.98757565e-01 3.03608000e-01
-8.14884752e-02 8.37521970e-01 -9.04728472e-01 4.49532978e-02
8.26742232e-01 8.25538695e-01 -8.55575681e-01 1.46814346e+00
-5.69908202e-01 9.61587548e-01 -5.09680986e-01 1.42287046e-01
3.82253230e-01 5.69973290e-02 4.15586472e-01 1.59395754e+00
-7.43012503e-02 -1.59952939e-01 5.63476503e-01 8.52543235e-01
-8.27778280e-02 -1.06219323e-02 -2.02708185e-01 3.08815628e-01
4.80900407e-01 1.45755863e+00 -1.13715827e+00 -5.05416214e-01
-4.89967436e-01 7.09681511e-01 5.13890505e-01 2.56166935e-01
-7.04216838e-01 -2.79068172e-01 3.67350131e-01 2.39042148e-01
6.23219252e-01 -1.08853772e-01 -3.20921987e-01 -1.03565168e+00
-3.26553345e-01 -5.93285620e-01 4.22983736e-01 -7.75860786e-01
-1.21503317e+00 5.05380690e-01 9.34670940e-02 -1.21556818e+00
3.26621771e-01 -6.54784381e-01 -3.39829206e-01 6.30028725e-01
-1.86512041e+00 -1.30144620e+00 -3.54958624e-01 5.02089143e-01
6.02221310e-01 1.01834729e-01 4.57762241e-01 3.58845979e-01
-8.94429266e-01 6.70838833e-01 3.20406497e-01 7.99117684e-02
1.00576162e+00 -1.40390503e+00 1.26329750e-01 7.91190028e-01
3.63438576e-01 3.76934916e-01 4.46542531e-01 -6.27462924e-01
-5.51476240e-01 -1.38233960e+00 5.52601576e-01 -1.56256482e-01
5.22933006e-01 -7.71535933e-01 -8.87332022e-01 4.74057823e-01
5.61621226e-02 4.58777487e-01 5.02025425e-01 7.63783753e-02
-6.81983352e-01 -1.05477065e-01 -1.06073701e+00 3.41831982e-01
1.03338873e+00 -5.20695210e-01 -5.76276302e-01 4.70695496e-01
8.83158028e-01 -2.93430150e-01 -3.76192719e-01 5.59659898e-01
2.73852557e-01 -9.02599156e-01 8.99400115e-01 -2.17309564e-01
1.54253505e-02 -6.94768667e-01 6.57596514e-02 -1.01265168e+00
-5.68375111e-01 -4.05928463e-01 -2.54172772e-01 1.28845334e+00
3.87490988e-01 -4.21354055e-01 1.21586406e+00 3.41428936e-01
-4.37520176e-01 -8.69416773e-01 -6.32703543e-01 -7.30699122e-01
-4.66288269e-01 -3.05689991e-01 1.87988356e-01 7.93065488e-01
-3.64421040e-01 2.26025701e-01 -4.20892715e-01 1.13531955e-01
7.62331784e-01 1.71507388e-01 6.88999057e-01 -1.33541679e+00
-3.48342180e-01 -1.76673546e-01 -2.74988592e-01 -1.37742603e+00
1.10484950e-01 -8.70639563e-01 3.48406762e-01 -1.32776737e+00
5.12584209e-01 -9.05628085e-01 -5.93565762e-01 8.03008556e-01
-3.60242516e-01 7.56909192e-01 9.89777222e-02 5.01312077e-01
-9.69048619e-01 7.68908083e-01 1.15334690e+00 -7.28050619e-02
-3.97992820e-01 -1.29685268e-01 -3.91350538e-01 9.72858310e-01
4.76509094e-01 -1.03378177e+00 -2.20185623e-01 1.30678471e-02
-3.48910123e-01 -6.35987461e-01 6.26471043e-02 -1.08271122e+00
2.87450790e-01 1.34888053e-01 4.37469333e-01 -1.09201360e+00
5.89324534e-01 -7.14197755e-01 -3.53550136e-01 2.45048612e-01
-2.51785696e-01 -4.14126366e-01 1.79627407e-02 7.87408412e-01
-2.32692882e-01 -4.58089948e-01 1.31526732e+00 8.75841826e-02
-7.96811998e-01 2.69575685e-01 -1.36596486e-01 -2.10717078e-02
1.33562267e+00 -4.41575885e-01 -2.99985051e-01 5.48333339e-02
-7.35114753e-01 2.17612788e-01 4.34756070e-01 7.91540891e-02
4.78963912e-01 -1.04848158e+00 -3.30913752e-01 2.67938584e-01
1.78080633e-01 4.01492059e-01 4.87872958e-02 7.51053214e-01
-3.39820445e-01 5.56042314e-01 5.02278171e-02 -7.10146487e-01
-1.32281494e+00 7.59935558e-01 1.69802710e-01 -4.84400630e-01
-9.51760173e-01 1.28000581e+00 6.35535121e-01 -3.16886514e-01
7.00299740e-01 -2.54656971e-01 -2.60202259e-01 2.13038281e-01
3.69113505e-01 7.84253255e-02 -3.77708562e-02 -7.54500508e-01
-3.83297175e-01 5.02468407e-01 -4.22599316e-01 -4.94998740e-03
1.21851361e+00 -1.99742336e-03 2.53582209e-01 3.34336698e-01
1.00461102e+00 -3.82921696e-02 -1.68928862e+00 -6.42412484e-01
2.19256625e-01 -4.38271284e-01 5.67406595e-01 -6.07362390e-01
-1.09373462e+00 8.68557692e-01 7.57897973e-01 -2.75955498e-01
1.13601816e+00 1.80021018e-01 5.55080354e-01 1.71737149e-01
3.60478908e-01 -1.15651917e+00 2.51583040e-01 4.11544234e-01
4.59313899e-01 -1.30109847e+00 2.74840474e-01 -8.16105604e-01
-6.14767194e-01 6.65865004e-01 9.79439497e-01 -2.19726369e-01
1.04125154e+00 2.98192173e-01 1.51837856e-01 -6.39677495e-02
-7.58504570e-01 -7.82317460e-01 5.17043054e-01 5.59232891e-01
9.86678228e-02 -1.04893394e-01 -2.03292891e-01 4.74544555e-01
2.08700046e-01 -2.25972667e-01 3.74861397e-02 8.28226209e-01
-8.06878030e-01 -1.06161630e+00 -6.04597270e-01 6.47958875e-01
-3.23688239e-01 -1.48905918e-01 -4.22577620e-01 7.12274194e-01
4.51857477e-01 8.27037334e-01 7.18098879e-02 -4.04613465e-01
4.93233725e-02 1.59395635e-01 1.49358273e-01 -1.09127843e+00
-7.54440784e-01 4.48164672e-01 -1.94877103e-01 -4.56022561e-01
-5.57434916e-01 -5.45508444e-01 -1.31415939e+00 1.27184421e-01
-1.05488098e+00 -5.80472164e-02 4.96173501e-01 8.03930223e-01
3.09734672e-01 3.46113324e-01 4.94545043e-01 -1.04259086e+00
-4.37881052e-01 -1.22101450e+00 -7.96577334e-01 3.38633090e-01
2.36635149e-01 -1.25239635e+00 -4.21670467e-01 -3.09142590e-01] | [9.171998023986816, 1.2698352336883545] |
1489d1fd-b3d9-404e-8078-3e2e440e46c6 | deep-attractor-network-for-single-microphone | 1611.08930 | null | http://arxiv.org/abs/1611.08930v2 | http://arxiv.org/pdf/1611.08930v2.pdf | Deep attractor network for single-microphone speaker separation | Despite the overwhelming success of deep learning in various speech
processing tasks, the problem of separating simultaneous speakers in a mixture
remains challenging. Two major difficulties in such systems are the arbitrary
source permutation and unknown number of sources in the mixture. We propose a
novel deep learning framework for single channel speech separation by creating
attractor points in high dimensional embedding space of the acoustic signals
which pull together the time-frequency bins corresponding to each source.
Attractor points in this study are created by finding the centroids of the
sources in the embedding space, which are subsequently used to determine the
similarity of each bin in the mixture to each source. The network is then
trained to minimize the reconstruction error of each source by optimizing the
embeddings. The proposed model is different from prior works in that it
implements an end-to-end training, and it does not depend on the number of
sources in the mixture. Two strategies are explored in the test time, K-means
and fixed attractor points, where the latter requires no post-processing and
can be implemented in real-time. We evaluated our system on Wall Street Journal
dataset and show 5.49\% improvement over the previous state-of-the-art methods. | ['Zhuo Chen', 'Nima Mesgarani', 'Yi Luo'] | 2016-11-27 | null | null | null | null | ['speaker-separation'] | ['speech'] | [-1.73231423e-01 -3.00412983e-01 2.29789868e-01 -6.09435551e-02
-9.08535361e-01 -5.34304917e-01 4.08155978e-01 6.02580085e-02
-4.55666155e-01 2.18029499e-01 3.26154262e-01 -2.27395490e-01
-5.88230550e-01 -3.24012756e-01 -5.00440598e-01 -1.10741770e+00
-2.45523825e-01 3.96404505e-01 1.02200463e-01 -4.48595453e-03
5.52681908e-02 4.70187336e-01 -1.43765473e+00 -3.25425677e-02
6.13622069e-01 1.03900778e+00 2.42736518e-01 9.53541517e-01
-2.71010399e-01 2.46689841e-01 -1.02426004e+00 -9.57086384e-02
3.25799644e-01 -5.67594469e-01 -1.32117733e-01 7.73499086e-02
2.02634871e-01 -1.92454271e-02 -5.45785069e-01 1.10996270e+00
1.08413720e+00 2.41228566e-01 7.86750674e-01 -1.20152009e+00
-3.26533526e-01 8.23905885e-01 -4.96404439e-01 5.39719284e-01
-1.27002791e-01 -2.84264058e-01 7.14330733e-01 -9.71340835e-01
-1.97704121e-01 1.17773592e+00 5.94582200e-01 4.09026563e-01
-1.00585353e+00 -9.68476057e-01 -7.28083029e-02 1.88070729e-01
-1.43420053e+00 -1.13679397e+00 1.09521616e+00 -4.25156087e-01
7.92973638e-01 2.01885238e-01 4.02929157e-01 8.05735409e-01
-1.12070516e-01 6.55753255e-01 6.91727996e-01 -5.13438523e-01
4.30898070e-01 2.23340780e-01 1.20834991e-01 3.88054907e-01
-5.63913137e-02 -1.73655018e-01 -6.83656573e-01 -1.95862100e-01
6.59996629e-01 -2.54445318e-02 -3.15850049e-01 -2.69900531e-01
-1.17880189e+00 7.42552698e-01 3.01310718e-01 4.65719581e-01
-2.21419424e-01 -4.15349714e-02 1.49454400e-01 2.95419507e-02
3.15876275e-01 1.50262251e-01 -2.59763271e-01 -6.16234690e-02
-1.20130575e+00 1.22787155e-01 8.59074175e-01 6.15395486e-01
3.35761160e-01 2.56536245e-01 2.30019763e-01 1.18803871e+00
7.32165873e-01 6.12364113e-01 9.50998127e-01 -7.14400172e-01
7.10412920e-01 1.77704319e-01 -1.33550629e-01 -1.03941405e+00
-2.81851888e-01 -7.39486814e-01 -9.70489264e-01 5.24950102e-02
5.41708350e-01 -4.59652424e-01 -9.23955560e-01 1.71679461e+00
3.40865791e-01 8.17165613e-01 3.43944967e-01 8.47610593e-01
7.33680606e-01 9.17769790e-01 -4.31082994e-01 -4.01611716e-01
1.16556585e+00 -1.05556095e+00 -9.71239865e-01 -2.51472741e-01
1.32702356e-02 -1.03465104e+00 5.41861773e-01 4.70052242e-01
-1.05479646e+00 -6.69318199e-01 -1.23906577e+00 3.91518205e-01
-3.72326821e-01 3.65712970e-01 1.99390694e-01 9.12085891e-01
-9.29631054e-01 5.09654820e-01 -7.62156188e-01 9.06855091e-02
2.06341073e-01 5.13998926e-01 -1.58293650e-01 4.85256195e-01
-1.18834066e+00 3.89517397e-01 6.41304329e-02 4.34579939e-01
-7.99638629e-01 -5.18940985e-01 -7.91452169e-01 2.27304056e-01
1.56068921e-01 -2.73763001e-01 1.13023829e+00 -8.66770148e-01
-1.78818643e+00 2.32673392e-01 -4.06252086e-01 -4.09306109e-01
2.24710077e-01 -2.02152297e-01 -8.09558928e-01 5.89997470e-02
8.40736032e-02 3.21732968e-01 1.27380717e+00 -1.05898547e+00
-7.03062594e-01 -3.37984979e-01 -4.98965114e-01 3.10970843e-01
-4.32320744e-01 -1.04437396e-02 -6.47893369e-01 -6.84337080e-01
5.26655495e-01 -9.30175126e-01 -1.61322042e-01 -3.44010383e-01
-4.63365316e-01 -3.47929984e-01 8.38105261e-01 -6.72071040e-01
1.15090919e+00 -2.73585153e+00 2.07603097e-01 9.92083475e-02
1.91661254e-01 4.38878924e-01 1.45036718e-02 2.25210994e-01
-1.76384777e-01 -1.34822115e-01 -1.05045155e-01 -8.10978591e-01
2.80740857e-01 -2.67353088e-01 -5.53507149e-01 6.80495977e-01
-7.86956176e-02 2.32091635e-01 -7.29179084e-01 -2.97260374e-01
4.06540215e-01 7.29631424e-01 -2.99114734e-01 4.24076378e-01
2.95874268e-01 2.39560783e-01 2.06690654e-01 2.02960610e-01
8.87424409e-01 1.89314798e-01 2.97291595e-02 -1.50472075e-01
-3.90869677e-02 5.52186370e-01 -1.89582205e+00 1.51415122e+00
-3.90097797e-01 8.80443275e-01 3.98025781e-01 -1.00312245e+00
8.52029502e-01 6.03986979e-01 3.75937968e-01 -3.32985699e-01
1.11070700e-01 4.75840539e-01 3.62826169e-01 -2.65598118e-01
3.27861488e-01 -1.18521988e-01 -6.06416352e-02 3.58646274e-01
3.03276896e-01 6.28758892e-02 1.38820291e-01 -3.13122459e-02
6.76351011e-01 -5.50484002e-01 -4.03476767e-02 -1.62618682e-01
5.46219110e-01 -7.86389768e-01 4.67648506e-01 5.58434010e-01
-2.85388887e-01 6.45342767e-01 2.40070179e-01 9.71292481e-02
-7.26576984e-01 -1.41859317e+00 -1.63267091e-01 7.68776536e-01
2.28746697e-01 -2.93052942e-01 -7.91611612e-01 -3.66372108e-01
-1.94080055e-01 4.57629025e-01 -1.44359022e-01 -9.40366462e-02
-6.10984802e-01 -8.07451665e-01 7.03929722e-01 2.95953721e-01
3.18363219e-01 -8.18206251e-01 -3.08042467e-01 3.38385820e-01
-2.48927012e-01 -1.08426416e+00 -6.11378789e-01 5.04961729e-01
-6.32777870e-01 -6.21196568e-01 -7.10752904e-01 -9.12116826e-01
5.44388473e-01 3.96903574e-01 4.67954755e-01 -5.50267041e-01
5.14275730e-02 1.23677135e-04 6.81243744e-03 -4.96113926e-01
-2.95517147e-01 -7.02575967e-02 4.91208702e-01 3.92515481e-01
5.13628125e-01 -6.66901767e-01 -5.57069540e-01 3.98307920e-01
-5.86465299e-01 -4.33898509e-01 4.62156981e-01 4.14498836e-01
4.10338461e-01 7.52597630e-01 8.11332881e-01 -1.54044718e-01
7.51650333e-01 -6.48337364e-01 -3.21237803e-01 -2.30944365e-01
-1.09906301e-01 1.33035198e-01 8.28249395e-01 -6.60879612e-01
-5.84664822e-01 2.42378097e-02 -3.13476890e-01 -5.71363568e-01
-2.58228958e-01 1.60680950e-01 -4.55733716e-01 3.56587917e-01
2.38352507e-01 2.79331446e-01 -1.84147269e-01 -7.87338912e-01
2.94578433e-01 1.11246526e+00 4.29263413e-01 8.31782669e-02
7.55703330e-01 2.68708020e-01 -3.87808502e-01 -1.20566487e+00
-4.38711464e-01 -6.28685176e-01 -6.01614177e-01 -2.38423236e-02
6.68016732e-01 -9.60667193e-01 -5.26828825e-01 6.68807209e-01
-1.06612420e+00 9.65266749e-02 -3.00724566e-01 8.98771524e-01
-1.74571708e-01 3.42281193e-01 -3.01570475e-01 -1.02050149e+00
-1.77672476e-01 -1.28768563e+00 9.30429399e-01 4.17654306e-01
-9.06392932e-02 -8.65172982e-01 2.47266337e-01 2.03627989e-01
2.85941154e-01 -2.95253009e-01 7.34044552e-01 -9.84742343e-01
-2.32261196e-01 -2.45089039e-01 2.80179501e-01 5.91070235e-01
4.03211325e-01 -1.20925404e-01 -1.26040208e+00 -3.13220739e-01
4.89526629e-01 3.34578961e-01 8.10545564e-01 6.40819371e-01
8.14292967e-01 -3.38218689e-01 -4.47126955e-01 5.82931638e-01
1.06913114e+00 5.66627979e-01 3.12993348e-01 -7.98462629e-02
6.54573321e-01 5.36539853e-01 -7.19570890e-02 2.33425751e-01
1.60093457e-01 7.38209963e-01 8.78929198e-02 -6.67902529e-02
-3.04777384e-01 1.99294891e-02 6.40045822e-01 1.45379114e+00
5.31162381e-01 -5.00854194e-01 -6.44424796e-01 8.55973423e-01
-1.67975223e+00 -9.89733756e-01 1.29964739e-01 2.40711904e+00
6.32535279e-01 2.43545502e-01 2.71588385e-01 7.29815483e-01
8.40073705e-01 3.04878503e-01 -4.81437415e-01 -3.43350112e-01
-2.24315412e-02 1.75616741e-01 2.41935700e-01 7.71220386e-01
-1.34357417e+00 4.71452624e-01 5.91306353e+00 1.04631054e+00
-1.22023618e+00 1.46826893e-01 3.01334232e-01 -4.23980653e-01
2.10750729e-01 -3.26995403e-01 -9.46806431e-01 7.49037564e-01
1.24034441e+00 1.00683495e-01 5.09390712e-01 4.69772756e-01
1.25601143e-01 8.65747631e-02 -1.13196778e+00 1.25412858e+00
3.42657000e-01 -8.72573614e-01 -3.53475869e-01 2.09043119e-02
2.90608436e-01 -3.75509746e-02 2.39940777e-01 2.26625562e-01
1.13962684e-03 -9.54868197e-01 7.99406767e-01 3.26773852e-01
2.62077659e-01 -9.87400293e-01 4.87198710e-01 4.48032230e-01
-1.25795770e+00 -2.00596884e-01 -3.90554667e-01 1.34370714e-01
1.99868277e-01 8.33483040e-01 -9.39348757e-01 2.98507690e-01
6.50585890e-01 2.68265218e-01 -3.10543597e-01 1.01715302e+00
-3.99662815e-02 7.41103351e-01 -5.90876877e-01 -9.66100767e-03
1.31013677e-01 -1.40623957e-01 8.41370285e-01 1.40860987e+00
5.62465847e-01 -3.57590616e-01 4.71953079e-02 5.70665240e-01
-7.59212598e-02 3.06272111e-03 -3.52977067e-01 -4.68201526e-02
7.53163993e-01 1.18134856e+00 -8.12898517e-01 -1.72478914e-01
-1.00329690e-01 7.94947565e-01 1.15990706e-01 4.15861428e-01
-8.58615398e-01 -8.81579101e-01 7.87769616e-01 8.75538588e-02
6.66106164e-01 -4.28718179e-01 -1.00309595e-01 -8.96708906e-01
1.09591916e-01 -7.66707361e-01 5.49782440e-02 -3.51687789e-01
-1.30423594e+00 8.57378542e-01 -1.98911801e-01 -1.28254414e+00
-4.08065021e-01 -5.23256660e-01 -5.26200175e-01 1.03522313e+00
-1.23533583e+00 -5.00293076e-01 1.00054011e-01 5.65142274e-01
6.70049965e-01 -6.06177986e-01 7.17988253e-01 6.24806523e-01
-7.30149925e-01 7.25207567e-01 6.12022758e-01 2.84733802e-01
5.56207120e-01 -1.22729015e+00 4.27729696e-01 1.12186694e+00
6.46617174e-01 6.34788334e-01 6.19066596e-01 -2.24852413e-01
-1.19405222e+00 -8.64336789e-01 8.79161716e-01 -2.58220285e-01
5.99494219e-01 -8.64601731e-01 -6.71893537e-01 2.32179239e-01
4.63060439e-01 -3.82028855e-02 1.01582873e+00 -1.14798091e-01
-1.88588530e-01 -3.65435362e-01 -7.68245041e-01 6.03210211e-01
6.31128013e-01 -4.41500366e-01 -6.08597815e-01 1.88943997e-01
7.06739426e-01 -2.55003095e-01 -2.74721593e-01 4.95037735e-02
3.23068559e-01 -9.74389791e-01 1.12654793e+00 -1.84014708e-01
4.36805822e-02 -6.14546001e-01 -2.40147114e-01 -1.63940048e+00
-4.01219219e-01 -7.66960323e-01 -2.89317429e-01 1.38656461e+00
4.20695782e-01 -5.89214444e-01 4.56833720e-01 7.41175413e-02
-1.40185133e-01 -6.14353538e-01 -1.33129656e+00 -7.37172306e-01
-1.01482280e-01 -4.38748330e-01 6.38957202e-01 5.66221774e-01
1.87859423e-02 5.96760869e-01 -3.29474837e-01 6.91765904e-01
6.59639299e-01 6.50947168e-02 6.55616462e-01 -1.06349373e+00
-5.33600628e-01 -6.57199502e-01 -4.29247677e-01 -1.27415276e+00
2.05906883e-01 -8.52370441e-01 2.92845339e-01 -1.43783462e+00
-3.79620135e-01 -4.92566824e-01 -5.30593514e-01 -5.23112640e-02
-1.52418703e-01 -2.41327323e-02 2.21055284e-01 2.12505966e-01
-4.12310719e-01 6.42002404e-01 5.55194438e-01 -2.33752951e-01
-4.71063942e-01 2.97744870e-01 -6.60712004e-01 8.42839599e-01
7.83833742e-01 -6.67154372e-01 -5.55342019e-01 -5.25390148e-01
-3.22022766e-01 -8.26462805e-02 1.18410647e-01 -1.43079638e+00
5.35229445e-01 3.77362728e-01 4.87527996e-01 -6.96879089e-01
6.93448007e-01 -1.05386829e+00 1.19970404e-01 1.95823222e-01
-2.23800048e-01 -2.05820560e-01 3.20361376e-01 6.26457989e-01
-2.41771117e-01 -5.47192752e-01 8.04809511e-01 2.32935533e-01
-9.26271603e-02 1.46297384e-02 -4.07867551e-01 -2.77495589e-02
9.32677150e-01 -2.33912133e-02 -3.07497266e-03 -4.11951095e-01
-5.87397635e-01 -2.08480954e-01 -3.23214024e-01 5.14712930e-01
5.21938026e-01 -1.43421960e+00 -6.28958464e-01 5.61446428e-01
-5.00411987e-01 1.50453672e-02 2.60781288e-01 6.84856355e-01
-1.97697371e-01 3.10071826e-01 1.80978417e-01 -6.51728451e-01
-1.15970337e+00 5.31893551e-01 5.92871130e-01 1.54581079e-02
-3.63727510e-01 1.05365217e+00 2.09031448e-01 -3.75300378e-01
6.65232420e-01 -2.90568113e-01 -4.15708929e-01 3.08778912e-01
6.32828414e-01 5.27913451e-01 1.07020341e-01 -8.25882733e-01
-5.64357698e-01 6.60431743e-01 -2.64535546e-02 -5.86751282e-01
1.29095840e+00 -1.22858435e-01 -1.00322165e-01 8.59408140e-01
1.42168915e+00 6.07844114e-01 -1.05495143e+00 -3.24300379e-01
-2.35208467e-01 -4.35378343e-01 1.58707783e-01 -5.52817762e-01
-1.14598918e+00 1.14342177e+00 9.97801960e-01 4.91630048e-01
1.05571544e+00 -8.48055184e-02 9.08380091e-01 1.21105276e-01
-1.96404815e-01 -1.01855564e+00 7.29464591e-02 4.22484845e-01
5.91084719e-01 -8.78848433e-01 -4.34358835e-01 -1.13932647e-01
-3.18864852e-01 1.03007281e+00 2.72454828e-01 -2.14046806e-01
1.07304347e+00 5.84785700e-01 3.29578549e-01 6.58982769e-02
-3.60270113e-01 7.95415938e-02 2.50792891e-01 4.24704552e-01
2.96360075e-01 8.69072378e-02 1.01901472e-01 9.46092486e-01
-4.75007504e-01 -5.72473884e-01 3.40155423e-01 5.66351235e-01
-4.54621166e-01 -9.45330441e-01 -7.06333935e-01 2.49935016e-01
-4.94660795e-01 -8.28879029e-02 -9.52200145e-02 5.88775277e-02
3.10650557e-01 1.42022991e+00 2.91622043e-01 -4.78882939e-01
4.34322715e-01 2.44486451e-01 1.39173657e-01 -4.47767615e-01
-3.50167155e-01 7.72166252e-01 -3.72544944e-01 5.60225472e-02
-2.75503010e-01 -7.56881952e-01 -1.21787369e+00 -3.29311118e-02
-5.72339237e-01 5.89620888e-01 9.05835927e-01 8.15748632e-01
5.23898959e-01 7.51085281e-01 8.40110600e-01 -9.84892368e-01
-5.34364164e-01 -1.09622145e+00 -7.06167936e-01 1.17582195e-01
8.90614510e-01 -4.77664471e-01 -8.51489127e-01 1.13747187e-01] | [14.970771789550781, 5.8420023918151855] |
13f0af40-878a-440f-9e1e-a0491af13fbd | tempo-lexical-context-driven-word-embedding | null | null | https://aclanthology.org/N18-1026 | https://aclanthology.org/N18-1026.pdf | Tempo-Lexical Context Driven Word Embedding for Cross-Session Search Task Extraction | Task extraction is the process of identifying search intents over a set of queries potentially spanning multiple search sessions. Most existing research on task extraction has focused on identifying tasks within a single session, where the notion of a session is defined by a fixed length time window. By contrast, in this work we seek to identify tasks that span across multiple sessions. To identify tasks, we conduct a global analysis of a query log in its entirety without restricting analysis to individual temporal windows. To capture inherent task semantics, we represent queries as vectors in an abstract space. We learn the embedding of query words in this space by leveraging the temporal and lexical contexts of queries. Embedded query vectors are then clustered into tasks. Experiments demonstrate that task extraction effectiveness is improved significantly with our proposed method of query vector embedding in comparison to existing approaches that make use of documents retrieved from a collection to estimate semantic similarities between queries. | ['Debasis Ganguly', 'Procheta Sen', 'Gareth Jones'] | 2018-06-01 | null | null | null | naacl-2018-6 | ['session-search'] | ['natural-language-processing'] | [ 6.69429421e-01 -3.05383921e-01 -5.81549585e-01 -3.26670527e-01
-1.03696394e+00 -9.31221187e-01 1.04171503e+00 4.41830188e-01
-7.57984400e-01 2.38967448e-01 4.39671785e-01 -7.23566189e-02
-4.25555259e-01 -4.29292709e-01 -1.95825025e-01 -1.79258406e-01
-3.00045639e-01 2.17585117e-01 4.62454140e-01 9.36286747e-02
6.24603808e-01 1.70593515e-01 -1.54024339e+00 3.30027819e-01
5.11783957e-01 9.56219375e-01 5.75354338e-01 6.56182706e-01
-4.70277607e-01 5.79625070e-01 -8.82725358e-01 5.46502732e-02
-6.63980246e-02 -1.44825846e-01 -1.22226930e+00 5.97330555e-02
1.28627270e-01 -1.08142376e-01 -5.88670790e-01 6.43494427e-01
2.02850193e-01 3.34897816e-01 6.68032944e-01 -1.48417449e+00
-3.84423733e-01 2.33154595e-01 -2.08637029e-01 6.14815593e-01
7.38203585e-01 -5.00921965e-01 1.62922776e+00 -1.02975154e+00
5.65873861e-01 8.65280330e-01 7.91329518e-02 1.70343012e-01
-1.59151447e+00 -6.60445094e-01 3.64134520e-01 9.29706395e-02
-1.32971847e+00 -3.74051571e-01 9.49507713e-01 -7.63998330e-01
1.40652156e+00 4.97204959e-01 2.56351054e-01 1.15247071e+00
1.29123420e-01 1.02325976e+00 5.36898911e-01 -4.64640886e-01
6.52590320e-02 1.25255197e-01 8.00957263e-01 2.97104299e-01
1.26304969e-01 -2.13077411e-01 -8.94751132e-01 -8.46844971e-01
3.44284087e-01 3.44600916e-01 -1.25658825e-01 -5.03503740e-01
-1.37662482e+00 1.01029658e+00 1.31317586e-01 5.70919991e-01
-5.78288376e-01 2.28974536e-01 7.13393211e-01 4.28969562e-01
6.76042020e-01 7.71126330e-01 -4.92877185e-01 -1.14956275e-01
-9.16819274e-01 5.59798121e-01 1.05323899e+00 1.14519548e+00
1.03748834e+00 -4.97255325e-01 -7.15204895e-01 8.88311505e-01
-3.39986407e-03 9.52232182e-02 6.83603644e-01 -7.38901198e-01
6.24010801e-01 8.25520635e-01 3.67787778e-01 -9.44243193e-01
-1.01720482e-01 -4.15219888e-02 1.66273993e-02 -5.90537608e-01
-6.09462112e-02 1.61881760e-01 -9.10308838e-01 1.67559040e+00
7.91202933e-02 -1.13867559e-01 -5.62419668e-02 5.17048717e-01
4.90654588e-01 4.40627575e-01 3.78006399e-01 -2.30186865e-01
1.79379821e+00 -8.39089990e-01 -9.31055903e-01 -4.72317547e-01
7.41724253e-01 -8.21754694e-01 1.13406026e+00 -8.18648636e-02
-6.07783258e-01 -4.13870662e-01 -8.96014929e-01 -2.25866452e-01
-7.26241529e-01 6.49761856e-02 7.00573206e-01 2.71421373e-01
-8.53193760e-01 -4.55202945e-02 -7.27053523e-01 -5.55242479e-01
-4.68346523e-03 4.24262226e-01 -1.55226171e-01 2.01058239e-01
-1.31242061e+00 4.00477946e-01 6.20561242e-01 -5.76101482e-01
-9.96801257e-01 -6.83296263e-01 -8.97062540e-01 9.34202224e-02
7.56922066e-01 -4.48134750e-01 1.69944251e+00 -3.88703406e-01
-5.64371824e-01 8.83813500e-01 -1.09465790e+00 -2.99977481e-01
-3.14712256e-01 -4.34818089e-01 -4.35769141e-01 1.17852822e-01
6.05761290e-01 4.06770527e-01 1.01534700e+00 -1.00946462e+00
-9.46921468e-01 -4.61023331e-01 3.14416707e-01 3.09183240e-01
-9.04327571e-01 4.73901361e-01 -9.29739237e-01 -8.02896380e-01
-3.33984867e-02 -9.63747919e-01 -8.12048372e-03 -5.59135139e-01
-6.08139217e-01 -7.51345754e-01 1.10909212e+00 -6.19289637e-01
1.91143429e+00 -2.13791203e+00 -7.50078559e-02 1.25047833e-01
5.41521370e-01 -3.20259094e-01 -2.82369349e-02 9.73695099e-01
-2.47126818e-02 3.80047530e-01 6.43071309e-02 -3.82588625e-01
1.47291720e-01 6.53341338e-02 -8.29676628e-01 1.56090677e-01
-1.38061792e-01 9.61253345e-01 -9.42436755e-01 -8.30728471e-01
-1.96469843e-01 9.89607051e-02 -2.77959406e-01 3.74356419e-01
-3.79313290e-01 -2.23568723e-01 -1.29287469e+00 5.48976064e-01
-8.26643221e-03 -5.03229856e-01 1.45394191e-01 -1.35800824e-01
2.14423649e-02 7.50595093e-01 -7.37429619e-01 1.98859620e+00
-7.89446533e-01 7.80120492e-01 -1.45170718e-01 -1.01453876e+00
4.69639957e-01 7.91566014e-01 1.06269503e+00 -6.52465284e-01
-5.18008888e-01 1.49576692e-02 -5.70693314e-01 -5.77768028e-01
8.06493104e-01 4.66194242e-01 -6.32743776e-01 1.04625249e+00
3.76346447e-02 3.51924717e-01 3.06364328e-01 4.49611992e-01
1.46005213e+00 -2.79957741e-01 3.96835148e-01 -1.12068258e-01
4.17610347e-01 4.98298258e-01 1.22671530e-01 9.73166406e-01
-1.38525456e-01 8.58965218e-02 2.69057900e-01 -2.13152617e-01
-1.00299919e+00 -9.53685343e-01 -4.05339599e-02 1.94611716e+00
1.33319959e-01 -9.68589604e-01 -2.82011867e-01 -1.03729641e+00
2.74551988e-01 3.33761722e-01 -7.34578431e-01 -5.02865575e-02
-5.45080662e-01 -3.71797949e-01 4.70968127e-01 3.94120336e-01
1.53043360e-01 -1.23522711e+00 -6.58451498e-01 3.07137847e-01
-4.24393296e-01 -1.22999001e+00 -1.17437863e+00 2.46719107e-01
-6.45149887e-01 -1.18258011e+00 -6.90411150e-01 -1.22013021e+00
3.99622768e-01 7.54481852e-01 1.30942047e+00 -2.47071624e-01
-5.86493373e-01 8.58782053e-01 -2.71940738e-01 -5.17400980e-01
1.45528972e-01 4.53142196e-01 3.92557830e-02 -5.47964312e-02
1.03196084e+00 -3.00023109e-01 -7.22741544e-01 3.45105112e-01
-1.22685683e+00 -3.23632896e-01 4.34739470e-01 6.67919219e-01
5.84834397e-01 8.15298408e-02 5.41056335e-01 -9.12396729e-01
1.22152781e+00 -8.57244790e-01 -2.79452294e-01 5.71353972e-01
-9.68058944e-01 2.78289586e-01 -1.48766395e-02 -7.92245209e-01
-1.07873261e+00 -2.75150873e-02 7.19664156e-01 -1.11459322e-01
-1.60103008e-01 5.85395217e-01 3.81086826e-01 3.45170081e-01
6.07561409e-01 6.09238923e-01 -3.32070082e-01 -5.65902352e-01
2.92620718e-01 9.78074789e-01 6.12282716e-02 -4.51941848e-01
7.62876034e-01 5.57225823e-01 -5.31686246e-01 -1.02044356e+00
-8.47706616e-01 -1.52334714e+00 -3.95414531e-01 1.58485714e-02
9.53464150e-01 -8.77838671e-01 -7.17880905e-01 -2.60463446e-01
-1.06649721e+00 1.07638407e-02 -2.40419712e-02 4.75627303e-01
-3.91666502e-01 3.12575012e-01 -3.52195323e-01 -8.84503603e-01
-2.70536512e-01 -8.82555664e-01 1.47881472e+00 -8.87030736e-02
-9.36407328e-01 -1.25018263e+00 3.60373020e-01 2.95237482e-01
3.12835693e-01 -1.04771458e-01 9.88990188e-01 -1.28772426e+00
-4.73456740e-01 -3.27794552e-01 -1.73647389e-01 -1.63670197e-01
6.99994981e-01 -7.82898009e-01 -9.73061979e-01 -5.10843515e-01
1.82457358e-01 -3.00511181e-01 8.52272570e-01 4.45725247e-02
1.23453677e+00 -4.11272764e-01 -9.65448022e-01 1.75661281e-01
1.39261448e+00 5.42263567e-01 -2.18604833e-01 5.09991109e-01
4.46343482e-01 8.61278117e-01 7.25253642e-01 4.45381582e-01
1.11148812e-01 8.10593188e-01 3.51552293e-02 2.08830178e-01
4.00442332e-01 -4.37985897e-01 2.71547139e-01 2.82678515e-01
5.54879725e-01 -2.26767987e-01 -9.62398946e-01 1.12044406e+00
-1.78179383e+00 -8.05836380e-01 4.67541873e-01 2.09410977e+00
9.08175409e-01 7.50359222e-02 9.53548029e-02 -5.17834201e-02
5.00156581e-01 6.18937731e-01 -5.55939615e-01 -5.15973791e-02
4.60380137e-01 4.70876805e-02 3.85432690e-01 3.61268103e-01
-1.18494666e+00 8.11499536e-01 6.83255863e+00 7.99266040e-01
-7.48812616e-01 1.12877740e-02 -1.24907769e-01 -1.73768178e-01
-5.64707577e-01 2.45191723e-01 -8.39844406e-01 3.05010378e-01
7.65120804e-01 -6.76191092e-01 2.13076696e-01 8.67965102e-01
5.66599108e-02 7.81837106e-03 -1.50825536e+00 9.21806335e-01
1.70594603e-01 -9.79786694e-01 1.74438506e-01 2.84263581e-01
4.75722164e-01 -2.15252772e-01 7.94484168e-02 5.10668159e-01
2.01981738e-01 -9.19330001e-01 2.06354648e-01 3.38413417e-01
6.01937175e-01 -3.79703522e-01 1.87164262e-01 5.15588284e-01
-1.72374570e+00 -1.89500973e-01 -1.16062080e-02 2.52679408e-01
-5.09222224e-02 2.55706251e-01 -1.35986078e+00 3.63840520e-01
6.26540542e-01 5.41322827e-01 -7.39995241e-01 7.95029521e-01
4.10613984e-01 6.29761815e-01 -1.47871319e-02 -1.23299778e-01
2.12059826e-01 -1.35668116e-02 6.81648850e-01 1.45242786e+00
1.24965943e-01 -2.89209574e-01 6.01687789e-01 7.07755566e-01
-5.65142259e-02 1.90633520e-01 -1.15861499e+00 -5.25391817e-01
8.53700578e-01 9.95314777e-01 -5.44552565e-01 -4.76206094e-01
-6.59080327e-01 1.24272931e+00 4.09729220e-03 8.59928370e-01
-4.59400147e-01 -8.38035583e-01 9.05885816e-01 2.33400948e-02
1.76483318e-01 -3.76378983e-01 5.72873373e-03 -1.01133657e+00
5.81770897e-01 -3.65943313e-01 7.20969200e-01 -3.87071222e-01
-1.15663111e+00 4.77152795e-01 5.01164556e-01 -1.09193552e+00
-7.45337069e-01 -4.88654599e-02 -5.10365963e-01 1.15320337e+00
-1.28344738e+00 -1.06824183e+00 -1.63589090e-01 7.78304100e-01
1.21491611e+00 -2.22722620e-01 9.80206549e-01 -6.90223351e-02
-7.82115757e-02 3.57997924e-01 -3.19684148e-02 2.40564093e-01
7.28343904e-01 -1.38023102e+00 4.65473443e-01 4.76803839e-01
6.30419374e-01 1.11615181e+00 4.80058640e-01 -8.15158665e-01
-1.41585696e+00 -9.46667075e-01 1.76256311e+00 -8.69464040e-01
8.75007629e-01 -7.28442729e-01 -9.34752405e-01 9.36359942e-01
2.72045046e-01 -4.00397569e-01 9.37122762e-01 3.78385574e-01
-5.19400477e-01 5.16023301e-02 -5.22134721e-01 5.04321635e-01
9.60411429e-01 -1.41749144e+00 -1.00408077e+00 6.75041318e-01
1.20828617e+00 2.30916589e-01 -8.71616542e-01 1.18525252e-01
4.96177167e-01 -2.21130237e-01 1.30128396e+00 -8.76339495e-01
-2.64147818e-01 6.83283582e-02 -2.11525649e-01 -8.21117461e-01
-9.97937769e-02 -7.01664925e-01 -4.31899250e-01 1.02520645e+00
5.10676265e-01 -4.30471450e-01 8.58434737e-01 5.98129332e-01
2.77958632e-01 -5.07570088e-01 -5.82329750e-01 -8.52760077e-01
-5.22300541e-01 -5.67229867e-01 4.37247217e-01 8.96817446e-01
1.28157109e-01 6.64436102e-01 -3.21497656e-02 8.62058923e-02
7.32054293e-01 4.85079348e-01 4.51230288e-01 -1.42223692e+00
-1.00620694e-01 -2.24437654e-01 5.43068200e-02 -1.25588834e+00
4.41509336e-01 -8.82913649e-01 5.41931577e-02 -1.50647783e+00
4.96662468e-01 -1.68672934e-01 -5.75673223e-01 2.60782927e-01
-2.74475634e-01 -1.22544371e-01 -2.42218703e-01 8.30961943e-01
-8.94204021e-01 2.29927421e-01 7.71734357e-01 -3.01609069e-01
-5.09500742e-01 2.14085475e-01 -6.57073319e-01 2.80028373e-01
7.05306470e-01 -7.72945940e-01 -1.00962317e+00 -5.09107113e-01
2.34607384e-01 4.20824997e-02 2.00213581e-01 -5.15371263e-01
4.86270428e-01 -3.24647397e-01 -9.41572152e-03 -5.76406598e-01
5.01826286e-01 -7.30952799e-01 -1.59237251e-01 7.40998164e-02
-9.44876969e-01 2.56742835e-01 -1.10145524e-01 1.05746913e+00
-7.41350830e-01 -7.31742457e-02 -1.85748547e-01 1.77409383e-03
-9.69284713e-01 3.76867503e-01 -5.19263625e-01 2.64365554e-01
1.05280709e+00 -1.68895677e-01 1.57328591e-01 -3.67140412e-01
-7.48386443e-01 4.51803744e-01 1.93386972e-01 8.71319413e-01
7.36462831e-01 -1.34945631e+00 -2.28247598e-01 4.57378812e-02
9.07419324e-01 -2.91003674e-01 -1.62160128e-01 3.71817559e-01
2.86076486e-01 1.29203105e+00 4.34792995e-01 -4.81246620e-01
-1.45182478e+00 7.11465359e-01 -2.87204474e-01 -6.61578894e-01
-3.47081929e-01 6.28124237e-01 5.74107826e-01 1.51976133e-02
7.03992486e-01 -1.36942595e-01 -6.34506226e-01 4.59613621e-01
3.76656950e-01 -9.42372717e-03 4.85805459e-02 -3.60991061e-01
-4.71903235e-01 2.78328419e-01 -3.62148076e-01 -6.18572176e-01
8.70749235e-01 -1.98507458e-01 -4.65034470e-02 7.29185462e-01
1.75266302e+00 -1.09257065e-01 -6.90146863e-01 -9.43206608e-01
9.10538256e-01 -3.62321168e-01 -1.72547419e-02 -3.68328482e-01
-4.28671062e-01 3.90163600e-01 3.81269366e-01 7.18777120e-01
1.09483659e+00 4.88013834e-01 8.55133712e-01 7.08262503e-01
1.46667659e-01 -1.15206993e+00 5.43073535e-01 5.00310004e-01
7.86552787e-01 -1.32285380e+00 -2.18434095e-01 -2.33377919e-01
-5.21288455e-01 8.07409942e-01 3.82451534e-01 1.65686950e-01
8.73867691e-01 -2.40796894e-01 -3.23701054e-01 -6.91585422e-01
-8.68697584e-01 -4.53067958e-01 7.29112148e-01 2.46987954e-01
6.20720029e-01 -1.90710068e-01 -3.25219572e-01 4.17496920e-01
7.46730939e-02 -2.17939392e-01 -3.87043729e-02 1.68329501e+00
-5.39011717e-01 -1.41831934e+00 -7.48203769e-02 7.77637720e-01
-8.00652921e-01 -2.35162050e-01 -6.35384858e-01 7.34320939e-01
-3.92991483e-01 1.11535716e+00 2.34743029e-01 -5.06910801e-01
3.68341148e-01 7.04762518e-01 -1.63959801e-01 -1.14608502e+00
-5.08397102e-01 1.02257386e-01 2.10866481e-01 -7.61128247e-01
-3.88510078e-01 -4.54755396e-01 -8.27842176e-01 4.60443318e-01
-1.36319175e-01 7.59770691e-01 7.86323011e-01 7.08714008e-01
5.02382040e-01 3.95238876e-01 7.56050885e-01 -2.71626592e-01
-6.93466425e-01 -1.10037684e+00 -5.09699583e-01 6.84658349e-01
6.91541672e-01 -6.63560212e-01 -3.15500140e-01 4.02376592e-01] | [11.567388534545898, 7.609540939331055] |
0df6b838-6f39-44d2-b1ce-a4bf554c59fd | the-whole-and-the-parts-the-mdl-principle-and | 2112.06853 | null | https://arxiv.org/abs/2112.06853v1 | https://arxiv.org/pdf/2112.06853v1.pdf | The whole and the parts: the MDL principle and the a-contrario framework | This work explores the connections between the Minimum Description Length (MDL) principle as developed by Rissanen, and the a-contrario framework for structure detection proposed by Desolneux, Moisan and Morel. The MDL principle focuses on the best interpretation for the whole data while the a-contrario approach concentrates on detecting parts of the data with anomalous statistics. Although framed in different theoretical formalisms, we show that both methodologies share many common concepts and tools in their machinery and yield very similar formulations in a number of interesting scenarios ranging from simple toy examples to practical applications such as polygonal approximation of curves and line segment detection in images. We also formulate the conditions under which both approaches are formally equivalent. | ['Gregory Randall', 'Ignacio Ramírez Paulino', 'Rafael Grompone von Gioi'] | 2021-12-13 | null | null | null | null | ['line-segment-detection'] | ['computer-vision'] | [ 2.13505790e-01 2.41083235e-01 -5.87864816e-02 -2.32822075e-02
-3.13062102e-01 -7.95609355e-01 6.20798826e-01 6.28091097e-01
-2.90678628e-02 4.71312374e-01 -3.36496741e-01 -3.62251431e-01
-5.93600094e-01 -8.79363000e-01 -4.03891623e-01 -7.65089571e-01
-1.20728657e-01 4.65728194e-01 6.75821304e-01 -3.45769912e-01
5.44360876e-01 9.77363825e-01 -1.51113462e+00 -7.11391866e-02
6.79885030e-01 7.08953023e-01 -1.01374649e-01 8.67136002e-01
7.32371118e-03 1.60301477e-01 -3.21227819e-01 -5.91012120e-01
4.05685335e-01 -1.06090933e-01 -8.93870056e-01 2.31306940e-01
4.25034404e-01 1.37786314e-01 -2.81966686e-01 1.30327284e+00
3.32178026e-01 -1.71250328e-01 1.11408925e+00 -1.24445105e+00
-1.81447089e-01 3.35769296e-01 -6.38557792e-01 2.83173889e-01
5.77415645e-01 -2.64042348e-01 7.89973855e-01 -7.81934917e-01
5.88357627e-01 1.13427126e+00 8.34544241e-01 1.66343853e-01
-1.30022025e+00 2.34703064e-01 -3.24122399e-01 1.09057993e-01
-1.34408545e+00 -2.93142259e-01 7.13939786e-01 -5.64275205e-01
5.63698053e-01 5.55636823e-01 3.68188500e-01 4.53506559e-01
2.32893541e-01 8.32680583e-01 9.53931093e-01 -7.00575113e-01
1.24446815e-02 2.24315122e-01 5.26202738e-01 7.38951743e-01
7.24247515e-01 1.53525174e-01 4.51559611e-02 -4.22154635e-01
9.15393651e-01 -2.24733755e-01 -1.07773028e-01 -8.70806217e-01
-9.17943835e-01 9.65719700e-01 -1.15468770e-01 5.10671318e-01
-1.16202429e-01 -4.30578589e-01 5.02456486e-01 3.21885675e-01
1.14579715e-01 3.73074651e-01 -1.40681088e-01 1.31646872e-01
-7.35910535e-01 5.45791328e-01 9.79612827e-01 1.16956067e+00
4.24772173e-01 -1.86427772e-01 2.54829317e-01 3.28699529e-01
2.50093669e-01 3.11434895e-01 -6.96244538e-02 -8.59843731e-01
4.78368014e-01 5.69625914e-01 1.69604614e-01 -1.08704877e+00
-4.96509165e-01 -3.48764628e-01 -6.26096249e-01 2.94409245e-01
6.34106338e-01 7.99970850e-02 -3.26737702e-01 1.33687043e+00
1.46191135e-01 -1.87586650e-01 1.19676977e-01 4.30800170e-01
3.88855398e-01 4.27870333e-01 -4.28192556e-01 -5.21095991e-01
1.03017461e+00 -4.42733854e-01 -4.30355161e-01 3.97158295e-01
7.60983109e-01 -9.71921861e-01 6.46035492e-01 5.77461898e-01
-1.16259086e+00 -3.86592120e-01 -1.01401389e+00 2.63919476e-02
-3.56796801e-01 -3.95005867e-02 4.39243853e-01 7.82743394e-01
-9.43890393e-01 8.40265155e-01 -5.55243433e-01 -5.88902056e-01
2.19895348e-01 1.90735489e-01 -2.69547552e-01 3.42322111e-01
-5.71838200e-01 6.52840555e-01 3.75858426e-01 2.97508538e-02
-3.55462551e-01 -1.13429338e-01 -5.91050565e-01 -2.25094799e-02
6.19947970e-01 -2.29188696e-01 1.12937891e+00 -4.36375380e-01
-1.05882382e+00 1.03998923e+00 -5.86758777e-02 -5.57170630e-01
1.01015067e+00 -8.75461474e-02 -3.49370927e-01 2.32641980e-01
-1.22130970e-02 -2.07415409e-03 4.78650868e-01 -1.15044212e+00
-5.10768473e-01 -3.94502759e-01 1.55246828e-03 -1.14097700e-01
1.59869999e-01 8.86377543e-02 -8.73461366e-02 -7.35212803e-01
4.07209098e-01 -8.16453159e-01 -1.58809736e-01 1.31266117e-01
-5.76597810e-01 -3.17262918e-01 1.11729515e+00 -2.98242480e-01
1.14928603e+00 -2.02563357e+00 1.86830118e-01 8.77389312e-01
2.67286539e-01 4.89210904e-01 1.21564411e-01 1.00197387e+00
-3.79347622e-01 1.40228227e-01 -4.69450027e-01 9.04487073e-02
-2.43093167e-02 4.25498247e-01 -4.55083102e-01 8.75387490e-01
1.39928712e-02 3.45657617e-01 -6.27950430e-01 -5.67524791e-01
3.45395416e-01 1.08567312e-01 -1.97616890e-01 -1.97598308e-01
-7.93473236e-03 2.80930817e-01 -5.20382941e-01 5.57119250e-01
7.97429383e-01 -1.01187145e-02 3.91098708e-02 -1.16944928e-02
-5.90858638e-01 9.02852416e-02 -1.47240138e+00 1.15533090e+00
2.03155160e-01 6.78655505e-01 3.93409990e-02 -1.45722365e+00
1.15503967e+00 3.95674258e-01 6.42790437e-01 -2.76189536e-01
1.00425482e-01 3.36937129e-01 -6.37864619e-02 -5.95279098e-01
3.08118343e-01 -8.87077898e-02 -1.15311578e-01 2.53702223e-01
-2.32389748e-01 -2.55938381e-01 6.44305766e-01 -1.36181433e-02
1.00573373e+00 1.79578718e-02 9.67663825e-01 -6.46245420e-01
8.51995707e-01 -8.94789398e-02 3.38153273e-01 8.09278488e-01
-3.46294969e-01 5.94470203e-01 9.69169021e-01 -3.10575485e-01
-1.38207483e+00 -1.23662126e+00 -7.42631674e-01 2.39658982e-01
3.12778383e-01 -5.37472963e-01 -6.93875492e-01 -6.29452229e-01
1.01633772e-01 5.25126755e-01 -4.25630689e-01 1.87408999e-01
-5.68302155e-01 -3.70397717e-01 5.80590069e-01 3.46413493e-01
3.17621648e-01 -6.73368573e-01 -7.65065789e-01 1.88156381e-01
2.24965721e-01 -1.08431017e+00 1.76991135e-01 -7.36759603e-02
-1.21346354e+00 -1.41481733e+00 -5.99074483e-01 -5.29367805e-01
5.39519250e-01 2.02035710e-01 1.22232759e+00 1.90537587e-01
-3.72007370e-01 5.17091155e-01 -3.59840810e-01 -5.56231260e-01
-8.01423371e-01 -2.40468130e-01 8.15628394e-02 -2.92060137e-01
3.64618093e-01 -4.78338540e-01 -1.27636999e-01 4.71609265e-01
-7.24972486e-01 -4.47463900e-01 4.03898686e-01 3.50884289e-01
5.63412428e-01 2.75413305e-01 2.34221265e-01 -7.35564351e-01
5.23182869e-01 -3.36265206e-01 -8.71042967e-01 1.86135888e-01
-7.40169406e-01 5.56059144e-02 7.17180192e-01 6.65574074e-02
-5.34548521e-01 -1.21294605e-02 -3.66709620e-01 -1.69361219e-01
-5.44313908e-01 8.10979456e-02 -3.36453974e-01 -1.50838941e-01
4.97045726e-01 2.01287612e-01 -1.92561656e-01 -6.45089388e-01
8.13607648e-02 4.42875236e-01 7.36694455e-01 -6.80170059e-01
9.11347270e-01 7.79028475e-01 5.87671936e-01 -1.21847641e+00
-4.22597051e-01 -6.44054651e-01 -1.14367378e+00 -1.62906811e-01
6.60601258e-01 -1.10535316e-01 -7.04856396e-01 3.38112026e-01
-1.32879817e+00 4.52397466e-01 -5.58371305e-01 3.63171935e-01
-1.08996749e+00 8.81507695e-01 -3.16586465e-01 -1.09454775e+00
2.07386091e-01 -1.05974340e+00 1.05875492e+00 8.01979080e-02
-7.18986318e-02 -1.17593539e+00 3.54424834e-01 -2.13465691e-01
-2.76887506e-01 6.69120967e-01 1.14173687e+00 -8.05073977e-01
-3.24267596e-01 -4.89824891e-01 -1.58526108e-01 1.65808544e-01
-8.67807716e-02 3.25130552e-01 -8.41666818e-01 -2.59439737e-01
1.24838769e-01 4.46263015e-01 6.86080754e-01 3.73629987e-01
8.31624687e-01 -3.13468397e-01 -4.68889713e-01 3.66016984e-01
1.68536580e+00 2.44910836e-01 7.50949502e-01 1.00086547e-01
3.27826798e-01 9.32430983e-01 6.52284384e-01 3.73606890e-01
-1.76566169e-01 7.70380855e-01 4.77597356e-01 4.37168591e-02
2.53546536e-01 2.73168720e-02 2.19059482e-01 6.16956472e-01
-4.06116903e-01 -2.89400011e-01 -8.73593390e-01 5.52396417e-01
-1.98743296e+00 -1.31176877e+00 -9.44038630e-01 2.42608309e+00
3.88328195e-01 5.06627977e-01 7.57189929e-01 6.91241443e-01
1.02052188e+00 1.11582831e-01 -8.76133442e-02 -7.26396084e-01
-4.08458620e-01 -1.12236865e-01 6.32914305e-01 5.11204839e-01
-1.02513278e+00 3.11960548e-01 7.21509600e+00 8.28844845e-01
-6.18182838e-01 -1.56608820e-01 1.82087094e-01 5.05252898e-01
-7.50521049e-02 2.71860421e-01 -6.95501029e-01 3.57718527e-01
7.40517676e-01 -2.90784448e-01 -2.77132630e-01 6.63307011e-01
7.65129924e-02 -2.22201079e-01 -1.15616131e+00 8.81975889e-01
2.53177676e-02 -1.24368906e+00 -1.22293189e-01 3.79236877e-01
5.34659863e-01 -2.56216437e-01 -1.20679930e-01 -3.27773571e-01
-5.84236197e-02 -6.88929021e-01 7.36519039e-01 7.97170818e-01
5.29581785e-01 -7.48646736e-01 7.48723149e-01 4.20067102e-01
-1.26788950e+00 9.66363400e-03 -4.39043730e-01 -1.36820212e-01
1.58227593e-01 7.17148900e-01 -5.91387570e-01 9.42552447e-01
1.42798707e-01 4.94828969e-01 -4.73658025e-01 1.35934770e+00
6.48955032e-02 4.52301711e-01 -4.48952079e-01 1.70043379e-01
1.73471779e-01 -6.65595949e-01 1.09094346e+00 1.23538113e+00
2.15479106e-01 -1.45432115e-01 -3.08985412e-02 8.04280579e-01
5.45095444e-01 2.24959552e-01 -1.06527865e+00 1.21052049e-01
1.02373853e-01 8.90020490e-01 -1.14660490e+00 -1.65634364e-01
-7.39920497e-01 5.18225968e-01 -1.24664687e-01 8.76286998e-02
-4.86018240e-01 -4.72101420e-01 3.74038279e-01 3.55683357e-01
2.51490295e-01 -3.85512471e-01 -3.35828125e-01 -9.32258010e-01
1.30409658e-01 -5.33535898e-01 4.89667416e-01 -1.81483403e-01
-1.21421504e+00 2.95271069e-01 4.17778075e-01 -1.45024288e+00
-3.43273461e-01 -1.01984966e+00 -9.44184184e-01 5.88900506e-01
-1.13364136e+00 -8.10219586e-01 1.14775054e-01 5.26848912e-01
4.71668214e-01 -1.18346438e-01 7.11775601e-01 -2.15672944e-02
-5.41380584e-01 1.64447427e-01 6.03552520e-01 1.25750273e-01
1.98398709e-01 -1.60928798e+00 4.04972941e-01 9.51821387e-01
1.73225626e-01 4.24833268e-01 1.25354230e+00 -5.49571753e-01
-1.02350616e+00 -5.42836070e-01 9.45352554e-01 -3.87305856e-01
5.72128117e-01 -1.52631029e-01 -1.10426128e+00 4.53865737e-01
-5.91234956e-03 -2.64139771e-01 3.58057439e-01 -1.25423282e-01
-1.90416858e-01 2.97591597e-01 -1.08356154e+00 3.91721934e-01
8.97684336e-01 -2.39302740e-01 -8.09110940e-01 5.07212102e-01
-1.80436298e-02 -6.27201721e-02 -1.01732278e+00 4.19573337e-01
4.85684246e-01 -1.59944832e+00 8.66932690e-01 -5.39676845e-01
-4.11284296e-03 -1.88635930e-01 -2.22724944e-01 -7.27449715e-01
-9.27963033e-02 -9.89839733e-01 -3.92246917e-02 1.11993492e+00
1.06596023e-01 -6.49599791e-01 5.16775250e-01 -5.21998554e-02
-1.81251183e-01 -7.76485980e-01 -1.03738546e+00 -1.13626587e+00
2.17170104e-01 -5.18951774e-01 3.97778213e-01 6.75553918e-01
3.69780068e-03 7.19574019e-02 9.08157080e-02 9.25906226e-02
7.80015528e-01 1.10256426e-01 8.20116401e-01 -1.85464668e+00
-2.47035637e-01 -7.20796108e-01 -9.18203235e-01 -9.59659994e-01
3.70875746e-02 -7.43378818e-01 -3.83080244e-01 -1.29931462e+00
-1.93462688e-02 -2.62733519e-01 2.33104788e-02 -1.56010628e-01
4.46684659e-01 -1.37975097e-01 3.69933583e-02 4.43688750e-01
-4.12382096e-01 1.17253542e-01 7.69347548e-01 3.60233337e-01
1.13175940e-02 5.86994767e-01 -3.87073338e-01 1.33439219e+00
5.25148153e-01 -3.03449422e-01 -2.55201995e-01 2.10854501e-01
4.97533470e-01 1.77643776e-01 5.25582016e-01 -9.59343731e-01
-3.59091349e-02 -1.28352061e-01 1.01839621e-02 -9.01764333e-01
-5.35748107e-03 -9.48115766e-01 -1.61862254e-01 4.92392451e-01
-1.34819433e-01 3.88224334e-01 -6.46773130e-02 5.98026097e-01
-1.10937893e-01 -9.41097677e-01 9.22204256e-01 -4.33711149e-02
-6.34150267e-01 1.37497380e-01 -1.82236597e-01 -1.28366590e-01
1.24807155e+00 -7.17162669e-01 -3.38674843e-01 -4.34578240e-01
-7.73001552e-01 -1.17574573e-01 6.74628735e-01 -5.70304431e-02
3.76067758e-01 -1.14547980e+00 -6.39560401e-01 3.83535773e-01
1.54433444e-01 -1.84169322e-01 3.74351032e-02 1.12783265e+00
-7.30691373e-01 7.23500073e-01 -2.13041604e-01 -9.17142510e-01
-1.26051939e+00 7.46125102e-01 3.61232281e-01 -2.08279893e-01
-8.52633417e-01 2.75193065e-01 2.40548223e-01 3.55464261e-04
-4.88184132e-02 -3.63058150e-01 -2.08474398e-01 -2.42446288e-02
3.73359531e-01 9.18699741e-01 9.66328233e-02 -9.57795978e-01
-2.63949484e-01 9.04420614e-01 4.05486748e-02 2.62268377e-03
1.09375417e+00 -3.12204361e-01 -2.10541356e-02 6.64764345e-01
1.00405610e+00 4.63542223e-01 -1.10855770e+00 -1.24174051e-01
4.28234547e-01 -3.65146935e-01 -3.19795370e-01 -1.41192287e-01
-4.96178985e-01 8.62308681e-01 4.95574743e-01 1.04293180e+00
1.00851417e+00 2.38845855e-01 4.77133095e-01 4.50370848e-01
2.64230698e-01 -1.11629915e+00 -3.50995362e-01 3.95564497e-01
9.00657833e-01 -8.13086033e-01 1.09024927e-01 -8.55395555e-01
-2.64585882e-01 1.73819780e+00 6.74149618e-02 -4.65133756e-01
7.00171828e-01 4.12051916e-01 -4.64087546e-01 -4.24091637e-01
-4.50862855e-01 -5.11083066e-01 3.05983573e-01 7.33392000e-01
3.43802303e-01 -1.78512931e-01 -8.53015304e-01 1.40987679e-01
-1.98428601e-01 -3.42214495e-01 8.13902795e-01 9.44556117e-01
-6.09182596e-01 -1.11338818e+00 -6.24098241e-01 3.07519525e-01
-3.73081148e-01 3.74948919e-01 -6.32804692e-01 1.37965035e+00
4.61699031e-02 7.32393742e-01 2.35040113e-01 -1.46644428e-01
5.96563458e-01 5.80685958e-03 7.03506768e-01 -4.13821876e-01
-2.02900946e-01 4.42055941e-01 3.95530500e-02 -1.96665257e-01
-7.25844324e-01 -1.07131195e+00 -1.06410003e+00 -3.73183846e-01
-5.11184156e-01 2.40357742e-01 4.42560941e-01 1.16849494e+00
-3.73166353e-02 3.75365876e-02 6.15096033e-01 -4.59944755e-01
-7.08650529e-01 -6.29162312e-01 -1.05650783e+00 2.37840116e-01
3.78849417e-01 -7.22162306e-01 -3.31475019e-01 -5.34781516e-02] | [7.371376037597656, 4.13771390914917] |
7508b96f-052d-4687-a68d-3292d6622971 | t-yolo-tiny-vehicle-detection-based-on-yolo | null | null | https://ieeexplore.ieee.org/document/9658533 | https://www.researchgate.net/publication/357258399_T-YOLO_Tiny_vehicle_detection_based_on_YOLO_and_multi-scale_convolutional_neural_networks/fulltext/61c3c877c48a3d26b74a7594/T-YOLO-Tiny-vehicle-detection-based-on-YOLO-and-multi-scale-convolutional-neural-networks.pdf?origin=publicationDetail&_sg%5B0%5D=cnnzAJDqU61aGdwfG6pFIoDItYCW7wHZ-sqVDeZx8G0-fZQmI1TSWEH31WCi93ShBilfyufWLSwBV-0i9-GFKQ.IPLAdUZnICAwnHhzl_bEi_ZFm_zAUTEJzIgsJ-zZdXNGJTY7Nc015oj5AFsJgbzoUD4X3Y-AIvFAXyEo0-SBzA&_sg%5B1%5D=nNwZJXww5diSEIC64lZ0mVnmXZscPJLB5NEujX0EXNbrV7Z9rYMu38Zm5xHWY6ODcsbexCa2xAj9zNEwJyq4PEVO3Xdbb_l5AIM8HfAdKQg9.IPLAdUZnICAwnHhzl_bEi_ZFm_zAUTEJzIgsJ-zZdXNGJTY7Nc015oj5AFsJgbzoUD4X3Y-AIvFAXyEo0-SBzA&_iepl=&_rtd=eyJjb250ZW50SW50ZW50IjoibWFpbkl0ZW0ifQ%3D%3D | T-YOLO: Tiny Vehicle Detection Based on YOLO and Multi-Scale Convolutional Neural Networks | To solve real-life problems for different smart city applications, using deep Neural Network, such as parking occupancy detection, requires fine-tuning of these networks. For large parking, it is desirable to use a cenital-plane camera located at a high distance that allows the monitoring of the entire parking space or a large parking area with only one camera. Today’s most popular object detection models, such as YOLO, achieve good precision scores at real-time speed. However, if we use our own data different from that of the general-purpose datasets, such as COCO and ImageNet, we have a large margin for improvisation. In this paper, we propose a modified, yet lightweight, deep object detection model based on the YOLO-v5 architecture. The proposed model can detect large, small, and tiny objects. Specifically, we propose the use of a multi-scale mechanism to learn deep discriminative feature representations at different scales and automatically determine the most suitable scales for detecting objects in a scene (i.e., in our case vehicles). The proposed multi-scale module reduces the number of trainable parameters compared to the original YOLO-v5 architecture. The experimental results also demonstrate that precision is improved by a large margin. In fact, as shown in the experiments, the results show a small reduction from 7.28 million parameters of the YOLO-v5-S profile to 7.26 million parameters in our model. In addition, we reduced the detection speed by inferring 30 fps compared to the YOLO-v5-L/X profiles. In addition, the tiny vehicle detection performance was significantly improved by 33% compared to the YOLO-v5-X profile. | ['Domenec Puig', 'Miguel Ángel García', 'Hatem RashwanHatem Rashwan', 'Daniel Padilla Carrasco'] | 2021-12-01 | null | null | null | ieee-access-2021-12 | ['parking-space-occupancy'] | ['computer-vision'] | [-5.83178878e-01 -3.06337863e-01 5.00340015e-03 -1.20712087e-01
-6.99796677e-01 -1.53188154e-01 2.40844026e-01 -2.56399959e-01
-8.07617724e-01 5.14043987e-01 -5.24738073e-01 -3.20670694e-01
3.34022552e-01 -9.92724359e-01 -8.64650309e-01 -7.42611766e-01
1.75517648e-01 5.28379261e-01 7.08487153e-01 -1.48532033e-01
-5.73085360e-02 6.18000925e-01 -1.85706484e+00 -1.54292420e-01
8.40372026e-01 1.27539051e+00 7.24474609e-01 5.58519721e-01
6.27740622e-02 4.61577713e-01 -5.28631985e-01 9.94202197e-02
2.97476262e-01 2.66334087e-01 -1.87487572e-01 2.53893435e-01
5.04772127e-01 -6.51474774e-01 -4.28864062e-01 9.72920299e-01
4.76133853e-01 -3.19822095e-02 2.98007220e-01 -1.39238155e+00
-2.50841439e-01 2.36690611e-01 -7.27202654e-01 3.53717268e-01
-3.41999143e-01 4.44982022e-01 7.27828085e-01 -1.18601537e+00
1.24909513e-01 1.21768177e+00 5.67804992e-01 2.95114845e-01
-9.44477141e-01 -9.11898017e-01 2.70792693e-02 5.22738516e-01
-1.77788889e+00 -4.17602986e-01 4.69016552e-01 -2.68967301e-01
7.54649043e-01 1.81395025e-03 3.76499921e-01 5.82616806e-01
1.13584630e-01 1.06486893e+00 6.07838094e-01 -1.42001539e-01
2.42345348e-01 3.34759384e-01 -9.75918919e-02 7.36408651e-01
6.73113227e-01 -1.30868051e-02 2.82644015e-02 2.50866294e-01
7.57195473e-01 3.12390059e-01 3.41886461e-01 -5.16630888e-01
-9.00245130e-01 1.09523261e+00 6.86171174e-01 2.76564986e-01
-1.61546052e-01 4.41882521e-01 5.30716419e-01 -3.22346628e-01
2.91865561e-02 5.53341443e-03 -3.71315658e-01 2.64537465e-02
-8.94115567e-01 1.66291386e-01 2.74360657e-01 1.02713037e+00
1.04908490e+00 2.03528970e-01 2.27800589e-02 6.59213603e-01
3.30611825e-01 8.68303955e-01 5.84961116e-01 -7.21882761e-01
6.23259425e-01 7.19684184e-01 3.32920432e-01 -8.55978251e-01
-6.59810543e-01 -3.77114058e-01 -7.25522816e-01 2.54316539e-01
4.31692153e-01 2.45205611e-02 -9.75515783e-01 1.47302926e+00
4.07134622e-01 2.57278960e-02 -9.16184634e-02 9.18952584e-01
8.46273482e-01 6.42270029e-01 3.31440307e-02 3.61067444e-01
1.73009467e+00 -1.15938902e+00 -2.04888135e-01 -7.32146800e-01
8.08220148e-01 -6.86887085e-01 1.25923932e+00 2.80215386e-02
-7.57815838e-01 -8.19644690e-01 -1.11613464e+00 -1.07933395e-01
-5.90949476e-01 7.41391063e-01 6.85928285e-01 7.37416923e-01
-8.75261307e-01 5.57901375e-02 -7.78424680e-01 -5.62491477e-01
4.76812333e-01 4.76028979e-01 -1.11232735e-01 -5.61509468e-02
-1.14012849e+00 8.01603198e-01 4.99878973e-01 3.83279994e-02
-1.08024716e+00 -3.17184836e-01 -8.39977205e-01 5.62646806e-01
4.68124658e-01 -3.07478458e-01 1.09306669e+00 -4.34649229e-01
-1.10573280e+00 6.54544055e-01 -4.20835726e-02 -7.28370130e-01
4.14309889e-01 -1.55602351e-01 -3.44778657e-01 4.74336185e-02
4.35412258e-01 9.73665893e-01 5.77751100e-01 -9.39774454e-01
-1.21033859e+00 -3.24249923e-01 3.24724168e-01 -1.22194842e-01
-2.38585949e-01 -1.57640532e-01 -9.67343032e-01 -3.08031756e-02
-1.59257501e-01 -1.14099085e+00 -3.99014801e-01 5.83238564e-02
-1.97911054e-01 -4.69274402e-01 1.13624251e+00 -1.81216583e-01
9.73975539e-01 -2.33970737e+00 -8.11286867e-01 -1.29491255e-01
2.86121964e-01 5.37744522e-01 4.99677993e-02 -1.02778554e-01
3.63960207e-01 -1.62806943e-01 1.33613631e-01 -3.78580719e-01
3.35544273e-02 3.02885115e-01 1.03068508e-01 5.33972025e-01
2.44443398e-02 1.08537173e+00 -5.96051514e-01 -6.70583129e-01
7.12578595e-01 3.93097103e-01 -4.56819028e-01 -8.94003958e-02
9.64122340e-02 -1.05198413e-01 -4.80813473e-01 8.67839217e-01
9.74339008e-01 -5.91658831e-01 -1.38479188e-01 -6.39197230e-02
-3.00232798e-01 1.91121414e-01 -1.14485085e+00 1.14979374e+00
-8.07675600e-01 9.11109865e-01 8.32667872e-02 -9.70899761e-01
1.04935241e+00 -1.37775600e-01 2.96624660e-01 -1.06800091e+00
2.78971255e-01 2.65685827e-01 -1.64042786e-01 -1.86132759e-01
8.73229980e-01 2.26937592e-01 -3.59297872e-01 7.77546912e-02
-2.58916795e-01 3.82895142e-01 4.44027573e-01 1.73198715e-01
1.01220882e+00 -3.97803903e-01 2.56591082e-01 -2.69195680e-02
7.07016885e-01 3.89543772e-02 5.78827083e-01 9.17612910e-01
-5.72937965e-01 2.86908299e-01 2.15896040e-01 -6.61736250e-01
-1.08594823e+00 -9.04085279e-01 -2.92860359e-01 8.97664547e-01
4.84694868e-01 -3.68810326e-01 -7.82081068e-01 -4.52919364e-01
2.81384289e-01 4.68401253e-01 -3.35321367e-01 -5.58313280e-02
-7.59239793e-01 -7.27487504e-01 6.06508970e-01 8.18029463e-01
8.18065524e-01 -9.63322520e-01 -1.04427850e+00 2.67842710e-01
2.23558191e-02 -1.65232968e+00 -2.90818363e-01 2.31400043e-01
-6.97814703e-01 -9.67317760e-01 -4.06809092e-01 -7.58861959e-01
5.10836661e-01 8.45805526e-01 8.64760280e-01 -1.46976337e-01
-4.59218293e-01 2.60273498e-02 -1.24557056e-01 -2.63197064e-01
5.58896549e-02 1.22767702e-01 1.84983283e-01 -1.07943051e-01
8.15468609e-01 -1.54637188e-01 -7.71332562e-01 6.17759943e-01
-6.57597959e-01 -1.79336235e-01 8.61096442e-01 7.48752117e-01
6.34222388e-01 1.58292681e-01 4.12335902e-01 -3.91808510e-01
-4.13781479e-02 -4.53546464e-01 -1.11878812e+00 -2.04638451e-01
-7.04644084e-01 -5.82990535e-02 8.03390324e-01 -3.99194092e-01
-5.67355275e-01 3.07727784e-01 -1.80244427e-02 -6.48906648e-01
-2.41876885e-01 7.94505142e-03 -1.97627500e-01 -5.26178628e-02
4.54636455e-01 3.28795791e-01 -1.35247901e-01 -4.75499153e-01
4.11338627e-01 7.46840656e-01 4.55601692e-01 -1.14157125e-01
8.71057689e-01 7.53608286e-01 -1.59437329e-01 -8.38511825e-01
-5.65640986e-01 -7.05610931e-01 -3.96899968e-01 -5.58773391e-02
8.39542687e-01 -1.36917925e+00 -1.16319525e+00 2.98968732e-01
-8.88796568e-01 -2.14554384e-01 -1.61109924e-01 4.42849696e-01
-3.56233299e-01 3.15439433e-01 -2.54756212e-01 -7.45448411e-01
-4.24078763e-01 -1.38554633e+00 1.40816927e+00 5.30651927e-01
3.45305681e-01 -5.12521505e-01 -4.03316110e-01 4.87798363e-01
4.66204584e-01 -1.87237218e-01 2.51553118e-01 -2.45427623e-01
-1.02493429e+00 -4.32875276e-01 -7.64173388e-01 1.88500240e-01
-1.49697959e-01 -3.53575945e-01 -9.27214503e-01 -3.99738967e-01
-3.03795934e-01 -1.97949812e-01 1.07444012e+00 5.12241423e-01
1.26358318e+00 1.00595608e-01 -5.61499119e-01 5.81122041e-01
1.44044757e+00 1.86066598e-01 6.28164709e-01 9.09975708e-01
6.65040970e-01 -6.26506135e-02 9.35647011e-01 4.70349282e-01
6.77303970e-01 9.87824798e-01 7.17517436e-01 -2.57191926e-01
-9.23523083e-02 -1.67621434e-01 5.06525397e-01 3.96414757e-01
2.92038172e-01 -2.79385280e-02 -7.86067069e-01 8.20978463e-01
-1.89087546e+00 -8.92216325e-01 -2.24792007e-02 2.19041276e+00
1.61480084e-01 4.81868476e-01 2.29044974e-01 7.12563097e-03
7.91036308e-01 1.19125113e-01 -6.06265068e-01 -2.24857375e-01
2.12410763e-02 -6.61807880e-02 1.24431884e+00 2.67781079e-01
-1.36454725e+00 1.19534671e+00 5.76315260e+00 1.12984943e+00
-1.28308499e+00 2.29485080e-01 4.74100649e-01 -3.99395645e-01
2.80255884e-01 -8.85571763e-02 -1.58161938e+00 7.54045486e-01
1.00574803e+00 -4.99403058e-03 1.73063204e-01 1.66586852e+00
5.00011206e-01 -3.26626986e-01 -7.49613941e-01 1.27198470e+00
2.67560370e-02 -1.47007668e+00 -3.71585280e-01 3.17102164e-01
3.51958483e-01 5.67911267e-01 -8.35358649e-02 7.25011528e-01
2.15304971e-01 -8.63984466e-01 6.76345646e-01 -1.40794039e-01
7.31354833e-01 -8.72209907e-01 8.08400452e-01 4.59795713e-01
-1.63790047e+00 -2.96989888e-01 -9.11489129e-01 6.16800934e-02
2.76877820e-01 5.93962073e-01 -9.76831377e-01 1.30294040e-02
1.07275009e+00 4.44572806e-01 -7.14897215e-01 1.00361323e+00
-7.65666515e-02 3.16333443e-01 -6.35056734e-01 -1.98468789e-01
5.41596353e-01 -3.36681004e-03 1.62648797e-01 1.05836010e+00
3.67914319e-01 -4.05629158e-01 3.25432658e-01 8.43562007e-01
-6.67895004e-02 -1.78341851e-01 -5.33484340e-01 4.30786014e-01
5.95063686e-01 1.65054953e+00 -6.67771876e-01 -4.51386362e-01
-6.98481977e-01 5.80440521e-01 2.63318896e-01 -8.77610222e-02
-1.24108601e+00 -3.29767615e-01 8.43437016e-01 3.69129211e-01
1.03131747e+00 -3.28685254e-01 -7.94069692e-02 -9.58973229e-01
5.21204956e-02 -3.65825295e-01 4.84792888e-02 -6.90627337e-01
-7.03327119e-01 4.67703015e-01 -2.98883826e-01 -1.25542557e+00
-2.43949722e-02 -7.81406105e-01 -5.91434002e-01 4.89959270e-01
-1.67039442e+00 -1.02163649e+00 -4.62285608e-01 5.43657362e-01
7.14655101e-01 -1.78515673e-01 3.26098740e-01 8.19226921e-01
-7.65896261e-01 8.12442839e-01 2.51051486e-01 2.30007023e-01
4.02908593e-01 -8.87704372e-01 5.54757118e-01 8.12633991e-01
-5.45890182e-02 2.30677456e-01 4.57417369e-01 -1.68499604e-01
-1.32332730e+00 -1.44288290e+00 8.37962627e-01 -2.58037001e-01
5.58943629e-01 -4.56281453e-01 -5.59620380e-01 6.12968028e-01
-3.63207638e-01 3.69874060e-01 5.11962064e-02 -1.83956102e-01
-3.70310433e-02 -6.01997077e-01 -1.11094797e+00 3.37351352e-01
6.94311321e-01 -3.23925585e-01 -7.25475401e-02 2.77850866e-01
5.84816039e-01 -3.62401515e-01 -4.11638975e-01 1.45971894e-01
5.28751493e-01 -8.98990035e-01 1.14898872e+00 -3.75449620e-02
8.80785212e-02 -6.57303214e-01 -3.72996926e-01 -6.24857903e-01
-4.20735866e-01 9.66519639e-02 -1.44135818e-01 1.02406549e+00
1.48450196e-01 -6.39476836e-01 1.22836196e+00 2.91817009e-01
-8.93520266e-02 -7.08698690e-01 -1.28011048e+00 -1.04407668e+00
-2.52364755e-01 -3.97132039e-01 6.32057607e-01 3.62583548e-01
-5.89179099e-01 3.43804777e-01 -4.01600540e-01 4.40429300e-01
6.27692819e-01 2.13270470e-01 1.30553699e+00 -1.07509768e+00
-1.13982320e-01 -2.64139980e-01 -8.23652089e-01 -1.44880164e+00
-1.93274528e-01 -6.26941085e-01 5.87393381e-02 -1.42196608e+00
2.14016572e-01 -6.88330770e-01 -3.99196923e-01 3.63990307e-01
-9.74091813e-02 3.17350686e-01 3.54071409e-01 3.12522262e-01
-8.77113760e-01 5.22755921e-01 8.54180157e-01 -2.26526171e-01
-9.39700380e-02 -3.37949302e-03 -5.49790323e-01 7.29303837e-01
9.33055639e-01 -4.28898484e-01 -8.70827362e-02 -3.73831570e-01
-1.17124274e-01 -3.38597715e-01 4.98603731e-01 -1.33265305e+00
2.40251526e-01 3.60697508e-02 4.74073201e-01 -1.16133869e+00
6.00953877e-01 -7.89347410e-01 -1.41950831e-01 6.91745520e-01
4.97852325e-01 -6.25221431e-02 3.87376994e-01 4.01677698e-01
-2.01386526e-01 1.17823463e-02 9.42002594e-01 -1.82287261e-01
-1.33697367e+00 2.47608602e-01 -4.87502873e-01 -2.56635249e-01
1.34466267e+00 -4.61459160e-01 -6.11326933e-01 -1.39377207e-01
-3.06912869e-01 4.80376005e-01 3.86482865e-01 5.40721059e-01
5.16789854e-01 -1.54946554e+00 -2.76879638e-01 1.82636246e-01
3.83334994e-01 1.70304060e-01 2.83357739e-01 8.62216175e-01
-5.43982863e-01 9.92854238e-01 -2.00050622e-01 -1.01694095e+00
-1.17733121e+00 6.55157208e-01 2.56953090e-01 -3.88498813e-01
-8.05698991e-01 5.08756101e-01 5.63755095e-01 -3.26953799e-01
-5.79175390e-02 -3.89378130e-01 -1.61297739e-01 -1.53345466e-01
5.24772227e-01 4.32664841e-01 6.60139844e-02 -7.76441634e-01
-6.94203019e-01 5.61833143e-01 -1.33838490e-01 5.04490793e-01
1.17101729e+00 -3.59570414e-01 4.25453365e-01 1.07384235e-01
9.82840538e-01 -9.72035304e-02 -1.32354927e+00 -2.15240657e-01
-1.48318276e-01 -3.77274603e-01 2.77570277e-01 -2.70344764e-01
-1.23325384e+00 7.68414199e-01 1.08224845e+00 9.53259543e-02
8.86536539e-01 3.33313912e-01 1.12995732e+00 4.49376851e-01
5.64810753e-01 -1.26407349e+00 -1.12358510e-01 5.02042234e-01
3.36422145e-01 -1.61952400e+00 3.37377354e-03 -2.04321891e-01
-4.47631449e-01 8.39853406e-01 9.95839953e-01 -2.28075445e-01
4.47262913e-01 1.82633176e-01 -1.63639411e-02 -2.78031141e-01
-5.45214176e-01 -5.94505548e-01 -1.63598433e-01 3.32392097e-01
-1.02635063e-01 5.85197031e-01 -5.06700762e-02 3.19701850e-01
-6.27083480e-02 -9.25635546e-02 4.91697997e-01 6.76829457e-01
-1.00656915e+00 -8.03413391e-01 -6.39049232e-01 4.30559963e-01
-1.91138275e-02 -4.58631739e-02 2.77753115e-01 1.11415780e+00
3.61065447e-01 8.87978733e-01 4.03846830e-01 -3.37137282e-01
3.33453536e-01 -3.38560075e-01 4.61861165e-03 -3.03633153e-01
-9.15785599e-03 -1.95220821e-02 -1.83962345e-01 -7.21686542e-01
-5.37632480e-02 -3.97769272e-01 -1.30560660e+00 -6.57552063e-01
-4.82521832e-01 -1.75963134e-01 8.08693707e-01 8.49947333e-01
5.39139450e-01 4.29250658e-01 6.92712307e-01 -1.01260591e+00
-4.38144654e-01 -8.46234143e-01 -6.96048975e-01 5.95887266e-02
1.98481545e-01 -1.04732752e+00 -3.54809731e-01 -4.14453179e-01] | [8.284208297729492, -0.8639853000640869] |
27454017-e1a2-4856-bd32-af2ce18182d3 | pirc-net-using-proposal-indexing | 1812.03213 | null | http://arxiv.org/abs/1812.03213v1 | http://arxiv.org/pdf/1812.03213v1.pdf | PIRC Net : Using Proposal Indexing, Relationships and Context for Phrase Grounding | Phrase Grounding aims to detect and localize objects in images that are
referred to and are queried by natural language phrases. Phrase grounding finds
applications in tasks such as Visual Dialog, Visual Search and Image-text
co-reference resolution. In this paper, we present a framework that leverages
information such as phrase category, relationships among neighboring phrases in
a sentence and context to improve the performance of phrase grounding systems.
We propose three modules: Proposal Indexing Network(PIN); Inter-phrase
Regression Network(IRN) and Proposal Ranking Network(PRN) each of which analyze
the region proposals of an image at increasing levels of detail by
incorporating the above information. Also, in the absence of ground-truth
spatial locations of the phrases(weakly-supervised), we propose knowledge
transfer mechanisms that leverages the framework of PIN module. We demonstrate
the effectiveness of our approach on the Flickr 30k Entities and ReferItGame
datasets, for which we achieve improvements over state-of-the-art approaches in
both supervised and weakly-supervised variants. | ['Rama Kovvuri', 'Ram Nevatia'] | 2018-12-07 | null | null | null | null | ['phrase-grounding'] | ['natural-language-processing'] | [ 2.16581523e-01 3.55563045e-01 -6.82748616e-01 -2.20606461e-01
-1.15275633e+00 -8.46729040e-01 9.42265391e-01 4.52487022e-01
-4.52667594e-01 4.59765911e-01 5.03258586e-01 -2.19185635e-01
-6.40174150e-02 -4.84854788e-01 -8.97512019e-01 -3.88982654e-01
3.34349275e-02 4.54999954e-01 7.73105204e-01 -6.56713247e-02
4.64291632e-01 3.93394113e-01 -1.38196421e+00 6.19454265e-01
1.96332350e-01 1.02929282e+00 3.04440618e-01 5.04097283e-01
-2.29335785e-01 8.49858344e-01 -2.08113611e-01 -4.71745968e-01
1.10649236e-01 6.16613254e-02 -1.04551041e+00 -8.96155834e-03
8.93389285e-01 -1.82550639e-01 -4.49017435e-01 9.24416423e-01
2.04013452e-01 3.99554633e-02 6.52763307e-01 -1.22052717e+00
-6.98973894e-01 6.72471523e-01 -1.03363895e+00 2.76897013e-01
4.76765215e-01 -2.11118951e-01 1.55438745e+00 -1.37317014e+00
9.74929571e-01 1.43120790e+00 5.47552109e-01 2.53486395e-01
-1.30911422e+00 -5.93795359e-01 4.33847934e-01 1.37944117e-01
-1.82963681e+00 -2.90028334e-01 5.12582719e-01 -3.65003824e-01
1.08745170e+00 1.18310869e-01 3.81894231e-01 8.43761384e-01
4.62097023e-03 9.78408694e-01 7.75727928e-01 -5.65428019e-01
-1.15620166e-01 3.51980239e-01 8.57460424e-02 9.07438457e-01
-1.99176893e-01 -1.75982207e-01 -1.04134703e+00 -2.76203990e-01
7.29732037e-01 -2.13547256e-02 -2.06744283e-01 -6.58752203e-01
-1.52338886e+00 7.59549916e-01 8.55518043e-01 1.64369494e-01
-3.71876806e-01 4.77925211e-01 1.69468865e-01 -3.66367161e-01
5.09535670e-01 4.28338945e-01 -3.35631460e-01 3.49387378e-01
-1.23559296e+00 4.56370384e-01 6.80288911e-01 1.43027449e+00
9.90320981e-01 -1.05492759e+00 -8.34582508e-01 6.55202925e-01
5.36318421e-01 5.08154988e-01 3.70861366e-02 -9.42203462e-01
8.55228364e-01 6.49898529e-01 2.90886045e-01 -1.37345827e+00
-2.05763265e-01 -2.34917447e-01 -2.90846497e-01 -5.53981364e-01
-2.75860149e-02 4.61148679e-01 -9.99116838e-01 1.48751771e+00
4.39258575e-01 3.63152564e-01 -2.06913263e-01 9.14672256e-01
1.24501705e+00 7.51976013e-01 3.27796787e-01 3.49128693e-02
1.75026643e+00 -1.27905762e+00 -5.06871939e-01 -1.19915225e-01
3.72392774e-01 -7.06666529e-01 1.01416492e+00 -2.68562555e-01
-9.94007230e-01 -4.08921957e-01 -6.18583441e-01 -4.09674168e-01
-6.59617424e-01 1.34627491e-01 2.59915590e-01 2.25420576e-02
-1.35216606e+00 2.32904181e-02 -4.40249532e-01 -8.31480503e-01
4.59083140e-01 3.83096278e-01 -3.99906725e-01 -4.14648131e-02
-9.98448014e-01 7.60111809e-01 6.04797244e-01 -6.34605810e-02
-7.66032755e-01 -9.25331354e-01 -8.43834221e-01 1.16860708e-02
6.65675044e-01 -7.98439026e-01 1.19401836e+00 -3.58842909e-01
-7.04447925e-01 1.33416605e+00 -3.13685805e-01 -6.57823443e-01
2.06559554e-01 -2.87927657e-01 -5.97234927e-02 5.90752602e-01
6.19359791e-01 1.63135624e+00 7.86950469e-01 -1.33422017e+00
-1.08502054e+00 -3.08010906e-01 3.74011338e-01 5.80005348e-01
1.58469081e-01 2.65408218e-01 -1.35169613e+00 -5.18631220e-01
3.51473540e-01 -1.11529064e+00 -5.06081544e-02 2.68651545e-01
-9.08632100e-01 -5.73909461e-01 9.16633010e-01 -5.91875374e-01
1.18116379e+00 -2.16847873e+00 2.15150431e-01 2.88033962e-01
3.19424331e-01 -1.94343403e-01 -1.29426464e-01 4.54740852e-01
2.73265690e-01 3.26485842e-01 1.63089946e-01 -4.71379489e-01
1.98985934e-01 3.34076345e-01 -6.51924610e-01 2.37169474e-01
4.28012013e-01 1.45334291e+00 -8.42430711e-01 -1.10961998e+00
2.41856687e-02 3.21588129e-01 -2.90071219e-01 -4.30577509e-02
-4.58941579e-01 2.22369000e-01 -6.62835896e-01 7.93574810e-01
2.31523186e-01 -6.07682347e-01 -3.44641626e-01 -6.69438183e-01
-1.68796629e-01 1.33966744e-01 -9.37593639e-01 2.09865451e+00
-1.16896920e-01 5.94468296e-01 -5.11348918e-02 -5.32698035e-01
5.76078296e-01 2.78109610e-01 3.34762245e-01 -7.85363019e-01
-4.59620595e-01 -3.56632024e-02 -7.70170271e-01 -3.59472752e-01
9.71722126e-01 4.52367216e-01 -2.96943843e-01 3.39714199e-01
8.41605142e-02 3.80103737e-02 1.93616226e-01 8.09890866e-01
1.16722453e+00 2.41178766e-01 3.49383384e-01 -7.02408329e-02
4.65681106e-01 2.29763433e-01 4.60626259e-02 1.27677143e+00
-1.65258691e-01 7.98167825e-01 4.21512961e-01 -2.32207403e-01
-1.03665864e+00 -1.14354432e+00 -4.67442013e-02 1.58632064e+00
6.19421065e-01 -6.99851990e-01 -3.46332341e-01 -8.38118017e-01
1.39379904e-01 3.52843672e-01 -8.13706279e-01 3.29503775e-01
-4.73551899e-01 -1.17725588e-01 4.35195565e-01 7.03713894e-01
5.66904902e-01 -1.05934966e+00 -5.15405715e-01 -1.49159029e-01
-3.89086932e-01 -1.53114104e+00 -6.34448230e-01 1.53530231e-02
-4.24309164e-01 -8.18770051e-01 -8.54402602e-01 -8.60719502e-01
6.90431535e-01 5.14280200e-01 1.45133305e+00 1.69835299e-01
-1.60449877e-01 9.23673511e-01 -3.31415355e-01 -2.26524279e-01
1.42034069e-01 3.77274215e-01 -2.69846439e-01 -2.37682551e-01
4.42946523e-01 -1.16499960e-01 -7.65281379e-01 4.93905038e-01
-6.96837664e-01 2.38304198e-01 8.66588295e-01 6.33855999e-01
1.14115679e+00 -1.79220602e-01 1.13604546e-01 -7.85148501e-01
3.84561002e-01 -3.80394608e-01 -6.15182579e-01 6.87492371e-01
-3.10418040e-01 1.10127240e-01 -3.22056115e-01 -1.81605339e-01
-7.63098717e-01 2.48943269e-01 4.02225137e-01 -5.00964403e-01
-1.94280729e-01 5.00977218e-01 1.69083565e-01 -3.68429720e-02
4.64313537e-01 1.04658037e-01 -5.29115617e-01 -2.34118447e-01
8.72672319e-01 4.27268982e-01 7.84319580e-01 -5.45983970e-01
9.43786323e-01 6.51622057e-01 -4.99872118e-02 -6.02090776e-01
-1.35300660e+00 -1.25933754e+00 -8.94604802e-01 -1.37485191e-01
1.22972190e+00 -1.27835774e+00 -3.98806602e-01 -4.10352647e-01
-1.43070900e+00 1.81211069e-01 -2.52601266e-01 1.28348216e-01
-5.50437510e-01 -6.13024868e-02 -4.33725208e-01 -5.33967316e-01
-3.75344753e-01 -9.57590282e-01 1.78758097e+00 2.21676335e-01
-2.50608511e-02 -6.32056594e-01 -1.12088248e-02 4.93703306e-01
1.13628991e-01 2.17479039e-02 7.02941597e-01 -6.43268704e-01
-1.01638162e+00 -2.78681330e-02 -8.12349200e-01 -3.88361812e-01
-1.61716864e-01 -2.61279732e-01 -9.30083394e-01 -6.20384216e-02
-7.26510048e-01 -3.99071246e-01 1.10004449e+00 3.68578166e-01
1.06226599e+00 -4.76496667e-01 -8.90227258e-01 3.73379827e-01
1.37441897e+00 -3.30196857e-01 2.99420774e-01 3.99730682e-01
9.17877913e-01 8.27927768e-01 8.27726662e-01 3.74527499e-02
6.82367742e-01 1.16780984e+00 4.82395113e-01 -3.67570251e-01
-1.30367771e-01 -6.03322864e-01 -1.14945896e-01 9.00586843e-02
5.92366271e-02 -2.20633581e-01 -1.07711101e+00 8.14612031e-01
-2.26969266e+00 -9.38969791e-01 1.67637110e-01 1.59669220e+00
7.98065305e-01 1.98328048e-02 1.81105986e-01 -4.71017659e-01
8.82859945e-01 2.43612111e-01 -4.45760846e-01 4.62540723e-02
-8.36915970e-02 -2.13549942e-01 7.24441946e-01 3.17188263e-01
-1.50431156e+00 1.32388389e+00 5.89362955e+00 8.50935578e-01
-6.03588700e-01 1.28693029e-01 5.88403642e-01 1.75480068e-01
-2.02859342e-01 8.06469377e-03 -1.23249674e+00 -6.23782985e-02
5.32851398e-01 1.78761587e-01 4.58363667e-02 6.74212575e-01
3.56423892e-02 -2.63304383e-01 -1.40121603e+00 1.03138101e+00
2.27122888e-01 -1.73921430e+00 1.91207290e-01 1.42823055e-01
6.95953488e-01 2.69937575e-01 2.94323355e-01 1.94938451e-01
1.73329189e-01 -1.07692564e+00 8.75469685e-01 5.54186583e-01
6.69689119e-01 -4.69921798e-01 7.91933119e-01 2.76868105e-01
-1.45317149e+00 1.32629583e-02 -2.81187743e-01 5.77693462e-01
2.07429782e-01 1.65409893e-01 -1.19721591e+00 5.51279008e-01
8.79983664e-01 8.72189045e-01 -8.88048589e-01 8.47070634e-01
-4.68102932e-01 4.39430386e-01 -3.82947028e-01 7.73415640e-02
5.16916335e-01 3.72409672e-01 5.90021610e-01 1.35090899e+00
-8.48811045e-02 2.22011462e-01 5.24657369e-01 1.09417772e+00
-2.48778328e-01 2.06485137e-01 -5.87604761e-01 8.71731043e-02
6.40160739e-01 1.35158694e+00 -1.09943998e+00 -1.33330107e-01
-5.53166509e-01 8.38877857e-01 3.53526413e-01 5.43576300e-01
-7.36929893e-01 4.92013879e-02 4.47063595e-02 2.40872696e-01
7.08615303e-01 -1.42171264e-01 1.21715188e-01 -9.30884361e-01
2.16710716e-01 -3.42762381e-01 4.86497313e-01 -1.18145216e+00
-1.34294152e+00 3.92559767e-01 5.12311578e-01 -9.56982613e-01
-8.71209204e-02 -3.82497460e-01 -2.33076766e-01 7.98767328e-01
-1.73789704e+00 -1.78972816e+00 -3.13558102e-01 5.91644466e-01
5.36532223e-01 4.93960083e-02 7.09197879e-01 3.63636250e-03
-2.40993232e-01 2.54839957e-01 -4.19147432e-01 3.05401385e-01
7.97568440e-01 -1.28088129e+00 2.78246641e-01 6.16116822e-01
8.79038095e-01 7.84995615e-01 4.57281440e-01 -5.89902461e-01
-1.09906888e+00 -1.28841186e+00 1.08378243e+00 -8.12144816e-01
8.39770913e-01 -5.68588078e-01 -5.87421179e-01 6.99380994e-01
2.71941155e-01 4.25508380e-01 5.67531407e-01 1.97500825e-01
-5.17661572e-01 6.87888488e-02 -8.39428723e-01 5.66586912e-01
1.03575063e+00 -8.91739190e-01 -7.67389953e-01 5.78217685e-01
1.14951491e+00 -6.47440374e-01 -6.19511366e-01 2.61205256e-01
5.42562425e-01 -4.59881902e-01 1.33089447e+00 -4.04579371e-01
5.03983259e-01 -5.76714993e-01 -4.38696712e-01 -5.61758637e-01
-2.46307403e-01 -2.30036199e-01 -3.01135212e-01 1.38477862e+00
7.81984568e-01 1.32888526e-01 9.29497600e-01 4.98782307e-01
2.06515834e-01 -7.50470340e-01 -8.84476840e-01 -3.23029220e-01
-4.27225798e-01 -3.78221393e-01 4.17437822e-01 5.74978948e-01
-1.21792875e-01 6.09690368e-01 -1.12392761e-01 4.87436622e-01
5.13716996e-01 2.47184515e-01 6.67240262e-01 -9.45714056e-01
-3.60594504e-02 -1.59254260e-02 -5.96332908e-01 -1.36336577e+00
4.43232387e-01 -7.90093958e-01 2.40574434e-01 -1.82749951e+00
6.99837446e-01 -4.30078089e-01 -4.45705205e-01 7.33299017e-01
-1.09619141e-01 6.56924427e-01 3.26296896e-01 6.97735488e-01
-1.61751425e+00 3.17256927e-01 7.47634768e-01 -4.76687849e-01
-1.70818478e-01 -2.46348023e-01 -5.84571779e-01 6.35793269e-01
1.75991461e-01 -6.32048547e-01 -4.13950443e-01 -5.88000178e-01
3.73017699e-01 -9.20356140e-02 8.16925764e-01 -6.83673084e-01
8.22189033e-01 1.28233716e-01 2.92843252e-01 -1.46599460e+00
3.85565698e-01 -9.78058875e-01 -2.53721535e-01 -1.07966222e-01
-8.13582957e-01 2.91117221e-01 1.98481798e-01 1.04415381e+00
-3.13619077e-01 -5.15653789e-02 1.39067456e-01 -2.45933816e-01
-1.04534590e+00 3.55899006e-01 6.76971301e-02 -8.84229303e-06
1.04449344e+00 -1.00015126e-01 -4.16507155e-01 -3.09525251e-01
-6.56151235e-01 5.62170208e-01 3.59241337e-01 5.14281094e-01
7.97132552e-01 -1.18755686e+00 -4.87772018e-01 -4.03478801e-01
7.59743154e-01 1.38742298e-01 -6.70147762e-02 9.23527896e-01
-3.45582843e-01 9.93446231e-01 3.12012017e-01 -1.24663401e+00
-1.31696391e+00 5.45875788e-01 1.37822824e-02 -4.32734013e-01
-7.03382969e-01 1.09740388e+00 5.22461772e-01 -3.31881374e-01
5.65509975e-01 -2.64398783e-01 -4.37551826e-01 8.14505070e-02
5.15358984e-01 -2.29981706e-01 -1.56587452e-01 -1.05506110e+00
-6.03013217e-01 8.84983301e-01 -5.00712097e-01 -3.87382537e-01
1.14195240e+00 -4.86743361e-01 -1.68697178e-01 3.35797250e-01
1.03481126e+00 -1.56717375e-01 -1.12959111e+00 -7.17598975e-01
4.02436525e-01 -2.63328075e-01 1.74150050e-01 -7.87056804e-01
-6.78983271e-01 4.96739239e-01 5.41041315e-01 2.42673401e-02
9.38522458e-01 8.50697935e-01 2.70563781e-01 6.68332636e-01
4.55551863e-01 -8.70062947e-01 3.37177217e-01 3.29187274e-01
8.54135931e-01 -1.35340130e+00 1.11298427e-01 -6.00706041e-01
-6.60709500e-01 7.27174997e-01 4.82958496e-01 1.07912853e-01
6.30960166e-01 -5.93498610e-02 -2.95355260e-01 -6.44910514e-01
-8.46895158e-01 -4.21667337e-01 8.95600200e-01 4.16250467e-01
2.02786461e-01 -2.56675571e-01 -1.16456091e-01 2.75558293e-01
1.24695271e-01 -3.12213689e-01 1.36827221e-02 9.09718215e-01
-4.63051379e-01 -7.17432439e-01 -3.54112595e-01 2.42171407e-01
-4.29331362e-01 -4.32943463e-01 -6.31792724e-01 8.40282381e-01
1.27821669e-01 7.84629703e-01 2.06493214e-01 -6.97576627e-02
2.89009422e-01 -1.39840215e-01 1.39224544e-01 -1.01473105e+00
-5.42566359e-01 1.48986682e-01 9.15055722e-02 -8.38724554e-01
-1.01476705e+00 -5.49632072e-01 -1.10962963e+00 4.05796617e-01
-5.50603747e-01 2.61710912e-01 5.91007352e-01 1.07734084e+00
4.73408222e-01 2.69383490e-01 7.35778213e-02 -7.99522817e-01
-2.83111352e-02 -7.90533602e-01 -4.58328009e-01 2.40141988e-01
2.47035742e-01 -6.22650862e-01 2.05781427e-03 2.68439710e-01] | [10.477545738220215, 1.469075083732605] |
5f6e5296-0a7d-4877-be37-86246b8d7183 | improving-solar-flare-prediction-by-time | 2206.07197 | null | https://arxiv.org/abs/2206.07197v1 | https://arxiv.org/pdf/2206.07197v1.pdf | Improving Solar Flare Prediction by Time Series Outlier Detection | Solar flares not only pose risks to outer space technologies and astronauts' well being, but also cause disruptions on earth to our hight-tech, interconnected infrastructure our lives highly depend on. While a number of machine-learning methods have been proposed to improve flare prediction, none of them, to the best of our knowledge, have investigated the impact of outliers on the reliability and those models' performance. In this study, we investigate the impact of outliers in a multivariate time series benchmark dataset, namely SWAN-SF, on flare prediction models, and test our hypothesis. That is, there exist outliers in SWAN-SF, removal of which enhances the performance of the prediction models on unseen datasets. We employ Isolation Forest to detect the outliers among the weaker flare instances. Several experiments are carried out using a large range of contamination rates which determine the percentage of present outliers. We asses the quality of each dataset in terms of its actual contamination using TimeSeriesSVC. In our best finding, we achieve a 279% increase in True Skill Statistic and 68% increase in Heidke Skill Score. The results show that overall a significant improvement can be achieved to flare prediction if outliers are detected and removed properly. | ['Rafal A. Angryk', 'Azim Ahmadzadeh', 'Md Reazul Islam', 'Junzhi Wen'] | 2022-06-14 | null | null | null | null | ['solar-flare-prediction'] | ['time-series'] | [-3.96649353e-02 -2.51512289e-01 5.74521944e-02 -9.09883678e-02
-5.25075555e-01 -5.98183930e-01 6.31029189e-01 -5.20543754e-02
1.45786628e-01 1.10049880e+00 6.99711293e-02 -1.58705279e-01
-6.55377269e-01 -7.08580732e-01 -6.87833726e-01 -6.99554920e-01
-1.22754611e-01 2.83883423e-01 2.58242577e-01 -3.03869545e-01
4.09087211e-01 7.75656939e-01 -1.72691369e+00 -2.09855780e-01
1.25905371e+00 1.12071729e+00 -4.18400943e-01 4.41692263e-01
2.92934000e-01 6.10853434e-01 -7.86236763e-01 3.93562242e-02
3.55646998e-01 -2.91935593e-01 -5.61969839e-02 -2.47380912e-01
2.41240174e-01 4.27885383e-01 -2.16383979e-01 8.60557675e-01
4.67917383e-01 3.70370020e-04 6.42825484e-01 -1.40669978e+00
1.86513662e-01 8.23535025e-02 -4.66182768e-01 7.08596945e-01
2.51852036e-01 2.85738498e-01 5.32439768e-01 -6.15133166e-01
3.85336637e-01 5.28417051e-01 8.40344012e-01 -3.06810051e-01
-9.53966081e-01 -8.79841924e-01 -3.36895257e-01 1.25377283e-01
-1.16588521e+00 -5.26821673e-01 5.34350932e-01 -7.22428501e-01
1.00349212e+00 4.17941749e-01 5.59996724e-01 6.76639616e-01
8.33105028e-01 9.80181843e-02 9.98913646e-01 -3.02405417e-01
2.33715788e-01 -9.30740535e-02 2.97147661e-01 1.62964806e-01
8.46442461e-01 4.79949713e-01 -6.25827730e-01 -2.20773458e-01
1.35397166e-01 -6.08597286e-02 -3.36175233e-01 -3.14969532e-02
-9.94434714e-01 7.31630743e-01 2.66148508e-01 6.28793001e-01
-3.47370267e-01 -2.24206537e-01 1.14831530e-01 4.93770540e-01
3.77163202e-01 6.09795809e-01 -5.30557275e-01 -1.45752057e-01
-9.56578255e-01 1.60293460e-01 6.99931920e-01 4.91790175e-01
4.87715185e-01 3.22454721e-01 -1.72791034e-01 2.41637632e-01
-1.13963269e-01 9.91964042e-01 5.02835631e-01 -3.47994566e-01
2.59772718e-01 6.28757417e-01 2.12877825e-01 -1.26340365e+00
-6.37645185e-01 -9.45538700e-01 -7.55986512e-01 1.44525766e-01
-2.53555588e-02 -3.88301164e-01 -8.33629310e-01 1.15326607e+00
4.14078720e-02 4.42642599e-01 3.12432144e-02 5.70236683e-01
5.17042398e-01 4.36267555e-01 -2.95926660e-01 -4.99701351e-01
5.50239325e-01 -4.71002489e-01 -8.20715249e-01 -4.00980920e-01
6.64596677e-01 -8.78164530e-01 3.92065257e-01 7.72154570e-01
-7.37660706e-01 -5.21170020e-01 -1.23795116e+00 6.63913786e-01
-3.24293435e-01 -7.86276460e-02 7.54453838e-01 7.71786988e-01
-3.35884631e-01 8.50354671e-01 -8.00795197e-01 -3.03362429e-01
1.62270516e-01 1.49763405e-01 -2.41524890e-01 1.24593481e-01
-9.81217980e-01 9.82956767e-01 2.94314504e-01 9.58609059e-02
-4.67326760e-01 -6.63017511e-01 -3.82662386e-01 -7.94050023e-02
2.25021854e-01 -4.79236126e-01 7.34797478e-01 -7.61222243e-01
-6.38538659e-01 2.41537064e-01 -1.87703401e-01 -6.66114271e-01
4.12880003e-01 -2.18385294e-01 -1.26416159e+00 -3.90997648e-01
1.12593345e-01 -1.42139643e-01 7.71315098e-01 -1.07793367e+00
-8.09811294e-01 -5.10117769e-01 -7.67936230e-01 -3.03373545e-01
1.30529374e-01 -7.51887262e-02 1.87261015e-01 -6.48705959e-01
4.08250540e-01 -1.07874572e+00 -1.37520954e-01 -6.92347169e-01
-1.15824595e-01 6.82736039e-02 7.73486078e-01 -6.63291276e-01
1.62696242e+00 -2.15521002e+00 -2.81438917e-01 7.28876770e-01
-2.44009286e-01 2.34841973e-01 3.42525005e-01 5.75774431e-01
-1.74660087e-01 -4.32515293e-02 -3.99694383e-01 2.25224212e-01
-1.68483883e-01 3.51498306e-01 -5.60930729e-01 6.24275088e-01
2.09800169e-01 5.41328371e-01 -4.22740847e-01 3.02630097e-01
2.10644260e-01 7.68384188e-02 -4.19113934e-01 7.69824535e-02
-1.12037413e-01 5.92834949e-01 -3.67940485e-01 6.31187737e-01
5.50340533e-01 2.34602705e-01 -1.62287086e-01 3.52932997e-02
-3.24768424e-01 -1.98087245e-02 -1.17630231e+00 1.16962159e+00
-8.65128487e-02 5.47280371e-01 -5.98838568e-01 -6.61210001e-01
1.32634914e+00 8.96409228e-02 4.56660956e-01 -1.06437087e+00
7.72431567e-02 5.14338911e-01 2.77576953e-01 -6.03976190e-01
5.03820360e-01 -3.16673577e-01 -8.38234127e-02 3.18420045e-02
-3.10898542e-01 -1.39880687e-01 3.54234487e-01 -4.48077582e-02
1.39358580e+00 -1.62297964e-01 1.31973907e-01 -4.63277280e-01
3.16876382e-01 1.33659959e-01 7.83006072e-01 6.18837655e-01
-1.09434202e-01 4.74307418e-01 3.38499606e-01 -2.91443110e-01
-7.33797133e-01 -1.01046109e+00 -4.65586424e-01 5.07566988e-01
-1.79993406e-01 -1.23822339e-01 -8.39079916e-02 -4.23442096e-01
4.37211215e-01 1.04820108e+00 -2.45239317e-01 -4.30713832e-01
-1.54906645e-01 -1.22649944e+00 4.32557851e-01 3.22111160e-01
2.17785001e-01 -1.11971462e+00 -5.99608660e-01 -3.49937752e-02
-1.95810106e-02 -8.18498492e-01 2.46075645e-01 4.16895449e-01
-1.03616893e+00 -1.22424650e+00 3.26984704e-01 -2.85110950e-01
3.08640867e-01 6.47723004e-02 1.16229808e+00 9.34023261e-02
-4.09674346e-01 -4.72781919e-02 -4.99278367e-01 -7.54630685e-01
-4.29180935e-02 -1.40041009e-01 2.27676272e-01 -1.57692045e-01
4.88666594e-01 -6.10523999e-01 -4.21233833e-01 3.79953414e-01
-9.02504444e-01 -6.54461205e-01 4.99669641e-01 3.44317257e-01
5.58844328e-01 6.40804470e-01 9.34658825e-01 -8.02903712e-01
2.50168145e-01 -8.72455657e-01 -6.32338345e-01 -1.11201152e-01
-8.09480548e-01 5.64968772e-02 4.20184851e-01 6.88736793e-03
-7.69211173e-01 -2.28772908e-01 1.01855218e-01 -3.36445242e-01
-2.38115937e-01 6.01871729e-01 5.33866659e-02 -1.75919071e-01
8.22173953e-01 -3.59362513e-01 -2.73367226e-01 -3.53494912e-01
-1.96487546e-01 5.91966450e-01 6.54703915e-01 -8.20877403e-02
1.11835480e+00 4.72873032e-01 2.40779757e-01 -8.21637392e-01
-9.77412403e-01 -7.55007327e-01 -3.87965411e-01 -3.09165090e-01
4.97286648e-01 -7.51711965e-01 -2.09596410e-01 3.22510362e-01
-4.67237443e-01 3.48801762e-02 -4.62280102e-02 5.89758635e-01
-4.00001436e-01 8.10834318e-02 2.49917701e-01 -9.64006305e-01
-2.57497430e-01 -6.11602187e-01 7.31925786e-01 2.85707980e-01
-2.06288159e-01 -7.77486145e-01 4.04573679e-01 2.75640517e-01
4.49608594e-01 8.67441297e-01 6.16252244e-01 -5.31077743e-01
-4.17812824e-01 -4.10535097e-01 7.59357512e-02 2.78034866e-01
4.43661623e-02 1.23373203e-01 -9.71352577e-01 -6.90197825e-01
2.77387828e-01 2.84814656e-01 7.50287592e-01 3.62092763e-01
8.13503087e-01 1.18411228e-01 -5.37609637e-01 5.49200535e-01
1.58743191e+00 4.22432482e-01 9.61433351e-01 4.58688796e-01
2.68006027e-01 2.38143042e-01 9.79607761e-01 5.87320089e-01
-4.37744737e-01 2.11351857e-01 5.07211626e-01 -1.00342790e-02
1.27034262e-01 9.52662379e-02 4.08041477e-01 7.74350107e-01
2.08192449e-02 -2.30093613e-01 -9.83542740e-01 6.08261585e-01
-1.84724569e+00 -1.12532258e+00 -7.04360187e-01 2.42765379e+00
3.93142849e-02 5.49735308e-01 -1.17851436e-01 5.73840320e-01
3.86489570e-01 6.19536452e-02 -5.52798688e-01 -1.25250190e-01
-2.92109787e-01 3.27124685e-01 7.70928383e-01 2.16221660e-01
-8.35320592e-01 3.55274737e-01 6.22773647e+00 3.00341278e-01
-1.13586116e+00 -2.41250142e-01 2.58454204e-01 -4.30310965e-01
-3.75845730e-01 7.03225732e-02 -5.79165876e-01 6.37679219e-01
1.21210694e+00 -2.78223902e-01 3.10058177e-01 3.87491405e-01
4.19014901e-01 -4.12376255e-01 -6.28538072e-01 8.40256751e-01
4.72527206e-01 -9.04493213e-01 -4.93788898e-01 -3.83180194e-02
9.52653706e-01 3.29484403e-01 -3.28110725e-01 3.00403148e-01
3.30089293e-02 -1.02148485e+00 5.07513940e-01 1.42168653e+00
1.53694361e-01 -1.21209884e+00 1.02113128e+00 4.02125657e-01
-7.15464771e-01 -3.58539224e-01 -2.34019652e-01 -3.35400254e-01
8.03788081e-02 1.43975592e+00 -5.97231686e-01 9.25737917e-01
9.51113641e-01 5.92712879e-01 -8.14534605e-01 1.52108681e+00
2.95562297e-02 7.91842937e-01 -4.67358291e-01 2.80881643e-01
-1.11481391e-01 -2.61710793e-01 6.23024821e-01 7.21116781e-01
8.54766488e-01 -1.96997859e-02 1.25173442e-02 2.05612525e-01
3.12963307e-01 1.10092899e-02 -8.99964929e-01 -9.76950768e-03
4.30297375e-01 1.03189254e+00 -3.56065452e-01 1.15478896e-01
-4.16115522e-01 2.92445660e-01 -2.44820565e-01 1.01733662e-01
-8.78574729e-01 -3.04428518e-01 6.41789973e-01 4.19327825e-01
3.27226907e-01 -1.33390874e-01 -4.75620866e-01 -9.01990235e-01
9.98505950e-02 -8.61114919e-01 6.22498035e-01 -1.00935113e+00
-1.40103996e+00 3.58645052e-01 -3.69861186e-01 -1.35506189e+00
-1.09749958e-01 -3.28419507e-01 -7.84238875e-01 6.55584157e-01
-1.42231774e+00 -4.79353786e-01 -4.83857244e-01 3.65027279e-01
8.56889114e-02 -3.75155866e-01 4.45096940e-01 2.93672919e-01
-4.64387923e-01 1.65536031e-01 5.05770564e-01 -4.35393453e-01
9.03614819e-01 -7.82577336e-01 1.16362780e-01 1.13736093e+00
6.35593459e-02 1.03544809e-01 1.05802143e+00 -1.18701541e+00
-1.17212248e+00 -1.24310875e+00 8.13952804e-01 -6.24927104e-01
8.89756799e-01 2.02160999e-01 -8.73110950e-01 6.82441473e-01
-7.62354806e-02 -1.40520066e-01 4.71166968e-01 3.70058507e-01
1.43409848e-01 -1.71705037e-01 -1.22627306e+00 2.06163138e-01
8.01355541e-01 -2.17685372e-01 -7.43124306e-01 3.26973677e-01
3.05521995e-01 -9.56740752e-02 -1.07632422e+00 1.20078552e+00
2.46934056e-01 -1.38258445e+00 6.65238917e-01 -5.22370934e-01
6.06515296e-02 -6.76780343e-01 -2.01495633e-01 -1.28959751e+00
-3.05443615e-01 -3.45244199e-01 4.73412536e-02 1.11240017e+00
4.16912585e-01 -6.71175063e-01 5.38643479e-01 1.68464780e-01
-4.21348631e-01 -2.14542180e-01 -9.03767645e-01 -8.92253399e-01
5.05875237e-03 -4.84908164e-01 6.14270985e-01 9.52229083e-01
-2.52938658e-01 2.56628066e-01 -1.23487324e-01 6.32114589e-01
4.76580501e-01 2.69140631e-01 8.04728329e-01 -1.80375791e+00
-1.57272428e-01 -9.67549980e-02 -4.11615670e-01 1.58981696e-01
-2.11184070e-01 -7.49644220e-01 5.75078055e-02 -1.24831545e+00
-1.78165972e-01 -4.28336203e-01 -3.93172741e-01 2.69668877e-01
-1.64675981e-01 2.45687336e-01 -2.88708974e-02 3.45989168e-01
-3.79475772e-01 3.77481908e-01 6.83284700e-01 1.81132257e-01
-1.53097287e-01 2.19240427e-01 -2.30219260e-01 7.61475682e-01
1.05289423e+00 -6.23661578e-01 -2.38201931e-01 -1.86014816e-01
5.04910052e-01 1.33731782e-01 2.55037099e-01 -1.88139009e+00
7.29790255e-02 -8.99007469e-02 5.97574770e-01 -8.46635222e-01
-1.29536062e-01 -8.72901440e-01 7.20562458e-01 4.90259409e-01
4.68829662e-01 2.37104654e-01 4.05915737e-01 5.13993621e-01
-1.89701989e-01 -3.51587087e-01 5.61928928e-01 2.77783006e-01
-4.39146012e-01 -9.98730678e-03 -3.45280439e-01 -1.72114819e-01
1.09446311e+00 -7.38049895e-02 -3.49866480e-01 -5.07012531e-02
-5.99006295e-01 2.25170583e-01 5.38224339e-01 5.37840307e-01
-7.15411901e-02 -1.15256417e+00 -5.41804373e-01 5.45325696e-01
1.94973961e-01 -3.59606624e-01 1.71947941e-01 1.00846457e+00
-2.47114152e-01 5.16067266e-01 -2.66261518e-01 -4.52068269e-01
-1.05325925e+00 4.44111943e-01 3.53347778e-01 -7.22483844e-02
-3.55168492e-01 2.21121326e-01 -4.05793220e-01 -2.15938434e-01
5.85838184e-02 -3.27136636e-01 -2.28666887e-01 2.82815903e-01
2.44610414e-01 6.10578120e-01 7.63840675e-01 -4.61993307e-01
-2.49219149e-01 4.91371393e-01 3.83896053e-01 3.70297939e-01
1.33719456e+00 8.63653123e-02 -2.00534701e-01 6.46227002e-01
4.44319785e-01 4.30229604e-01 -7.95801699e-01 2.43867457e-01
3.36132735e-01 -6.01591706e-01 2.45749131e-01 -1.00332391e+00
-9.49438930e-01 2.68039316e-01 8.97102833e-01 5.18265843e-01
1.23041880e+00 -3.09065789e-01 5.93944967e-01 3.88702273e-01
4.80650574e-01 -9.64709222e-01 -2.05802187e-01 5.50584853e-01
7.77422786e-01 -1.15742946e+00 3.80867086e-02 -1.52691498e-01
-4.90701944e-01 6.91859901e-01 3.53429854e-01 -3.59569222e-01
6.49561822e-01 3.20452482e-01 1.57321356e-02 -3.33621204e-01
-7.80532241e-01 -3.73363435e-01 5.14359027e-02 3.38305026e-01
1.60741687e-01 1.90521136e-01 -4.62114692e-01 5.24937689e-01
-5.20071507e-01 6.47744983e-02 5.24130940e-01 8.16035450e-01
-8.96864176e-01 -7.16475010e-01 -7.57378101e-01 1.01096535e+00
-1.97908431e-01 2.13374481e-01 -2.76102573e-01 7.41300166e-01
3.49846482e-01 1.21251762e+00 2.86686301e-01 -1.66909471e-01
7.31669724e-01 2.67630398e-01 1.58792987e-01 -1.48296937e-01
-6.33319736e-01 -1.32955670e-01 3.72397304e-01 -4.78007615e-01
-2.93319136e-01 -8.70151103e-01 -1.13965106e+00 -4.91631150e-01
-5.14757216e-01 3.80317450e-01 6.03201449e-01 1.08302402e+00
2.90264457e-01 8.40883315e-01 7.58446753e-01 -3.55881125e-01
-2.20915645e-01 -1.01325774e+00 -7.41820097e-01 5.12034476e-01
1.42232135e-01 -8.84478509e-01 -7.91768730e-01 -3.85954231e-01] | [6.7673659324646, 2.8233373165130615] |
295821b5-6ed5-45d1-affd-bc4550abda19 | expression-empowered-residen-network-for | 1806.04957 | null | http://arxiv.org/abs/1806.04957v1 | http://arxiv.org/pdf/1806.04957v1.pdf | Expression Empowered ResiDen Network for Facial Action Unit Detection | The paper explores the topic of Facial Action Unit (FAU) detection in the
wild. In particular, we are interested in answering the following questions:
(1) how useful are residual connections across dense blocks for face analysis?
(2) how useful is the information from a network trained for categorical Facial
Expression Recognition (FER) for the task of FAU detection? The proposed
network (ResiDen) exploits dense blocks along with residual connections and
uses auxiliary information from a FER network. The experiments are performed on
the EmotionNet and DISFA datasets. The experiments show the usefulness of
facial expression information for AU detection. The proposed network achieves
state-of-art results on the two databases. Analysis of the results for cross
database protocol shows the effectiveness of the network. | ['Abhinav Dhall', 'Shreyank Jyoti'] | 2018-06-13 | null | null | null | null | ['action-unit-detection', 'facial-action-unit-detection'] | ['computer-vision', 'computer-vision'] | [ 5.78740984e-02 1.40277758e-01 4.04408909e-02 -5.76360703e-01
-3.54885042e-01 2.23005321e-02 3.36567461e-01 -6.18435919e-01
-2.91932702e-01 5.56508541e-01 2.11924121e-01 4.61163014e-01
2.13719279e-01 -4.77615863e-01 -5.39477766e-01 -8.74604464e-01
-4.08172399e-01 -2.62395591e-01 -2.41308734e-01 -6.36476696e-01
-1.83479041e-01 8.69976461e-01 -1.79399490e+00 5.72320640e-01
6.41944036e-02 1.88575709e+00 -5.67276180e-01 2.52030522e-01
1.94919050e-01 1.19845951e+00 -5.29650807e-01 -4.91467655e-01
3.42684299e-01 -4.74676341e-01 -7.47192264e-01 2.27567941e-01
6.12255633e-01 -5.18478334e-01 -2.28668705e-01 1.04170966e+00
6.36938214e-01 3.72350067e-02 6.00960732e-01 -1.46392202e+00
-1.12192392e-01 8.75636488e-02 -7.44672120e-01 3.25587422e-01
2.59432673e-01 -1.20692044e-01 7.75893033e-01 -1.13611162e+00
7.05585241e-01 1.53946733e+00 5.42418480e-01 8.32555175e-01
-8.62905324e-01 -1.07247329e+00 -2.03672230e-01 3.54115665e-01
-1.59792519e+00 -1.04311609e+00 7.86640942e-01 -2.48470649e-01
9.80845094e-01 3.16070318e-02 3.61045837e-01 1.28943753e+00
-8.67544189e-02 8.79788101e-01 9.79039371e-01 -4.80054080e-01
-4.67334986e-02 1.06814474e-01 1.65974379e-01 1.00160551e+00
-4.59480673e-01 1.12683447e-02 -6.23139620e-01 -7.80309737e-02
5.62886119e-01 -3.48003834e-01 -1.85849667e-01 1.24219164e-01
-3.62559587e-01 1.00358081e+00 3.28527480e-01 5.33861578e-01
-5.98617077e-01 2.06627160e-01 7.94220388e-01 5.81551492e-01
8.65329862e-01 7.51999170e-02 -2.48428345e-01 -1.95802420e-01
-8.70348155e-01 -4.13941704e-02 6.32215321e-01 6.77680850e-01
7.53441572e-01 5.46438813e-01 -1.39343888e-01 1.21775079e+00
2.46081665e-01 2.31233940e-01 2.76137978e-01 -1.05052638e+00
-4.85300692e-03 5.11359990e-01 -2.18781725e-01 -1.06919980e+00
-2.33570501e-01 1.11325324e-01 -8.17919374e-01 5.94288707e-01
1.77966967e-01 -6.13280535e-01 -7.55510926e-01 1.97407627e+00
2.04057604e-01 1.54194757e-01 -3.43603715e-02 7.23578036e-01
1.19777906e+00 5.37251651e-01 2.40303859e-01 -2.55752504e-01
1.23176503e+00 -8.10588121e-01 -8.20346951e-01 2.12186635e-01
6.07212305e-01 -5.42008162e-01 4.06075954e-01 2.80439734e-01
-9.71725285e-01 -7.37662554e-01 -8.07749927e-01 1.34915411e-01
-3.89253259e-01 6.15461171e-01 6.79339707e-01 7.40074635e-01
-1.38825715e+00 2.86158264e-01 -4.74183023e-01 -7.09723890e-01
8.94069314e-01 8.40285361e-01 -9.15414214e-01 1.51484668e-01
-1.15487456e+00 6.02287173e-01 -1.30201178e-02 3.85511339e-01
-9.30225909e-01 -1.88683599e-01 -9.51065242e-01 8.79873186e-02
1.09818481e-01 4.93867472e-02 9.07591760e-01 -2.01470542e+00
-1.52718604e+00 1.11162245e+00 -2.17384383e-01 -3.37161958e-01
3.20090801e-01 -1.83280453e-01 -5.19030690e-01 6.14572823e-01
-9.15664211e-02 9.74000514e-01 1.00150979e+00 -7.04585910e-01
-5.33287823e-01 -5.91480970e-01 7.86776692e-02 -1.54280603e-01
-4.13174331e-01 6.01451635e-01 -3.84800047e-01 -6.51994526e-01
-3.10048074e-01 -8.61783087e-01 4.33217049e-01 2.36052230e-01
2.30892608e-03 -6.39699161e-01 1.17230022e+00 -8.52432549e-01
9.96497333e-01 -2.55648899e+00 -9.00465548e-02 2.87517756e-01
-6.72102869e-02 3.95484120e-01 -4.11747277e-01 -1.19573865e-02
-5.78079522e-01 -5.01812659e-02 1.53440803e-01 -2.92266041e-01
-1.98431656e-01 1.73535809e-01 -1.08204730e-01 5.67545533e-01
5.94246268e-01 7.64749169e-01 -3.75312567e-01 -5.60236394e-01
-6.01153560e-02 4.19911534e-01 -5.99956632e-01 4.51712728e-01
8.60476941e-02 1.91145688e-01 -4.76291299e-01 1.06311214e+00
7.16621995e-01 3.37348700e-01 -8.47009420e-02 -6.30074680e-01
1.72730416e-01 -4.64319438e-01 -8.45083356e-01 1.32517695e+00
-1.53696910e-01 8.20272088e-01 5.92572331e-01 -1.03717077e+00
1.03821969e+00 6.03952765e-01 5.86670578e-01 -7.04867601e-01
7.52484977e-01 -1.75164808e-02 -1.79959517e-02 -5.86689591e-01
4.47788555e-03 -1.28147542e-01 2.65395820e-01 2.17548922e-01
8.36405873e-01 5.55663586e-01 2.11177379e-01 -6.10895082e-02
1.21012080e+00 1.50047556e-01 2.63544858e-01 -4.53755409e-01
8.71111989e-01 -5.33796847e-01 5.41417301e-01 2.49568298e-01
-6.45334244e-01 2.50019103e-01 9.45086777e-01 -5.71912467e-01
-6.24614060e-01 -6.81125343e-01 -2.14843899e-01 1.44139850e+00
-3.43950480e-01 -4.10034090e-01 -9.36644495e-01 -8.49201441e-01
-1.15507945e-01 1.61316901e-01 -1.34297132e+00 6.24589249e-02
-3.90661389e-01 -6.56884551e-01 8.72514904e-01 6.00885153e-01
9.62196529e-01 -1.34677708e+00 -3.93165410e-01 -2.40996912e-01
-1.37242123e-01 -1.23647749e+00 -8.65700468e-02 8.37974548e-02
-3.50311607e-01 -1.08255148e+00 -5.04432619e-01 -7.08272040e-01
5.77893138e-01 -1.37135819e-01 8.86875093e-01 2.35812529e-03
-3.00240368e-01 6.27497554e-01 -3.79576862e-01 -3.75521064e-01
-2.88714945e-01 -2.68275082e-01 5.41106202e-02 8.19057107e-01
8.59300911e-01 -4.90323752e-01 -2.80348718e-01 4.38629806e-01
-6.45033836e-01 -5.80798388e-01 4.34160799e-01 8.37040901e-01
3.04178298e-01 -1.57292813e-01 5.65565526e-01 -7.43370235e-01
5.33440411e-01 -4.58170652e-01 -3.27599317e-01 8.60911310e-02
2.59188898e-02 -1.50510266e-01 1.87415898e-01 -2.75557727e-01
-1.15204692e+00 1.30199850e-01 -5.45181632e-01 -8.68927002e-01
-4.12744343e-01 -4.51060422e-02 -4.39248115e-01 -3.32553059e-01
6.27824843e-01 -2.49272376e-01 3.59476894e-01 -1.95964321e-01
2.72669340e-03 7.72216678e-01 2.96980709e-01 -5.16884863e-01
2.03185566e-02 6.19918346e-01 1.52545720e-01 -1.19977474e+00
-7.58423805e-01 -4.21140254e-01 -5.81800044e-01 -5.09858191e-01
1.12759852e+00 -1.05798864e+00 -9.19572413e-01 5.36412835e-01
-9.84069824e-01 4.88092080e-02 -7.49832764e-02 3.94711792e-02
-5.21436632e-01 1.83468550e-01 -7.84789622e-01 -9.03786421e-01
-5.15710115e-01 -1.15011716e+00 1.07812071e+00 3.96849215e-02
-3.53768945e-01 -6.31378353e-01 -1.94231197e-02 1.95994779e-01
3.07007372e-01 6.20786667e-01 5.08148193e-01 -5.67193031e-01
1.20224692e-01 -2.46039420e-01 -3.45074743e-01 9.19653475e-01
1.76151454e-01 1.42873615e-01 -1.58318138e+00 2.96003670e-02
8.49833488e-02 -9.88434434e-01 9.48088408e-01 2.20464289e-01
1.19817460e+00 -2.45249093e-01 1.26397148e-01 6.04624331e-01
1.02416468e+00 4.40948717e-02 8.89434695e-01 -1.24734372e-01
3.99081022e-01 9.11201596e-01 4.61433500e-01 4.39139038e-01
-4.42029595e-01 6.81901634e-01 6.05544865e-01 -2.06047431e-01
-7.15942904e-02 2.57049590e-01 6.16536915e-01 1.79505751e-01
-3.77556086e-01 1.03602961e-01 -5.13221502e-01 4.22036111e-01
-1.81454563e+00 -1.24449992e+00 3.08317602e-01 1.34850752e+00
4.51352149e-01 -3.51830333e-01 2.77515113e-01 5.93377883e-03
5.83051622e-01 3.72774482e-01 -1.71048522e-01 -9.61986482e-01
-3.30782533e-01 5.64037204e-01 -1.61110442e-02 1.11439891e-01
-1.38760710e+00 1.01429427e+00 6.28033590e+00 1.03920555e+00
-1.30780661e+00 3.59076947e-01 1.16327047e+00 -1.33956909e-01
6.82128787e-01 -6.28993332e-01 -7.92856991e-01 2.39525244e-01
9.99868631e-01 3.81529808e-01 1.26290500e-01 1.12644279e+00
7.44997337e-02 -3.37280244e-01 -9.80491936e-01 1.21942556e+00
4.93275136e-01 -1.04311395e+00 1.32419169e-01 -8.16212595e-02
4.84423339e-01 5.97285107e-02 8.38464871e-02 4.17122871e-01
1.04158178e-01 -1.26939595e+00 4.62250650e-01 4.69996572e-01
9.60093498e-01 -1.03964615e+00 1.03315258e+00 -2.41765261e-01
-9.95530963e-01 -1.14414565e-01 -5.40545702e-01 2.56265514e-02
-4.00755107e-01 -7.72571773e-04 -6.92254722e-01 1.25565499e-01
9.15701330e-01 9.20913994e-01 -5.25261641e-01 2.05896571e-01
-2.34959766e-01 7.40014791e-01 -4.49076772e-01 9.06439051e-02
3.88668716e-01 -1.68073401e-01 2.78925717e-01 1.36177683e+00
1.54213592e-01 1.70878068e-01 -4.72802699e-01 7.03126609e-01
-4.68246162e-01 2.79294461e-01 -7.78679371e-01 -8.57417583e-02
-2.28468910e-01 1.58098543e+00 -4.07903492e-01 -8.56372416e-02
-6.27159417e-01 1.04542327e+00 5.15014589e-01 3.10559005e-01
-5.90100050e-01 -8.14074278e-02 1.03517652e+00 1.48799866e-01
3.78736168e-01 3.73659164e-01 4.73943055e-01 -7.89706945e-01
-1.52795389e-01 -8.83051097e-01 6.60606563e-01 -9.36727345e-01
-1.19158852e+00 9.91372645e-01 -1.35811821e-01 -8.96222293e-01
-4.05077815e-01 -1.01184833e+00 -6.33024812e-01 6.71810210e-01
-1.24961531e+00 -1.33498669e+00 -4.10619766e-01 1.09933913e+00
4.09022927e-01 -5.85253239e-01 1.03006124e+00 4.23252642e-01
-6.80570126e-01 8.35224807e-01 -3.66976738e-01 8.67250502e-01
6.38384104e-01 -6.72393620e-01 -4.43715781e-01 6.01467967e-01
5.66223077e-02 2.67261863e-01 4.65928227e-01 -1.51192859e-01
-1.13920367e+00 -1.16763401e+00 4.34466571e-01 -2.30416462e-01
4.77464944e-01 -3.68735462e-01 -5.82356751e-01 6.58559740e-01
3.56917024e-01 4.49285835e-01 7.51912832e-01 2.75493935e-02
-4.60793048e-01 -3.24322969e-01 -1.40891182e+00 1.41899437e-01
7.84136295e-01 -5.26424229e-01 -1.62878886e-01 2.94376593e-02
-2.14663167e-02 5.48844524e-02 -6.77015007e-01 6.73521459e-01
7.56978393e-01 -1.08880234e+00 6.00085378e-01 -9.75145400e-01
5.85742295e-01 1.72585055e-01 -5.43966413e-01 -9.66963351e-01
-2.08136830e-02 -4.81390715e-01 2.81905066e-02 1.20802629e+00
1.55860290e-01 -2.16099679e-01 7.80953586e-01 4.89539295e-01
3.97275448e-01 -8.96466553e-01 -1.08125341e+00 -3.97517234e-01
-2.66174793e-01 -3.87150139e-01 1.72213152e-01 9.16698575e-01
-2.50261098e-01 6.08503461e-01 -5.75738192e-01 -2.74545848e-01
2.30003059e-01 -1.43234581e-01 7.51661301e-01 -1.01220000e+00
1.54374540e-01 -1.67968631e-01 -1.02729404e+00 -4.09610301e-01
7.27599800e-01 -6.38760507e-01 -1.85130373e-01 -7.49336123e-01
1.12965956e-01 2.63368636e-01 -3.34931076e-01 8.73323143e-01
3.12118322e-01 7.95726418e-01 1.57959327e-01 -1.09793179e-01
-6.16928220e-01 7.66416192e-01 6.48135900e-01 -8.65473896e-02
2.80531645e-01 -1.77827418e-01 -3.88650477e-01 9.06515539e-01
6.37910843e-01 -1.20369650e-01 -1.07124589e-01 -1.99271180e-02
-3.97226959e-02 -1.02472395e-01 2.86035746e-01 -8.66074800e-01
-1.82650104e-01 3.25065553e-01 6.82742417e-01 -3.42033267e-01
7.00185359e-01 -8.97729456e-01 -3.15708309e-01 2.29218621e-02
-9.34551433e-02 8.48064646e-02 5.55655301e-01 1.28919512e-01
-7.16897786e-01 -2.58567613e-02 1.20932889e+00 -4.18666974e-02
-1.01172709e+00 4.26654279e-01 -4.46011633e-01 -1.54759094e-01
1.04548383e+00 -1.99918777e-01 4.71435580e-03 -6.23154581e-01
-1.13566470e+00 -1.67854041e-01 -4.88230139e-02 4.08715695e-01
4.92359608e-01 -1.51619792e+00 -8.17651272e-01 4.05352652e-01
2.89390564e-01 -7.70580351e-01 4.24495749e-02 9.16455925e-01
-3.53939146e-01 2.28885502e-01 -7.42641091e-01 -5.05353451e-01
-1.80637681e+00 2.19329283e-01 7.17276812e-01 9.20592919e-02
-8.04033875e-03 1.16550338e+00 5.30402541e-01 -5.75606376e-02
2.83905417e-01 2.74946243e-01 -5.34424365e-01 5.38635671e-01
6.93118036e-01 2.89448619e-01 8.64730552e-02 -1.33715856e+00
-3.24613929e-01 2.98923284e-01 -8.11503977e-02 1.50416359e-01
1.56092143e+00 9.97914523e-02 -5.17063975e-01 1.60959274e-01
1.68147004e+00 -1.50228247e-01 -1.04249835e+00 -1.39375895e-01
-2.28832647e-01 -2.84124166e-01 6.31116703e-02 -5.80808163e-01
-1.56278062e+00 1.04068732e+00 1.07274735e+00 -3.05051833e-01
1.51614892e+00 4.75195274e-02 1.87902704e-01 3.27108532e-01
1.76484957e-01 -1.13760018e+00 2.30478808e-01 5.39878964e-01
1.17307019e+00 -1.20553029e+00 -3.80808294e-01 -3.50768298e-01
-6.29328251e-01 1.28908336e+00 8.85650456e-01 -2.63698280e-01
1.13484406e+00 3.00494701e-01 1.20825641e-01 -4.98751462e-01
-6.35362625e-01 -2.74916291e-01 1.79996386e-01 2.82939017e-01
5.57005644e-01 -2.77871549e-01 -3.90928946e-02 5.38845181e-01
8.30118284e-02 6.11506514e-02 7.12530762e-02 8.06238174e-01
-2.61736840e-01 -7.16337800e-01 -1.56495571e-01 4.63850826e-01
-9.10808802e-01 2.07063153e-01 -9.41252351e-01 1.00880134e+00
3.50503504e-01 8.20838869e-01 1.94762275e-02 -2.30633378e-01
2.41706893e-01 3.62137258e-01 5.90767145e-01 -3.19455534e-01
-7.33999193e-01 1.77794561e-01 3.29788148e-01 -1.08122909e+00
-8.82365048e-01 -5.37149072e-01 -8.87319863e-01 -2.23829418e-01
-4.43618447e-02 3.89925167e-02 3.14888716e-01 8.36642802e-01
3.75201344e-01 2.23146319e-01 7.82214224e-01 -6.26390874e-01
-4.21709381e-02 -1.35921097e+00 -8.69539559e-01 5.72505713e-01
2.11467177e-01 -8.07652295e-01 -3.53676260e-01 1.08855300e-01] | [13.599461555480957, 1.7990878820419312] |
908ee594-631f-4781-a76d-a4da0e8cd1de | coreference-resolution-through-a-seq2seq | 2211.12142 | null | https://arxiv.org/abs/2211.12142v1 | https://arxiv.org/pdf/2211.12142v1.pdf | Coreference Resolution through a seq2seq Transition-Based System | Most recent coreference resolution systems use search algorithms over possible spans to identify mentions and resolve coreference. We instead present a coreference resolution system that uses a text-to-text (seq2seq) paradigm to predict mentions and links jointly. We implement the coreference system as a transition system and use multilingual T5 as an underlying language model. We obtain state-of-the-art accuracy on the CoNLL-2012 datasets with 83.3 F1-score for English (a 2.3 higher F1-score than previous work (Dobrovolskii, 2021)) using only CoNLL data for training, 68.5 F1-score for Arabic (+4.1 higher than previous work) and 74.3 F1-score for Chinese (+5.3). In addition we use the SemEval-2010 data sets for experiments in the zero-shot setting, a few-shot setting, and supervised setting using all available training data. We get substantially higher zero-shot F1-scores for 3 out of 4 languages than previous approaches and significantly exceed previous supervised state-of-the-art results for all five tested languages. | ['Michael Collins', 'Chris Alberti', 'Bernd Bohnet'] | 2022-11-22 | null | null | null | null | ['coreference-resolution'] | ['natural-language-processing'] | [-8.05407390e-02 3.22874784e-01 -7.28403211e-01 -2.26851389e-01
-1.55363691e+00 -8.12772512e-01 9.75141048e-01 1.22918017e-01
-8.05044413e-01 1.16398478e+00 6.71242237e-01 -1.07851923e-01
-2.16099858e-01 -3.33442122e-01 -6.79294527e-01 -3.00528467e-01
-1.28014341e-01 1.16771841e+00 3.91884446e-01 -7.19393194e-01
2.09678084e-01 1.53590545e-01 -1.39641511e+00 5.51197708e-01
8.89470339e-01 1.91133693e-01 3.13035287e-02 5.70370734e-01
-2.57432848e-01 6.20684683e-01 -4.19155687e-01 -7.40999937e-01
-5.16867861e-02 -2.03211591e-01 -1.48358250e+00 -8.51161063e-01
8.26077104e-01 4.85355169e-01 -3.91188592e-01 1.15420723e+00
6.71556354e-01 2.38267988e-01 2.99781829e-01 -9.51491654e-01
-1.79881394e-01 1.32894325e+00 -5.70935905e-01 4.08930004e-01
8.18521142e-01 -2.91301399e-01 1.37494171e+00 -8.00523043e-01
1.10733533e+00 1.52931762e+00 5.57138503e-01 6.99725926e-01
-1.24863732e+00 -1.10691047e+00 -3.44885997e-02 5.63153327e-01
-1.36813819e+00 -7.78540075e-01 2.33415380e-01 3.95385809e-02
1.54123998e+00 3.29814136e-01 1.31629318e-01 1.31905222e+00
-1.13871686e-01 7.88605630e-01 7.63337970e-01 -7.01336920e-01
-3.05169642e-01 -4.02625471e-01 3.13333392e-01 2.95222253e-01
1.48189729e-02 2.67710030e-01 -9.31797922e-01 -1.20494448e-01
8.84230286e-02 -7.02594876e-01 -4.43436444e-01 2.35145867e-01
-1.35294759e+00 7.57276893e-01 3.27641606e-01 5.82611084e-01
-2.92121410e-01 -4.08477485e-01 4.71451610e-01 5.36614954e-01
1.11234270e-01 5.86580038e-01 -6.81650937e-01 -2.84112960e-01
-1.07139516e+00 3.96788806e-01 1.19804084e+00 1.08871400e+00
3.02792639e-01 -3.89331609e-01 -2.30013371e-01 1.08586526e+00
-5.44781610e-03 6.21607125e-01 3.04075629e-01 -1.34940970e+00
8.07013035e-01 2.38104954e-01 1.43827602e-01 -5.00259221e-01
-5.46945512e-01 -4.29304481e-01 -5.00827730e-01 -3.80637884e-01
6.58218920e-01 -7.24024698e-02 -4.38867569e-01 2.19103432e+00
6.97980002e-02 1.55669972e-01 7.18593359e-01 7.35770702e-01
1.18446970e+00 5.72605789e-01 1.25992283e-01 -6.93338692e-01
1.52141738e+00 -1.12043953e+00 -8.23877037e-01 -3.36067498e-01
8.64385486e-01 -9.71517265e-01 6.41762614e-01 4.97632831e-01
-1.15506196e+00 -2.42670670e-01 -8.02190483e-01 3.75062227e-02
-4.94263545e-02 -2.61504382e-01 4.10847872e-01 1.75012276e-01
-8.42139840e-01 6.26769900e-01 -5.00281572e-01 -9.30728674e-01
-1.99897274e-01 4.08345371e-01 -7.17313766e-01 -1.96731418e-01
-1.83614230e+00 1.31985092e+00 7.65727401e-01 -5.23582339e-01
-5.59753060e-01 -8.92905414e-01 -8.01145077e-01 -3.01209316e-02
6.71683431e-01 -2.26493016e-01 1.40569973e+00 -4.25297141e-01
-1.12586224e+00 1.23514736e+00 -2.37626702e-01 -6.30338907e-01
1.50066048e-01 -5.67832410e-01 -9.88242328e-01 1.42968908e-01
3.53390217e-01 8.72968018e-01 -2.95778245e-01 -1.05695856e+00
-8.71101499e-01 -5.53581454e-02 -9.60623287e-03 3.44167173e-01
2.68872112e-01 7.02935934e-01 -5.83253741e-01 -2.92387068e-01
-1.14788540e-01 -9.68271732e-01 1.50931686e-01 -7.59537697e-01
-5.05616724e-01 -5.69495440e-01 4.34034258e-01 -5.99246144e-01
1.24921227e+00 -1.74232721e+00 1.02070563e-01 -2.05606103e-01
-3.10215563e-01 6.05544746e-01 -5.99476993e-01 6.80514812e-01
-4.19201463e-01 4.63098101e-02 -1.09539114e-01 -1.87615946e-01
-1.20803908e-01 1.58871233e-01 -1.10324673e-01 2.58403003e-01
5.29203378e-02 6.81699812e-01 -1.19828975e+00 -7.54658461e-01
-1.33552223e-01 3.15853953e-01 -6.98125124e-01 3.33214663e-02
-8.17028582e-02 3.33436072e-01 -8.20993856e-02 3.70039701e-01
3.35228592e-01 3.66482854e-01 6.66108012e-01 -3.03039789e-01
-5.79461694e-01 7.32556283e-01 -1.08907211e+00 2.17345572e+00
-3.49496454e-01 5.19957900e-01 1.81142285e-01 -9.09331322e-01
1.03006935e+00 6.52118802e-01 3.54327649e-01 -8.50625992e-01
-1.05214883e-02 4.19812202e-01 4.13562834e-01 -3.03572536e-01
5.76639831e-01 -1.93001017e-01 -5.22544742e-01 2.45958090e-01
4.21661317e-01 1.94162488e-01 5.07133424e-01 5.59748054e-01
1.05673814e+00 5.64637110e-02 2.63962060e-01 -7.23812521e-01
6.81161940e-01 2.86271453e-01 9.61068273e-01 6.14754856e-01
-6.50386363e-02 4.57452029e-01 5.16529799e-01 3.06607541e-02
-6.19350195e-01 -7.82794833e-01 -1.75542757e-01 1.24938881e+00
-1.03092166e-02 -7.54514217e-01 -6.35406733e-01 -5.82254827e-01
-2.90003508e-01 1.25280559e+00 -4.21881407e-01 6.96389154e-02
-9.31750476e-01 -3.50593895e-01 1.14194763e+00 3.58598918e-01
2.85820216e-01 -1.26759636e+00 -2.20234305e-01 2.65340328e-01
-1.02890527e+00 -1.29610991e+00 -7.14235723e-01 1.69642955e-01
-3.74042183e-01 -1.41528642e+00 -3.14889312e-01 -8.82089913e-01
-2.27866352e-01 -1.94340006e-01 1.39335096e+00 -1.45767078e-01
-2.71556824e-02 -3.98473702e-02 -4.05027866e-01 -1.48362860e-01
-3.03483665e-01 2.38946006e-01 3.91567856e-01 -5.62916934e-01
8.45185161e-01 -2.50561744e-01 -7.23472610e-02 9.65994820e-02
-1.76590458e-01 -2.20719188e-01 2.92105883e-01 1.01824117e+00
4.03042048e-01 -5.18994093e-01 5.17482698e-01 -9.72432137e-01
3.47947538e-01 -4.03585017e-01 -2.81730056e-01 4.88952875e-01
-5.62470078e-01 1.51667222e-01 3.03652436e-01 -3.00814986e-01
-1.29593170e+00 -2.46465847e-01 -2.03953579e-01 -2.60834038e-01
-2.14030236e-01 4.67995793e-01 -2.82528520e-01 5.06922007e-01
6.09331727e-01 -1.88457340e-01 -4.40920472e-01 -7.77559638e-01
4.67954993e-01 7.82663226e-01 1.41842496e+00 -9.75744128e-01
1.84936866e-01 -2.33556032e-02 -2.54438162e-01 -6.61621451e-01
-1.16976929e+00 -6.57894373e-01 -9.71732259e-01 1.29828006e-01
7.65804172e-01 -9.94464815e-01 -9.34880674e-01 1.16979547e-01
-1.47721136e+00 -1.44536525e-01 -2.81599425e-02 7.49116361e-01
-3.84117365e-01 3.70301664e-01 -6.97845340e-01 -4.73199844e-01
-8.93891394e-01 -8.59508693e-01 6.35311127e-01 2.70638108e-01
-7.42285788e-01 -7.66248167e-01 6.01073086e-01 4.60050434e-01
1.57616083e-02 8.83162692e-02 6.89042687e-01 -1.21217978e+00
2.95719266e-01 1.99285254e-01 -1.07491530e-01 -3.78345639e-01
-2.50372916e-01 -1.37123674e-01 -7.10515618e-01 -3.92921418e-01
-5.54045260e-01 -4.92764145e-01 8.75143468e-01 1.02748938e-01
1.38385102e-01 -2.57835478e-01 -6.84708118e-01 3.54697406e-01
1.23995399e+00 1.99982688e-01 5.71723700e-01 6.42756641e-01
3.27864349e-01 7.61360645e-01 9.40848112e-01 1.18551113e-01
6.43943131e-01 1.08714604e+00 9.53835472e-02 3.28398317e-01
-2.62990773e-01 -4.49582517e-01 4.22534317e-01 9.82663751e-01
2.69185901e-02 -2.78253436e-01 -1.31860077e+00 1.00267756e+00
-2.04792690e+00 -1.37784898e+00 -5.05463719e-01 2.04567766e+00
1.43463230e+00 1.69926539e-01 1.26901582e-01 -1.95277765e-01
1.06022978e+00 -3.81872952e-02 -1.13866888e-01 -5.33854485e-01
-4.38231975e-01 5.32983124e-01 3.40260357e-01 1.01463127e+00
-1.15507066e+00 1.59704387e+00 6.18731260e+00 5.57688534e-01
-8.54082584e-01 7.44714215e-02 -1.91426858e-01 -4.36296105e-01
-7.46123195e-02 2.67709970e-01 -1.09762037e+00 2.32141450e-01
1.45068860e+00 -4.33240563e-01 3.52415860e-01 1.48470730e-01
-1.46648645e-01 1.43865831e-02 -1.11292601e+00 8.79901648e-01
1.84209704e-01 -1.10145617e+00 -2.03086659e-01 -2.98509777e-01
6.32084012e-01 7.64866769e-01 -6.12053216e-01 5.58386981e-01
8.85729313e-01 -1.04305828e+00 5.38051844e-01 5.59502602e-01
1.01436865e+00 -9.82182741e-01 8.91236544e-01 4.38009888e-01
-1.13640177e+00 2.88856626e-01 -2.06282914e-01 2.27294549e-01
5.64057648e-01 1.22261897e-01 -3.55431050e-01 1.06606209e+00
1.02958632e+00 7.08826780e-01 1.11605590e-02 9.56369817e-01
-5.85785270e-01 5.77557802e-01 -5.02658784e-01 2.14210853e-01
2.34597191e-01 4.03906971e-01 8.73332739e-01 1.73863101e+00
1.17107294e-01 5.56697488e-01 2.63083577e-01 3.11802953e-01
-2.98207432e-01 1.54137686e-01 -2.49603301e-01 3.13420832e-01
1.25844550e+00 1.11243308e+00 3.48702120e-03 -2.21383259e-01
-3.87165129e-01 6.74370944e-01 5.67209125e-01 -1.12579139e-02
-5.28143406e-01 -8.89693856e-01 6.59575820e-01 -4.89855081e-01
3.22472453e-01 1.70157954e-01 3.77137572e-01 -1.14660966e+00
-7.12050557e-01 -1.16019619e+00 9.80850875e-01 -4.79439199e-01
-1.22016156e+00 6.48689210e-01 3.33108783e-01 -8.29138815e-01
-5.59893429e-01 -2.25109950e-01 -8.03259194e-01 8.88168514e-01
-1.38308239e+00 -1.11917651e+00 2.14851692e-01 6.38979435e-01
4.29004610e-01 -3.02189022e-01 1.20689762e+00 3.96567076e-01
-4.88592237e-01 8.34340215e-01 -1.42908767e-01 4.92358267e-01
1.52823913e+00 -1.02771294e+00 4.12836224e-01 7.96796024e-01
3.35860670e-01 7.31792986e-01 8.73859644e-01 -6.69847488e-01
-1.21280003e+00 -7.73941815e-01 1.84641993e+00 -4.86279160e-01
7.87526369e-01 -1.50099481e-02 -9.76523042e-01 8.84054065e-01
8.58342290e-01 -3.27527463e-01 6.33634567e-01 7.03015625e-01
-6.91517711e-01 1.27758458e-01 -1.20367634e+00 3.62773359e-01
1.35422313e+00 -4.42694396e-01 -1.41415274e+00 -9.26302075e-02
7.63207257e-01 -5.05080998e-01 -1.35800445e+00 6.46991313e-01
4.88252342e-01 -7.01899767e-01 7.94662356e-01 -9.93303180e-01
1.41904250e-01 -2.04667891e-03 -5.46296775e-01 -1.44768047e+00
-5.16685188e-01 -7.95999706e-01 9.83250588e-02 1.68149960e+00
6.92950308e-01 -2.55429506e-01 2.55991787e-01 1.74888626e-01
-2.44721130e-01 -1.64588228e-01 -1.17745018e+00 -8.08081627e-01
5.81694007e-01 -3.04879010e-01 3.85301679e-01 1.49928260e+00
8.09925973e-01 9.46306109e-01 -2.61378914e-01 5.71211614e-03
1.00388801e+00 2.03610405e-01 4.41853017e-01 -1.50953746e+00
4.34026727e-03 -4.98026699e-01 1.97358325e-01 -6.50026739e-01
8.00406814e-01 -1.19866574e+00 9.45724174e-02 -1.29730022e+00
4.22235459e-01 -2.69102305e-01 -4.26381260e-01 9.08826709e-01
4.21069637e-02 1.57757044e-01 4.17024583e-01 2.81351030e-01
-8.53766263e-01 3.11755687e-01 7.83060253e-01 -1.80391371e-01
7.38702789e-02 -5.91924548e-01 -8.13650727e-01 4.84441727e-01
6.05595291e-01 -5.51426351e-01 1.44230023e-01 -3.99024487e-01
-1.58859938e-01 3.83717328e-01 -2.79699117e-01 -5.79639196e-01
7.49768317e-01 -2.37444595e-01 -2.04806492e-01 -8.06926966e-01
1.85270622e-01 -1.90448433e-01 9.05110091e-02 6.19492948e-01
-5.16256392e-01 3.59861962e-02 5.56919873e-01 8.46972615e-02
-3.32364023e-01 -2.29656860e-01 6.27732515e-01 -7.57022947e-02
-1.01735675e+00 -7.72736445e-02 -8.82946923e-02 8.86083603e-01
6.27901435e-01 4.54334527e-01 -7.64383256e-01 -1.41342938e-01
-7.47606933e-01 8.03495824e-01 1.66891068e-01 9.58872139e-01
8.49196464e-02 -1.25002694e+00 -1.42484295e+00 -1.90047786e-01
2.39383027e-01 -4.44237083e-01 2.83309035e-02 9.78043020e-01
5.37431613e-02 9.23789382e-01 -2.54742384e-01 -3.74372005e-01
-1.59695423e+00 2.81240523e-01 2.62825966e-01 -4.57686186e-01
-4.95549381e-01 7.85776675e-01 -3.18943650e-01 -8.89985979e-01
3.28959107e-01 4.34127927e-01 -5.31621218e-01 3.87527794e-01
6.60736501e-01 3.90234798e-01 -1.61366910e-02 -1.00886977e+00
-1.05543292e+00 4.99796838e-01 -4.09321308e-01 -3.41179520e-01
1.42510629e+00 6.00876026e-02 -1.09180033e-01 2.68294871e-01
9.46022868e-01 1.91993296e-01 -5.65846682e-01 -5.90373993e-01
5.86116016e-01 8.87051132e-03 -1.86830461e-02 -1.29831064e+00
-8.83134663e-01 6.25740469e-01 1.25175372e-01 -3.49502981e-01
5.22527456e-01 4.60079223e-01 5.53772092e-01 4.45510805e-01
5.55679739e-01 -1.02298427e+00 -4.84387636e-01 1.18339407e+00
9.02990460e-01 -9.90615785e-01 -2.03156546e-01 -2.86622047e-01
-7.86464632e-01 7.85689712e-01 6.88332438e-01 7.75434077e-02
2.70623773e-01 5.58599949e-01 1.15343980e-01 -1.23364709e-01
-1.27112091e+00 -3.17666888e-01 2.84323692e-01 4.13366795e-01
9.85938013e-01 5.32910712e-02 -6.65144205e-01 1.10782576e+00
-4.24312919e-01 -2.22425833e-01 3.06394428e-01 4.83529121e-01
-4.38012034e-01 -1.52421176e+00 -2.45449021e-01 3.86111811e-02
-8.72038543e-01 -3.79478455e-01 -5.85474730e-01 9.82653379e-01
-8.57551247e-02 1.23694098e+00 2.42920727e-01 -4.01502579e-01
6.99065685e-01 2.77392030e-01 6.83446169e-01 -5.71556151e-01
-8.25098634e-01 -5.15610985e-02 8.81162047e-01 -7.16434002e-01
-7.39728987e-01 -1.11512923e+00 -1.58136415e+00 -6.34732366e-01
-5.07141948e-01 6.77559793e-01 7.42737055e-02 1.05182183e+00
2.38438323e-01 1.80793032e-01 1.20978929e-01 -4.58531618e-01
-2.93503672e-01 -1.27541232e+00 -2.83466458e-01 3.83751571e-01
-5.52139655e-02 -7.12227821e-01 -1.49517193e-01 -3.76838475e-01] | [9.245495796203613, 9.589848518371582] |
8842ecd6-ea26-4047-9af5-acf9e9e59ca7 | learning-to-search-for-and-detect-objects-in | 2304.05741 | null | https://arxiv.org/abs/2304.05741v1 | https://arxiv.org/pdf/2304.05741v1.pdf | Learning to search for and detect objects in foveal images using deep learning | The human visual system processes images with varied degrees of resolution, with the fovea, a small portion of the retina, capturing the highest acuity region, which gradually declines toward the field of view's periphery. However, the majority of existing object localization methods rely on images acquired by image sensors with space-invariant resolution, ignoring biological attention mechanisms. As a region of interest pooling, this study employs a fixation prediction model that emulates human objective-guided attention of searching for a given class in an image. The foveated pictures at each fixation point are then classified to determine whether the target is present or absent in the scene. Throughout this two-stage pipeline method, we investigate the varying results obtained by utilizing high-level or panoptic features and provide a ground-truth label function for fixation sequences that is smoother, considering in a better way the spatial structure of the problem. Finally, we present a novel dual task model capable of performing fixation prediction and detection simultaneously, allowing knowledge transfer between the two tasks. We conclude that, due to the complementary nature of both tasks, the training process benefited from the sharing of knowledge, resulting in an improvement in performance when compared to the previous approach's baseline scores. | ['Plinio Moreno', 'Beatriz Paula'] | 2023-04-12 | null | null | null | null | ['object-localization'] | ['computer-vision'] | [ 3.73206526e-01 -6.01031482e-02 3.74519788e-02 1.24844117e-02
-1.78526789e-01 -6.29421949e-01 4.24146503e-01 3.11552912e-01
-8.34238946e-01 5.86185992e-01 -7.96447173e-02 1.58384088e-02
-8.34290087e-02 -6.78320289e-01 -6.86045527e-01 -8.48082125e-01
1.47789970e-01 -8.82352144e-02 6.62966728e-01 1.65365726e-01
6.66268885e-01 5.88792384e-01 -2.14683652e+00 5.04267633e-01
4.78127658e-01 1.23121917e+00 9.71580565e-01 5.96316874e-01
2.85416543e-01 5.89259923e-01 -3.86191458e-01 -2.65693724e-01
2.38141432e-01 -3.90649319e-01 -7.31792033e-01 1.16706096e-01
7.04751074e-01 -1.66802973e-01 -9.01059732e-02 1.12376654e+00
3.85404617e-01 2.00482726e-01 5.46347260e-01 -8.32398295e-01
-5.50085604e-01 -1.04599362e-02 -4.75132406e-01 6.91789806e-01
5.10191977e-01 4.22441274e-01 9.70497012e-01 -1.03678906e+00
6.18509769e-01 9.29536998e-01 1.11287005e-01 3.48216623e-01
-1.41682494e+00 -1.41801775e-01 2.67754823e-01 3.58763635e-01
-1.47651243e+00 -4.91513312e-01 4.81628656e-01 -6.96494520e-01
1.09767306e+00 1.00774458e-03 7.96484888e-01 8.70262504e-01
3.42532605e-01 5.10975182e-01 1.46761549e+00 -6.74107075e-01
2.29492649e-01 4.24901277e-01 -4.46931869e-02 5.59949934e-01
3.08752030e-01 2.65294671e-01 -7.16110349e-01 3.39918584e-01
8.82381797e-01 1.18447445e-01 -6.84154689e-01 -5.92980862e-01
-1.36773765e+00 3.02299589e-01 8.25763702e-01 4.86506552e-01
-7.61959612e-01 -2.73018569e-01 -1.83457956e-01 -3.20215896e-02
1.59635738e-01 5.43213010e-01 -1.75280184e-01 2.24299356e-01
-1.12992513e+00 4.61941399e-02 3.26638967e-01 5.62111855e-01
9.60419238e-01 -3.91904473e-01 -6.86357319e-01 3.67912561e-01
4.13591802e-01 4.78017718e-01 5.05837917e-01 -8.84990931e-01
6.37105182e-02 6.56238496e-01 4.77419049e-01 -9.36989069e-01
-4.44586068e-01 -7.79590845e-01 -2.28538662e-01 7.78447926e-01
7.34952450e-01 1.42534107e-01 -6.95917845e-01 1.79869425e+00
1.36880562e-01 -3.48615274e-02 -2.05998182e-01 1.28721225e+00
2.34778345e-01 2.39237234e-01 7.20684156e-02 -2.73171395e-01
1.73702443e+00 -9.04476225e-01 -4.79749203e-01 -4.28197384e-01
1.22033447e-01 -6.50846124e-01 1.06693840e+00 4.07364070e-01
-1.28239071e+00 -1.04553556e+00 -9.45656896e-01 -1.21368110e-01
-5.09980142e-01 3.22673351e-01 4.12053078e-01 4.57305610e-01
-1.52932727e+00 3.75839055e-01 -4.84500200e-01 -5.72867215e-01
4.91305053e-01 1.55969277e-01 -3.53977859e-01 -5.10386862e-02
-9.14721429e-01 1.23049080e+00 3.61612499e-01 6.66684583e-02
-8.02149355e-01 -5.26427746e-01 -5.19624233e-01 4.98714238e-01
2.37506583e-01 -9.11630988e-01 9.83918607e-01 -1.21462274e+00
-1.14752269e+00 1.22113502e+00 -6.35362566e-01 -4.52594668e-01
3.05802882e-01 -9.31332558e-02 -6.87892139e-02 4.86126870e-01
6.80821836e-02 1.06884086e+00 1.04064143e+00 -1.15596092e+00
-9.57282662e-01 -5.87699890e-01 1.67961746e-01 1.79927438e-01
3.68550457e-02 -3.29510085e-02 -4.06696379e-01 -1.61029682e-01
-1.02565311e-01 -6.61736131e-01 3.98915373e-02 1.62143409e-01
1.01909734e-01 -2.53400624e-01 3.90021861e-01 -4.72272098e-01
1.20231164e+00 -2.35882711e+00 1.13113418e-01 -3.13016213e-02
2.64813006e-01 4.19003546e-01 -1.48335591e-01 2.26565272e-01
-1.93386540e-01 -2.08591402e-01 -4.07120213e-02 -3.69134218e-01
-3.48621756e-01 -2.29086071e-01 -2.30666429e-01 4.68153298e-01
3.88086975e-01 8.75462532e-01 -9.64919209e-01 -3.25320154e-01
3.41370881e-01 3.66733640e-01 -3.47057164e-01 2.03513324e-01
-9.42941755e-02 5.42368293e-01 3.45172249e-02 5.18136919e-01
6.21404290e-01 -5.30402780e-01 -1.59599051e-01 -1.56632185e-01
-5.73245943e-01 5.75908199e-02 -8.61855268e-01 1.97759044e+00
-2.00302348e-01 9.88227487e-01 6.95322378e-05 -6.38833165e-01
6.95690989e-01 1.21753044e-01 1.20458633e-01 -1.02293825e+00
3.51631761e-01 1.26069993e-01 3.97551060e-01 -4.63484526e-01
4.49307203e-01 1.99798524e-01 6.01358771e-01 2.88878858e-01
1.30845755e-01 3.72601688e-01 3.63927871e-01 -1.68980241e-01
8.72261167e-01 3.19704533e-01 4.95569378e-01 -1.93051010e-01
6.35467052e-01 -1.19641624e-01 9.28661004e-02 8.87363493e-01
-5.77357173e-01 8.95669639e-01 2.94660538e-01 -1.51020586e-01
-7.83577561e-01 -1.10983968e+00 -2.82037944e-01 9.50705409e-01
1.83751911e-01 9.59655419e-02 -9.71849024e-01 -2.38514051e-01
-1.17914870e-01 5.28776824e-01 -1.00300682e+00 -1.34419486e-01
-1.36906654e-01 -1.71689078e-01 -1.26233265e-01 1.14275761e-01
5.11259615e-01 -1.44068730e+00 -1.61227429e+00 -1.84125274e-01
-1.96237698e-01 -8.63883793e-01 -1.49949372e-01 1.26566783e-01
-6.91742897e-01 -1.23382390e+00 -9.35032606e-01 -7.06730962e-01
7.14686871e-01 7.26359487e-01 1.05936849e+00 -6.99480399e-02
-5.54161668e-01 6.02058291e-01 -1.31921083e-01 -3.10620189e-01
2.51241148e-01 -1.38098851e-01 -1.94464386e-01 5.35906494e-01
6.28364682e-01 -2.87081838e-01 -1.05160475e+00 7.52238482e-02
-6.77020252e-01 2.19522975e-02 9.23133492e-01 4.99378830e-01
5.54443717e-01 -2.48614863e-01 4.53116372e-02 -3.11246425e-01
3.32174718e-01 -4.05654192e-01 -8.00486624e-01 2.13691771e-01
-4.34497714e-01 -1.04338676e-01 2.11369336e-01 -3.07984114e-01
-1.03348219e+00 4.37193550e-02 4.65807498e-01 -6.10324562e-01
-5.60120881e-01 2.19988525e-01 7.38798529e-02 -2.08788127e-01
9.41002011e-01 3.18294913e-01 1.54011518e-01 -1.03918009e-01
2.46077646e-02 4.40302640e-01 4.82149869e-01 -9.97772999e-03
2.97779202e-01 6.08343303e-01 7.63287991e-02 -7.15635955e-01
-1.09673071e+00 -7.24131942e-01 -9.40819025e-01 -5.25805831e-01
1.07149208e+00 -8.18078160e-01 -1.09537256e+00 3.40952694e-01
-1.31488383e+00 -1.96495429e-01 -3.50532919e-01 8.22075129e-01
-5.63689947e-01 8.70151073e-02 -1.09306261e-01 -9.33409572e-01
1.53899565e-01 -9.58096087e-01 1.04237080e+00 4.27675843e-01
1.45911484e-03 -7.99255788e-01 1.27747953e-01 1.98406845e-01
4.43513036e-01 -1.16756625e-01 7.39094257e-01 -3.41946870e-01
-9.18823719e-01 -2.51395665e-02 -5.87164402e-01 2.89804250e-01
-6.02064952e-02 -1.46912649e-01 -1.43430352e+00 -2.55606085e-01
1.95819542e-01 -4.37130705e-02 1.05809736e+00 8.57501686e-01
8.90072167e-01 3.13167453e-01 -4.86123443e-01 3.67721826e-01
1.51768720e+00 1.34394750e-01 8.86737525e-01 4.24874514e-01
1.20285049e-01 1.02581775e+00 7.12077677e-01 2.33874977e-01
6.28080666e-02 7.31557965e-01 6.75792694e-01 -2.64256001e-01
-4.80870545e-01 1.06212996e-01 1.65516630e-01 -7.31093585e-02
-2.54712820e-01 -2.21596584e-01 -7.93962538e-01 7.11550653e-01
-1.72240508e+00 -1.07187462e+00 -5.85700572e-02 2.66516685e+00
5.41486919e-01 1.31512895e-01 1.74876265e-02 -6.29510032e-04
7.60527611e-01 -8.28971863e-02 -6.28756404e-01 -4.83680367e-02
-2.40446210e-01 2.08321258e-01 4.19101328e-01 3.70716214e-01
-9.20495450e-01 7.44778395e-01 6.86652708e+00 3.60883176e-01
-1.44085872e+00 3.06191836e-02 3.93036723e-01 -3.55055213e-01
2.42004707e-01 -9.44542363e-02 -8.52604866e-01 5.42920828e-01
9.33395982e-01 1.51958838e-01 5.50896585e-01 4.04672652e-01
2.82448262e-01 -6.91767812e-01 -1.19476521e+00 1.00528312e+00
3.27041447e-01 -9.30445731e-01 -1.49010941e-01 2.03952029e-01
4.37731683e-01 2.66721435e-02 4.53704834e-01 -1.22158758e-01
-3.52158606e-01 -9.74407494e-01 8.00144911e-01 1.11029029e+00
4.97950017e-01 -2.78015643e-01 4.50435758e-01 3.70519012e-01
-7.41088450e-01 -3.54665995e-01 -3.75089049e-01 -4.09869581e-01
-1.63086459e-01 4.26352739e-01 -6.94815576e-01 2.42810801e-01
7.62687922e-01 5.91033697e-01 -9.92050529e-01 1.56544709e+00
-2.68178523e-01 1.56555235e-01 1.28019229e-01 7.31073096e-02
3.16414461e-02 -1.32921755e-01 5.31616986e-01 1.06175888e+00
4.75800067e-01 -9.19701681e-02 -1.40783280e-01 1.28440976e+00
3.26239705e-01 -6.25809580e-02 -5.96334577e-01 3.57764363e-01
4.52881753e-01 1.29632223e+00 -7.50162303e-01 -1.00909844e-01
-5.89754164e-01 9.36529815e-01 4.81944293e-01 7.53424823e-01
-5.69090784e-01 -3.70684832e-01 3.82191092e-01 3.92504424e-01
6.20386243e-01 -7.46590942e-02 -1.05717801e-01 -9.08195019e-01
1.27621042e-02 -4.11815017e-01 2.45849639e-02 -1.42064667e+00
-7.85644531e-01 6.77765071e-01 -2.30338484e-01 -1.16096532e+00
-1.49733812e-01 -1.07002473e+00 -1.82412907e-01 1.44355774e+00
-1.77145064e+00 -9.26970184e-01 -5.49848497e-01 6.51440382e-01
2.57447392e-01 1.73974894e-02 7.32038736e-01 1.46594876e-02
-2.77254492e-01 2.37136066e-01 -1.80282220e-01 -2.03020275e-01
8.15626442e-01 -1.12666798e+00 -1.66306943e-01 1.14108074e+00
1.20006308e-01 8.34870875e-01 4.70255047e-01 -3.13077033e-01
-9.05954957e-01 -6.37636602e-01 9.35812175e-01 -6.26764774e-01
3.99529129e-01 -3.29384774e-01 -1.02694440e+00 3.60437036e-01
3.63210201e-01 1.31224021e-01 4.83854145e-01 -9.42888483e-03
-2.46551603e-01 -1.85624331e-01 -9.62886989e-01 4.71147478e-01
8.62015247e-01 -6.00739300e-01 -6.93872750e-01 4.68126079e-03
3.48299801e-01 -8.67710710e-02 -4.10648614e-01 1.70400161e-02
5.43922246e-01 -1.49509192e+00 8.76849592e-01 -2.93344319e-01
3.64918977e-01 -5.18885195e-01 3.90131846e-02 -1.03134644e+00
-8.04728925e-01 -6.59371316e-02 -1.73820164e-02 7.87713349e-01
2.82184422e-01 -5.81712186e-01 4.75262880e-01 1.81406021e-01
6.30974993e-02 -3.54782939e-01 -7.88650036e-01 -3.69215995e-01
-4.59084958e-01 -9.50276777e-02 6.87738955e-02 4.33600992e-01
-2.03104675e-01 3.31846148e-01 4.06550989e-02 2.26527616e-01
5.22557139e-01 2.08233342e-01 4.43616837e-01 -1.39634180e+00
-2.60914952e-01 -7.42627859e-01 -5.79723299e-01 -1.08964109e+00
-5.55151887e-02 -7.27957010e-01 1.12250589e-01 -1.34758925e+00
2.41276503e-01 1.46664053e-01 -7.04287052e-01 2.02102214e-01
-1.94963858e-01 3.36194575e-01 2.24406809e-01 4.59959388e-01
-6.53305411e-01 -6.19464852e-02 1.41345894e+00 2.56821781e-01
-1.90709569e-02 -1.92475859e-02 -6.45760119e-01 4.41080809e-01
4.02753145e-01 -3.21210138e-02 -4.09701824e-01 -5.16729474e-01
1.91089749e-01 -2.85612326e-02 7.54791379e-01 -1.26122177e+00
5.33911467e-01 6.57887850e-03 6.77248359e-01 -3.95871639e-01
3.38030577e-01 -9.93350387e-01 -1.68106139e-01 5.84939718e-01
-3.45580667e-01 -1.52186334e-01 2.83644766e-01 6.33763731e-01
-3.34436953e-01 -4.64219272e-01 1.01005220e+00 -3.01470220e-01
-8.10100138e-01 2.35197730e-02 -4.99185979e-01 -2.61611253e-01
1.09737599e+00 -5.16736269e-01 -4.48633552e-01 1.06998138e-01
-9.01136100e-01 -1.99528992e-01 6.73379838e-01 2.36668557e-01
5.47911584e-01 -8.49214911e-01 -5.04225791e-01 3.76605183e-01
9.82795954e-02 -3.02071184e-01 4.03904557e-01 1.19838655e+00
-1.95218131e-01 8.43076587e-01 -6.09818697e-01 -9.09571171e-01
-9.92746115e-01 1.02249408e+00 5.03087401e-01 5.12961484e-02
-1.30451709e-01 8.10045600e-01 6.05761945e-01 4.01263356e-01
2.60294974e-01 -3.33910197e-01 -7.02369452e-01 2.71863878e-01
8.22922528e-01 1.83314055e-01 6.17175885e-02 -6.18499637e-01
-3.13607454e-01 6.23096883e-01 9.17466283e-02 -8.32332298e-02
7.89004445e-01 -5.00752568e-01 -2.52100140e-01 7.57529736e-01
7.49885678e-01 1.08044734e-02 -1.60025203e+00 -3.21347147e-01
-2.62316525e-01 -7.86694646e-01 7.47009292e-02 -1.03138304e+00
-7.13573158e-01 1.30028510e+00 1.01893389e+00 2.77745217e-01
1.36992311e+00 -7.31282160e-02 -7.46275783e-02 9.98789966e-02
4.15108532e-01 -6.80566192e-01 1.79887146e-01 3.09919387e-01
7.22766697e-01 -1.38205826e+00 -3.03554952e-01 -1.67233825e-01
-4.94419277e-01 9.28959727e-01 5.40421724e-01 -1.19869433e-01
2.89728820e-01 -3.73969346e-01 -1.52562946e-01 -2.19487920e-01
-7.17467904e-01 -7.86812961e-01 5.38554311e-01 8.43191445e-01
3.50685775e-01 -2.67663985e-01 2.65746266e-02 2.40263507e-01
1.10627003e-01 -3.62011744e-03 2.47112617e-01 7.41046488e-01
-9.25964832e-01 -5.60496092e-01 -3.82016093e-01 2.00030461e-01
-2.89240330e-01 -3.02142948e-01 -1.56655729e-01 5.80830336e-01
3.86928171e-01 9.32365537e-01 4.98079896e-01 1.59950942e-01
3.53499532e-01 1.41076609e-01 8.42082679e-01 -7.57074714e-01
-4.68077242e-01 6.84940740e-02 -5.93187571e-01 -6.15542829e-01
-6.03044093e-01 -8.10851038e-01 -8.81345987e-01 2.28950575e-01
-1.98125064e-01 -4.44294289e-02 6.14274621e-01 7.73750722e-01
4.69646007e-01 6.29136026e-01 3.54438037e-01 -1.08516383e+00
-1.74311414e-01 -9.04870689e-01 -6.84946418e-01 4.37634706e-01
5.37974119e-01 -9.69881535e-01 -3.17668378e-01 2.20543340e-01] | [10.084147453308105, 1.6663780212402344] |
7b407a3c-2dca-473d-8d2e-521df2641d76 | summary-oriented-vision-modeling-for | 2212.07672 | null | https://arxiv.org/abs/2212.07672v2 | https://arxiv.org/pdf/2212.07672v2.pdf | Summary-Oriented Vision Modeling for Multimodal Abstractive Summarization | Multimodal abstractive summarization (MAS) aims to produce a concise summary given the multimodal data (text and vision). Existing studies mainly focus on how to effectively use the visual features from the perspective of an article, having achieved impressive success on the high-resource English dataset. However, less attention has been paid to the visual features from the perspective of the summary, which may limit the model performance, especially in the low- and zero-resource scenarios. In this paper, we propose to improve the summary quality through summary-oriented visual features. To this end, we devise two auxiliary tasks including vision to summary task and masked image modeling task. Together with the main summarization task, we optimize the MAS model via the training objectives of all these tasks. By these means, the MAS model can be enhanced by capturing the summary-oriented visual features, thereby yielding more accurate summaries. Experiments on 44 languages, covering mid-high-, low-, and zero-resource scenarios, verify the effectiveness and superiority of the proposed approach, which achieves state-of-the-art performance under all scenarios. Additionally, we will contribute a large-scale multilingual multimodal abstractive summarization (MM-Sum) dataset. | ['Jie zhou', 'Yufeng Chen', 'Jiaan Wang', 'Jinan Xu', 'Fandong Meng', 'Yunlong Liang'] | 2022-12-15 | null | null | null | null | ['abstractive-text-summarization'] | ['natural-language-processing'] | [ 2.02923536e-01 -3.27905059e-01 -2.02166542e-01 -2.39917547e-01
-1.16055477e+00 -2.94683814e-01 8.27916086e-01 2.61717290e-01
-5.07498026e-01 7.21540749e-01 6.86246037e-01 8.86990223e-03
2.09463820e-01 -3.85863215e-01 -6.24818027e-01 -5.32080710e-01
4.29849833e-01 -7.62043670e-02 -1.21908098e-01 -1.61579587e-02
4.96539921e-01 1.67132318e-01 -1.58042991e+00 4.98698324e-01
1.36496067e+00 7.88591623e-01 5.34853160e-01 6.32529855e-01
-2.28716567e-01 6.79333925e-01 -7.44376123e-01 -7.14703083e-01
-3.02657127e-01 -4.90549237e-01 -4.47104901e-01 3.58584881e-01
6.83507025e-01 -4.74386007e-01 -4.04838145e-01 1.10333121e+00
7.82594979e-01 2.03730151e-01 7.65237749e-01 -1.16859853e+00
-8.52640390e-01 4.23645020e-01 -8.74090791e-01 5.48713561e-03
4.95024234e-01 2.23413855e-01 1.17478061e+00 -1.23144710e+00
5.01507103e-01 1.44184828e+00 1.63414866e-01 2.91901559e-01
-8.58635366e-01 -4.06983048e-01 3.78904104e-01 3.27787429e-01
-1.29899061e+00 -6.88612342e-01 8.41502130e-01 -1.52081341e-01
7.64620900e-01 4.39064175e-01 4.86999273e-01 9.36233282e-01
-7.99232256e-03 1.28771687e+00 8.95824373e-01 -3.74482125e-01
-2.22874433e-01 2.20390260e-01 4.57797805e-03 7.44023740e-01
3.21569681e-01 -6.33929074e-01 -7.20742285e-01 1.21708699e-01
1.85160786e-01 5.08963466e-02 -5.64719439e-01 -4.13179174e-02
-1.50424576e+00 6.29613996e-01 3.08986574e-01 1.00869417e-01
-6.04631662e-01 -2.24990532e-01 5.26907265e-01 -6.90836161e-02
5.39260447e-01 9.50050876e-02 2.56360918e-02 -6.44196523e-03
-1.11820722e+00 2.09226474e-01 4.15897280e-01 1.18727970e+00
7.56373227e-01 1.65260896e-01 -8.35966170e-01 1.10976911e+00
3.23350608e-01 8.12603354e-01 1.97043717e-01 -8.29090238e-01
1.00441670e+00 7.22748935e-01 1.01934984e-01 -1.22171688e+00
-3.03422660e-01 -5.55117607e-01 -9.92869020e-01 -3.91165137e-01
-1.16108462e-01 -6.56575859e-02 -6.91965818e-01 1.59679401e+00
1.37695894e-01 -2.36379385e-01 4.66749340e-01 9.80774164e-01
1.61313045e+00 1.04404306e+00 7.95201510e-02 -5.72833121e-01
1.53874290e+00 -1.28823376e+00 -9.52993989e-01 -2.98014224e-01
3.73825073e-01 -9.72781062e-01 1.35022044e+00 1.19470283e-01
-1.31912410e+00 -5.41877925e-01 -1.05800068e+00 -2.41872862e-01
-1.14684224e-01 7.83067822e-01 2.03805596e-01 1.41822770e-01
-8.34174275e-01 1.14758812e-01 -4.07931805e-01 -5.17784059e-01
4.17125374e-01 -3.26317549e-02 -4.26116943e-01 -3.84554595e-01
-9.92628634e-01 8.40091407e-01 4.74668235e-01 2.47202918e-01
-5.64539790e-01 -3.34419608e-01 -1.00604725e+00 1.33818537e-01
4.63065624e-01 -9.85361636e-01 1.14018559e+00 -7.38398433e-01
-1.13510537e+00 6.38043404e-01 -4.91235375e-01 -1.19419828e-01
5.58191836e-01 -1.44675404e-01 -3.68734479e-01 5.03313720e-01
2.91185588e-01 8.47303212e-01 7.34733641e-01 -1.66764784e+00
-8.87997389e-01 -4.19680685e-01 1.81780793e-02 8.25106978e-01
-6.97698474e-01 -3.02330647e-02 -1.10975707e+00 -8.20130646e-01
-2.74798065e-01 -7.01095641e-01 1.51717097e-01 -3.18704933e-01
-5.36798358e-01 -1.80887565e-01 6.06912196e-01 -1.20199680e+00
1.46075130e+00 -2.13719893e+00 4.50375080e-01 -3.51640254e-01
2.04023659e-01 3.02061617e-01 -3.13723445e-01 6.95676565e-01
3.95746022e-01 -1.87143274e-02 -4.40941304e-01 -7.57633090e-01
4.38758582e-02 -2.14260770e-04 -1.31213620e-01 2.70340264e-01
2.89291561e-01 1.02253962e+00 -7.04278529e-01 -9.86129165e-01
1.46580264e-01 2.61252165e-01 -2.48009145e-01 3.17623287e-01
-2.86675934e-02 4.11506802e-01 -5.03050089e-01 7.08035648e-01
6.88212931e-01 -1.12870924e-01 -1.47749320e-01 -5.66335678e-01
-1.90380901e-01 -2.63110101e-01 -8.02993417e-01 1.73979354e+00
-5.31505346e-01 6.42263651e-01 -9.66182351e-03 -8.38009417e-01
7.41300464e-01 5.30592725e-02 2.42984086e-01 -1.01210594e+00
-7.47031868e-02 9.53314155e-02 -2.92542070e-01 -7.04977393e-01
1.03846216e+00 8.05509910e-02 -2.19389275e-01 1.14997849e-01
-1.82726711e-01 -6.15217462e-02 4.51933324e-01 6.28474772e-01
4.07861590e-01 -2.36531403e-02 4.19599056e-01 2.47975007e-01
8.27762127e-01 3.20911594e-02 2.75512636e-01 6.96239889e-01
-9.01264176e-02 7.51936197e-01 6.22514904e-01 1.61964908e-01
-8.44654620e-01 -8.64696622e-01 2.62230784e-01 9.15387571e-01
4.25227851e-01 -6.28545880e-01 -6.91511214e-01 -7.09388435e-01
-2.54838973e-01 9.28523481e-01 -3.25429976e-01 -2.48741075e-01
-3.95227998e-01 -6.60330117e-01 3.95889729e-01 3.78547013e-01
7.79299676e-01 -1.06548226e+00 -4.42292631e-01 -2.93063149e-02
-6.94633067e-01 -1.34542751e+00 -8.61588001e-01 -5.74442327e-01
-6.10459805e-01 -7.09059954e-01 -1.34150183e+00 -7.63745189e-01
5.73949814e-01 6.60376489e-01 8.47170353e-01 -4.91560288e-02
-4.64915596e-02 5.85981190e-01 -3.40100080e-01 -4.27537709e-01
-2.43600339e-01 8.36216658e-02 -2.52654016e-01 2.49388009e-01
-4.09665145e-02 -6.72942540e-03 -6.27589762e-01 -8.82825628e-02
-1.01403058e+00 4.99216616e-01 1.01552367e+00 8.28412712e-01
6.27839267e-01 -1.71037033e-01 7.58557141e-01 -4.46806163e-01
9.77815390e-01 -2.60185093e-01 -5.01488149e-02 6.42414153e-01
-3.40860397e-01 -6.42974377e-02 6.28616393e-01 -2.73342252e-01
-1.29919577e+00 -2.00611994e-01 -1.21964868e-02 -2.91275650e-01
1.58410385e-01 9.58568811e-01 -5.14380813e-01 3.07547569e-01
7.84219503e-02 7.12393880e-01 -1.30540222e-01 -2.86046773e-01
5.14865518e-01 9.62439179e-01 6.80736780e-01 -3.85876775e-01
5.03008366e-01 3.36729884e-01 -4.79002334e-02 -1.23459423e+00
-9.09856379e-01 -4.33144480e-01 -2.88043261e-01 -3.08193535e-01
7.96825886e-01 -1.15111876e+00 -4.63343352e-01 4.85562831e-01
-1.27957177e+00 2.99307108e-01 1.29743591e-01 5.26553631e-01
-4.42815095e-01 1.01224768e+00 -1.38098180e-01 -8.87860239e-01
-7.47101188e-01 -1.31647468e+00 1.34881270e+00 4.30999398e-01
6.26166910e-02 -6.66859806e-01 -3.06978732e-01 5.48936427e-01
2.27592409e-01 8.78581256e-02 8.89300585e-01 -5.34289002e-01
-4.15672362e-01 -9.78045240e-02 -7.75358379e-01 4.20767635e-01
1.52579220e-02 1.93817720e-01 -7.37781823e-01 -2.68637925e-01
-2.25682393e-01 -2.45018601e-01 1.33891857e+00 4.35912371e-01
1.08043766e+00 -4.29088265e-01 -5.48731051e-02 3.28690320e-01
1.12402320e+00 -2.64958069e-02 5.45862734e-01 1.92146897e-01
9.14718390e-01 7.58920074e-01 1.12020326e+00 5.31671345e-01
9.17763412e-01 6.35569870e-01 4.24978435e-01 -2.24322587e-01
-3.70826960e-01 -2.11722121e-01 3.92634153e-01 1.13718545e+00
-5.08135818e-02 -5.05428255e-01 -6.74741507e-01 5.91550529e-01
-2.18279934e+00 -9.36620951e-01 -1.75486073e-01 1.96713293e+00
7.25410998e-01 -2.08048999e-01 7.72962496e-02 -2.14867786e-01
8.52042496e-01 5.68048239e-01 -3.96500885e-01 -2.75639415e-01
-4.44904238e-01 -5.77604711e-01 3.87934707e-02 2.18841150e-01
-1.05082023e+00 9.49316978e-01 4.93316269e+00 1.13604081e+00
-1.03776991e+00 -5.37241958e-02 5.08354604e-01 -3.68581973e-02
-4.70729142e-01 -2.15563148e-01 -6.91234529e-01 5.76025844e-01
5.04312992e-01 -3.16335440e-01 2.25142017e-01 5.02329528e-01
3.78836393e-01 -2.96991944e-01 -8.21589708e-01 1.37921739e+00
7.10489690e-01 -1.22833979e+00 6.61763132e-01 -1.64125934e-01
7.29682386e-01 -3.67655694e-01 2.64641434e-01 4.36459810e-01
-5.24613738e-01 -8.35076690e-01 9.28559721e-01 7.10768044e-01
1.01696289e+00 -1.01045489e+00 8.25890660e-01 4.41568881e-01
-1.26728761e+00 8.30062032e-02 -2.50020504e-01 3.58688712e-01
2.39466488e-01 2.76057512e-01 -4.74827468e-01 1.04306722e+00
4.15150583e-01 9.66607809e-01 -9.64785576e-01 1.05205524e+00
-1.83434397e-01 1.34979874e-01 2.64075398e-01 -2.00416297e-01
3.56517702e-01 -2.27300778e-01 6.73657894e-01 1.49663639e+00
4.57846642e-01 -2.88744625e-02 2.37940878e-01 5.18147886e-01
-3.14166933e-01 5.11437833e-01 -4.90164101e-01 -2.56072015e-01
3.12129110e-01 1.35077214e+00 -3.53984118e-01 -5.67913532e-01
-7.47056425e-01 1.02326930e+00 3.22432756e-01 5.31037629e-01
-7.89045632e-01 -5.23129463e-01 1.72741413e-01 -2.76535034e-01
2.67337710e-01 -1.16003081e-01 -2.62491256e-01 -1.53352964e+00
3.11077595e-01 -1.03688228e+00 4.08249170e-01 -1.03148115e+00
-1.06293964e+00 4.94683862e-01 1.77062556e-01 -1.26355672e+00
-3.66131240e-03 -2.07238138e-01 -6.15114033e-01 8.33314359e-01
-1.68422794e+00 -1.65017653e+00 -5.60644865e-01 5.22871077e-01
9.59252834e-01 -2.20209539e-01 3.65983695e-01 2.58805335e-01
-8.31999123e-01 5.24980366e-01 1.36168107e-01 -2.63117582e-01
1.09404457e+00 -9.80304599e-01 -2.25607753e-02 1.04530001e+00
2.90694255e-02 3.98368776e-01 7.89174318e-01 -7.03264713e-01
-1.54163492e+00 -1.20265365e+00 8.65955949e-01 8.76101013e-03
4.11007732e-01 9.77631565e-03 -7.58169532e-01 3.54346693e-01
5.64174354e-01 -4.27345544e-01 5.18465698e-01 -2.07143053e-01
-6.08653799e-02 -1.62330359e-01 -8.13935161e-01 1.07761574e+00
8.64540100e-01 -3.83686632e-01 -6.75989032e-01 1.05294161e-01
6.68660939e-01 -2.81807452e-01 -6.54163599e-01 4.78875667e-01
6.20097697e-01 -6.61992073e-01 8.94878983e-01 -3.80338162e-01
9.54659164e-01 -3.48532945e-01 -3.64988148e-01 -1.49862671e+00
-2.52667367e-02 -2.25524426e-01 -3.48431498e-01 1.66313434e+00
2.07420409e-01 -3.46696168e-01 1.54762223e-01 1.67844236e-01
-3.72603685e-01 -6.74131572e-01 -6.27247572e-01 -3.34314793e-01
-2.76332170e-01 -5.71742207e-02 5.49916923e-01 5.57620704e-01
-1.76129133e-01 7.41501510e-01 -7.30081141e-01 1.13262415e-01
5.20050645e-01 3.99797320e-01 1.01661110e+00 -7.24990726e-01
1.05666354e-01 -6.25999689e-01 -1.20673530e-01 -1.15467727e+00
3.61377299e-01 -7.11650312e-01 -3.09091937e-02 -2.22819567e+00
8.29913199e-01 3.23419154e-01 -6.55369237e-02 2.08370999e-01
-8.51724267e-01 1.92299068e-01 5.04191101e-01 3.27242643e-01
-1.12120748e+00 1.05382109e+00 1.55280638e+00 -4.46557075e-01
4.03139554e-03 -1.82585061e-01 -1.02284229e+00 6.29499733e-01
5.45086622e-01 -2.74660867e-02 -3.38286757e-01 -5.99281728e-01
5.94182126e-03 2.94709921e-01 3.32225651e-01 -5.77061176e-01
2.58968115e-01 -2.38692984e-01 4.02884632e-01 -1.20653307e+00
5.05929649e-01 -6.00206137e-01 -2.52269953e-01 1.60832152e-01
-3.80845875e-01 1.93430334e-01 8.02717134e-02 6.98401988e-01
-5.98379910e-01 -2.05614805e-01 5.67629039e-01 -8.50313827e-02
-8.51860881e-01 3.08220655e-01 -8.09191540e-02 7.70862550e-02
8.94495666e-01 -1.54246837e-01 -5.73791146e-01 -6.25655115e-01
-3.64539653e-01 5.91277957e-01 5.90166211e-01 4.33911622e-01
9.35192704e-01 -1.35113084e+00 -1.14846623e+00 -1.87240005e-01
6.16700113e-01 -8.29663351e-02 6.63743615e-01 1.13695955e+00
-4.20487553e-01 4.61289138e-01 -1.18969813e-01 -6.93274260e-01
-1.54314661e+00 6.18022859e-01 -1.90097466e-01 -1.60581172e-01
-4.40226555e-01 2.21121252e-01 4.88780916e-01 -4.92246598e-02
3.25335771e-01 2.41443589e-02 -6.67842746e-01 4.98144001e-01
6.54720247e-01 4.88514811e-01 -2.81344112e-02 -1.06455517e+00
-2.81700492e-01 6.03193223e-01 -9.61203352e-02 -1.24006718e-01
1.14707661e+00 -6.63774490e-01 -1.15710072e-01 3.78057897e-01
1.21491992e+00 2.11230680e-01 -1.02437830e+00 -2.88400680e-01
-2.91697055e-01 -4.26095515e-01 -9.96143278e-03 -6.24609411e-01
-9.14740384e-01 1.10291100e+00 1.82524219e-01 -2.00865306e-02
1.24477482e+00 -6.46990165e-02 8.05041790e-01 2.75252938e-01
-8.57665241e-02 -9.55754280e-01 2.97311276e-01 2.94683665e-01
1.25253892e+00 -1.48420393e+00 3.32131416e-01 -2.29807138e-01
-1.36261189e+00 9.39565778e-01 4.58980441e-01 2.76088059e-01
-1.60728335e-01 -3.41966361e-01 -2.33221695e-01 1.63396972e-03
-6.09218180e-01 -3.23601842e-01 8.36169899e-01 4.09849197e-01
3.51258010e-01 1.26621604e-01 -5.07699370e-01 8.17253709e-01
1.27184957e-01 -3.73341829e-01 4.88264114e-01 7.18760908e-01
-5.40845692e-01 -5.25310218e-01 -4.40634876e-01 5.34587622e-01
-4.39832360e-01 -2.22604349e-01 -5.19802928e-01 7.44739294e-01
-3.52119505e-01 1.17769873e+00 -2.45343998e-01 -9.15781632e-02
5.40464103e-01 -2.51471788e-01 4.20391738e-01 -5.03438771e-01
-3.06044102e-01 2.73025542e-01 3.93268853e-01 -2.06084400e-01
-6.31766379e-01 -6.04370236e-01 -1.00784731e+00 -3.56557399e-01
-1.92377567e-01 -4.63396013e-02 6.64609253e-01 8.93878043e-01
3.20627272e-01 6.56363726e-01 6.43633842e-01 -9.34778392e-01
-5.42183280e-01 -9.93987441e-01 -4.77857500e-01 3.15623760e-01
2.98186421e-01 -4.40637171e-01 -1.99107662e-01 5.76778166e-02] | [10.684871673583984, 0.7446709871292114] |
debc8b98-67a9-4ee7-ab40-b90ee18e7c7b | dynamic-inter-treatment-information-sharing | 2305.15984 | null | https://arxiv.org/abs/2305.15984v1 | https://arxiv.org/pdf/2305.15984v1.pdf | Dynamic Inter-treatment Information Sharing for Heterogeneous Treatment Effects Estimation | Existing heterogeneous treatment effects learners, also known as conditional average treatment effects (CATE) learners, lack a general mechanism for end-to-end inter-treatment information sharing, and data have to be split among potential outcome functions to train CATE learners which can lead to biased estimates with limited observational datasets. To address this issue, we propose a novel deep learning-based framework to train CATE learners that facilitates dynamic end-to-end information sharing among treatment groups. The framework is based on \textit{soft weight sharing} of \textit{hypernetworks}, which offers advantages such as parameter efficiency, faster training, and improved results. The proposed framework complements existing CATE learners and introduces a new class of uncertainty-aware CATE learners that we refer to as \textit{HyperCATE}. We develop HyperCATE versions of commonly used CATE learners and evaluate them on IHDP, ACIC-2016, and Twins benchmarks. Our experimental results show that the proposed framework improves the CATE estimation error via counterfactual inference, with increasing effectiveness for smaller datasets. | ['David A. Clifton', 'Ghadeer Ghosheh', 'Soheila Molaei', 'Jiandong Zhou', 'Vinod Kumar Chauhan'] | 2023-05-25 | null | null | null | null | ['counterfactual-inference'] | ['miscellaneous'] | [ 6.05369397e-02 2.53968209e-01 -1.05697227e+00 -6.47335052e-01
-8.25754285e-01 -2.01980144e-01 4.79906440e-01 -3.87678631e-02
-6.73057318e-01 1.16316676e+00 6.43099606e-01 -4.47782189e-01
-6.69764936e-01 -7.86641061e-01 -9.34199989e-01 -7.07484543e-01
-2.05711812e-01 5.27683318e-01 -1.32408708e-01 9.52342078e-02
-2.75622122e-02 -3.03299844e-01 -1.05904555e+00 3.57060790e-01
1.39101636e+00 8.64468575e-01 -2.45266914e-01 -9.82391015e-02
1.18538486e-02 6.25317574e-01 -1.72259793e-01 -7.60047972e-01
1.82907522e-01 -5.35108559e-02 -5.81645012e-01 -4.75881785e-01
3.64638120e-01 -5.94589055e-01 -4.24772531e-01 7.89211750e-01
9.74828362e-01 1.88923895e-01 9.35054779e-01 -1.68123686e+00
-6.52142882e-01 1.43508410e+00 -8.13950896e-01 -4.18347679e-02
-2.09310964e-01 3.09858114e-01 8.65428925e-01 -5.31884789e-01
4.75571156e-01 1.56055319e+00 8.82854939e-01 7.45626211e-01
-1.28396297e+00 -1.47237360e+00 3.28337103e-01 2.67955750e-01
-9.95989203e-01 -2.63734102e-01 4.71653342e-01 -2.45854437e-01
4.91700113e-01 -1.97228184e-03 4.05251235e-01 1.80333185e+00
4.13859874e-01 1.20549929e+00 1.18463004e+00 2.03230437e-02
4.85754848e-01 -1.77565262e-01 1.73910394e-01 1.32517368e-01
4.24615294e-01 6.83016121e-01 -3.69410843e-01 -3.12895834e-01
7.33073294e-01 2.69398808e-01 -4.35309649e-01 -7.14442372e-01
-1.10737169e+00 1.22676528e+00 9.09724474e-01 -2.83191949e-01
-6.39069676e-01 4.20102805e-01 9.12638664e-01 2.48930275e-01
8.19934249e-01 2.59095840e-02 -6.17444098e-01 3.76977511e-02
-3.60814720e-01 4.89911675e-01 4.98851925e-01 6.75550878e-01
2.25798801e-01 -1.53294146e-01 -7.36612499e-01 1.09172320e+00
1.66914657e-01 4.59890246e-01 4.24503267e-01 -9.35166180e-01
9.67524230e-01 2.51166821e-01 2.56861269e-01 -4.02678728e-01
-6.66252494e-01 -5.05042493e-01 -1.34323978e+00 -9.61395912e-03
3.14312011e-01 -8.07733774e-01 -9.87424016e-01 2.25616765e+00
5.55206001e-01 4.93725002e-01 5.68417273e-03 8.14536393e-01
7.57203579e-01 5.15811801e-01 7.14918315e-01 -2.04684734e-01
9.25615907e-01 -6.25895500e-01 -9.59540069e-01 -7.24747172e-03
1.01394105e+00 -1.45913064e-01 8.46662045e-01 7.80125707e-02
-1.00554132e+00 -2.85868580e-03 -7.75009334e-01 2.80567825e-01
-2.93562233e-01 -2.79941291e-01 9.60244477e-01 6.98586404e-01
-7.36654997e-01 6.08512759e-01 -4.94877189e-01 5.34937643e-02
1.21204793e+00 6.39662147e-01 -1.78967446e-01 -1.95008352e-01
-1.71814406e+00 6.62733316e-01 8.00860703e-01 -1.25906840e-01
-1.08616114e+00 -1.74840319e+00 -9.52845693e-01 5.20695806e-01
8.01836789e-01 -1.15815139e+00 1.40055323e+00 -7.83521414e-01
-1.69997251e+00 2.19332352e-01 5.46008468e-01 -7.86565781e-01
7.35396028e-01 -8.20640400e-02 -1.60708323e-01 -4.41343695e-01
-1.55304223e-01 9.09931362e-01 4.02502149e-01 -1.17556703e+00
-6.80847526e-01 -4.54778969e-01 4.28058282e-02 3.58160168e-01
-5.18039465e-01 -2.56950974e-01 8.32559913e-02 -7.31997013e-01
-6.69835865e-01 -8.76568079e-01 -2.40656078e-01 -3.31943762e-03
-3.62764060e-01 -6.19219959e-01 4.70907032e-01 -2.77388871e-01
1.20333159e+00 -1.97314095e+00 -3.31046954e-02 8.62282887e-02
5.34067690e-01 3.80538344e-01 -2.47709259e-01 2.31228918e-01
-4.29008991e-01 2.31144518e-01 -3.48418981e-01 -2.06914037e-01
4.13131505e-01 8.88581872e-02 -4.94616367e-02 4.77836519e-01
-3.45519274e-01 9.19241846e-01 -9.87033844e-01 -4.93466437e-01
5.18729270e-01 3.05880159e-01 -9.83804941e-01 3.16970795e-02
-2.07564130e-01 1.15181737e-01 -7.07426906e-01 -2.97810417e-02
1.21078408e+00 -6.26516119e-02 2.02343911e-01 8.46857652e-02
5.83323240e-02 3.13317120e-01 -1.15422928e+00 1.43126833e+00
-8.45127046e-01 3.62155885e-02 -8.61302167e-02 -1.32513702e+00
3.40207577e-01 5.46477079e-01 5.02698362e-01 -6.93285704e-01
3.26302350e-01 1.35449365e-01 2.01378301e-01 -1.61976933e-01
-1.08737327e-01 -3.43170315e-01 -1.50820345e-01 5.12357831e-01
3.07553619e-01 2.56048322e-01 -1.64193153e-01 1.49909809e-01
6.54893279e-01 -1.46996632e-01 4.13939893e-01 -6.04690433e-01
3.88203524e-02 -6.71101391e-01 9.24127698e-01 1.04088688e+00
-1.09530248e-01 2.86449432e-01 6.67087555e-01 -4.04046297e-01
-7.23489523e-01 -1.14148211e+00 -4.72131461e-01 1.17424631e+00
-8.75020996e-02 3.28757837e-02 -6.70405924e-01 -9.61607635e-01
3.80479425e-01 1.13343883e+00 -1.04110789e+00 -5.26903212e-01
-3.10077574e-02 -1.15125585e+00 2.65739828e-01 7.23840535e-01
8.20424199e-01 -1.07391012e+00 -1.00533269e-01 1.81679726e-01
-1.53557047e-01 -2.42667317e-01 -7.98577666e-01 3.81741636e-02
-7.94095159e-01 -8.25641096e-01 -9.93565321e-01 -1.35785341e-01
-2.23811623e-02 -2.35738233e-03 1.13083756e+00 -4.58681673e-01
2.54611790e-01 1.42509133e-01 -8.76970142e-02 -1.01952064e+00
-7.85802305e-02 1.66064143e-01 1.21431975e-02 -3.28776054e-02
3.41659874e-01 -7.70101368e-01 -1.09489691e+00 3.05284113e-01
-9.90847111e-01 -8.84786844e-02 5.75617850e-01 1.25308657e+00
1.34218544e-01 -3.62284511e-01 1.15144181e+00 -1.46115780e+00
6.33242190e-01 -1.03800714e+00 -6.77238584e-01 3.30484778e-01
-7.80533314e-01 1.24591984e-01 5.79979062e-01 -5.32214642e-01
-1.73777485e+00 -4.55945402e-01 -4.36403155e-02 -3.51895809e-01
2.61549741e-01 8.17251444e-01 -2.95376718e-01 2.11794630e-01
6.97880030e-01 -2.54662514e-01 -1.29773349e-01 -3.35398167e-01
5.39319515e-01 8.25075448e-01 2.81865418e-01 -6.26132667e-01
2.22665042e-01 2.63473004e-01 -1.50551111e-01 1.03525139e-01
-1.17562163e+00 -9.88285914e-02 5.29522933e-02 1.36172235e-01
7.15708494e-01 -1.08300984e+00 -1.34289384e+00 5.36107063e-01
-7.89240181e-01 -6.98559523e-01 -2.90589213e-01 9.06186521e-01
-6.01457357e-01 -8.28667954e-02 -4.44191635e-01 -5.33056855e-01
-5.73933840e-01 -1.01158035e+00 8.91950190e-01 2.68580467e-01
1.93927854e-01 -1.28970981e+00 1.38800830e-01 2.11506009e-01
4.97771293e-01 2.06544623e-01 1.04576695e+00 -4.64564979e-01
-5.06823063e-02 5.78502379e-02 -4.29474771e-01 2.24383585e-02
-4.40649539e-02 -2.84731537e-01 -8.36722314e-01 -3.83569688e-01
-4.51403022e-01 -6.55546069e-01 1.29432070e+00 1.25061858e+00
1.85597014e+00 -3.87795150e-01 -6.31231487e-01 6.26572132e-01
1.23616171e+00 1.51357368e-01 6.26144588e-01 -3.53854299e-02
4.89472181e-01 5.36962807e-01 3.06291819e-01 7.89961994e-01
6.89088166e-01 6.06148362e-01 6.71017349e-01 -2.97127932e-01
1.48938060e-01 -3.61985087e-01 8.70247409e-02 3.10741961e-01
7.85708129e-02 -5.25413871e-01 -6.70351148e-01 7.24428952e-01
-2.16654158e+00 -9.47609782e-01 7.60002360e-02 2.48056245e+00
9.77311611e-01 -5.13672121e-02 2.08414569e-01 -5.60568273e-01
9.93339896e-01 1.15342386e-01 -1.01360631e+00 -3.72897983e-01
1.40913457e-01 4.26516205e-01 9.65909004e-01 1.71820536e-01
-1.27396083e+00 5.35025358e-01 5.60085630e+00 1.41074002e+00
-9.20194089e-01 4.56080049e-01 8.98609459e-01 -1.13639079e-01
-5.58641732e-01 -2.91482598e-01 -3.81976902e-01 6.73153043e-01
1.00444806e+00 -5.70490539e-01 -1.28786936e-02 6.86713099e-01
3.67044240e-01 1.88866436e-01 -1.19155109e+00 7.78794110e-01
-5.51294088e-01 -1.54409313e+00 8.88703624e-04 2.04089046e-01
1.29606509e+00 1.34337470e-01 4.87516463e-01 9.31773782e-01
1.32489932e+00 -1.06037676e+00 2.58200198e-01 2.00296000e-01
7.74141550e-01 -1.14560676e+00 1.12768209e+00 1.47972703e-01
-7.81495631e-01 -4.79270548e-01 -6.20692074e-01 1.83487162e-01
2.34086756e-02 5.88198066e-01 -5.77625275e-01 9.91447628e-01
5.51288605e-01 8.21928740e-01 2.69833878e-02 1.33804321e+00
-2.10206714e-02 8.82914245e-01 -2.79424012e-01 7.31913894e-02
4.29125875e-01 3.08986697e-02 9.63938087e-02 8.39744270e-01
4.64370936e-01 1.92614973e-01 1.92704752e-01 7.47652352e-01
-6.99268103e-01 6.29409775e-02 -5.27307808e-01 3.60136300e-01
6.48505747e-01 8.18523467e-01 -1.12171002e-01 -4.37333584e-01
-3.87344688e-01 2.30104655e-01 4.93499964e-01 2.37962902e-01
-9.14937377e-01 -3.05873789e-02 6.73382044e-01 -2.27576599e-01
1.42161816e-01 8.13942254e-01 -2.98244536e-01 -1.14724088e+00
-4.27057952e-01 -8.39993894e-01 1.10539937e+00 -5.26539147e-01
-1.85072184e+00 -1.96783811e-01 5.40709555e-01 -1.12087274e+00
-2.34327301e-01 -3.25562388e-01 -7.50930309e-01 6.20727241e-01
-1.31694162e+00 -9.60629642e-01 1.13098435e-01 7.22987592e-01
3.32169056e-01 1.43949881e-01 5.96745193e-01 4.44787294e-01
-7.42400646e-01 1.10141718e+00 5.85622728e-01 -1.59537613e-01
1.07245171e+00 -1.41326797e+00 -1.87367257e-02 1.39966458e-01
-5.56295633e-01 3.10500085e-01 4.18648303e-01 -4.85055923e-01
-9.03653681e-01 -1.35359967e+00 3.95552039e-01 -7.21681565e-02
5.29712737e-01 -3.50368887e-01 -7.63345003e-01 9.20997918e-01
5.49785435e-01 -2.48480625e-02 6.82398498e-01 7.72367895e-01
-3.64917785e-01 -3.39121789e-01 -1.25202131e+00 7.20662236e-01
1.25398099e+00 3.30772647e-03 -5.26065350e-01 5.41972756e-01
9.59245324e-01 -4.80130881e-01 -9.77495313e-01 5.90029657e-01
5.31440198e-01 -7.53060102e-01 9.93506312e-01 -9.17700171e-01
8.99747729e-01 5.11817813e-01 4.78395313e-01 -2.09406972e+00
-2.59465098e-01 -5.73223293e-01 2.51371503e-01 1.05914795e+00
5.24488509e-01 -9.93874907e-01 6.37647748e-01 8.28071654e-01
-1.70839041e-01 -6.66334271e-01 -1.08218849e+00 -6.08301818e-01
7.41189182e-01 -2.85243213e-01 8.99828017e-01 1.34214175e+00
1.66521855e-02 2.08908349e-01 -7.75923908e-01 -2.40113083e-02
1.02312875e+00 -7.24864602e-02 4.21298116e-01 -1.34197176e+00
-4.35728207e-02 -5.59699059e-01 5.86048849e-02 -7.07553446e-01
5.39086461e-01 -1.03862751e+00 -2.03861192e-01 -1.30799019e+00
7.39545047e-01 -5.50149620e-01 -7.37848997e-01 5.06095648e-01
-5.77765584e-01 -1.83187097e-01 8.05440024e-02 -3.82375836e-01
-6.07807636e-01 1.10668743e+00 1.34132743e+00 -1.22455962e-01
-1.07996374e-01 2.81476349e-01 -6.66244030e-01 4.93491739e-01
6.49484992e-01 -7.60274589e-01 -7.41614997e-01 -3.22858185e-01
-1.00058526e-01 5.06329417e-01 2.21624330e-01 -5.11613667e-01
1.24761038e-01 -3.44694734e-01 2.86739767e-01 -3.85274202e-01
-1.14180952e-01 -6.53496444e-01 1.45930216e-01 4.75489199e-01
-7.93478251e-01 -4.35942769e-01 1.60786018e-01 7.07866371e-01
5.08904122e-02 -4.51932438e-02 7.06658065e-01 8.14638212e-02
-2.94905722e-01 7.74634004e-01 -2.09816515e-01 1.85290232e-01
9.33888137e-01 4.05733645e-01 -5.11729538e-01 -5.94395459e-01
-4.85161752e-01 9.78174329e-01 -1.98552147e-01 2.93049037e-01
1.96803927e-01 -1.61008227e+00 -1.02457190e+00 -3.39668572e-01
3.68489444e-01 2.96839606e-02 1.12090957e+00 8.55025351e-01
1.71193391e-01 4.47650015e-01 -2.53921121e-01 -3.52076381e-01
-8.38118672e-01 7.89208531e-01 4.55690801e-01 -6.91614151e-01
-4.57815647e-01 7.55761385e-01 1.11863244e+00 -1.01865065e+00
2.92422861e-01 -2.09674209e-01 -1.75727233e-01 -7.55341202e-02
4.18849856e-01 3.88986826e-01 -1.89676642e-01 1.77367210e-01
3.06525528e-02 -9.34455916e-02 -3.74971479e-01 1.10814899e-01
1.53472400e+00 -3.75459343e-03 2.22173959e-01 1.89786926e-01
1.19493711e+00 -7.56266773e-01 -1.44956422e+00 -5.61837852e-01
-2.98814088e-01 -5.91013193e-01 4.69443947e-01 -1.17663789e+00
-1.33517003e+00 7.95044780e-01 9.01389360e-01 -1.79106534e-01
1.12171459e+00 -3.84740502e-01 5.68289042e-01 2.49567091e-01
2.43509308e-01 -1.13570631e+00 -3.61912787e-01 -7.56934704e-03
8.69831443e-01 -1.57410073e+00 -1.32680848e-01 -2.98491657e-01
-6.92784607e-01 5.38934827e-01 7.60192215e-01 -1.06619827e-01
9.10073936e-01 6.93567544e-02 -1.84709176e-01 -1.09694272e-01
-1.08172631e+00 1.17856815e-01 2.26158619e-01 5.85621655e-01
3.96999747e-01 4.39257771e-01 -8.33409727e-01 7.92336226e-01
1.72631592e-01 4.61101681e-01 4.13210541e-01 3.67484510e-01
2.76753157e-01 -1.01051176e+00 -2.60748575e-03 8.82045031e-01
-5.38447499e-01 -1.79851964e-01 -1.31536480e-02 1.12661827e+00
-6.45292252e-02 6.23278856e-01 1.12102009e-01 8.02485943e-02
3.33521128e-01 -6.82941824e-02 2.99477488e-01 -1.66633949e-01
-7.03822672e-01 -1.32070825e-01 8.58158767e-02 -5.91170669e-01
-4.73456860e-01 -5.10778010e-01 -6.94269419e-01 -6.86654806e-01
-3.51727158e-01 1.60661563e-01 2.64912903e-01 9.67887580e-01
4.51486856e-01 6.91798747e-01 8.08889568e-01 -6.43190861e-01
-1.11566782e+00 -1.22277057e+00 -5.66349924e-01 4.91957903e-01
6.09462708e-02 -1.09957922e+00 -1.47277012e-01 -6.68779850e-01] | [8.0557222366333, 5.417015075683594] |
7e486e9d-0636-4616-9b6a-988b82c03367 | inspire-creativity-with-oriba-transform | 2306.09776 | null | https://arxiv.org/abs/2306.09776v1 | https://arxiv.org/pdf/2306.09776v1.pdf | Inspire creativity with ORIBA: Transform Artists' Original Characters into Chatbots through Large Language Model | This research delves into the intersection of illustration art and artificial intelligence (AI), focusing on how illustrators engage with AI agents that embody their original characters (OCs). We introduce 'ORIBA', a customizable AI chatbot that enables illustrators to converse with their OCs. This approach allows artists to not only receive responses from their OCs but also to observe their inner monologues and behavior. Despite the existing tension between artists and AI, our study explores innovative collaboration methods that are inspiring to illustrators. By examining the impact of AI on the creative process and the boundaries of authorship, we aim to enhance human-AI interactions in creative fields, with potential applications extending beyond illustration to interactive storytelling and more. | ['Ze Gao', 'Xingyu Li', 'Yuqian Sun'] | 2023-06-16 | null | null | null | null | ['chatbot', 'chatbot'] | ['methodology', 'natural-language-processing'] | [ 1.25861377e-01 5.64328790e-01 8.98174047e-02 2.02870056e-01
2.56750733e-01 -1.08659899e+00 1.07299650e+00 -4.81432170e-01
6.53411523e-02 4.92057860e-01 5.58734894e-01 -1.12047538e-01
7.47325793e-02 -4.49604362e-01 -1.77230537e-01 -1.91006005e-01
4.90708768e-01 7.20534563e-01 -4.53860283e-01 -2.53344417e-01
7.62535453e-01 5.91672242e-01 -1.10523510e+00 5.11044383e-01
1.10472167e+00 3.45049113e-01 2.59698983e-02 9.96794283e-01
-4.74515945e-01 1.58229649e+00 -1.33523464e+00 -6.40277684e-01
5.18970229e-02 -7.89397776e-01 -8.79801869e-01 -1.21268377e-01
4.26518798e-01 -3.87835056e-01 -3.14949036e-01 6.49182200e-01
2.02833578e-01 -3.40018392e-01 5.62193155e-01 -1.64534032e+00
-1.41882384e+00 1.20411181e+00 -2.35438779e-01 -1.65890977e-01
5.35313249e-01 9.70676661e-01 6.61875784e-01 -3.14892799e-01
1.20773149e+00 1.44319332e+00 7.07419515e-01 6.32761180e-01
-1.18516183e+00 -8.65093589e-01 -2.22653881e-01 7.63498247e-02
-8.98828506e-01 -4.28417623e-01 1.10220027e+00 -8.11738551e-01
3.21787596e-01 3.73087287e-01 1.30383873e+00 1.37523258e+00
-2.96236753e-01 8.47110689e-01 1.17032552e+00 -3.05302680e-01
1.29082039e-01 5.61394870e-01 -6.43620128e-03 3.93049717e-01
6.35267943e-02 -3.91208440e-01 -7.03222632e-01 -1.64987355e-01
1.39443028e+00 -3.64712834e-01 2.89633535e-02 3.22164416e-01
-1.83878255e+00 7.09612548e-01 4.32370096e-01 9.36516762e-01
-7.53615081e-01 8.89448822e-01 1.22691594e-01 2.63583779e-01
2.42803171e-01 1.74354208e+00 3.11007440e-01 -9.35553253e-01
-2.10454285e-01 1.48373321e-01 1.17143023e+00 8.14168394e-01
2.05117062e-01 1.24977022e-01 -6.00897014e-01 6.00782692e-01
9.73182768e-02 2.59205729e-01 -1.14267424e-01 -1.89295650e+00
-2.62467623e-01 1.14116275e+00 3.22954863e-01 -1.13898063e+00
1.95516691e-01 -1.78128108e-01 -3.51068109e-01 7.01008141e-01
6.05343878e-01 -3.51496309e-01 -2.08240196e-01 1.29962075e+00
-4.88028824e-02 -4.94562387e-02 -2.04281971e-01 1.24265790e+00
1.06736434e+00 4.81449336e-01 1.85056269e-01 2.26827085e-01
1.32172394e+00 -1.20519483e+00 -1.06004202e+00 -1.61764577e-01
5.43962955e-01 -6.95976138e-01 1.59389567e+00 3.60397786e-01
-1.63235211e+00 -3.39718640e-01 -8.33983183e-01 -2.56564826e-01
-1.73738509e-01 2.94521064e-01 1.01374745e+00 4.80559081e-01
-1.05435383e+00 7.29916275e-01 -3.21430624e-01 -4.58544135e-01
8.07558775e-01 6.25279248e-02 -2.70727351e-02 5.38047016e-01
-6.55384600e-01 1.02724755e+00 -3.19502264e-01 -3.69294323e-02
-4.27193135e-01 -9.56492543e-01 1.05785564e-01 -2.74191201e-01
1.73163578e-01 -8.77866507e-01 1.40654886e+00 -1.82853901e+00
-1.96600997e+00 1.04626143e+00 2.85001338e-01 -1.53913945e-01
9.20979559e-01 -2.91974455e-01 -1.20735966e-01 3.27820837e-01
1.09040834e-01 8.35673869e-01 2.55942911e-01 -1.58348846e+00
6.45572320e-02 3.03558130e-02 3.27195555e-01 1.01005010e-01
-2.88428038e-01 2.57217020e-01 1.34733379e-01 -8.05482090e-01
-4.52014357e-01 -7.38663018e-01 1.22869924e-01 6.53544247e-01
-5.50831974e-01 -4.06302273e-01 1.00914371e+00 -3.41007203e-01
1.00110173e+00 -2.21783876e+00 3.96460354e-01 -1.11719146e-01
9.88481283e-01 3.23625416e-01 -8.30813646e-02 1.14959168e+00
7.79270411e-01 4.42852974e-01 2.60134161e-01 -2.01253906e-01
3.60447288e-01 3.58422734e-02 -5.60100935e-02 7.50444755e-02
-1.77526608e-01 1.42950571e+00 -9.80689943e-01 -5.11035442e-01
-7.38726556e-02 4.67159331e-01 -2.73224413e-01 3.02098095e-01
-4.04427648e-01 1.05018783e+00 -5.38040102e-01 5.90826035e-01
2.74773538e-01 -6.47212744e-01 1.12337962e-01 3.02732170e-01
-5.27063191e-01 6.10639155e-02 -4.41129088e-01 1.52891445e+00
-4.31323797e-01 1.54476464e+00 3.66212517e-01 -2.02930629e-01
1.06945777e+00 2.79069185e-01 8.50294381e-02 -3.80351752e-01
1.99959069e-01 1.48012802e-01 5.05258203e-01 -6.34526312e-01
3.10551047e-01 9.91928130e-02 2.12789960e-02 1.29979646e+00
-6.03223741e-01 -6.56902015e-01 -1.26460910e-01 5.63677728e-01
1.27518260e+00 2.72508591e-01 -2.39986293e-02 -1.17779322e-01
-1.03621623e-02 6.08727634e-01 -8.73076990e-02 9.69828546e-01
-1.80585518e-01 1.44696236e-01 7.76732504e-01 -6.14060819e-01
-1.25530887e+00 -9.83167768e-01 4.11422044e-01 9.66954410e-01
2.42192939e-01 -2.43736953e-01 -1.00000083e+00 -2.18419284e-01
6.88742548e-02 1.16052914e+00 -7.67664731e-01 1.44718185e-01
-4.94342387e-01 5.55915415e-01 1.01508164e+00 3.04457635e-01
8.34987104e-01 -1.85437918e+00 -1.26362550e+00 1.31944388e-01
-2.31623054e-01 -8.79387438e-01 -2.48297781e-01 -8.42027247e-01
-5.33243120e-01 -7.29816794e-01 -7.52368748e-01 -4.31741387e-01
4.95382011e-01 5.38206026e-02 1.15874648e+00 5.39505184e-01
-3.33168805e-01 6.00159407e-01 -4.44760919e-01 -6.79585695e-01
-9.57910180e-01 -8.30150619e-02 -5.16039729e-01 -1.14553787e-01
4.77184318e-02 -1.12494910e+00 -5.84986985e-01 4.17151511e-01
-5.16846299e-01 9.37664807e-01 6.29915893e-01 1.34715870e-01
-5.62334538e-01 -8.64694595e-01 4.74378467e-01 -9.16063249e-01
9.94283557e-01 -4.21120524e-01 2.14344695e-01 2.61288881e-01
-2.19874248e-01 -4.18673277e-01 7.35576093e-01 -1.08354449e+00
-1.01669919e+00 -4.95409578e-01 8.20363581e-01 -4.34398830e-01
-2.81442791e-01 3.51461656e-02 4.52382445e-01 -2.91213542e-01
8.05366874e-01 -2.39169881e-01 3.83047342e-01 -2.38010451e-01
6.59384370e-01 7.78727531e-01 9.22254920e-01 -5.91564000e-01
7.19344378e-01 5.72435856e-01 -3.02275032e-01 -9.46909249e-01
2.05554534e-02 2.89299935e-01 -3.51959765e-01 -1.07029712e+00
7.41085112e-01 -3.12282830e-01 -1.71313787e+00 3.20181400e-01
-1.69837809e+00 -6.52340353e-01 -6.33716285e-01 4.21436518e-01
-5.16037583e-01 -1.19752392e-01 -7.36572683e-01 -9.70240057e-01
-2.64712632e-01 -7.21755564e-01 7.37483859e-01 5.89730740e-01
-1.25471842e+00 -6.75718248e-01 5.47510684e-02 7.43389726e-01
8.60486925e-01 6.00750148e-01 8.40049684e-01 -4.79883909e-01
-8.08071792e-01 -4.52488177e-02 -5.00996232e-01 -4.76140171e-01
-5.10398485e-02 3.53494376e-01 -7.54508495e-01 7.30700016e-01
-4.76138800e-01 -5.21063864e-01 -2.78166860e-01 -1.47363394e-01
5.20728528e-01 -8.61854196e-01 -2.64231682e-01 2.01391533e-01
8.03646743e-01 6.17966473e-01 8.97866905e-01 5.59624970e-01
5.64774811e-01 8.56360376e-01 3.62942666e-01 6.37706101e-01
1.22649409e-01 4.06623393e-01 1.16380140e-01 -9.85179916e-02
-3.79864067e-01 -3.96869183e-01 2.23199174e-01 3.75034302e-01
-8.06509495e-01 -2.59048522e-01 -1.03232515e+00 2.33226404e-01
-2.01921272e+00 -9.38943386e-01 -7.60819912e-01 1.29624379e+00
6.89335704e-01 -3.93363535e-01 2.22495735e-01 -2.96009272e-01
6.08659983e-01 5.35284467e-02 -5.18893898e-01 -9.06970203e-01
-4.50317860e-02 1.22199595e-01 -2.40428314e-01 1.71462759e-01
-4.55100648e-02 1.18614912e+00 6.76240301e+00 3.34221244e-01
-8.88735890e-01 1.45997539e-01 4.75795627e-01 -1.76362813e-01
-6.24435425e-01 2.47560903e-01 2.84096122e-01 5.04410088e-01
3.12207133e-01 -2.46344686e-01 1.06153023e+00 5.65033257e-01
3.89291972e-01 -3.23880911e-01 -1.09688389e+00 5.95145822e-01
6.94552716e-03 -1.72250462e+00 -2.75238343e-02 1.76001713e-02
6.94483101e-01 -7.92856514e-01 1.53943605e-03 -1.84763059e-01
6.30035520e-01 -1.34386909e+00 7.83733428e-01 7.79584110e-01
6.57751679e-01 -1.76812053e-01 1.43616542e-01 3.43349017e-02
-4.07461226e-01 -1.39500752e-01 3.85584325e-01 -8.51226985e-01
3.25125813e-01 -1.36379272e-01 -9.67167914e-01 -2.97208697e-01
3.63460213e-01 5.58923423e-01 -3.58650267e-01 5.70916653e-01
-3.49869013e-01 3.61497670e-01 -5.04085049e-02 -8.30755591e-01
3.20825189e-01 -7.49089956e-01 8.60134959e-01 1.06128025e+00
2.07584426e-02 6.64975941e-01 -5.68232238e-01 1.87282240e+00
2.21334193e-02 -2.06035867e-01 -8.12281728e-01 -1.23013556e+00
6.50891125e-01 1.25342047e+00 -9.41652179e-01 -3.75466317e-01
2.72555381e-01 1.20262587e+00 8.37378055e-02 4.54347968e-01
-8.20333779e-01 -3.12766552e-01 6.42178118e-01 3.18376958e-01
-4.64322984e-01 -4.12225872e-01 -1.15443075e+00 -4.46764648e-01
-4.38602477e-01 -1.06820834e+00 -5.76829314e-01 -1.58685422e+00
-1.19907165e+00 3.91265333e-01 -2.09520057e-01 -6.69358432e-01
-8.57520569e-03 -7.25218132e-02 -1.37085164e+00 3.56049716e-01
-2.87772804e-01 -1.67217195e+00 -9.07929242e-01 -2.52474807e-02
3.74192566e-01 -1.35389760e-01 5.60864627e-01 -2.53291875e-01
-2.11395413e-01 3.11077535e-01 -3.11269462e-01 1.05087541e-01
5.76367259e-01 -7.22763777e-01 4.42212999e-01 -1.69335790e-02
1.13638133e-01 5.92326760e-01 8.37384641e-01 -8.14735234e-01
-1.84909022e+00 1.33215174e-01 7.26878285e-01 -6.97211862e-01
8.23519826e-01 -2.84870863e-01 -3.18746358e-01 7.14224815e-01
8.98628116e-01 -7.38015890e-01 4.41629320e-01 -3.55452262e-02
-8.58212784e-02 5.13863623e-01 -1.06533778e+00 1.35860956e+00
1.68341458e+00 -4.60401386e-01 -5.48064172e-01 3.83763522e-01
6.06549263e-01 1.13821819e-01 -6.89942122e-01 -3.36932361e-01
1.13471186e+00 -9.42331791e-01 8.02347124e-01 -5.67565799e-01
1.20415676e+00 -8.06330964e-02 6.18308067e-01 -1.02774203e+00
-3.31838042e-01 -1.34382820e+00 2.02117100e-01 1.52260411e+00
2.50034422e-01 -7.87157297e-01 6.14301801e-01 1.20187366e+00
-1.05856135e-01 -1.49130300e-01 -4.23602104e-01 -3.48319143e-01
-1.79090142e-01 -1.51730198e-02 4.43685323e-01 1.42776728e+00
5.53466201e-01 2.71022022e-01 -2.79966772e-01 -5.95782340e-01
1.44596025e-01 1.01052798e-01 1.34854293e+00 -1.50568581e+00
-4.99442816e-01 -1.06661510e+00 -1.00075036e-01 -5.42298436e-01
-1.54506475e-01 -7.42452919e-01 -3.13990384e-01 -1.65552235e+00
1.99398160e-01 -3.85156989e-01 8.48691463e-01 3.98759693e-01
4.27895039e-01 6.01040661e-01 7.87621319e-01 7.28812754e-01
-5.52975476e-01 1.17123164e-01 1.97667754e+00 1.53018773e-01
-4.96597826e-01 -5.98002493e-01 -8.97919118e-01 6.95276201e-01
6.89709067e-01 -1.33785352e-01 -1.03736520e-01 -4.58286792e-01
5.41590929e-01 -2.90931910e-02 9.03296709e-01 -8.41978610e-01
4.43519801e-01 -4.33193177e-01 7.48027042e-02 5.31834643e-03
4.18403625e-01 -7.38189578e-01 9.38014269e-01 5.27650118e-01
-6.27658546e-01 -4.32674699e-02 -3.96765657e-02 1.86862517e-02
1.73006520e-01 -1.87092081e-01 3.97079974e-01 -3.46549243e-01
7.84981064e-03 -5.62415540e-01 -1.02965021e+00 2.60071810e-02
1.29132199e+00 -8.26630414e-01 -9.79034603e-01 -9.73789513e-01
-4.50042725e-01 4.85446677e-02 9.90203381e-01 1.15285859e-01
2.36513987e-01 -1.18191433e+00 -6.83306634e-01 -2.87942380e-01
-2.50260174e-01 -3.19333404e-01 -7.79571235e-02 7.22635508e-01
-1.05678117e+00 5.40303998e-02 -6.22189403e-01 -9.47005823e-02
-1.03449607e+00 1.36111319e-01 -8.55689961e-03 3.08754921e-01
-8.68321776e-01 7.49599874e-01 2.90519267e-01 -1.96235254e-02
2.38164932e-01 2.28049487e-01 -8.43805224e-02 -1.00862369e-01
3.05085003e-01 9.42353010e-01 -1.27367437e+00 -1.73933774e-01
5.30490875e-02 4.11132306e-01 4.82457459e-01 -6.80341303e-01
1.14961016e+00 5.29615581e-02 -4.31454569e-01 7.11796463e-01
5.99642456e-01 1.20731898e-01 -1.03295612e+00 3.06216836e-01
-1.96489930e-01 -9.01379108e-01 -5.95159531e-01 -1.50687075e+00
-6.12435937e-01 8.69427979e-01 -1.76933065e-01 4.58278835e-01
4.56611633e-01 3.56133193e-01 7.84530044e-01 3.03084135e-01
5.42414561e-02 -1.08522689e+00 8.32379162e-01 3.18616629e-02
1.51059020e+00 -6.41393483e-01 -1.83006465e-01 -3.49719405e-01
-1.23901868e+00 1.25737429e+00 7.91112721e-01 -3.06454629e-01
-4.85999472e-02 4.40957457e-01 2.63632625e-01 -6.21621490e-01
-7.61611462e-01 1.65444583e-01 -6.77476600e-02 7.11474419e-01
4.60056424e-01 1.20275341e-01 -6.31317914e-01 5.75264037e-01
-4.12122220e-01 4.85000402e-01 7.10564971e-01 6.36638820e-01
-1.12656966e-01 -7.62024522e-01 -5.01581490e-01 3.67961973e-02
-1.30146801e-01 2.70774484e-01 -1.96846056e+00 1.01830637e+00
1.66247860e-01 9.87217784e-01 4.63737249e-01 -2.75846511e-01
4.15781550e-02 -5.15905619e-02 3.09288204e-01 -1.72515571e-01
-1.58111525e+00 -4.21819389e-01 6.27495229e-01 -3.14372122e-01
-3.40186298e-01 -5.07104218e-01 -1.15496314e+00 -1.03620851e+00
1.41725326e-02 1.59618258e-01 7.44959652e-01 6.37178957e-01
8.11458349e-01 3.96766663e-01 4.82632101e-01 -8.99261057e-01
4.60398912e-01 -1.25374031e+00 -1.80225268e-01 3.05278361e-01
-2.40693226e-01 -1.75493136e-01 -2.88565844e-01 -1.20076358e-01] | [9.403361320495605, 6.347986221313477] |
f3a4aaf9-9140-48e3-8ab6-20db805584e5 | complex-deep-learning-models-for-denoising-of | 1908.10417 | null | https://arxiv.org/abs/1908.10417v3 | https://arxiv.org/pdf/1908.10417v3.pdf | Complex Deep Learning Models for Denoising of Human Heart ECG signals | Effective and powerful methods for denoising real electrocardiogram (ECG) signals are important for wearable sensors and devices. Deep Learning (DL) models have been used extensively in image processing and other domains with great success but only very recently have been used in processing ECG signals. This paper presents several DL models namely Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM), Restricted Boltzmann Machine (RBM) together with the more conventional filtering methods (low pass filtering, high pass filtering, Notch filtering) and the standard wavelet-based technique for denoising EEG signals. These methods are trained, tested and evaluated on different synthetic and real ECG datasets taken from the MIT PhysioNet database and for different simulation conditions (i.e. various lengths of the ECG signals, single or multiple records). The results show the CNN model is a performant model that can be used for off-line denoising ECG applications where it is satisfactory to train on a clean part of an ECG signal from an ECG record, and then to test on the same ECG signal, which would have some high level of noise added to it. However, for real-time applications or near-real time applications, this task becomes more cumbersome, as the clean part of an ECG signal is very probable to be very limited in size. Therefore the solution put forth in this work is to train a CNN model on 1 second ECG noisy artificial multiple heartbeat data (i.e. ECG at effort), which was generated in a first instance based on few sequences of real signal heartbeat ECG data (i.e. ECG at rest). Afterwards it would be possible to use the trained CNN model in real life situations to denoise the ECG signal. | ['Corneliu Arsene'] | 2019-08-27 | null | null | null | null | ['ecg-denoising', 'electrocardiography-ecg'] | ['medical', 'methodology'] | [ 4.67000544e-01 -1.64729096e-02 7.35352635e-01 -1.98332429e-01
-4.78515267e-01 -2.22188234e-01 1.96258634e-01 1.03867851e-01
-8.84327888e-01 1.00777435e+00 -1.51125461e-01 -2.15780318e-01
-3.93942863e-01 -7.46845722e-01 -4.24603611e-01 -8.56206417e-01
-3.66395175e-01 3.62851471e-01 -8.55496079e-02 -2.74497241e-01
-8.53372589e-02 6.09486520e-01 -1.47550225e+00 3.56286615e-01
6.25539064e-01 7.61099339e-01 1.59160018e-01 8.82607698e-01
3.51041168e-01 4.55347896e-01 -1.00982404e+00 -1.32192329e-01
1.29917160e-01 -7.72510469e-01 -6.21935010e-01 -2.55802482e-01
-1.28037572e-01 -1.41539991e-01 -4.36559096e-02 9.93602633e-01
1.08715153e+00 3.39693129e-02 4.71484601e-01 -6.10659599e-01
1.83190241e-01 3.59379798e-01 -2.66838193e-01 5.47735751e-01
2.99896330e-01 1.69673741e-01 4.76093292e-02 -4.85723525e-01
4.48435009e-01 6.77654207e-01 1.16764200e+00 6.03160739e-01
-1.34643424e+00 -5.22733569e-01 -5.31995893e-01 2.82880306e-01
-1.23649061e+00 -2.89274007e-01 7.25824535e-01 -1.34208068e-01
8.61162603e-01 3.38746786e-01 9.08572853e-01 1.45306480e+00
6.61685646e-01 2.51819372e-01 1.38347113e+00 -3.05423379e-01
3.44518691e-01 -5.55817410e-02 1.93264022e-01 8.19284841e-02
2.16922432e-01 6.95820823e-02 -1.18737660e-01 -2.50378013e-01
7.86261141e-01 3.27841789e-02 -4.41572517e-01 3.31277281e-01
-1.11848450e+00 5.34798145e-01 3.33695412e-01 8.91685545e-01
-1.00203514e+00 1.35027483e-01 6.80559516e-01 7.70695090e-01
4.43170726e-01 3.84234190e-01 -5.50029576e-01 -3.51281106e-01
-1.17786181e+00 3.43141705e-01 9.07019496e-01 8.13572630e-02
2.83405393e-01 2.55637318e-01 -1.34556040e-01 8.29026580e-01
-4.43525836e-02 1.60353333e-01 8.61549795e-01 -4.01197463e-01
1.70910746e-01 -8.63296986e-02 5.15553774e-03 -9.16957319e-01
-7.54739821e-01 -6.50366724e-01 -1.46954131e+00 3.59049559e-01
5.80928445e-01 -4.92758274e-01 -9.45757210e-01 1.60322821e+00
-7.95280337e-02 5.46591222e-01 1.60656273e-01 9.14541423e-01
7.96049297e-01 7.71633446e-01 7.84038380e-02 -5.84452093e-01
1.26019931e+00 -3.69621627e-02 -1.03476000e+00 -6.23819716e-02
3.40955555e-01 -6.04557157e-01 6.17097020e-01 1.06692183e+00
-1.08425891e+00 -8.73172939e-01 -1.05902255e+00 4.07491446e-01
-3.20984066e-01 -5.37936613e-02 2.47560903e-01 8.24472487e-01
-1.05648506e+00 1.19403231e+00 -8.49782526e-01 -3.20251256e-01
3.44275504e-01 6.66932285e-01 -4.93878692e-01 4.76712249e-02
-1.46039140e+00 9.83972192e-01 3.43336433e-01 6.47299349e-01
-8.42763126e-01 -3.61431539e-01 -5.65530717e-01 1.23601906e-01
-6.47901371e-02 -5.84771693e-01 8.67139041e-01 -1.33419824e+00
-1.44248319e+00 7.12223589e-01 2.56340384e-01 -8.92640054e-01
5.87711155e-01 -6.72951862e-02 -4.63313252e-01 -5.28858639e-02
-4.81479675e-01 6.70777261e-02 1.13234353e+00 -9.40340042e-01
1.74213186e-01 -3.71866792e-01 -2.87883162e-01 -6.71846122e-02
-3.61032560e-02 4.56719361e-02 2.86832541e-01 -7.68983006e-01
1.04123570e-01 -6.63167596e-01 -3.08857173e-01 -6.59934819e-01
-8.31960887e-02 8.08615834e-02 4.56968695e-01 -1.09236920e+00
9.81063068e-01 -2.19707775e+00 2.00938806e-01 4.89138097e-01
6.01527020e-02 5.69555819e-01 -9.00812000e-02 4.61466759e-01
-6.34938300e-01 7.46496990e-02 -4.19758737e-01 -2.36462146e-01
-4.36729401e-01 3.51690114e-01 1.57301873e-01 6.58762574e-01
-9.10499692e-02 6.45597160e-01 -5.67233622e-01 -1.16705298e-01
2.41171032e-01 8.22527289e-01 9.71145637e-04 3.44975054e-01
3.03101718e-01 8.89643550e-01 -4.31780741e-02 -1.17201351e-01
6.36906207e-01 4.42494512e-01 4.41563912e-02 -4.67428327e-01
8.97026360e-02 -5.17439693e-02 -1.28825200e+00 1.79429829e+00
-4.09242898e-01 4.69846368e-01 4.69592996e-02 -1.42409408e+00
1.14506996e+00 9.61006820e-01 4.86264646e-01 -7.49627173e-01
5.32084763e-01 2.37970427e-01 4.02098596e-01 -7.77416468e-01
-2.57728040e-01 -6.83548629e-01 2.21671030e-01 4.04058933e-01
3.00030291e-01 -1.10479213e-01 3.10676675e-02 -4.75934446e-01
1.29800940e+00 4.64684293e-02 2.58798122e-01 -3.00377756e-01
7.12267339e-01 -6.43468380e-01 3.93274635e-01 7.52918363e-01
-2.67398376e-02 9.80439544e-01 4.16100204e-01 -9.04118657e-01
-9.83210087e-01 -6.56605899e-01 -1.29368335e-01 2.42176950e-01
-3.86768878e-01 -1.67984605e-01 -9.65864718e-01 -2.44050860e-01
-5.99245191e-01 5.35631299e-01 -5.90711474e-01 -4.40661699e-01
-7.40539610e-01 -1.12634063e+00 6.56952441e-01 1.96983293e-01
4.25548702e-01 -1.68520057e+00 -1.16590023e+00 6.71960711e-01
-1.43016666e-01 -5.82833052e-01 5.03308296e-01 6.72454238e-01
-1.23095357e+00 -6.70221150e-01 -9.86167610e-01 -5.06559432e-01
2.01384887e-01 -5.17748296e-01 1.12985671e+00 2.00852484e-01
-5.68842113e-01 1.04313418e-01 -3.03209037e-01 -7.20244050e-01
-3.23467553e-01 -1.53263494e-01 7.91366920e-02 1.70765191e-01
2.98034966e-01 -1.09349728e+00 -5.16951025e-01 -1.90621644e-01
-1.11167407e+00 -1.52728200e-01 6.13955081e-01 9.87735510e-01
4.61270392e-01 3.72121900e-01 9.88903821e-01 -9.04830337e-01
1.15656805e+00 -2.91171610e-01 -1.57324299e-01 -1.45087078e-01
-2.89788246e-01 -2.67129421e-01 8.46353173e-01 -5.22588313e-01
-8.09106290e-01 -2.28026345e-01 -9.75237489e-01 -3.38025004e-01
-5.71891010e-01 7.27283835e-01 -2.30355840e-02 2.85948068e-02
9.10054147e-01 2.99867958e-01 5.43787703e-02 -6.53472006e-01
-3.05487096e-01 6.11911893e-01 4.11503464e-01 -2.46541753e-01
4.11794096e-01 7.82377720e-02 1.92352273e-02 -9.69650507e-01
-3.14793110e-01 -5.36110289e-02 -7.22208619e-01 -2.38490343e-01
1.02202368e+00 -5.50896168e-01 -6.01138234e-01 6.71351671e-01
-1.18204308e+00 -4.44601744e-01 -4.48803872e-01 5.88721335e-01
-5.01649320e-01 2.25832537e-01 -8.20394635e-01 -8.92739415e-01
-7.38426149e-01 -8.69894147e-01 6.00694120e-01 1.66749805e-01
-3.08270216e-01 -8.70633721e-01 1.74110919e-01 -6.88606873e-02
5.39902508e-01 7.77543783e-01 8.30112159e-01 -5.70987523e-01
2.20589876e-01 -6.08818173e-01 3.09350163e-01 9.25018072e-01
-7.42856413e-03 -3.96808654e-01 -1.07220113e+00 -3.94454181e-01
8.06884825e-01 -2.24898726e-01 7.03372955e-01 6.67604983e-01
1.29873574e+00 3.48183885e-02 8.89980569e-02 5.70484817e-01
1.57768226e+00 4.22151566e-01 1.30452204e+00 2.75762528e-02
3.00992727e-01 3.41898143e-01 1.48634985e-01 2.90383399e-01
-3.26669455e-01 3.11995655e-01 2.61317909e-01 -4.35019493e-01
5.19583896e-02 5.73758483e-01 2.02302724e-01 8.53948653e-01
-6.54274881e-01 -2.03044623e-01 -8.24212968e-01 4.01232123e-01
-1.59390903e+00 -9.75307465e-01 -4.51587975e-01 2.52303505e+00
8.08404028e-01 3.98986042e-01 4.60953414e-02 8.88686717e-01
4.26139295e-01 -2.13135377e-01 -2.50699729e-01 -7.80575812e-01
-1.27142981e-01 1.11326802e+00 9.91228148e-02 -2.76319403e-02
-8.85926843e-01 9.15699825e-03 5.30868578e+00 6.52925491e-01
-1.37055337e+00 3.69933635e-01 6.89907193e-01 1.31081745e-01
3.14811021e-01 -4.83793586e-01 -1.86902899e-02 4.87840652e-01
1.54945207e+00 2.04556495e-01 4.72333997e-01 2.85247833e-01
4.26595062e-01 -2.95344621e-01 -1.02476263e+00 1.46163642e+00
-9.98667777e-02 -8.86727929e-01 -4.29459929e-01 -2.84708977e-01
3.45565826e-01 7.55213061e-03 -2.33020410e-01 3.03271323e-01
-7.12034643e-01 -1.46753383e+00 3.12754720e-01 9.68903720e-01
8.69945228e-01 -9.63423371e-01 1.51445591e+00 7.25189567e-01
-7.15061009e-01 -1.39311269e-01 -4.73574728e-01 -3.50352377e-01
7.45353922e-02 9.04209495e-01 -2.05684572e-01 7.21130908e-01
9.31998789e-01 6.36945903e-01 -4.37012106e-01 1.19813561e+00
6.86272308e-02 9.16561604e-01 -3.93445909e-01 2.49210224e-01
1.55016661e-01 -2.71967411e-01 4.69213784e-01 1.28862906e+00
4.37759668e-01 2.95574695e-01 -3.84551734e-01 7.19818354e-01
2.03577995e-01 2.74234004e-02 -6.98662877e-01 3.00602496e-01
-3.06841433e-01 1.38498604e+00 -9.11165237e-01 -3.75708431e-01
-1.67497978e-01 1.00435424e+00 -2.23590001e-01 5.11271298e-01
-7.44923770e-01 -7.05741405e-01 1.12737060e-01 7.39832446e-02
-1.08007550e-01 8.50828737e-02 -3.08194101e-01 -9.94810820e-01
-1.02854922e-01 -1.13340521e+00 2.34112293e-01 -7.58941710e-01
-1.26555359e+00 1.02070785e+00 -6.43140823e-02 -1.01503837e+00
-3.28776985e-01 -3.70648444e-01 -8.05585742e-01 1.38754058e+00
-9.76065755e-01 -4.63665247e-01 -4.05724138e-01 8.13395143e-01
3.41447443e-01 1.28616616e-01 1.20322073e+00 6.83248281e-01
-2.38295242e-01 1.05520509e-01 -1.17197372e-01 9.70653966e-02
4.45517838e-01 -1.15538430e+00 6.78248182e-02 6.14282429e-01
7.86797479e-02 5.66017568e-01 9.23153520e-01 -3.98781925e-01
-9.99672234e-01 -8.17822099e-01 7.91340709e-01 1.79453678e-02
-9.50242672e-03 -3.81116062e-01 -1.23389137e+00 2.30747223e-01
5.15378177e-01 1.59163579e-01 4.74354506e-01 -3.25102746e-01
4.51391697e-01 -4.97428238e-01 -1.40144598e+00 2.00979590e-01
4.70707238e-01 -2.77975887e-01 -9.44065273e-01 2.19684407e-01
-1.55691341e-01 -2.31824324e-01 -1.02642536e+00 5.42097926e-01
6.80334270e-01 -1.23154545e+00 8.55375350e-01 -2.63083637e-01
-2.24939585e-02 -1.81148738e-01 5.15446007e-01 -1.52643132e+00
-5.12230247e-02 -6.90694571e-01 8.65423232e-02 8.51097822e-01
1.00208305e-01 -7.45904744e-01 3.88569176e-01 1.35752603e-01
1.15339331e-01 -9.14372027e-01 -1.16010988e+00 -5.16464949e-01
-1.72509223e-01 -5.51544785e-01 3.32907081e-01 6.14777029e-01
-3.50351423e-01 3.65867406e-01 -5.92959881e-01 -1.44738436e-01
4.96122003e-01 -2.73239374e-01 3.79233450e-01 -1.55046487e+00
-3.29000950e-01 -8.03144649e-02 -6.21397197e-01 -2.48862788e-01
-1.57036304e-01 -5.69857717e-01 1.13792516e-01 -1.55444682e+00
-1.98386058e-01 -3.57448637e-01 -6.89942956e-01 3.06648970e-01
3.61119807e-02 5.49032569e-01 5.36330417e-02 -3.04710239e-01
7.10148588e-02 1.61841169e-01 8.68149042e-01 3.35795768e-02
-3.41132253e-01 4.43901271e-01 -4.35493477e-02 7.21060514e-01
9.06919539e-01 -7.41719007e-01 -3.73205125e-01 -1.61996111e-02
3.12442899e-01 4.63963836e-01 4.84683454e-01 -1.53522146e+00
-2.72553414e-02 6.59604132e-01 9.37346220e-01 -4.09314781e-01
3.40470403e-01 -9.95997488e-01 8.14293742e-01 8.51759374e-01
-1.57663643e-01 1.35368809e-01 2.73684770e-01 2.98853159e-01
-3.30283374e-01 -5.28243482e-01 8.04952204e-01 -5.40619910e-01
-1.75996959e-01 1.42423972e-01 -7.24761546e-01 -4.10026193e-01
7.88953245e-01 -4.68581378e-01 3.97650838e-01 -4.29761231e-01
-1.40342689e+00 -4.41876173e-01 -1.21716268e-01 9.70958266e-03
7.91155815e-01 -1.04040563e+00 -8.76526952e-01 4.19281185e-01
-3.31028074e-01 -8.69807601e-02 5.15567601e-01 1.36528778e+00
-6.18317425e-01 -6.41986653e-02 -6.71638727e-01 -5.17661393e-01
-1.29868650e+00 5.30670106e-01 6.05690122e-01 -3.46096665e-01
-8.99330974e-01 7.22381353e-01 -3.24907005e-01 5.13017811e-02
2.95052171e-01 -4.84200448e-01 -7.52824366e-01 2.40320832e-01
6.21823013e-01 9.27748755e-02 5.83859503e-01 -2.77299136e-01
-2.15665758e-01 5.65427482e-01 4.49548900e-01 -2.69818883e-02
1.80142331e+00 1.79934263e-01 -3.28739047e-01 7.94876873e-01
9.21349585e-01 -3.83372724e-01 -8.05660725e-01 2.74527192e-01
2.74409242e-02 4.97211963e-02 1.64842948e-01 -8.54759157e-01
-1.14624774e+00 1.21587241e+00 1.22726095e+00 4.85639125e-01
1.68420327e+00 -6.91752374e-01 7.39415824e-01 3.22857171e-01
5.77380478e-01 -8.88483882e-01 -2.89871067e-01 7.11369663e-02
9.65390861e-01 -7.14035034e-01 -8.23501050e-02 2.47759208e-01
-3.98971736e-01 1.58252192e+00 -1.94569528e-02 -5.50624728e-01
8.56377125e-01 4.44297194e-01 1.27435058e-01 -1.28306776e-01
-4.99108374e-01 -1.82277337e-02 1.05033845e-01 7.00217485e-01
5.90740979e-01 -1.34546205e-01 -7.35817075e-01 7.41631210e-01
1.02852173e-01 4.52425838e-01 6.21818244e-01 6.95980310e-01
4.19766409e-03 -1.04644144e+00 -5.61915040e-01 8.27659369e-01
-1.01184297e+00 -1.72697052e-01 8.75448529e-03 5.84752858e-01
6.41900003e-01 8.45262706e-01 -2.78877288e-01 -3.09391946e-01
3.89644176e-01 3.87975127e-01 6.82911873e-01 -5.29005766e-01
-1.24180818e+00 2.30684116e-01 -7.43143307e-03 -4.98020798e-01
-5.11141121e-01 -5.16434669e-01 -9.76109266e-01 6.33530319e-02
-1.95175871e-01 9.51178223e-02 8.55802417e-01 1.00014281e+00
-5.94425537e-02 9.62047100e-01 1.75407395e-01 -1.27653003e+00
-3.91170621e-01 -1.46216810e+00 -9.15752590e-01 6.46944344e-01
5.12644291e-01 -2.87783265e-01 -2.97530681e-01 2.26439431e-01] | [14.22307014465332, 3.2845358848571777] |
168edc11-f3e6-4507-b02b-935327d99b7f | knowledge-distillation-for-efficient-audio | 2306.09947 | null | https://arxiv.org/abs/2306.09947v1 | https://arxiv.org/pdf/2306.09947v1.pdf | Knowledge Distillation for Efficient Audio-Visual Video Captioning | Automatically describing audio-visual content with texts, namely video captioning, has received significant attention due to its potential applications across diverse fields. Deep neural networks are the dominant methods, offering state-of-the-art performance. However, these methods are often undeployable in low-power devices like smartphones due to the large size of the model parameters. In this paper, we propose to exploit simple pooling front-end and down-sampling algorithms with knowledge distillation for audio and visual attributes using a reduced number of audio-visual frames. With the help of knowledge distillation from the teacher model, our proposed method greatly reduces the redundant information in audio-visual streams without losing critical contexts for caption generation. Extensive experimental evaluations on the MSR-VTT dataset demonstrate that our proposed approach significantly reduces the inference time by about 80% with a small sacrifice (less than 0.02%) in captioning accuracy. | ['Wenwu Wang', 'Volkan Kılıç', 'Xubo Liu', 'Özkan Çaylı'] | 2023-06-16 | null | null | null | null | ['audio-visual-video-captioning', 'video-captioning'] | ['computer-vision', 'computer-vision'] | [ 4.06746894e-01 1.68639228e-01 -1.62929282e-01 -3.78272980e-01
-1.29318011e+00 -4.38491791e-01 3.70866537e-01 6.91639110e-02
-4.02146608e-01 7.16778576e-01 2.37734616e-01 -1.28247380e-01
2.55667180e-01 -3.47246975e-01 -9.09434080e-01 -5.73397338e-01
3.16233188e-01 1.54509470e-01 1.73280165e-01 2.39589810e-01
-4.62317877e-02 -1.80086738e-03 -1.62183106e+00 5.25870144e-01
7.37774193e-01 1.37525213e+00 3.59126329e-01 6.87727928e-01
-1.68084070e-01 8.38891685e-01 -6.07626855e-01 -6.70317709e-01
-3.09440792e-02 -2.92035431e-01 -4.75269079e-01 -3.35156806e-02
6.27321303e-01 -6.50208831e-01 -7.06511736e-01 1.02433717e+00
7.90717185e-01 1.98486328e-01 3.43220621e-01 -1.25708675e+00
-5.49835384e-01 7.56753981e-01 -6.49805903e-01 1.54095203e-01
1.99435100e-01 1.26424208e-01 1.07914388e+00 -8.39941800e-01
1.24604285e-01 1.25686181e+00 4.24780279e-01 6.43836439e-01
-1.06680715e+00 -9.74222362e-01 2.55334347e-01 7.93752432e-01
-1.46695423e+00 -7.14259923e-01 8.16449463e-01 -1.37853161e-01
7.91073740e-01 2.90705085e-01 3.67769450e-01 1.28503191e+00
-5.32871485e-01 1.02006257e+00 7.38005340e-01 -1.79457918e-01
1.21649720e-01 2.73410946e-01 -1.75642893e-01 3.63392830e-01
8.93997177e-02 -5.41323900e-01 -8.41471851e-01 -1.43242255e-01
6.05713546e-01 -2.13059425e-01 -3.73844236e-01 8.29510391e-02
-1.02235162e+00 6.16281688e-01 1.91376090e-01 -7.27942511e-02
-3.12483311e-01 4.92320895e-01 5.94583988e-01 -1.61121428e-01
4.53634977e-01 5.05273454e-02 -3.07198167e-01 -5.33279598e-01
-1.08316219e+00 8.23014677e-02 4.64801878e-01 1.08232605e+00
4.81880337e-01 1.95349097e-01 -4.99617606e-01 1.04853618e+00
1.95078790e-01 6.24770880e-01 4.73110288e-01 -9.59809959e-01
8.84308338e-01 1.98179945e-01 1.21965557e-01 -1.03682399e+00
-2.68889647e-02 -4.56412166e-01 -8.64662349e-01 -6.12751782e-01
2.69352525e-01 -3.47771764e-01 -7.45083332e-01 1.77362347e+00
2.60933816e-01 6.87917531e-01 5.12160510e-02 8.94577503e-01
8.24514687e-01 1.07122135e+00 2.02766478e-01 -1.97957039e-01
1.62853277e+00 -9.66006339e-01 -8.05861056e-01 -4.74660069e-01
1.60633937e-01 -8.18827152e-01 9.79997158e-01 2.54317164e-01
-1.06942403e+00 -5.83595574e-01 -8.89485598e-01 -2.95659631e-01
1.81565717e-01 4.27615047e-01 5.23689330e-01 4.77027893e-01
-8.48481357e-01 3.11157823e-01 -6.89014673e-01 -5.68217747e-02
6.74450219e-01 3.82247537e-01 -6.98421672e-02 -6.00031158e-03
-1.20077431e+00 2.96448141e-01 3.48732322e-01 1.90037474e-01
-9.96132195e-01 -9.10372436e-01 -7.42101490e-01 5.07155478e-01
5.81557572e-01 -4.89048719e-01 1.66290224e+00 -8.98703039e-01
-1.68665659e+00 3.41942400e-01 -3.83695841e-01 -5.44176340e-01
4.92409229e-01 -4.53568459e-01 -3.81887287e-01 5.64877450e-01
-2.03440145e-01 9.35490429e-01 1.15372574e+00 -1.06902683e+00
-7.06558228e-01 -3.18449102e-02 9.31182131e-02 3.36610556e-01
-8.79125774e-01 -6.32198602e-02 -7.10231483e-01 -7.40205467e-01
-1.44770041e-01 -8.88387620e-01 1.96715355e-01 9.79208946e-02
-4.73391622e-01 -1.36661336e-01 9.79112625e-01 -7.91265905e-01
1.16189778e+00 -2.29838443e+00 -2.68345326e-01 -2.23143205e-01
-3.32210474e-02 5.03395557e-01 -1.16409689e-01 7.57433847e-02
1.63955972e-01 -1.40625969e-01 -1.17256343e-01 -5.11883795e-01
1.82070192e-02 -1.82442013e-02 -6.44911885e-01 1.98762894e-01
3.72180670e-01 6.91352129e-01 -8.03819239e-01 -6.98762655e-01
2.29756907e-01 9.26885903e-01 -5.28003216e-01 3.38226557e-01
-3.76770496e-01 1.99925855e-01 -4.57958370e-01 3.60630751e-01
5.59837639e-01 -3.83374244e-01 1.60345674e-01 -3.54553461e-01
2.27012083e-01 4.69430774e-01 -9.70783532e-01 1.75984895e+00
-5.68356335e-01 1.02025175e+00 1.00244872e-01 -9.57232654e-01
8.71978343e-01 6.83628857e-01 2.17827618e-01 -5.90309322e-01
2.44469523e-01 1.34057805e-01 -2.87908942e-01 -6.56739771e-01
5.36704600e-01 5.82887866e-02 -3.08201034e-02 2.51841635e-01
6.99547008e-02 2.58243978e-01 1.84305608e-02 8.14883187e-02
7.91855812e-01 -1.41516691e-02 2.03364175e-02 2.77734935e-01
5.46766400e-01 -2.93589115e-01 5.43684542e-01 4.85579401e-01
-1.85773566e-01 7.89763153e-01 3.35460663e-01 -1.44732997e-01
-1.06822157e+00 -8.19127083e-01 7.31430054e-02 1.22000277e+00
-2.41770241e-02 -3.46813351e-01 -9.36943352e-01 -2.49574557e-01
-2.66850978e-01 6.93013191e-01 -2.60953546e-01 -7.89016709e-02
-6.85462296e-01 -6.03596687e-01 7.94488609e-01 5.51578760e-01
6.71962559e-01 -9.90967214e-01 -5.71594417e-01 2.26405978e-01
-6.31339371e-01 -1.65592289e+00 -5.05621731e-01 -3.09121609e-01
-7.94438303e-01 -6.14182532e-01 -1.17316377e+00 -8.28574359e-01
4.75729942e-01 3.55390161e-01 7.51091778e-01 -3.71802807e-01
-1.25380650e-01 1.94828793e-01 -2.27906868e-01 -3.00099969e-01
-2.24111989e-01 2.73307890e-01 -1.61953911e-01 4.27691728e-01
4.55876589e-01 -6.97277844e-01 -6.97182655e-01 1.02898002e-01
-7.60483682e-01 3.58594984e-01 6.97762430e-01 6.40736103e-01
5.70995271e-01 -1.93514928e-01 7.73387730e-01 -5.07949173e-01
4.17478681e-01 -4.94420409e-01 -6.04593575e-01 1.42830357e-01
-1.29740521e-01 2.58763996e-03 8.16274464e-01 -7.76229918e-01
-1.25025940e+00 8.14730152e-02 1.58931091e-01 -7.19856560e-01
-1.57021776e-01 2.31112793e-01 -1.92536637e-01 3.06985229e-01
2.77888685e-01 4.66354400e-01 -2.16994718e-01 -6.49754584e-01
3.86302322e-01 9.87586677e-01 8.56118202e-01 -5.19385099e-01
7.03369796e-01 4.38594490e-01 -1.33348748e-01 -8.72390091e-01
-8.83077919e-01 -3.34871173e-01 -4.22598422e-02 -1.77811384e-01
8.07333946e-01 -1.41849077e+00 -1.03712034e+00 2.91952610e-01
-1.33681881e+00 1.36429174e-02 4.70405854e-02 5.66681862e-01
-5.62510610e-01 3.62593412e-01 -5.89126766e-01 -9.84132469e-01
-7.49368012e-01 -1.10285187e+00 9.86308277e-01 3.96148086e-01
-4.83906157e-02 -3.54648113e-01 -3.99718881e-01 7.75857985e-01
4.65064675e-01 -4.69221734e-02 8.24474156e-01 -6.22290194e-01
-7.18781710e-01 -1.23973608e-01 -6.35574996e-01 3.51180792e-01
-6.83585629e-02 -3.46946806e-01 -1.52677190e+00 -1.12927161e-01
-2.53276557e-01 -4.58118051e-01 7.12092698e-01 2.99147338e-01
1.68062270e+00 -5.29612482e-01 -5.59711605e-02 3.95661473e-01
1.17734849e+00 2.31640860e-01 3.14116120e-01 3.73961478e-02
8.26590061e-01 5.44339716e-01 5.09389341e-01 7.28761017e-01
3.30245048e-01 7.89028466e-01 4.66765970e-01 1.23249136e-01
-3.37092251e-01 -4.71843988e-01 5.03264844e-01 1.06425679e+00
2.94045806e-01 -3.22553724e-01 -7.42764652e-01 8.30634534e-01
-1.91966212e+00 -9.26578820e-01 1.26276106e-01 2.07010102e+00
1.19527781e+00 2.44747009e-02 -4.06590253e-02 1.58278748e-01
1.06192553e+00 2.51369029e-01 -6.74000680e-01 -1.64084032e-01
6.65701255e-02 1.09262802e-01 2.71102518e-01 3.18588585e-01
-9.54726696e-01 7.36180067e-01 5.57105255e+00 1.18469048e+00
-9.57693398e-01 1.08147107e-01 8.02559555e-01 -4.49277788e-01
-5.06110638e-02 -2.99025655e-01 -8.86886120e-01 6.48239195e-01
1.17734516e+00 -1.90653563e-01 3.93138200e-01 8.57311904e-01
3.72414857e-01 5.44407703e-02 -1.06367660e+00 1.36439288e+00
1.65259972e-01 -1.21617973e+00 2.34071553e-01 -3.32079887e-01
5.35054624e-01 -1.79050401e-01 2.11943015e-01 2.34679580e-01
-2.58577555e-01 -9.02474821e-01 8.01550686e-01 1.30041003e-01
1.00749362e+00 -9.49956417e-01 6.80730999e-01 5.61319478e-02
-1.15111065e+00 -1.14184260e-01 -3.12228680e-01 -4.65548187e-02
2.77549475e-01 6.41442597e-01 -1.05983496e+00 2.12630987e-01
7.40242958e-01 3.94881725e-01 -2.85857052e-01 1.03574312e+00
-1.82672635e-01 1.02198732e+00 -4.61884737e-01 -1.39294893e-01
2.09365711e-01 2.38432214e-01 5.25580943e-01 1.27192366e+00
4.04686928e-01 3.02496161e-02 -1.97633013e-01 5.98921657e-01
-4.83481258e-01 1.74924862e-02 -1.38291568e-01 -1.65573269e-01
8.45165730e-01 1.11177146e+00 -2.17347428e-01 -5.40821493e-01
-4.13591385e-01 8.39385033e-01 2.33619511e-02 4.15587991e-01
-1.14508331e+00 -5.70268035e-01 7.63307095e-01 9.79033783e-02
5.30246556e-01 7.66599691e-03 5.88448942e-02 -9.76513624e-01
3.46236467e-01 -8.54556143e-01 6.88928142e-02 -9.36921120e-01
-1.01885974e+00 5.43694675e-01 1.03752591e-01 -1.15971518e+00
-3.38125348e-01 -3.30969065e-01 -2.46874496e-01 7.24115551e-01
-1.66450942e+00 -8.58021021e-01 -5.86845875e-01 5.95758200e-01
7.79559910e-01 -8.52028430e-02 6.81828976e-01 6.31861687e-01
-6.90509856e-01 9.37352002e-01 1.70497417e-01 3.09575558e-01
7.92248905e-01 -7.86823452e-01 3.18687588e-01 6.02656364e-01
1.06432416e-01 2.07623735e-01 8.20462883e-01 -1.81569576e-01
-1.25709391e+00 -1.22653937e+00 9.27713096e-01 1.34237587e-01
5.68392754e-01 -5.37880599e-01 -1.00352812e+00 3.91272932e-01
3.24385136e-01 7.82578215e-02 6.84087276e-01 -2.64920980e-01
-5.15155733e-01 -4.44980174e-01 -8.30343902e-01 6.82762861e-01
6.54439628e-01 -6.84222519e-01 -3.77138227e-01 2.17713118e-01
1.01728761e+00 -3.11859816e-01 -6.18433774e-01 1.32307038e-01
5.94895840e-01 -6.42159760e-01 9.60531712e-01 -1.86374187e-01
3.58445048e-01 -3.40338379e-01 -1.80394962e-01 -8.88739169e-01
3.67582738e-02 -9.93620098e-01 -4.41995740e-01 1.63986194e+00
3.44465584e-01 -2.84604192e-01 9.06844318e-01 4.63539839e-01
7.08750933e-02 -5.50482094e-01 -1.19648707e+00 -6.58747911e-01
-4.92321163e-01 -4.84354883e-01 4.62393910e-01 7.33293355e-01
-1.01192594e-01 5.68984807e-01 -6.73254073e-01 1.09813266e-01
6.93102360e-01 -6.10007681e-02 6.74715340e-01 -9.93452132e-01
-5.22947669e-01 -5.95984571e-02 -2.67227679e-01 -1.37494946e+00
1.92056239e-01 -5.69686592e-01 2.35551342e-01 -1.41934454e+00
3.56027931e-01 -1.35032877e-01 -2.15239212e-01 6.37489140e-01
-1.89087704e-01 2.84186065e-01 3.88737828e-01 1.13975354e-01
-6.99752569e-01 7.90019333e-01 9.45000708e-01 -2.22003579e-01
-6.50587380e-02 -2.19745055e-01 -6.24494731e-01 6.12689734e-01
9.33324814e-01 -6.63670957e-01 -6.03044212e-01 -8.23925436e-01
9.95430946e-02 1.50285333e-01 3.34016442e-01 -1.09944284e+00
1.54482827e-01 1.41808167e-01 2.18246326e-01 -4.66856986e-01
8.60421956e-01 -8.50585222e-01 2.70886719e-02 4.02923487e-02
-5.73155165e-01 -1.09390147e-01 5.31799018e-01 7.56098092e-01
-3.48052830e-01 -2.20843747e-01 6.60293937e-01 1.94355890e-01
-4.29218620e-01 1.39583245e-01 -3.45013469e-01 1.87565446e-01
8.66932690e-01 1.49433436e-02 -4.10991430e-01 -7.57287920e-01
-3.21701735e-01 1.16405658e-01 -1.23436868e-01 5.56355417e-01
4.67206717e-01 -1.34797275e+00 -6.99848473e-01 -5.46862483e-02
-1.31811276e-02 2.16148615e-01 5.44765949e-01 6.00091875e-01
-2.63540626e-01 6.86389983e-01 9.61478725e-02 -5.80636799e-01
-1.32691193e+00 2.97036797e-01 -1.38809964e-01 1.51005515e-03
-5.63109577e-01 8.24868679e-01 3.00406098e-01 3.14230561e-01
6.16919816e-01 -1.83819771e-01 -1.44068792e-01 1.52321368e-01
7.30470359e-01 4.80900198e-01 -4.18717749e-02 -6.52165592e-01
-2.96245307e-01 4.23184782e-01 -2.58670002e-01 -2.15743229e-01
1.11514068e+00 -2.13478044e-01 3.78087997e-01 2.41024375e-01
1.33111155e+00 -3.29037517e-01 -1.51083374e+00 -3.94155294e-01
-3.26318681e-01 -2.60562599e-01 3.45350504e-01 -5.63999116e-01
-1.08750725e+00 1.25877523e+00 4.21998233e-01 9.92169007e-05
1.30947006e+00 -1.07993640e-01 1.22607434e+00 4.25559551e-01
2.95741633e-02 -9.90280449e-01 1.62649959e-01 1.83363989e-01
5.73941529e-01 -1.09349406e+00 -1.46543637e-01 -3.92325491e-01
-6.77163780e-01 9.95785892e-01 5.38229883e-01 1.81420282e-01
2.40003720e-01 -1.35699548e-02 8.47319979e-03 3.58954817e-01
-9.33442652e-01 5.93865402e-02 2.29925334e-01 5.17920792e-01
2.40962803e-01 -2.10549701e-02 8.39337111e-02 7.57493317e-01
-1.51596844e-01 -8.99690986e-02 5.05432189e-01 5.93608856e-01
-4.99144554e-01 -7.57499516e-01 -4.00640160e-01 2.74369806e-01
-7.87386894e-01 -3.16033036e-01 1.19160609e-02 4.55821186e-01
-2.40771100e-01 1.20179641e+00 3.49192739e-01 -2.48811409e-01
2.06619695e-01 1.62193373e-01 2.19259202e-01 -3.21929961e-01
-4.21052128e-01 2.37932727e-01 3.76206636e-02 -4.35734272e-01
-4.44735169e-01 -4.66908067e-01 -1.09821117e+00 -1.17675684e-01
-1.34720802e-01 1.83606744e-01 6.95414543e-01 7.01051414e-01
7.98029423e-01 4.75989103e-01 4.64984328e-01 -8.13413799e-01
-6.47877038e-01 -1.05882788e+00 -1.70005575e-01 1.91908166e-01
4.02132064e-01 -4.22505707e-01 -3.32436234e-01 3.46779227e-01] | [15.15964126586914, 4.816662311553955] |
005608d4-1b4e-47fb-b77a-f725b9aa885e | trustworthy-multi-phase-liver-tumor | 2305.05344 | null | https://arxiv.org/abs/2305.05344v2 | https://arxiv.org/pdf/2305.05344v2.pdf | Trustworthy Multi-phase Liver Tumor Segmentation via Evidence-based Uncertainty | Multi-phase liver contrast-enhanced computed tomography (CECT) images convey the complementary multi-phase information for liver tumor segmentation (LiTS), which are crucial to assist the diagnosis of liver cancer clinically. However, the performances of existing multi-phase liver tumor segmentation (MPLiTS)-based methods suffer from redundancy and weak interpretability, % of the fused result, resulting in the implicit unreliability of clinical applications. In this paper, we propose a novel trustworthy multi-phase liver tumor segmentation (TMPLiTS), which is a unified framework jointly conducting segmentation and uncertainty estimation. The trustworthy results could assist the clinicians to make a reliable diagnosis. Specifically, Dempster-Shafer Evidence Theory (DST) is introduced to parameterize the segmentation and uncertainty as evidence following Dirichlet distribution. The reliability of segmentation results among multi-phase CECT images is quantified explicitly. Meanwhile, a multi-expert mixture scheme (MEMS) is proposed to fuse the multi-phase evidences, which can guarantee the effect of fusion procedure based on theoretical analysis. Experimental results demonstrate the superiority of TMPLiTS compared with the state-of-the-art methods. Meanwhile, the robustness of TMPLiTS is verified, where the reliable performance can be guaranteed against the perturbations. | ['Xinde Li', 'Shenghong Ju', 'Wenbo Xiao', 'Yuancheng Wang', 'Quchen Zou', 'Ying Cui', 'Tianyi Xia', 'Chuanfei Hu'] | 2023-05-09 | null | null | null | null | ['tumor-segmentation'] | ['computer-vision'] | [-3.91512126e-01 -1.45587578e-01 -1.91100240e-01 -8.02230984e-02
-1.11694646e+00 -1.16833806e-01 2.66418427e-01 2.46521860e-01
1.14406668e-01 7.64107943e-01 1.63042128e-01 -2.69472629e-01
-6.13139212e-01 -3.71977746e-01 -1.33004501e-01 -1.54721260e+00
4.09093350e-02 7.20865786e-01 2.85723716e-01 3.51775259e-01
-1.47306040e-01 1.61813602e-01 -7.73389637e-01 1.97771832e-01
1.46763468e+00 1.07707632e+00 2.46420085e-01 2.11604998e-01
2.10435558e-02 6.69668496e-01 -3.41948867e-01 -3.87346148e-02
-3.43875103e-02 -4.53126997e-01 -5.58327138e-01 3.58002812e-01
-6.76560760e-01 -1.88903213e-01 -8.06552824e-03 1.45254683e+00
4.72675860e-01 -4.69575047e-01 8.43848407e-01 -1.25638509e+00
-4.38574612e-01 8.24211121e-01 -7.63733029e-01 2.65777618e-01
1.14843249e-01 3.31790656e-01 1.66689053e-01 -9.60113466e-01
1.58869743e-01 7.59979725e-01 5.57083726e-01 -1.92092881e-02
-5.46868563e-01 -5.92963755e-01 -2.63210982e-01 3.70932996e-01
-1.61114502e+00 -5.80485128e-02 6.19737625e-01 -4.48816180e-01
8.49270821e-02 2.79667974e-01 9.91688013e-01 3.65892857e-01
8.13203096e-01 8.59703362e-01 1.50906324e+00 -1.99123800e-01
5.81493191e-02 1.57722265e-01 8.92568678e-02 7.97861040e-01
8.01427662e-01 1.49406344e-01 -2.60868669e-02 -2.97265232e-01
6.25890553e-01 9.53000132e-03 -8.10631037e-01 -1.09706677e-01
-1.51890087e+00 3.96412283e-01 5.22888243e-01 6.10801518e-01
-7.37144172e-01 -9.83159766e-02 4.55470145e-01 -2.97573656e-01
4.62551355e-01 -3.78818691e-01 -2.93653477e-02 3.41274321e-01
-1.12717736e+00 -3.05473834e-01 5.88962913e-01 9.87933755e-01
1.25064969e-01 -7.14705288e-02 -4.09391224e-01 -2.91480813e-02
1.03065276e+00 7.79321671e-01 6.97809160e-01 -5.39222002e-01
-2.92337745e-01 3.14794868e-01 1.16415387e-02 -7.18560636e-01
-1.67901039e-01 -6.24301016e-01 -1.40894902e+00 -2.74086177e-01
2.37541765e-01 1.73784599e-01 -8.74457717e-01 1.07938397e+00
5.73980749e-01 6.57631159e-01 1.29311860e-01 1.11938930e+00
9.73470330e-01 3.48183721e-01 2.07955122e-01 -9.10535574e-01
1.67616737e+00 -7.26644218e-01 -1.17110002e+00 3.89040768e-01
2.06810743e-01 -8.87950003e-01 2.83773184e-01 4.81224597e-01
-1.03853810e+00 -6.78618476e-02 -9.73216116e-01 7.68169940e-01
3.82452339e-01 2.73382931e-04 4.78903979e-01 7.99362004e-01
-8.98050427e-01 1.14075713e-01 -1.16996729e+00 2.19744682e-01
4.59080368e-01 2.73423016e-01 -1.86351925e-01 -4.47916359e-01
-1.25650191e+00 1.14479744e+00 5.47604442e-01 5.17472625e-01
-1.01234686e+00 -8.24615240e-01 -7.60704219e-01 -2.62992024e-01
3.97642940e-01 -8.44852448e-01 1.14027381e+00 -3.53607655e-01
-1.34963250e+00 4.40214455e-01 -1.70405865e-01 -1.54231980e-01
6.73367798e-01 6.38099611e-01 -2.62501836e-01 8.66302967e-01
1.13265298e-01 2.16711327e-01 5.23036063e-01 -1.47279394e+00
-4.98478025e-01 -2.69578725e-01 -7.32993841e-01 2.56925553e-01
1.58394694e-01 -6.41091447e-03 -1.97015196e-01 -4.51188773e-01
7.04865336e-01 -8.27164531e-01 -3.36854160e-01 -8.56164470e-02
-4.21301067e-01 -2.77474105e-01 6.17385089e-01 -9.43925858e-01
1.09360743e+00 -1.81816769e+00 6.49698675e-02 4.69667584e-01
3.79811078e-01 -4.19318452e-02 5.07968962e-01 -1.57280460e-01
1.74870118e-01 2.38019571e-01 -4.78286624e-01 -1.37551546e-01
-1.00792393e-01 4.87546653e-01 3.20512831e-01 1.09896684e+00
-1.67667642e-01 1.07239771e+00 -1.09729767e+00 -1.30858552e+00
4.13755208e-01 4.06974912e-01 9.37604010e-02 -6.06401004e-02
6.20324127e-02 1.08720219e+00 -7.54377127e-01 1.00619745e+00
1.01782298e+00 -5.05363762e-01 1.67949006e-01 -4.83348280e-01
-2.53933575e-02 -3.52765560e-01 -8.56486380e-01 1.46964061e+00
3.62913869e-02 -1.40296638e-01 2.91517198e-01 -7.39756048e-01
4.67338383e-01 1.01132524e+00 1.11867225e+00 -2.27469102e-01
4.23089445e-01 6.17212474e-01 3.59937139e-02 -7.70048976e-01
-1.59553558e-01 -6.74930632e-01 1.08974054e-01 3.66078049e-01
-3.02747041e-01 -5.48272431e-01 -1.17477596e-01 1.89412579e-01
5.66681206e-01 -2.17085332e-01 5.34598887e-01 -7.57825613e-01
7.85381198e-01 7.32903183e-02 8.24199200e-01 1.75835013e-01
-7.10088909e-01 3.28418165e-01 1.63701952e-01 -1.98155399e-02
-6.46635056e-01 -1.02552354e+00 -6.43028915e-01 -3.13792557e-01
6.87278092e-01 1.19716913e-01 -5.74979901e-01 -6.92531526e-01
-2.19441980e-01 5.26817381e-01 -3.62228572e-01 9.24780220e-03
-2.17081115e-01 -1.38468182e+00 4.19740260e-01 3.14874977e-01
8.43747258e-01 -4.09078568e-01 -5.24748325e-01 1.74926847e-01
-7.18146801e-01 -9.72618282e-01 -3.21069986e-01 -3.20087880e-01
-8.56779218e-01 -1.34748244e+00 -1.07245374e+00 -4.57372040e-01
7.84547269e-01 3.17188203e-01 9.29815888e-01 4.82243121e-01
-3.54823843e-02 5.09328187e-01 -3.34495842e-01 -1.70269489e-01
-7.18469024e-01 -7.34247983e-01 1.19221590e-01 -1.29064515e-01
4.01860215e-02 -3.92940789e-01 -9.24002349e-01 5.05211592e-01
-9.67985213e-01 1.76691353e-01 8.69892418e-01 1.07637250e+00
9.26014602e-01 7.80259371e-01 5.30205488e-01 -4.81030256e-01
4.00698513e-01 -7.58912921e-01 -3.94851536e-01 6.02701604e-01
-9.59740877e-01 -2.99866796e-01 1.41147196e-01 -1.98293746e-01
-1.22736514e+00 -5.18591821e-01 4.89987805e-02 -5.95714569e-01
9.96232182e-02 1.06145859e+00 6.21647248e-03 -4.18301195e-01
1.25792071e-01 6.40933633e-01 2.50757396e-01 1.87021568e-01
-2.96393018e-02 6.13384068e-01 4.20582891e-01 -7.06342459e-01
3.63155842e-01 4.73822832e-01 2.53284514e-01 -2.84485668e-01
-5.78496277e-01 -6.36179388e-01 -2.27395460e-01 -5.42209268e-01
6.62524581e-01 -8.85175228e-01 -7.51249075e-01 7.15951025e-01
-9.69194770e-01 2.31640443e-01 -7.01862350e-02 1.07499945e+00
-4.47684467e-01 9.34462309e-01 -1.14606535e+00 -8.99676859e-01
-5.02384186e-01 -1.86357391e+00 8.85077178e-01 3.81415606e-01
4.25740927e-01 -1.08665979e+00 -2.99490839e-01 4.32319522e-01
4.11011100e-01 4.29822177e-01 7.81393945e-01 -5.68624556e-01
-9.95842755e-01 -1.05944127e-01 -3.09849828e-01 1.63037449e-01
1.95378941e-02 -6.07426045e-03 -5.66817999e-01 -2.97252089e-01
7.97775507e-01 1.77706674e-01 5.97385168e-01 1.04255176e+00
8.23048413e-01 -2.25286648e-01 -4.48256850e-01 3.98320109e-01
1.42819095e+00 7.51583800e-02 5.95632315e-01 3.09909582e-02
2.22616598e-01 3.42605591e-01 8.60624671e-01 6.79792881e-01
6.42121553e-01 3.57709490e-02 6.00918770e-01 1.03232510e-01
-6.78903311e-02 1.60192311e-01 -4.83235307e-02 1.37901115e+00
-1.22121148e-01 -2.03640461e-01 -1.07164240e+00 6.12552106e-01
-1.87391436e+00 -6.37940347e-01 -5.35649598e-01 1.95818901e+00
1.00136006e+00 -1.56840891e-01 -4.46156174e-01 2.35807821e-01
8.74428093e-01 -3.18975031e-01 -3.67577910e-01 7.96653092e-01
-1.34776235e-02 -1.71518371e-01 3.73887479e-01 3.02033246e-01
-7.36592948e-01 1.20498970e-01 5.62656212e+00 1.38240874e+00
-8.33207607e-01 5.35856485e-01 7.63670087e-01 6.36345923e-01
-7.38585591e-01 1.40637606e-01 -1.63093120e-01 7.68410146e-01
5.14671683e-01 -6.24901474e-01 -9.87111554e-02 2.96562016e-01
3.87744337e-01 -7.20358551e-01 -5.19930243e-01 8.55057120e-01
-5.43162599e-02 -1.01979697e+00 -2.24678144e-01 2.43367434e-01
7.81933784e-01 -2.47150987e-01 4.72653545e-02 -1.51779443e-01
1.57987908e-01 -8.51323307e-01 4.95075554e-01 9.24525440e-01
8.49775672e-01 -5.82112789e-01 1.36153674e+00 5.86189508e-01
-1.19736445e+00 1.95451036e-01 -2.21790239e-01 6.19227946e-01
5.29990256e-01 1.36922455e+00 -9.48700726e-01 1.53010428e+00
4.28473264e-01 6.90787435e-01 -2.54866660e-01 1.31196809e+00
-4.24274057e-01 4.53054368e-01 -4.66532916e-01 2.05662087e-01
-2.66371388e-02 -3.91613036e-01 8.64982545e-01 8.06006968e-01
7.03354239e-01 5.26138186e-01 5.04723966e-01 7.54114449e-01
4.45876002e-01 6.39983043e-02 9.27003473e-02 1.67187214e-01
6.43989742e-01 1.46566153e+00 -1.11962986e+00 -7.45874822e-01
-4.79881093e-03 4.50563401e-01 -6.20034575e-01 1.19232766e-01
-9.91263986e-01 4.86138105e-01 -3.29006732e-01 -9.36236978e-02
-2.57692933e-01 5.39486669e-02 -4.17516232e-01 -1.36689806e+00
-1.93237960e-01 -5.56922734e-01 6.04088366e-01 -7.94514596e-01
-1.51608872e+00 6.19560659e-01 2.49487936e-01 -1.49404311e+00
6.66814372e-02 -1.92683309e-01 -5.87333262e-01 8.61906111e-01
-1.91135824e+00 -1.61367726e+00 -4.65959847e-01 5.40896177e-01
2.13974118e-01 1.58688515e-01 5.71084797e-01 1.43127710e-01
-3.60291541e-01 1.86671302e-01 -1.93093326e-02 -1.29209921e-01
5.76677322e-01 -1.14802468e+00 -8.95700097e-01 9.22830760e-01
-6.77271545e-01 3.38377059e-01 7.26905167e-01 -9.35259938e-01
-1.16600740e+00 -7.93514490e-01 4.36823517e-01 -3.01123429e-02
5.76916039e-01 9.24558640e-01 -8.47308159e-01 4.33307916e-01
3.76414239e-01 4.19585347e-01 8.86209786e-01 -7.06924558e-01
3.25861901e-01 1.90451756e-01 -1.68085921e+00 1.64462775e-01
1.91122428e-01 -9.33171138e-02 -6.15454674e-01 3.60788673e-01
5.60411811e-01 -7.16617584e-01 -1.50965297e+00 9.51003075e-01
4.32896525e-01 -9.80088294e-01 8.39418173e-01 2.35856488e-01
1.91065222e-01 -7.26064861e-01 -1.75069794e-02 -1.21371746e+00
-1.94406360e-01 -1.38143495e-01 -1.37470663e-01 6.84760511e-01
-5.99226542e-02 -8.11141074e-01 3.71197462e-01 6.53450072e-01
-6.26413763e-01 -8.23503733e-01 -1.26380718e+00 -4.34548140e-01
1.53825153e-02 -2.19035745e-01 5.96460938e-01 9.77944791e-01
3.17627788e-01 -4.32818174e-01 3.68955694e-02 5.79389870e-01
1.27015877e+00 8.72363299e-02 -1.47803172e-01 -1.08438981e+00
5.60918711e-02 -4.76400614e-01 -2.14069203e-01 -5.91350138e-01
-1.17767546e-02 -8.75096977e-01 2.40944058e-01 -1.72466385e+00
6.19340062e-01 -8.46000493e-01 -5.77083409e-01 1.89133227e-01
-4.46219414e-01 3.43539231e-02 -7.98070654e-02 7.42101908e-01
-6.52758479e-01 6.89759374e-01 1.87200880e+00 -2.79509485e-01
3.59603822e-01 1.51074752e-01 -3.95958483e-01 6.98952377e-01
4.63005245e-01 -4.68161762e-01 -2.86200792e-01 8.88495427e-03
1.53728072e-02 9.25923228e-01 5.24793804e-01 -9.66649354e-01
7.47615635e-01 -1.72709912e-01 3.23821187e-01 -9.45101321e-01
-1.97947444e-03 -1.08829629e+00 5.62839568e-01 1.02574992e+00
3.51324916e-01 -2.56927431e-01 2.11665034e-02 7.29596317e-01
-4.16450322e-01 -4.56610322e-01 8.26386094e-01 -4.45867807e-01
-1.95108458e-01 6.06764913e-01 -2.64179707e-01 -3.46513093e-01
1.36519206e+00 -1.08374126e-01 -2.44364709e-01 -5.18633388e-02
-7.40328908e-01 5.81749797e-01 2.58425206e-01 -6.27105117e-01
9.28373814e-01 -1.42865551e+00 -1.15315163e+00 3.14089544e-02
-1.77844554e-01 4.62680668e-01 8.22041571e-01 2.05767322e+00
-6.29196286e-01 3.88308942e-01 2.06833631e-01 -1.17517936e+00
-8.82980585e-01 6.58846498e-01 5.06931961e-01 -6.69183314e-01
-5.43681681e-01 4.23434943e-01 8.86716321e-02 1.09912075e-01
-4.39596623e-01 -3.37128103e-01 -5.70348240e-02 -2.08794951e-01
3.11169386e-01 2.81968355e-01 1.02113120e-01 -6.14212394e-01
-4.88644183e-01 3.81879598e-01 1.49256289e-01 -1.89493656e-01
8.35442841e-01 -5.26393354e-01 -6.75028741e-01 1.28601879e-01
3.81286532e-01 -4.61252891e-02 -9.13310945e-01 -2.58090138e-01
-1.85608655e-01 -3.67220461e-01 4.62769359e-01 -8.91687810e-01
-1.07425594e+00 6.85447752e-01 3.46456975e-01 1.44607872e-01
1.27751529e+00 -1.15311623e-01 8.78755271e-01 -4.28197652e-01
7.91114628e-01 -4.79550689e-01 -1.15692906e-01 -1.79594979e-01
5.59795320e-01 -1.21182227e+00 3.97759765e-01 -8.02585363e-01
-7.55912364e-01 1.15603673e+00 8.38990062e-02 1.74336076e-01
1.02115071e+00 5.01469791e-01 -1.18903749e-01 -1.69375524e-01
-5.25961936e-01 2.18408648e-02 2.50804394e-01 3.16930473e-01
3.82367671e-02 3.68166089e-01 -5.89343667e-01 9.71266627e-01
5.09351850e-01 2.91760772e-01 3.78288329e-01 7.84870148e-01
-5.69377840e-01 -6.50357008e-01 -8.30842137e-01 1.90209910e-01
-4.60839331e-01 -1.77910239e-01 5.28483331e-01 6.43037796e-01
1.19777121e-01 9.79885042e-01 -5.46546400e-01 1.58529095e-02
-2.57123798e-01 -1.33254707e-01 6.00790799e-01 -1.98342487e-01
-3.58783662e-01 7.77396977e-01 -4.37205881e-01 -3.68747227e-02
-9.68498409e-01 -6.09403610e-01 -1.39791632e+00 -2.25857154e-01
-8.27466249e-01 6.92896783e-01 5.22056699e-01 1.13825226e+00
-9.39093903e-02 7.04180241e-01 7.19508708e-01 -5.40800929e-01
-7.81901300e-01 -1.04227722e+00 -7.10412741e-01 -4.33112010e-02
6.90644309e-02 -7.41345882e-01 -6.01839185e-01 3.45150456e-02] | [14.32898998260498, -2.6747381687164307] |
38c40391-68dd-42c2-a097-e418194f5326 | logically-at-the-factify-2022-multimodal-fact | 2112.09253 | null | https://arxiv.org/abs/2112.09253v2 | https://arxiv.org/pdf/2112.09253v2.pdf | Logically at Factify 2022: Multimodal Fact Verification | This paper describes our participant system for the multi-modal fact verification (Factify) challenge at AAAI 2022. Despite the recent advance in text based verification techniques and large pre-trained multimodal models cross vision and language, very limited work has been done in applying multimodal techniques to automate fact checking process, particularly considering the increasing prevalence of claims and fake news about images and videos on social media. In our work, the challenge is treated as multimodal entailment task and framed as multi-class classification. Two baseline approaches are proposed and explored including an ensemble model (combining two uni-modal models) and a multi-modal attention network (modeling the interaction between image and text pair from claim and evidence document). We conduct several experiments investigating and benchmarking different SoTA pre-trained transformers and vision models in this work. Our best model is ranked first in leaderboard which obtains a weighted average F-measure of 0.77 on both validation and test set. Exploratory analysis of dataset is also carried out on the Factify data set and uncovers salient patterns and issues (e.g., word overlapping, visual entailment correlation, source bias) that motivates our hypothesis. Finally, we highlight challenges of the task and multimodal dataset for future research. | ['Anil Bandhakavi', 'David Kiskovski', 'Stylianos Oikonomou', 'Hella-Franziska Hoffmann', 'Jie Gao'] | 2021-12-16 | null | null | null | null | ['visual-entailment'] | ['reasoning'] | [ 1.97850034e-01 2.24070236e-01 -6.43661171e-02 -2.84167498e-01
-1.31394279e+00 -8.31076682e-01 1.12732494e+00 3.19165975e-01
-3.91127974e-01 4.99896646e-01 5.22562683e-01 -4.68425572e-01
1.59794793e-01 -9.20218080e-02 -9.76428926e-01 -1.93812594e-01
2.47322679e-01 4.22536880e-01 4.71266499e-03 -1.94813475e-01
5.86172879e-01 5.76747805e-02 -1.36685967e+00 1.18808866e+00
4.64078784e-01 1.06133020e+00 -4.14648443e-01 7.29593813e-01
1.14430942e-01 1.29025888e+00 -6.47378027e-01 -1.51656175e+00
1.71443626e-01 -3.15294802e-01 -1.13236701e+00 2.82926410e-01
1.39688289e+00 -3.20161074e-01 -2.73894221e-01 1.18656075e+00
5.14481843e-01 -3.35846961e-01 5.75988173e-01 -1.40657771e+00
-1.29692769e+00 6.01021647e-01 -8.80428672e-01 4.37948763e-01
5.38877130e-01 1.81342646e-01 1.04249656e+00 -1.13042057e+00
1.00842381e+00 1.37175727e+00 7.46171832e-01 2.85379201e-01
-9.64117527e-01 -2.89755791e-01 -7.21599087e-02 6.31287396e-01
-9.80064332e-01 -4.59805191e-01 7.42692888e-01 -6.99832439e-01
7.97497034e-01 1.60410628e-01 2.84837544e-01 1.48634434e+00
1.24857076e-01 1.22726202e+00 1.42274153e+00 -3.71075392e-01
-4.23610419e-01 6.18062735e-01 3.00499052e-01 8.89979899e-01
8.94061252e-02 -8.23610462e-03 -8.32637131e-01 -3.19049090e-01
5.62997945e-02 -2.84024775e-01 -2.10895911e-01 -3.95451218e-01
-1.51037788e+00 1.08304143e+00 1.23631209e-01 3.64057779e-01
-3.12810510e-01 -4.08773050e-02 9.46239412e-01 5.21637201e-01
4.14612770e-01 1.86545804e-01 -9.48239937e-02 7.54010528e-02
-1.06156504e+00 4.44000632e-01 6.45316184e-01 5.19469261e-01
2.42785707e-01 -3.00747216e-01 -4.20958400e-01 7.12950885e-01
3.47303987e-01 4.77470666e-01 2.63578117e-01 -7.83674240e-01
9.77329969e-01 6.21404469e-01 3.26120526e-01 -1.55685413e+00
-1.93154693e-01 -8.68358687e-02 -5.70140779e-01 1.69898704e-01
4.79074746e-01 2.37041172e-02 -5.77636123e-01 1.26861513e+00
2.87399292e-01 -5.43164052e-02 1.86125964e-01 1.05539930e+00
1.03552258e+00 2.77606428e-01 1.23291440e-01 1.46170199e-01
1.91809690e+00 -9.65617120e-01 -7.18563676e-01 -1.05465621e-01
4.15726453e-01 -1.09130311e+00 8.75634789e-01 3.33465278e-01
-1.24262142e+00 -4.86534148e-01 -1.10992122e+00 -3.18826586e-01
-7.50997961e-01 4.69265997e-01 4.24405873e-01 7.92014778e-01
-7.73948789e-01 1.14288293e-01 -1.64664209e-01 -4.08662796e-01
8.11967671e-01 -1.09717324e-01 -6.01427436e-01 -2.06468552e-01
-1.18879092e+00 1.27508020e+00 3.03408056e-01 1.52518973e-01
-1.01674318e+00 -3.71957630e-01 -7.22019553e-01 -4.28491980e-01
3.98464054e-01 -4.83371943e-01 9.86720204e-01 -1.28445733e+00
-8.89567196e-01 1.61094439e+00 7.56914765e-02 -7.88083613e-01
7.66665876e-01 -1.97686195e-01 -7.43376613e-01 4.42560315e-01
1.72673270e-01 5.40982425e-01 9.19504702e-01 -1.61440861e+00
-5.34833610e-01 -4.49171752e-01 3.27332854e-01 -5.40420674e-02
-1.40328154e-01 4.94991332e-01 -6.28169328e-02 -4.69470441e-01
-2.09129333e-01 -9.14104402e-01 5.83937049e-01 -2.91994125e-01
-5.71163416e-01 -1.36356935e-01 1.13117456e+00 -1.14785779e+00
9.14058685e-01 -1.91452301e+00 -3.63107808e-02 -2.51051337e-02
3.22199523e-01 2.22373098e-01 -2.10976213e-01 5.11101782e-01
-1.62074178e-01 2.08906397e-01 -8.33835453e-02 -5.00581324e-01
1.71374798e-01 -2.14148238e-01 -5.20394385e-01 7.07023442e-01
3.21621001e-01 1.18562448e+00 -6.68490708e-01 -6.46052480e-01
-2.65171453e-02 3.28443170e-01 -2.08554730e-01 -1.13943495e-01
-1.89620137e-01 3.12622786e-01 6.37091771e-02 1.24109602e+00
7.89447188e-01 -3.72669607e-01 1.80180535e-01 -8.13012183e-01
-1.47991371e-03 -8.77352804e-02 -7.80846596e-01 1.61152291e+00
-3.11414264e-02 9.43586230e-01 9.84122381e-02 -1.11133337e+00
4.89628583e-01 5.26878059e-01 1.24050684e-01 -6.41190469e-01
3.36223722e-01 -4.13431935e-02 -1.52175620e-01 -1.05378008e+00
8.01459849e-01 -9.30685326e-02 -1.62379518e-01 5.72080255e-01
1.65770814e-01 2.25154430e-01 1.15107596e-01 5.56008160e-01
7.37615108e-01 2.19064653e-01 -1.34252474e-01 1.25541687e-01
5.91774583e-01 4.60066527e-01 -1.18156046e-01 8.87704670e-01
-7.78577149e-01 5.09196758e-01 6.76106095e-01 -5.15651107e-01
-1.16129208e+00 -6.85697436e-01 4.01357487e-02 9.03120816e-01
1.96760595e-01 -2.66744226e-01 -4.67772067e-01 -1.08087492e+00
6.97236583e-02 5.53012609e-01 -8.31411421e-01 2.19653323e-01
-2.60907322e-01 -7.47576773e-01 1.01278913e+00 1.19001411e-01
7.27174461e-01 -9.34811473e-01 -6.07919633e-01 -1.99406341e-01
-7.56516576e-01 -1.44819582e+00 -2.40291297e-01 -5.43597341e-01
-2.74977207e-01 -1.55154443e+00 -4.95627671e-01 -6.00518227e-01
1.49171010e-01 3.37489098e-01 1.16775477e+00 1.78766876e-01
-2.44682372e-01 8.97393942e-01 -2.57020414e-01 -4.91885722e-01
-4.83523041e-01 -5.55951416e-01 -3.41240078e-01 4.85989153e-01
5.38681567e-01 2.44890228e-01 -4.91961420e-01 1.08033553e-01
-9.33945239e-01 5.61799072e-02 5.68969369e-01 1.05920649e+00
1.55899331e-01 -2.89687604e-01 4.08107132e-01 -6.15862787e-01
8.86873960e-01 -6.57093585e-01 -2.26029322e-01 6.36474133e-01
-3.72696757e-01 -5.64836144e-01 -1.33466616e-01 -4.33918327e-01
-1.07292831e+00 -2.57592261e-01 3.75068486e-01 -5.12317836e-01
-2.31942251e-01 6.65715456e-01 1.83222190e-01 -1.73337087e-01
5.54762721e-01 -1.39316532e-03 1.40627980e-01 -1.69127211e-02
6.04090929e-01 7.51311779e-01 6.86587512e-01 -3.72369319e-01
6.57386839e-01 7.62406051e-01 -3.08419555e-01 -5.34427524e-01
-1.08125579e+00 -3.94027293e-01 -5.67308247e-01 -5.96103966e-01
1.00919628e+00 -1.00630474e+00 -1.09094882e+00 4.04704899e-01
-1.56101382e+00 3.01243931e-01 3.55172843e-01 3.06431413e-01
-3.05199236e-01 8.18230271e-01 -7.77703285e-01 -1.11944628e+00
-2.96240300e-01 -1.09164846e+00 1.14818192e+00 -3.41266483e-01
-1.42056450e-01 -8.95749092e-01 3.00471690e-02 1.45197761e+00
2.32052878e-01 5.93445659e-01 8.46336305e-01 -8.84985626e-01
-5.92959762e-01 -5.55329680e-01 -6.57583892e-01 2.36963436e-01
-5.49171627e-01 -5.43014845e-03 -1.18701339e+00 -2.53096893e-02
-8.03272054e-02 -9.98822629e-01 1.04633260e+00 5.55446260e-02
6.14317179e-01 -3.18106979e-01 -1.46861479e-01 -3.14263135e-01
1.39469016e+00 -3.22238237e-01 6.32175267e-01 6.13191724e-01
6.95169032e-01 8.93698394e-01 6.17083549e-01 1.41193330e-01
7.91590452e-01 6.66344285e-01 8.31363738e-01 4.62331399e-02
-9.93104503e-02 -7.69074038e-02 4.45325851e-01 1.70377985e-01
-8.89154002e-02 -2.52939492e-01 -1.04837120e+00 7.46636033e-01
-2.10085273e+00 -1.63979948e+00 -5.11402130e-01 1.84690464e+00
4.12715793e-01 -1.06468223e-01 4.37689304e-01 7.15903938e-02
8.59158099e-01 7.82883912e-02 -3.78966704e-02 -4.44155484e-01
-7.16483116e-01 -5.36515296e-01 2.26394445e-01 4.46774065e-01
-1.56752193e+00 7.01527357e-01 6.03338146e+00 7.44478464e-01
-8.32686365e-01 5.39807141e-01 7.86276817e-01 -5.26817106e-02
-2.58050531e-01 -2.97833271e-02 -4.10026491e-01 3.61338288e-01
8.26918781e-01 4.65472251e-01 2.03007504e-01 4.53447551e-01
-7.31293932e-02 -3.39122057e-01 -8.31009924e-01 1.01274669e+00
8.43495131e-01 -1.62680757e+00 4.54974994e-02 5.84084727e-02
7.75753021e-01 2.26527885e-01 1.89401612e-01 3.04063380e-01
4.04367782e-02 -9.14203286e-01 1.21618259e+00 4.53941911e-01
4.34717447e-01 -5.23266613e-01 1.07162416e+00 4.16885875e-02
-5.48493862e-01 -1.91759750e-01 6.96348771e-02 2.40583032e-01
3.27955276e-01 4.60206568e-01 -6.02979124e-01 7.95811415e-01
7.52115667e-01 5.93285978e-01 -8.66945982e-01 7.23801076e-01
7.67822564e-02 3.28504264e-01 1.02946155e-01 1.11367427e-01
3.52324605e-01 8.71946141e-02 6.06674135e-01 1.44554210e+00
-2.76729912e-02 -3.36315691e-01 1.65528253e-01 8.35000515e-01
-1.89784661e-01 -2.45206803e-02 -6.73510492e-01 -2.55706549e-01
-4.11705151e-02 1.51040757e+00 -4.85833257e-01 -6.32670581e-01
-8.17666113e-01 1.21798933e+00 3.87713104e-01 2.86931604e-01
-1.27466440e+00 2.27045432e-01 3.16547342e-02 -1.48801222e-01
3.67733717e-01 -7.96923041e-02 -3.17566514e-01 -1.36341596e+00
7.95011148e-02 -1.17196298e+00 7.50222504e-01 -1.13243711e+00
-1.57676196e+00 6.08233273e-01 -1.27679855e-01 -1.05824482e+00
2.15081364e-01 -7.30820417e-01 -4.11426723e-01 5.22499263e-01
-1.57678044e+00 -1.79013908e+00 -1.61053702e-01 8.24934423e-01
2.67391801e-01 -3.53335798e-01 6.04015768e-01 4.47395444e-01
-4.04301912e-01 5.55792928e-01 -2.78455496e-01 2.36455157e-01
8.59838367e-01 -9.47282374e-01 -2.00346470e-01 8.08624744e-01
4.07365263e-01 4.86943305e-01 7.42899954e-01 -7.58551896e-01
-1.54857183e+00 -6.87356770e-01 1.11714900e+00 -1.03147376e+00
1.18544352e+00 8.71737394e-03 -6.92074895e-01 6.75894856e-01
8.51869524e-01 -2.81091362e-01 8.43195677e-01 7.27494955e-02
-8.76639903e-01 4.33479995e-01 -1.36486697e+00 4.69995409e-01
5.35021007e-01 -1.00171781e+00 -7.43058980e-01 5.97324014e-01
6.22815341e-02 -3.45772982e-01 -7.74541855e-01 3.10799718e-01
7.97332525e-01 -1.25598979e+00 9.48430121e-01 -9.24719512e-01
1.04812443e+00 -2.74922132e-01 -3.66395354e-01 -7.31141329e-01
3.02767269e-02 -2.76600778e-01 -2.74521351e-01 1.39010775e+00
5.99750221e-01 -2.59414643e-01 4.23518866e-01 4.66364443e-01
1.92469954e-01 -4.74631935e-01 -9.85762417e-01 -4.60164636e-01
-1.56618848e-01 -5.48154473e-01 -7.28469193e-02 1.35023916e+00
1.74941078e-01 5.45222282e-01 -8.06009233e-01 2.62230486e-01
5.71198225e-01 4.11849499e-01 8.33002567e-01 -8.04849327e-01
-1.72305629e-01 -3.53381485e-01 -4.80197996e-01 -4.98010486e-01
1.61743134e-01 -8.94114435e-01 -4.29076552e-01 -1.22667360e+00
7.88306952e-01 3.52667689e-01 2.68216264e-02 4.13858324e-01
-6.34513944e-02 6.50724411e-01 3.98031443e-01 4.00224745e-01
-9.01001871e-01 1.36276871e-01 9.84183133e-01 -5.88015318e-01
5.94466627e-01 -4.26695526e-01 -7.10465968e-01 5.97371042e-01
6.68395162e-01 -1.64415032e-01 4.09278572e-02 -4.85494286e-01
3.66999179e-01 -2.07511131e-02 1.23908615e+00 -5.14041960e-01
4.08529751e-02 7.86281675e-02 3.65024239e-01 -9.15583074e-01
4.43369031e-01 -8.70143652e-01 -1.73698425e-01 2.69564718e-01
-4.23360676e-01 3.27914536e-01 2.28288531e-01 7.67415583e-01
-3.40615362e-01 -1.85571164e-01 4.20038819e-01 -3.28846544e-01
-6.77048802e-01 -1.26109079e-01 -3.14966500e-01 -7.32396916e-02
9.31107938e-01 -2.78175801e-01 -9.53433990e-01 -5.39897621e-01
-7.29280353e-01 1.20833077e-01 1.55019328e-01 6.08283460e-01
6.30573213e-01 -1.42719340e+00 -1.05799818e+00 -3.37557137e-01
4.63495255e-01 -1.03125823e+00 4.82831061e-01 1.31425154e+00
-1.84699178e-01 4.90122229e-01 -9.10859182e-02 -4.65249151e-01
-1.61067808e+00 7.91484892e-01 2.55102426e-01 -2.95694023e-01
-1.86950728e-01 6.82830811e-01 -4.80172873e-01 -2.70097613e-01
7.60557950e-02 1.55201405e-01 -2.19676673e-01 3.92190844e-01
6.03953660e-01 4.19570506e-01 1.19233243e-02 -1.24320281e+00
-4.17280018e-01 3.62402827e-01 -1.36005327e-01 -3.17261696e-01
1.06346118e+00 -4.33141738e-01 -2.65713841e-01 4.05900955e-01
1.16858959e+00 4.29846644e-02 -5.98941565e-01 -7.88592771e-02
1.51647404e-01 -3.95023793e-01 -9.76784974e-02 -1.32988667e+00
-8.83476853e-01 8.01034391e-01 7.89495587e-01 6.15030408e-01
5.53281665e-01 2.73282170e-01 6.39894426e-01 2.27988854e-01
-9.62214991e-02 -1.24308574e+00 3.79236370e-01 4.52193737e-01
1.31408930e+00 -1.94319010e+00 4.06155251e-02 -7.52670318e-02
-1.22877896e+00 9.64718461e-01 4.67824399e-01 1.97673067e-02
2.04469293e-01 -1.79780751e-01 3.25268328e-01 -9.77193117e-01
-5.47426224e-01 -6.31810427e-02 6.30303800e-01 4.74653542e-01
5.64825833e-01 -2.22310886e-01 -3.78186285e-01 4.41815972e-01
1.21696919e-01 -1.26515582e-01 4.70697314e-01 7.24373639e-01
-1.35932133e-01 -5.90088487e-01 -7.12591946e-01 1.89394191e-01
-9.64283526e-01 -1.13129079e-01 -8.23685706e-01 8.28757048e-01
2.74320275e-01 1.32720375e+00 -3.00076723e-01 -3.76294494e-01
1.85775772e-01 2.88556814e-01 6.05737090e-01 -1.29350908e-02
-9.72160161e-01 -1.88808143e-01 6.66854024e-01 -5.37162960e-01
-1.02802062e+00 -7.75345385e-01 -5.49947977e-01 -4.02222514e-01
-5.11316180e-01 -3.03793699e-01 9.95029449e-01 1.10270429e+00
3.69056135e-01 9.14391130e-02 4.65807617e-02 -7.86289036e-01
-3.10219854e-01 -8.80434632e-01 -1.22688830e-01 9.80422556e-01
3.73421639e-01 -5.26473463e-01 -2.16117442e-01 3.92091691e-01] | [8.272506713867188, 10.332731246948242] |
0deadd79-2060-4d39-9abb-d0601b636ded | interpretable-gait-recognition-by-granger | 2206.06714 | null | https://arxiv.org/abs/2206.06714v3 | https://arxiv.org/pdf/2206.06714v3.pdf | Interpretable Gait Recognition by Granger Causality | Which joint interactions in the human gait cycle can be used as biometric characteristics? Most current methods on gait recognition suffer from the lack of interpretability. We propose an interpretable feature representation of gait sequences by the graphical Granger causal inference. Gait sequence of a person in the standardized motion capture format, constituting a set of 3D joint spatial trajectories, is envisaged as a causal system of joints interacting in time. We apply the graphical Granger model (GGM) to obtain the so-called Granger causal graph among joints as a discriminative and visually interpretable representation of a person's gait. We evaluate eleven distance functions in the GGM feature space by established classification and class-separability evaluation metrics. Our experiments indicate that, depending on the metric, the most appropriate distance functions for the GGM are the total norm distance and the Ky-Fan 1-norm distance. Experiments also show that the GGM is able to detect the most discriminative joint interactions and that it outperforms five related interpretable models in correct classification rate and in Davies-Bouldin index. The proposed GGM model can serve as a complementary tool for gait analysis in kinesiology or for gait recognition in video surveillance. | ['Claudia Plant', 'Petr Sojka', 'Katerina Hlavackova-Schindler', 'Michal Balazia'] | 2022-06-14 | null | null | null | null | ['gait-recognition'] | ['computer-vision'] | [-1.88129544e-01 -1.97404861e-01 -1.88575804e-01 -1.21567830e-01
-1.94522321e-01 -2.97882766e-01 7.09766746e-01 -3.41377384e-03
-2.27522179e-01 7.83195555e-01 4.97572184e-01 -1.88195437e-01
-7.51717448e-01 -5.43938100e-01 -2.91142285e-01 -8.27736497e-01
-8.24581921e-01 3.91704649e-01 1.73505366e-01 -2.82352298e-01
2.04325378e-01 6.19482636e-01 -1.61464798e+00 -8.55726376e-02
6.62156403e-01 4.74342912e-01 -2.76081949e-01 8.42962444e-01
4.97594476e-01 2.85613775e-01 -6.18919313e-01 -1.83532119e-01
-1.25403732e-01 -7.49874413e-01 -6.28970981e-01 1.35966808e-01
7.32052252e-02 5.37763536e-02 -1.90266669e-01 7.35585213e-01
2.58938074e-01 -1.92382764e-02 1.30562437e+00 -1.63388240e+00
-4.31096286e-01 2.42799148e-02 -4.77953106e-01 3.61997098e-01
8.41738462e-01 3.81789245e-02 8.78816485e-01 -8.02064300e-01
4.85174388e-01 1.63881433e+00 6.15744233e-01 2.50017077e-01
-1.20569038e+00 -6.61822259e-02 -3.12988371e-01 8.38667452e-01
-1.22230148e+00 7.01670721e-02 6.59212708e-01 -7.68957496e-01
6.26938999e-01 6.65613294e-01 9.77096677e-01 9.20128942e-01
7.83738852e-01 5.41095972e-01 9.60544407e-01 -5.49344718e-01
1.67394891e-01 -8.74960482e-01 4.02903020e-01 9.69987929e-01
8.48035872e-01 -6.24768846e-02 -6.27637744e-01 -4.24225599e-01
9.95805323e-01 -1.95123479e-01 -1.85553327e-01 -2.29779333e-01
-1.24339807e+00 6.79369211e-01 1.06489293e-01 4.13624644e-01
-3.66744578e-01 5.54285467e-01 2.78006911e-01 1.54707342e-01
1.88211873e-01 -5.42625450e-02 3.07874769e-01 -3.29275906e-01
-6.84024751e-01 5.58061063e-01 6.93741143e-01 3.74493927e-01
3.18770617e-01 1.30053744e-01 -1.14248782e-01 2.11501345e-01
7.57642746e-01 7.16154099e-01 3.51504058e-01 -6.46526575e-01
3.15141588e-01 8.57805848e-01 2.85268463e-02 -1.41103971e+00
-6.07633471e-01 9.19390991e-02 -7.44068503e-01 3.11784625e-01
7.16910362e-01 1.78936824e-01 -6.46004200e-01 1.34009218e+00
1.67884707e-01 4.09037113e-01 -4.28195864e-01 1.03984189e+00
3.46426308e-01 2.70711869e-01 1.44063354e-01 -8.86948705e-02
1.67643368e+00 -1.98472831e-02 -8.38002026e-01 2.82692522e-01
5.84405303e-01 -6.40722990e-01 7.34552622e-01 3.58485937e-01
-7.29873776e-01 -3.72716159e-01 -1.32470822e+00 3.00401747e-01
-2.87709326e-01 4.70846832e-01 2.43491337e-01 8.70497942e-01
-8.61554980e-01 7.18043447e-01 -1.13007808e+00 -6.38147891e-01
-1.82687402e-01 2.12653905e-01 -7.63312638e-01 2.19775021e-01
-1.17781627e+00 1.02006555e+00 4.33259517e-01 6.17366016e-01
-4.14700091e-01 1.66834459e-01 -7.54356325e-01 -2.69645512e-01
-1.56599596e-01 -7.82936215e-01 3.42039376e-01 -2.34184802e-01
-1.00479424e+00 6.63192630e-01 -1.24271438e-01 -3.73546481e-01
6.65662944e-01 -2.41244107e-01 -5.15744686e-01 3.38351607e-01
6.75491169e-02 -8.56795684e-02 9.95059550e-01 -1.10102367e+00
-3.67725223e-01 -6.19976878e-01 -3.37227553e-01 6.64226785e-02
6.94780890e-03 -2.49174610e-01 -1.12000637e-01 -8.22389841e-01
4.02414232e-01 -1.15096641e+00 -4.00059437e-03 -1.26916215e-01
-5.87942719e-01 -4.35379416e-01 1.08654225e+00 -9.54234958e-01
1.56887889e+00 -1.89521658e+00 3.48664165e-01 6.07582390e-01
3.57032150e-01 4.82669473e-02 1.17500998e-01 5.51197648e-01
-3.15524012e-01 1.52564839e-01 -2.61008531e-01 -5.36128543e-02
1.31812796e-01 5.30357420e-01 9.17887539e-02 1.07861936e+00
2.98635066e-01 6.32715046e-01 -1.03248310e+00 -7.23692060e-01
3.75568062e-01 3.70515674e-01 -1.37089446e-01 -4.18554656e-02
3.73459995e-01 4.96131808e-01 -5.91320932e-01 4.44217831e-01
3.57389659e-01 1.32162601e-01 2.39870146e-01 -3.03353578e-01
1.22062251e-01 -2.14405760e-01 -1.39261544e+00 1.19483531e+00
1.37127087e-01 7.62099504e-01 -6.30459666e-01 -1.19909394e+00
1.16564894e+00 4.16823626e-01 5.82203746e-01 -3.46698672e-01
-2.01023817e-01 3.80934834e-01 -5.19445278e-02 -8.14315557e-01
2.89316624e-01 1.17822647e-01 -2.70170480e-01 1.65559798e-01
-8.13769847e-02 5.55234611e-01 4.70240504e-01 -2.22793162e-01
1.21304500e+00 2.82709181e-01 6.73777938e-01 -6.27085686e-01
7.88602173e-01 -2.56044745e-01 4.66976106e-01 4.01173353e-01
-2.68150717e-01 4.63240743e-01 7.85446405e-01 -5.18442810e-01
-8.61649692e-01 -1.61501467e+00 1.32043967e-02 4.82901365e-01
1.55687714e-02 -3.65886152e-01 -6.72103882e-01 -4.49012429e-01
2.03139216e-01 2.73710430e-01 -6.58087611e-01 -2.88841993e-01
-8.35796237e-01 -8.31207037e-01 9.47453976e-01 5.30973971e-01
5.34585893e-01 -5.84091425e-01 -8.65829110e-01 4.43553664e-02
-4.77509528e-01 -6.69959724e-01 -2.07243666e-01 -2.82858342e-01
-1.02039981e+00 -1.43148196e+00 -6.53393090e-01 -3.63397002e-01
4.97681737e-01 -1.47276312e-01 9.31145191e-01 1.57005742e-01
-5.16821265e-01 6.25944674e-01 -4.09276515e-01 2.34443545e-01
-3.25484216e-01 -5.58906972e-01 4.40411299e-01 3.05163354e-01
3.81486952e-01 -6.32355034e-01 -8.50323617e-01 5.40786922e-01
-6.47460282e-01 -3.63242477e-01 2.33959347e-01 7.33517408e-01
2.34775484e-01 1.57815702e-02 3.90684158e-01 -8.75493512e-02
8.14783037e-01 -5.74244618e-01 -9.98319760e-02 2.28889152e-01
-3.72258633e-01 4.12550420e-01 3.17539662e-01 2.60088537e-02
-7.13707030e-01 -1.35737360e-01 2.08365321e-01 -1.49785534e-01
-1.47972763e-01 6.53987527e-01 -2.13628739e-01 7.64411241e-02
7.62136817e-01 1.13143481e-01 -2.54143886e-02 -5.70980608e-01
3.15334231e-01 2.65005589e-01 5.98878324e-01 -6.67770982e-01
6.78373992e-01 5.97530007e-01 8.76663983e-01 -1.27796078e+00
3.73435855e-01 -5.44872224e-01 -9.52909708e-01 -8.10022175e-01
1.12731433e+00 -3.56733561e-01 -1.09050274e+00 2.38910496e-01
-1.27977788e+00 3.96541089e-01 3.78639847e-02 7.84944534e-01
-8.28680456e-01 8.65368128e-01 -4.33537573e-01 -1.16591179e+00
4.49488126e-02 -1.04875004e+00 1.32087040e+00 -2.00133175e-01
-6.84101880e-01 -1.28268504e+00 2.28756517e-01 1.80542424e-01
-4.53589082e-01 1.15625155e+00 1.11675727e+00 -2.69108027e-01
-2.11241066e-01 -5.02773821e-01 1.33243054e-01 -4.99195643e-02
2.39132121e-01 3.45430940e-01 -4.05585676e-01 -1.44840265e-02
-1.61708489e-01 4.73506331e-01 8.36105824e-01 5.10266483e-01
4.73110855e-01 -3.48174661e-01 -4.52231675e-01 4.70592380e-02
1.26349819e+00 4.42795694e-01 7.88929999e-01 3.58360112e-01
7.38214016e-01 7.97825336e-01 7.79949486e-01 4.53766197e-01
1.90031990e-01 9.55970347e-01 3.62028152e-01 2.22198457e-01
-6.16495349e-02 -2.02833846e-01 6.38882577e-01 7.99083531e-01
-9.33418989e-01 -2.00658128e-01 -1.01364779e+00 3.47055644e-01
-2.36445141e+00 -1.15366316e+00 -9.87558663e-01 2.40825963e+00
1.47401199e-01 2.43365392e-02 6.38470948e-01 7.27298617e-01
8.90009224e-01 -1.14579350e-01 8.96691754e-02 -3.35346818e-01
-5.22084273e-02 -3.81353335e-03 3.95789206e-01 5.75287998e-01
-1.01448095e+00 6.18142299e-02 6.27180338e+00 6.22814476e-01
-5.71776450e-01 -1.71989948e-01 1.02721661e-01 5.27536273e-01
-7.33415410e-02 -3.85725349e-02 -3.80726099e-01 6.02141440e-01
9.57796812e-01 -3.01827610e-01 -3.74288678e-01 4.34260845e-01
8.89943600e-01 -3.85646164e-01 -1.08639228e+00 1.07661521e+00
-1.80163160e-01 -9.19618607e-01 3.77082139e-01 4.64616001e-01
1.03139073e-01 -8.83017421e-01 1.67217571e-02 -5.36191821e-01
-1.92042798e-01 -1.17197037e+00 5.30462861e-01 1.11814463e+00
4.76354212e-01 -7.29920208e-01 6.72182858e-01 -3.36067528e-02
-1.83385146e+00 1.34303138e-01 1.10302664e-01 -1.93201393e-01
2.69761533e-01 3.97676170e-01 -8.28951478e-01 9.54839349e-01
2.99984843e-01 8.64245474e-01 -7.74574220e-01 1.23865378e+00
-1.37531370e-01 7.88179219e-01 -3.20974857e-01 -2.80654371e-01
1.68237254e-01 -6.44261837e-01 9.80126619e-01 1.41384840e+00
4.47921634e-01 -1.12250391e-02 -1.49724782e-01 6.65074170e-01
8.07610273e-01 -6.58607259e-02 -7.25826502e-01 5.93413459e-03
1.78625152e-01 6.31943345e-01 -1.13456309e+00 -8.99270102e-02
3.97131927e-02 1.02274382e+00 -3.86888355e-01 2.81197071e-01
-7.33256340e-01 -4.87771556e-02 7.70437479e-01 2.06825912e-01
-2.61018574e-01 -7.33405173e-01 -3.00467432e-01 -1.07537138e+00
2.95906812e-01 -3.72789532e-01 6.33177757e-01 -6.36471331e-01
-1.23346353e+00 3.11261296e-01 6.47088230e-01 -1.67893791e+00
-6.90627098e-01 -9.20450449e-01 -6.04856133e-01 8.27529192e-01
-4.83052164e-01 -1.01910233e+00 -1.43487260e-01 7.00013876e-01
2.76213139e-01 -5.91162592e-02 8.33351612e-01 2.31475353e-01
-3.52179229e-01 1.43764287e-01 1.41956909e-02 1.62390277e-01
2.63204515e-01 -1.57486260e+00 4.19426978e-01 9.77794051e-01
1.07489899e-01 6.69644415e-01 1.28503144e+00 -8.84149730e-01
-1.42981803e+00 -6.94330215e-01 9.71771300e-01 -6.36541605e-01
6.85705721e-01 1.70117486e-02 -6.89647973e-01 5.72210014e-01
-4.19510514e-01 -2.05660746e-01 4.67963517e-01 -1.11253656e-01
5.86825050e-02 1.21331345e-02 -9.22989488e-01 7.79398024e-01
1.13700843e+00 -2.26207644e-01 -9.86468852e-01 1.55939460e-01
-6.07111081e-02 2.44788215e-01 -1.14142978e+00 2.18659148e-01
9.24324036e-01 -1.04028893e+00 1.11582208e+00 -5.25978863e-01
1.67407542e-01 -6.69397593e-01 6.83455765e-02 -1.08966041e+00
-3.84853125e-01 -3.86639684e-01 -1.73301607e-01 8.45334649e-01
-2.02520460e-01 -6.67260766e-01 8.38896990e-01 1.54139906e-01
2.20882133e-01 -4.64689136e-01 -1.50217628e+00 -1.13492024e+00
-3.01931143e-01 -5.11917889e-01 1.72904119e-01 5.71696579e-01
3.11226428e-01 4.81512696e-02 -4.87525791e-01 1.80905148e-01
1.27111042e+00 -4.32774782e-01 5.64635992e-01 -1.40922892e+00
-1.93923607e-01 -3.69080633e-01 -1.66200829e+00 -4.50124055e-01
-1.38748333e-01 -5.85084260e-01 -2.25914791e-01 -1.56454039e+00
-2.86160022e-01 5.99260330e-02 -1.27800945e-02 1.05449155e-01
-2.69002840e-02 2.12145969e-01 2.86606789e-01 3.09374839e-01
-8.15099478e-02 3.65580529e-01 1.01670325e+00 1.08927347e-01
4.18420248e-02 3.43632847e-02 3.58246028e-01 1.00221992e+00
4.22919303e-01 -2.44198352e-01 -3.97146106e-01 2.53826767e-01
9.32773054e-02 4.72763270e-01 9.03756499e-01 -1.13225877e+00
-9.56136882e-02 -1.30228952e-01 3.83794546e-01 -5.97823441e-01
4.28563029e-01 -7.60456562e-01 6.62089288e-01 8.49501312e-01
1.88598186e-01 4.65978831e-01 -1.74598709e-01 8.98086071e-01
-2.48792574e-01 -1.15318790e-01 3.42586160e-01 1.84529871e-01
-9.30742025e-01 -1.26916379e-01 -7.48600841e-01 -3.97295773e-01
8.27312708e-01 -8.30919921e-01 -1.32114336e-01 -3.37549895e-01
-1.11416900e+00 -1.70094743e-01 9.31945667e-02 3.56596082e-01
9.21028376e-01 -1.76020086e+00 -7.08335042e-01 -1.57594115e-01
1.43363193e-01 -8.91419351e-01 5.17508686e-02 1.08551395e+00
-9.40588593e-01 5.97547948e-01 -5.67429721e-01 -8.68078232e-01
-1.54986322e+00 -2.72155050e-02 3.45329702e-01 -3.02849501e-01
-5.66673458e-01 1.31049007e-01 -1.37712225e-01 3.64654034e-01
-3.08290958e-01 -4.90676701e-01 -5.41986942e-01 9.13556144e-02
2.56316066e-01 1.10449016e+00 -9.34637114e-02 -1.10874724e+00
-7.32422113e-01 7.26537168e-01 7.43955135e-01 -4.40371484e-01
8.97002161e-01 -1.34340882e-01 -2.13723749e-01 8.56384695e-01
1.12393606e+00 -3.90387982e-01 -9.96103346e-01 5.75122654e-01
5.88685513e-01 -4.61781383e-01 -6.47044122e-01 -2.84933537e-01
-4.40481514e-01 6.99578047e-01 9.53257799e-01 5.78797281e-01
1.00848603e+00 -2.95274645e-01 2.53204793e-01 1.59362555e-01
4.71402735e-01 -7.28816748e-01 4.09152173e-02 1.21741273e-01
1.11788535e+00 -5.91391504e-01 1.04708374e-01 -5.90680718e-01
-3.56463909e-01 1.76219273e+00 -1.45753892e-02 -5.89139462e-01
8.15601289e-01 -1.70268312e-01 -3.93579751e-01 -2.99122840e-01
-2.29264915e-01 -4.53819782e-01 7.77822554e-01 7.97005534e-01
5.12340724e-01 5.54346859e-01 -1.27132547e+00 2.91266084e-01
-2.20047250e-01 1.75731257e-02 5.21160603e-01 7.97272086e-01
-4.08901662e-01 -9.30908322e-01 -9.68886852e-01 2.64801651e-01
3.13809030e-02 4.76178557e-01 -2.71020442e-01 1.05596781e+00
1.26784667e-01 1.07025373e+00 6.02156948e-03 -5.56604743e-01
5.28287172e-01 1.50388449e-01 6.58814371e-01 -4.82349917e-02
1.05384029e-01 -1.33396760e-01 3.32226157e-01 -5.97436547e-01
-7.43497610e-01 -8.49578977e-01 -1.30577469e+00 -4.46720541e-01
1.24405585e-01 2.21983477e-01 4.17307347e-01 1.40290093e+00
-7.73924664e-02 7.55120665e-02 3.05045635e-01 -8.12877297e-01
-1.60390496e-01 -7.81219482e-01 -8.72870564e-01 6.83657944e-01
3.64533931e-01 -1.31332278e+00 -2.99424350e-01 5.25569737e-01] | [14.207072257995605, 1.4591526985168457] |
de5227f3-dc76-478f-bc5e-c18367c5660e | amee-a-robust-framework-for-explanation | 2306.05501 | null | https://arxiv.org/abs/2306.05501v1 | https://arxiv.org/pdf/2306.05501v1.pdf | AMEE: A Robust Framework for Explanation Evaluation in Time Series Classification | This paper aims to provide a framework to quantitatively evaluate and rank explanation methods for the time series classification task, which deals with a prevalent data type in critical domains such as healthcare and finance. The recent surge of research interest in explanation methods for time series classification has provided a great variety of explanation techniques. Nevertheless, when these explanation techniques disagree on a specific problem, it remains unclear which of them to use. Comparing the explanations to find the right answer is non-trivial. Two key challenges remain: how to quantitatively and robustly evaluate the informativeness (i.e., relevance for the classification task) of a given explanation method, and how to compare explanation methods side-by-side. We propose AMEE, a Model-Agnostic Explanation Evaluation framework for quantifying and comparing multiple saliency-based explanations for time series classification. Perturbation is added to the input time series guided by the saliency maps (i.e., importance weights for each point in the time series). The impact of perturbation on classification accuracy is measured and used for explanation evaluation. The results show that perturbing discriminative parts of the time series leads to significant changes in classification accuracy. To be robust to different types of perturbations and different types of classifiers, we aggregate the accuracy loss across perturbations and classifiers. This allows us to objectively quantify and rank different explanation methods. We provide a quantitative and qualitative analysis for synthetic datasets, a variety of UCR benchmark datasets, as well as a real-world dataset with known expert ground truth. | ['Georgiana Ifrim', 'Thach Le Nguyen', 'Thu Trang Nguyen'] | 2023-06-08 | null | null | null | null | ['time-series-classification'] | ['time-series'] | [ 4.89834309e-01 3.69970016e-02 -2.57468075e-01 -3.35818380e-01
-4.60865170e-01 -6.67128444e-01 6.94459856e-01 6.80282891e-01
6.23734184e-02 4.76074308e-01 3.40849042e-01 -3.91559094e-01
-4.88527834e-01 -3.65099162e-01 -4.72758234e-01 -6.71349764e-01
-2.23269895e-01 1.58319965e-01 -7.86035322e-03 -3.63567561e-01
6.57090068e-01 3.70300353e-01 -1.89074135e+00 2.46491164e-01
9.96739089e-01 1.08895910e+00 -4.85570915e-02 4.50021029e-01
1.31875932e-01 3.46935004e-01 -8.26362252e-01 -2.86514372e-01
8.55104327e-02 -5.00828564e-01 -7.99518347e-01 2.89109610e-02
1.03987016e-01 3.29796761e-01 3.74920368e-02 9.90884662e-01
3.35328668e-01 3.00819367e-01 8.07721138e-01 -1.78125930e+00
-6.76654816e-01 5.25792718e-01 -2.17919588e-01 7.19288290e-01
4.29491580e-01 1.81086119e-02 1.06443322e+00 -6.77242875e-01
3.31136465e-01 1.06360376e+00 6.15943253e-01 5.76889992e-01
-1.25670862e+00 -4.96203154e-01 2.64453888e-01 6.38931334e-01
-8.08536470e-01 3.23899053e-02 9.00993049e-01 -3.99336010e-01
6.63949728e-01 7.28276849e-01 5.82343638e-01 1.05067456e+00
4.17404443e-01 5.97267747e-01 1.02009261e+00 -1.90310717e-01
4.72817093e-01 5.91822602e-02 4.73388463e-01 2.89929271e-01
2.53725320e-01 1.48161098e-01 -5.93409240e-01 -1.62685454e-01
3.12376618e-01 3.94231081e-01 -5.83222091e-01 -2.49406531e-01
-1.42741513e+00 8.67340624e-01 5.29780567e-01 3.13602954e-01
-4.18061733e-01 8.38478878e-02 5.17791569e-01 3.02665114e-01
4.50764656e-01 1.02644753e+00 -6.71173096e-01 -3.02161902e-01
-5.69430888e-01 3.60681623e-01 4.19824094e-01 5.98442316e-01
4.38423276e-01 9.18620378e-02 -6.07228100e-01 4.13640797e-01
-2.30437726e-01 3.54414970e-01 9.22497928e-01 -6.87760592e-01
2.24252969e-01 7.41353393e-01 1.73713028e-01 -1.31224406e+00
-5.22951245e-01 -6.35753453e-01 -7.86061764e-01 1.01721041e-01
4.14750457e-01 1.06506333e-01 -9.63666737e-01 1.96368039e+00
1.58121839e-01 4.28470045e-01 6.16591685e-02 1.32623398e+00
8.09365273e-01 4.57639575e-01 -9.65585038e-02 -3.06949198e-01
1.46717131e+00 -9.86415088e-01 -7.75584102e-01 -2.10684568e-01
4.79337662e-01 -5.98172545e-01 1.42455077e+00 1.80172890e-01
-8.74653459e-01 -3.85550112e-01 -1.04796088e+00 2.22734615e-01
-6.20990038e-01 -2.52483875e-01 5.64479709e-01 1.87789395e-01
-4.77760077e-01 8.44746411e-01 -8.04282725e-01 -3.64734501e-01
1.26477972e-01 1.88653350e-01 -1.65579095e-01 2.07835287e-01
-1.38544106e+00 9.81247485e-01 9.04014707e-02 -1.97413564e-01
-5.69562972e-01 -9.97694433e-01 -7.60500669e-01 4.25757825e-01
1.17670991e-01 -6.07891798e-01 1.34616590e+00 -1.18811619e+00
-8.64750743e-01 4.30231541e-01 -1.79618880e-01 -7.65778422e-01
3.65235329e-01 -1.98059306e-01 -5.00334382e-01 -5.39636463e-02
3.57574314e-01 4.44224417e-01 8.28697860e-01 -9.73748982e-01
-4.47430640e-01 -1.39730081e-01 9.67247337e-02 -4.62962547e-03
-2.24072933e-01 -1.78678676e-01 -9.98831466e-02 -1.06815839e+00
3.47725332e-01 -8.95158827e-01 -1.79281905e-01 -2.61340559e-01
-6.21008098e-01 -7.75815099e-02 9.29932535e-01 -4.37961310e-01
1.19396937e+00 -2.07323575e+00 2.10242942e-01 1.33785412e-01
2.68864632e-01 -1.79159358e-01 -1.36046950e-03 3.62603515e-01
-6.98598623e-01 3.80919337e-01 -5.34943283e-01 -2.83169925e-01
6.33690879e-02 6.31782189e-02 -8.46265733e-01 3.49197298e-01
2.95692474e-01 7.56583333e-01 -1.14353526e+00 -1.12941183e-01
2.79862225e-01 4.46375847e-01 -3.79543155e-01 1.20097645e-01
-1.75487563e-01 6.33877039e-01 -3.87624443e-01 5.68785310e-01
-6.28318521e-04 -5.63573658e-01 -4.18868423e-01 -1.46184400e-01
9.47517604e-02 3.89156610e-01 -9.07672167e-01 1.38419342e+00
-2.42084548e-01 9.93034363e-01 -9.57134545e-01 -1.11353564e+00
7.98325121e-01 3.98374736e-01 4.58014101e-01 -6.84279919e-01
1.34402350e-01 3.54359537e-01 1.68025002e-01 -3.88079405e-01
4.26043957e-01 -1.44822463e-01 -1.60732344e-01 4.20462817e-01
-3.32883447e-01 -3.84429917e-02 8.81378800e-02 -2.22439207e-02
1.23496306e+00 -4.10085827e-01 5.40694952e-01 -2.76370674e-01
2.83750355e-01 2.59144366e-01 4.82236832e-01 6.06804550e-01
-3.28463882e-01 1.05862379e+00 5.16868830e-01 -8.65005314e-01
-8.96025300e-01 -8.88757825e-01 -1.66073978e-01 7.62918532e-01
5.05473852e-01 -2.20222622e-01 -5.58194816e-01 -6.65761232e-01
1.19937025e-01 1.04440618e+00 -1.17029786e+00 -6.16047800e-01
-1.96177691e-01 -7.47212350e-01 9.13094282e-02 4.78750229e-01
2.88946122e-01 -1.15109086e+00 -1.25809014e+00 4.79905158e-02
-4.94359314e-01 -8.75857711e-01 -4.95883167e-01 2.88400859e-01
-1.03027642e+00 -1.22624397e+00 -5.69838166e-01 -2.87654191e-01
7.67824113e-01 4.84825462e-01 1.25215304e+00 2.44015291e-01
-1.56417638e-01 4.84770507e-01 -4.79213625e-01 -7.35208392e-01
-8.01696330e-02 -1.08365655e-01 4.74185944e-02 -3.57372388e-02
2.48729393e-01 -5.06679535e-01 -8.27151895e-01 4.81307328e-01
-9.00929511e-01 8.38146955e-02 2.75044322e-01 9.11851823e-01
6.97450399e-01 2.66943928e-02 5.85313559e-01 -5.85665584e-01
8.41006041e-01 -6.77229583e-01 -2.31366128e-01 2.42517442e-01
-8.94611061e-01 3.51525247e-01 6.66815341e-01 -6.37305081e-01
-3.33803475e-01 -3.07966530e-01 4.57524747e-01 -7.13984370e-01
-1.06065925e-02 7.25984097e-01 3.01238239e-01 3.15644413e-01
9.00801837e-01 1.52934030e-01 -1.76002070e-01 -2.00318664e-01
2.86781639e-01 2.09503621e-01 6.68139875e-01 -2.45446339e-01
7.97789991e-01 4.48487848e-01 6.48028962e-03 -4.20247793e-01
-1.03736091e+00 -4.30321664e-01 -1.67487159e-01 -3.05619091e-01
7.52427936e-01 -2.72659183e-01 -5.78525364e-01 -1.23999916e-01
-1.22931373e+00 -8.35249200e-02 -5.40768981e-01 4.69038248e-01
-6.48310781e-01 2.89861672e-02 5.87133318e-02 -5.96342385e-01
-2.60299176e-01 -1.24022162e+00 1.18102300e+00 3.05512756e-01
-7.34461188e-01 -1.17618275e+00 5.21632284e-02 -6.70007989e-02
4.73733693e-01 6.66627288e-01 1.00796330e+00 -1.01021862e+00
-3.45654577e-01 -3.31091821e-01 -1.30518720e-01 -1.52635649e-01
3.77243191e-01 -4.61280756e-02 -1.02387464e+00 7.27164969e-02
2.24643394e-01 2.86926866e-01 7.31500149e-01 5.56162000e-01
1.50198638e+00 -5.78979313e-01 -3.22480172e-01 2.49997690e-01
1.05523717e+00 1.78978637e-01 5.01529098e-01 5.10075569e-01
4.47143257e-01 7.70541370e-01 9.49511230e-01 4.68046665e-01
1.97483018e-01 8.49152744e-01 7.09352911e-01 -1.45213589e-01
1.76094949e-01 6.16275743e-02 1.63423002e-01 4.88066196e-01
3.48347537e-02 -3.18232596e-01 -1.05049670e+00 6.83368862e-01
-2.10874081e+00 -1.19227612e+00 -1.29824668e-01 2.33589673e+00
5.15001774e-01 1.46865115e-01 5.76048680e-02 8.05251718e-01
7.17361450e-01 3.62352445e-03 -7.86815882e-01 -3.74519765e-01
-4.51057516e-02 -1.99321404e-01 1.59650460e-01 2.88598448e-01
-1.00383639e+00 3.40715826e-01 6.28885078e+00 4.91034418e-01
-1.55543995e+00 -1.84602499e-01 7.17266321e-01 2.63566431e-03
-5.22423029e-01 -7.39326775e-02 -5.00505753e-02 6.35404348e-01
8.59480441e-01 -7.91360557e-01 3.89857739e-01 7.33776629e-01
3.94986242e-01 1.83304667e-01 -1.42945552e+00 9.93302345e-01
-5.52313179e-02 -1.51726627e+00 6.41941428e-02 -3.23083550e-01
6.08171344e-01 -2.52652973e-01 4.16972607e-01 8.35746378e-02
-2.65442163e-01 -1.17845988e+00 1.00063574e+00 5.71211159e-01
2.99058378e-01 -4.45293754e-01 9.47961271e-01 8.92962962e-02
-1.17035735e+00 -2.18458518e-01 9.66797397e-02 -2.24470988e-01
-5.58411218e-02 6.22426987e-01 -7.86737502e-01 5.30618310e-01
8.40706050e-01 8.52124810e-01 -5.68506658e-01 1.06887054e+00
-2.72434533e-01 5.69012284e-01 -4.66652252e-02 -9.64979157e-02
4.78170775e-02 2.87794590e-01 8.74340355e-01 9.30570960e-01
5.17766654e-01 1.39674902e-01 -1.20954938e-01 9.94447947e-01
2.70715505e-01 -2.05499469e-03 -5.85991561e-01 -9.00989845e-02
6.29763365e-01 1.04005742e+00 -1.01516569e+00 -3.30095738e-01
-7.45163187e-02 8.69185746e-01 -9.48554054e-02 5.15267670e-01
-1.18169963e+00 -2.84926951e-01 8.62138391e-01 4.81243581e-02
-3.12423930e-02 1.32895023e-01 -8.07906568e-01 -9.78724897e-01
1.55902490e-01 -8.35285306e-01 5.36638916e-01 -1.06892288e+00
-1.22515869e+00 6.35536134e-01 2.13992760e-01 -1.71641624e+00
-3.29914182e-01 -3.01475644e-01 -9.88738894e-01 7.48628736e-01
-1.44282532e+00 -4.08304721e-01 -7.23538518e-01 2.61421829e-01
7.01751888e-01 1.81280784e-02 6.89065337e-01 1.89103857e-02
-5.36815882e-01 5.04582345e-01 -1.75757930e-01 -2.29117051e-01
4.26247329e-01 -1.47263145e+00 4.23502952e-01 7.73549199e-01
2.41892546e-01 6.68072939e-01 1.62843585e+00 -3.33514094e-01
-1.16499555e+00 -1.19517648e+00 9.04915452e-01 -7.46914387e-01
4.76955742e-01 4.10967730e-02 -1.16016114e+00 3.88390094e-01
-3.06588300e-02 -3.87564837e-03 5.25342226e-01 -2.26134676e-02
-2.83769518e-01 5.34324460e-02 -1.19312441e+00 8.58171761e-01
7.58918822e-01 -4.15732861e-01 -8.49738955e-01 3.53941679e-01
1.12219381e+00 -3.87535125e-01 -6.05706811e-01 5.70678651e-01
4.81874496e-01 -1.01229346e+00 9.67531502e-01 -6.96862996e-01
8.00916731e-01 -3.92695516e-01 -1.78231895e-01 -1.54235411e+00
-1.61822915e-01 -4.46033657e-01 -1.42971858e-01 8.57008934e-01
5.10586083e-01 -7.75129080e-01 5.07662833e-01 6.68093264e-01
-3.13802332e-01 -1.00970018e+00 -8.32287371e-01 -8.86549652e-01
-3.21928501e-01 -5.41985393e-01 9.98869538e-01 1.07688749e+00
2.95364290e-01 2.03693748e-01 1.61842294e-02 4.28599507e-01
3.86750281e-01 4.77170527e-01 5.28375268e-01 -1.28001785e+00
5.11900485e-02 -7.68904686e-01 -8.43404770e-01 -4.15610522e-01
-8.13813955e-02 -7.35329628e-01 1.51911639e-02 -1.41753399e+00
2.89197136e-02 -1.10178649e-01 -5.08758008e-01 5.62974691e-01
-4.91343141e-01 2.42734626e-02 1.05069689e-01 4.81839299e-01
-2.93110192e-01 7.09297538e-01 9.54820395e-01 -2.93180346e-01
-2.00308040e-01 -7.35344812e-02 -7.74909914e-01 7.08611190e-01
8.74467492e-01 -6.94358230e-01 -6.30469382e-01 -1.89833298e-01
9.35494080e-02 1.11806624e-01 8.05814803e-01 -1.03317761e+00
-6.93359878e-03 -3.43876392e-01 -2.08573826e-02 -3.14935893e-01
1.06441446e-01 -8.48471105e-01 1.61126673e-01 6.87503815e-01
-6.83394551e-01 5.83160520e-01 3.63373309e-01 6.91972017e-01
-4.13319319e-01 1.26580700e-01 6.11294389e-01 3.52327257e-01
-6.80321097e-01 2.10534871e-01 3.22850123e-02 1.78832218e-01
1.01121020e+00 -3.84991229e-01 -4.30640429e-01 -6.74599588e-01
-6.13391578e-01 2.47220665e-01 3.61153126e-01 7.71384537e-01
5.59541464e-01 -1.50895488e+00 -5.82124352e-01 -2.99089886e-02
5.57801783e-01 -4.35414970e-01 4.73341085e-02 9.40178037e-01
7.37811858e-03 4.69547778e-01 -2.24652469e-01 -8.89422774e-01
-1.14365160e+00 7.16961443e-01 5.11223674e-01 -1.15795441e-01
-6.03688896e-01 4.51412976e-01 3.41322750e-01 -1.05309367e-01
2.81227529e-01 -8.48474145e-01 -4.17053252e-01 -6.27540126e-02
4.20136571e-01 3.29866022e-01 6.37075379e-02 -3.91120851e-01
-5.40751278e-01 5.43963075e-01 3.54847282e-01 -1.93440586e-01
1.28347111e+00 2.19468269e-02 3.94607544e-01 8.34292948e-01
9.34163749e-01 -4.01911110e-01 -1.02234781e+00 4.46370766e-02
2.53344148e-01 -4.53202963e-01 -2.55295604e-01 -8.79822195e-01
-1.01450849e+00 9.30688381e-01 7.62037814e-01 8.65007460e-01
1.35091376e+00 -1.07591726e-01 4.97122467e-01 8.83694738e-02
2.36195579e-01 -6.80114508e-01 2.86561787e-01 2.15991840e-01
1.48724949e+00 -1.41875875e+00 -8.70414823e-02 -3.89028311e-01
-6.84262574e-01 9.80495512e-01 2.99728334e-01 4.81342226e-02
5.28905392e-01 -2.65614003e-01 -7.91153964e-03 -4.35968757e-01
-9.35771525e-01 -8.19307715e-02 8.36072743e-01 3.54028076e-01
4.98189837e-01 5.10154925e-02 -3.39664310e-01 9.03567910e-01
-5.10461211e-01 -3.91205788e-01 5.00368059e-01 6.50687814e-01
-2.62254447e-01 -5.70546567e-01 -4.36831355e-01 5.21795511e-01
-2.68526763e-01 6.27829880e-02 -4.14376289e-01 6.60836041e-01
-8.28726366e-02 1.05397642e+00 9.36840251e-02 -5.32048821e-01
4.74794567e-01 -3.98303606e-02 -9.43460166e-02 -4.85787988e-01
-6.24437571e-01 -4.68538165e-01 -2.57400721e-01 -5.98633587e-01
-3.21801871e-01 -8.41991663e-01 -1.60560226e+00 -1.80018693e-01
-3.13093990e-01 3.56563687e-01 6.41297936e-01 1.13539100e+00
5.01954198e-01 8.74090552e-01 7.09599316e-01 -8.25779378e-01
-3.99848104e-01 -6.61314011e-01 -1.41335919e-01 8.76861870e-01
6.81789041e-01 -1.07939374e+00 -9.12816226e-01 7.72056207e-02] | [7.363716125488281, 3.3044629096984863] |
8374b77e-72c4-41c4-85da-be557aca3234 | beyond-arabic-software-for-perso-arabic | 2301.11406 | null | https://arxiv.org/abs/2301.11406v1 | https://arxiv.org/pdf/2301.11406v1.pdf | Beyond Arabic: Software for Perso-Arabic Script Manipulation | This paper presents an open-source software library that provides a set of finite-state transducer (FST) components and corresponding utilities for manipulating the writing systems of languages that use the Perso-Arabic script. The operations include various levels of script normalization, including visual invariance-preserving operations that subsume and go beyond the standard Unicode normalization forms, as well as transformations that modify the visual appearance of characters in accordance with the regional orthographies for eleven contemporary languages from diverse language families. The library also provides simple FST-based romanization and transliteration. We additionally attempt to formalize the typology of Perso-Arabic characters by providing one-to-many mappings from Unicode code points to the languages that use them. While our work focuses on the Arabic script diaspora rather than Arabic itself, this approach could be adopted for any language that uses the Arabic script, thus providing a unified framework for treating a script family used by close to a billion people. | ['Richard Sproat', 'Brian Roark', 'Raiomond Doctor', 'Cibu Johny', 'Alexander Gutkin'] | 2023-01-26 | null | null | null | null | ['transliteration'] | ['natural-language-processing'] | [ 1.48195818e-01 -2.31142730e-01 9.73326787e-02 -2.99731612e-01
2.79414430e-02 -1.31653333e+00 9.87259090e-01 -1.60054088e-01
-4.35688645e-01 4.26912099e-01 1.34699449e-01 -7.39107847e-01
1.64170668e-01 -8.23454738e-01 -6.25322759e-02 -4.68432277e-01
2.96417087e-01 5.54610848e-01 5.17325282e-01 -9.38498020e-01
3.21919411e-01 9.06947613e-01 -1.23331654e+00 5.24718463e-02
6.91710949e-01 1.05548208e-03 3.21542546e-02 7.54730284e-01
-3.27611923e-01 3.63482147e-01 -6.79975867e-01 -8.14884603e-01
1.32085904e-01 -7.59145379e-01 -7.41329193e-01 -2.35702749e-02
9.87889394e-02 -2.11812317e-01 -4.09256488e-01 1.10292113e+00
2.71398039e-03 -1.38126478e-01 8.74479890e-01 -7.66205370e-01
-6.60039008e-01 8.50055754e-01 -1.24607466e-01 -4.18984443e-02
6.73531353e-01 -1.79700658e-01 5.85259914e-01 -6.65078044e-01
7.79650629e-01 1.23643434e+00 6.06503248e-01 3.23956132e-01
-8.92985106e-01 -2.06781909e-01 -3.35329950e-01 -1.55761763e-01
-1.48704934e+00 -4.64248508e-01 2.88182348e-01 -4.52758700e-01
1.03477144e+00 6.36373878e-01 4.90372002e-01 7.68398583e-01
1.69881716e-01 -5.66211231e-02 1.10581660e+00 -1.03665459e+00
-7.61706606e-02 2.75087655e-01 2.11871743e-01 8.60357583e-01
5.28813362e-01 -6.12629950e-01 -1.81025624e-01 -1.35465249e-01
1.11104190e+00 -5.17795384e-01 1.67714015e-01 7.97377378e-02
-1.18707263e+00 6.66957319e-01 -3.95099610e-01 8.05679321e-01
9.77476463e-02 -1.99421510e-01 5.21770060e-01 2.68132180e-01
-2.16415659e-01 3.20370823e-01 -2.64866948e-01 -4.80428338e-01
-8.72765362e-01 -4.62216660e-02 1.16604602e+00 1.03356290e+00
5.31728923e-01 4.90697265e-01 4.77371752e-01 7.94047475e-01
3.91282022e-01 7.16277599e-01 4.22404349e-01 -9.23774719e-01
2.27280289e-01 5.36011279e-01 5.86849190e-02 -7.30610490e-01
-1.94350421e-01 2.85882533e-01 -4.11808401e-01 1.94391295e-01
8.63059521e-01 -1.66871011e-01 -8.76294613e-01 1.39948416e+00
1.22134916e-01 -7.75623858e-01 2.72583157e-01 4.02434051e-01
5.05166352e-01 8.33496332e-01 -1.43070176e-01 -3.22672635e-01
1.70892382e+00 -5.83216786e-01 -7.91857481e-01 -2.69210506e-02
5.55670619e-01 -1.09774268e+00 1.30354118e+00 3.77726883e-01
-1.20636225e+00 -3.94036621e-02 -1.19473815e+00 -3.14511880e-02
-7.16405571e-01 1.18463509e-01 4.89527166e-01 1.37686038e+00
-1.21830583e+00 4.26314950e-01 -1.11186993e+00 -1.01621830e+00
-3.86473119e-01 2.96608835e-01 -4.50540453e-01 5.54219902e-01
-9.70914185e-01 1.43183744e+00 4.22802210e-01 -7.09833801e-02
-1.60960004e-01 2.80388117e-01 -9.71506596e-01 -9.20720100e-02
5.10007329e-02 -3.25263757e-03 1.25284481e+00 -1.31734204e+00
-1.95656872e+00 1.00318992e+00 -2.79829316e-02 4.55032811e-02
3.12984228e-01 1.54104561e-01 -8.28554273e-01 5.84536828e-02
-4.59962264e-02 -1.25199165e-02 5.66498101e-01 -9.65474725e-01
-4.55669105e-01 -2.80664533e-01 3.14946026e-02 2.68911749e-01
-5.61641872e-01 1.14953673e+00 -8.63436043e-01 -1.03567564e+00
2.16393724e-01 -9.33430254e-01 1.42452056e-02 -3.12227994e-01
-3.02265197e-01 2.71203130e-01 5.56755900e-01 -1.28959107e+00
1.63925922e+00 -1.99100208e+00 2.74835706e-01 6.24670327e-01
-4.92407352e-01 1.93258926e-01 -8.63038674e-02 9.21510577e-01
-6.75759977e-03 5.88267483e-02 -5.16253948e-01 1.26583189e-01
1.92817181e-01 6.96007729e-01 -1.52355939e-01 5.85333467e-01
2.89532449e-02 5.70514679e-01 -5.69229007e-01 -4.05377060e-01
-6.97412863e-02 4.27715212e-01 -8.03379938e-02 -3.55611682e-01
-6.29883558e-02 -1.36228830e-01 -4.51325215e-02 6.25094652e-01
4.91163224e-01 4.66112256e-01 8.87910366e-01 4.13312912e-01
-7.15575099e-01 3.84295791e-01 -1.37803710e+00 1.53671885e+00
-2.74601072e-01 3.91002446e-01 1.24872342e-01 -4.80682880e-01
1.01771510e+00 2.95061171e-01 -3.53926793e-02 -2.46375382e-01
3.31618994e-01 6.27604723e-01 1.46900982e-01 -2.89182425e-01
9.02966261e-01 2.73705348e-02 -4.82459754e-01 7.03125179e-01
1.60605222e-01 -3.19100559e-01 8.50905120e-01 2.02708423e-01
6.09109402e-01 6.93932533e-01 8.69017303e-01 -5.48678935e-01
7.16536283e-01 -1.35788526e-02 3.50665182e-01 3.77673328e-01
2.11632788e-01 4.80592877e-01 8.64417732e-01 -2.24319860e-01
-1.28665984e+00 -1.22201228e+00 -1.77495778e-01 1.23836994e+00
-2.42440298e-01 -6.16881490e-01 -1.04092062e+00 -3.24048966e-01
-5.70045710e-01 9.08282518e-01 -3.95922422e-01 1.96606323e-01
-8.57348680e-01 -7.04368770e-01 1.26617146e+00 3.11376989e-01
1.87699422e-01 -1.06798780e+00 -6.96160078e-01 1.26675785e-01
8.63422751e-02 -9.17730391e-01 -4.26216125e-01 1.49764419e-01
-5.27934551e-01 -9.01839972e-01 -7.36450374e-01 -1.02303028e+00
6.80470228e-01 -2.90479690e-01 5.21970630e-01 -5.64518794e-02
1.21534929e-01 3.67149472e-01 -5.88518560e-01 -2.43924335e-01
-1.17019212e+00 1.43794313e-01 1.75538629e-01 -1.32101223e-01
1.31186604e-01 -2.94958264e-01 4.24024880e-01 1.18784130e-01
-1.05517840e+00 -4.75703180e-02 -3.38400044e-02 2.18110532e-01
-1.22719899e-01 -3.78677279e-01 -7.90702924e-02 -9.64997113e-01
3.71345669e-01 -1.07432023e-01 -7.48248100e-01 5.90649486e-01
-3.75031948e-01 3.24430987e-02 8.22777092e-01 -1.94412738e-01
-1.34469485e+00 2.74321020e-01 -1.80795982e-01 5.55122197e-01
-1.03914760e-01 5.90098381e-01 -4.03733492e-01 -3.67301814e-02
6.45406067e-01 4.96391296e-01 9.51201543e-02 -3.79267782e-01
5.42585492e-01 9.82220650e-01 1.00914955e+00 -6.27413511e-01
9.20346618e-01 2.85990626e-01 -1.55856043e-01 -1.08858967e+00
2.41651356e-01 -8.78007617e-03 -1.03282547e+00 -1.56902120e-01
8.80094111e-01 -5.37230074e-01 -2.79020488e-01 9.00525749e-01
-1.29272151e+00 -2.91005045e-01 -1.70293257e-01 8.55622515e-02
-3.55554849e-01 7.51302898e-01 -9.54928339e-01 -6.17583394e-01
-9.50667784e-02 -1.01067209e+00 5.76484978e-01 2.74311990e-01
-5.94946861e-01 -9.95435834e-01 4.11477506e-01 -3.73653561e-01
1.85462400e-01 4.65223901e-02 1.18922865e+00 -7.93488741e-01
2.25644216e-01 -6.37390837e-02 5.58848381e-02 1.24455802e-01
3.24675590e-01 9.42752838e-01 -5.09777367e-01 -1.78325549e-01
-2.57252485e-01 2.94866771e-01 1.63546622e-01 -3.71330470e-01
-5.05528832e-03 -3.05274993e-01 1.77168638e-01 4.89378273e-01
1.35631549e+00 6.05384648e-01 8.29284608e-01 7.30799913e-01
5.28629065e-01 6.68117464e-01 2.60537237e-01 4.77893442e-01
5.03897429e-01 6.67273164e-01 2.72474345e-02 1.10094778e-01
-1.45866526e-02 1.39468327e-01 1.00268292e+00 1.15916944e+00
-4.64489371e-01 -2.11848095e-02 -1.16992474e+00 4.25740570e-01
-1.41936457e+00 -8.54283392e-01 -6.88972294e-01 2.27854800e+00
9.44930732e-01 -2.15051964e-01 3.60558480e-01 2.41592526e-01
7.70704687e-01 6.66061714e-02 2.06135154e-01 -1.21086693e+00
-5.46631575e-01 6.23057544e-01 5.03611207e-01 8.59311342e-01
-8.90671194e-01 1.50023043e+00 7.34112406e+00 4.64017123e-01
-1.27046525e+00 1.22890666e-01 -1.31712794e-01 3.26015532e-01
-3.80658329e-01 3.75940859e-01 -7.06289470e-01 2.20607013e-01
1.12658775e+00 -1.98169246e-01 8.07512224e-01 3.17314535e-01
2.70632505e-01 -9.92783308e-02 -6.02738738e-01 4.52409148e-01
2.99595594e-01 -1.10425472e+00 2.29892209e-01 -7.71607757e-02
7.46083021e-01 -1.99731171e-01 -3.58240366e-01 -1.70906171e-01
3.25360268e-01 -6.50159836e-01 1.31256044e+00 3.60913962e-01
1.29323900e+00 -6.00269079e-01 2.88218766e-01 -1.81902751e-01
-1.09624219e+00 1.14546962e-01 -1.11014552e-01 -2.15171114e-01
1.74832851e-01 -2.18231276e-01 -4.85404581e-01 6.24808192e-01
4.73923117e-01 3.47601533e-01 -8.97939384e-01 5.46613514e-01
-4.53595757e-01 7.18994379e-01 -4.52344239e-01 -1.16056621e-01
1.77989185e-01 -8.37900579e-01 4.67453808e-01 1.72161186e+00
6.81139290e-01 -2.04091907e-01 -3.02536339e-01 1.95184439e-01
5.73107123e-01 7.11098492e-01 -4.12656039e-01 -2.79916734e-01
4.20243561e-01 1.12546420e+00 -1.37045908e+00 -4.59821075e-01
-7.43798792e-01 1.43117130e+00 -2.20864818e-01 3.50307316e-01
-8.24121416e-01 -9.42138076e-01 4.95118439e-01 1.87410817e-01
2.57934481e-01 -7.20098913e-01 -3.43538910e-01 -1.16974497e+00
-1.11282974e-01 -1.29306507e+00 5.20302117e-01 -7.01000452e-01
-5.61322689e-01 7.72357702e-01 1.08856611e-01 -8.87997448e-01
-4.48967069e-01 -9.16104674e-01 -6.09871924e-01 7.55499005e-01
-5.86964667e-01 -1.49267173e+00 2.47473076e-01 5.45527875e-01
1.06977113e-01 -5.09997070e-01 1.34521627e+00 1.04538821e-01
-6.81982696e-01 4.86626983e-01 4.66723353e-01 3.50286603e-01
6.82558179e-01 -1.20642531e+00 6.43787980e-01 1.50942183e+00
2.32194841e-01 9.80492234e-01 7.60303080e-01 -7.37321794e-01
-1.18479836e+00 -4.96874839e-01 1.35703361e+00 -3.52861553e-01
1.02975631e+00 -5.84965050e-01 -6.34108782e-01 1.10741770e+00
7.20674813e-01 -6.99744403e-01 5.75575233e-01 -2.70602435e-01
-5.07368803e-01 2.60930151e-01 -8.21622431e-01 1.15362871e+00
7.67183244e-01 -6.06797516e-01 -7.47154832e-01 8.68908986e-02
1.78448692e-01 -1.99675187e-01 -8.19890857e-01 -3.48100960e-01
8.15279722e-01 -7.90533245e-01 4.52513009e-01 -4.65540826e-01
2.09881380e-01 -6.13471568e-01 -1.58497408e-01 -1.03954291e+00
-3.04312468e-01 -1.16116452e+00 4.81199145e-01 1.62788641e+00
4.02246684e-01 -9.53501463e-01 1.91334456e-01 2.89739072e-01
-2.25488737e-01 2.64463216e-01 -9.42963064e-01 -6.99525893e-01
2.20800087e-01 -3.40219378e-01 5.67140698e-01 1.19017041e+00
5.27176142e-01 7.01631978e-02 -2.09611252e-01 8.45957473e-02
-2.37104706e-02 -3.09474349e-01 5.25732636e-01 -1.09126794e+00
-3.37356210e-01 -7.37501025e-01 -4.38993156e-01 -3.65836740e-01
-1.36916470e-02 -9.92112935e-01 -1.03437267e-01 -1.37005615e+00
-1.87787458e-01 -1.47202695e-02 2.96529144e-01 8.04753840e-01
4.39555198e-01 5.29411733e-01 3.88981372e-01 1.58648506e-01
-1.16362438e-01 1.60177685e-02 5.81203640e-01 2.94817060e-01
-3.32942933e-01 -2.80312836e-01 -5.58762908e-01 1.01985884e+00
7.06723332e-01 -5.40505290e-01 6.76534921e-02 -4.72219586e-01
4.65765029e-01 -2.58169472e-01 2.89934687e-03 -7.48439848e-01
6.74731135e-02 -4.65821862e-01 4.92581241e-02 -5.13961837e-02
7.30530992e-02 -6.34419322e-01 4.35369194e-01 7.74344444e-01
2.18546480e-01 5.90814471e-01 2.35873222e-01 -4.59542096e-01
1.05589196e-01 -7.40811348e-01 9.11752343e-01 -1.65237054e-01
-7.76497364e-01 -4.83842731e-01 -1.41575825e+00 -3.12636554e-01
1.04950976e+00 -5.90722978e-01 -3.10125142e-01 -3.29095572e-01
-6.10217333e-01 -4.88980204e-01 1.05557573e+00 3.33818287e-01
2.42317896e-02 -1.09316957e+00 -6.47401631e-01 3.28133166e-01
1.01747550e-01 -9.74122584e-01 -4.26043689e-01 6.11461818e-01
-1.58855140e+00 3.87962490e-01 -8.73683512e-01 1.18657596e-01
-1.42728615e+00 1.42251253e-01 2.46848494e-01 9.27026197e-02
-2.29943722e-01 1.66060060e-01 -4.03452724e-01 -7.34793127e-01
-2.52102047e-01 -2.06911132e-01 -3.36681217e-01 1.94475606e-01
4.50949311e-01 5.00877321e-01 5.99999540e-02 -1.34168196e+00
-5.78270197e-01 5.66002309e-01 3.09679449e-01 -7.71205127e-01
1.06664670e+00 -1.43291175e-01 -8.46367061e-01 6.47851229e-01
4.13415700e-01 8.21932435e-01 -5.34630179e-01 2.14972749e-01
1.98919609e-01 -7.14796782e-02 -7.10055590e-01 -7.17504084e-01
-4.56076086e-01 5.36341429e-01 2.89002001e-01 1.58678547e-01
9.58717644e-01 -2.37396389e-01 2.43370995e-01 4.58678782e-01
3.20263952e-01 -1.29164588e+00 -6.29076302e-01 1.18134069e+00
7.27888703e-01 -1.73220962e-01 -3.37977931e-02 -4.43294495e-01
-7.48220325e-01 1.59158897e+00 1.32301301e-01 7.80164748e-02
3.62795800e-01 5.70711732e-01 5.51754177e-01 2.02445716e-01
-8.07267874e-02 -2.20915273e-01 -7.57669881e-02 6.90001845e-01
7.44690418e-01 1.28440976e-01 -1.21962452e+00 6.99846983e-01
-6.01159632e-01 -3.16068292e-01 1.20976245e+00 1.11854398e+00
-2.18409494e-01 -1.66863954e+00 -9.55797791e-01 1.14145033e-01
-4.43275660e-01 -3.21467996e-01 -5.66813827e-01 1.05673373e+00
1.43084735e-01 7.14387178e-01 3.31835926e-01 -2.80924350e-01
2.31028229e-01 4.54354435e-01 6.85259581e-01 -5.87922633e-01
-9.76080000e-01 3.44657928e-01 4.23260599e-01 -8.62210710e-03
-1.24978788e-01 -1.20945537e+00 -1.60000575e+00 -5.16368389e-01
8.34149122e-02 -1.89287812e-02 7.57413387e-01 8.50325704e-01
-2.58221209e-01 3.30407284e-02 2.14404702e-01 -7.30272710e-01
-3.33300740e-01 -8.27159822e-01 -9.47006524e-01 9.71516874e-03
-4.28071171e-01 -8.60291198e-02 5.33950189e-03 5.45426607e-01] | [10.486077308654785, 10.455249786376953] |
293ec8d6-ae9e-4a80-93c0-fa63af8ec7af | on-certified-generalization-in-structured | 2306.09112 | null | https://arxiv.org/abs/2306.09112v1 | https://arxiv.org/pdf/2306.09112v1.pdf | On Certified Generalization in Structured Prediction | In structured prediction, target objects have rich internal structure which does not factorize into independent components and violates common i.i.d. assumptions. This challenge becomes apparent through the exponentially large output space in applications such as image segmentation or scene graph generation. We present a novel PAC-Bayesian risk bound for structured prediction wherein the rate of generalization scales not only with the number of structured examples but also with their size. The underlying assumption, conforming to ongoing research on generative models, is that data are generated by the Knothe-Rosenblatt rearrangement of a factorizing reference measure. This allows to explicitly distill the structure between random output variables into a Wasserstein dependency matrix. Our work makes a preliminary step towards leveraging powerful generative models to establish generalization bounds for discriminative downstream tasks in the challenging setting of structured prediction. | ['Christoph Schnörr', 'Bastian Boll'] | 2023-06-15 | null | null | null | null | ['scene-graph-generation', 'structured-prediction', 'generalization-bounds'] | ['computer-vision', 'methodology', 'methodology'] | [ 8.25539529e-01 6.93222940e-01 -1.27185941e-01 -5.47556698e-01
-9.75273848e-01 -8.43045592e-01 5.45057476e-01 -1.23376124e-01
-1.94419637e-01 5.95414162e-01 1.83142483e-01 -2.35389188e-01
-2.48780370e-01 -5.37568510e-01 -9.68963921e-01 -1.05597365e+00
-7.55381435e-02 8.83024454e-01 1.21132448e-01 4.09288466e-01
8.83674845e-02 4.08293575e-01 -1.18615520e+00 -4.92162257e-02
8.83747995e-01 6.93802357e-01 1.35196090e-01 1.05490327e+00
3.45028341e-01 6.67680383e-01 -2.48842314e-01 -6.29832149e-01
5.01582921e-01 -6.13209665e-01 -6.68371975e-01 4.35710311e-01
6.35509193e-01 -1.94417030e-01 -2.22228914e-01 1.09805715e+00
1.40442029e-01 4.88719612e-01 1.29415178e+00 -1.39354801e+00
-7.57033288e-01 6.52268767e-01 -5.01102924e-01 8.53829235e-02
-2.33628467e-01 3.46258506e-02 1.47829390e+00 -5.91647863e-01
5.81598103e-01 1.31511712e+00 6.84586108e-01 4.78582084e-01
-1.67495871e+00 -4.01851326e-01 3.38223726e-01 -2.96580046e-01
-1.06902611e+00 -9.76637527e-02 3.94515336e-01 -7.83024907e-01
7.66923249e-01 2.71530360e-01 4.73251522e-01 1.20849431e+00
3.01644921e-01 1.05333126e+00 7.72198200e-01 -3.17897975e-01
3.36433560e-01 7.76690245e-02 2.39153236e-01 8.19559872e-01
4.21331555e-01 1.30441606e-01 -2.91217983e-01 -2.90931165e-01
7.55630791e-01 1.38809048e-02 -3.75254452e-02 -8.20612788e-01
-7.19524801e-01 1.04188228e+00 1.69967279e-01 -1.59086719e-01
8.17119554e-02 5.50952435e-01 -4.38239872e-02 -5.25443722e-03
4.69692737e-01 4.81690198e-01 -4.91644949e-01 -2.02106331e-02
-9.38104808e-01 2.52580076e-01 9.01632130e-01 1.35741377e+00
6.94490373e-01 -1.03267781e-01 -3.00295115e-01 4.80332345e-01
5.91666460e-01 7.11725712e-01 1.23634167e-01 -1.15717733e+00
5.31265914e-01 1.18038893e-01 4.07614522e-02 -6.09221816e-01
-1.19843736e-01 -6.09281898e-01 -8.06194425e-01 -5.92019707e-02
6.39099836e-01 -3.24054807e-01 -1.27478802e+00 1.86229956e+00
3.44453216e-01 1.57784179e-01 -1.93940893e-01 4.82413918e-01
-2.02177614e-02 6.72966838e-01 3.18429977e-01 -1.36659697e-01
8.88810933e-01 -9.12886560e-01 -2.14491144e-01 -2.91538537e-01
6.44219756e-01 -5.91223657e-01 8.91734958e-01 3.59483868e-01
-8.32227051e-01 -3.63815010e-01 -7.74370611e-01 -2.25512192e-01
1.77723348e-01 -9.98052359e-02 8.61675203e-01 9.00241911e-01
-1.00871682e+00 7.04918802e-01 -1.08546650e+00 -6.40743747e-02
7.48267233e-01 3.84265959e-01 -2.46242389e-01 -9.89473984e-02
-4.77121323e-01 5.74627340e-01 5.02104819e-01 -1.25794232e-01
-1.26954615e+00 -8.76991749e-01 -7.40516126e-01 1.83415636e-01
3.09983075e-01 -9.81431425e-01 1.35323679e+00 -8.32828462e-01
-1.27719533e+00 6.12108052e-01 -1.58781394e-01 -7.02127337e-01
6.36905909e-01 -4.05809343e-01 3.56579632e-01 1.05533451e-01
1.42473981e-01 8.12597990e-01 1.20752764e+00 -1.17363465e+00
-6.25720084e-01 -4.62587893e-01 -2.65289962e-01 4.17212278e-01
-5.40215932e-02 -3.88939083e-01 -3.53180081e-01 -6.02826297e-01
-6.52859744e-04 -1.33593392e+00 -8.64737272e-01 -4.27358691e-03
-7.64155030e-01 -1.89719677e-01 5.11723042e-01 -3.09383452e-01
8.07680070e-01 -2.11801696e+00 4.21189964e-01 3.68348688e-01
4.44953561e-01 -2.51727611e-01 -1.17354959e-01 2.67167479e-01
9.13087428e-02 3.19851696e-01 -3.31130534e-01 -4.42708075e-01
1.80446818e-01 3.28622520e-01 -5.39373636e-01 5.37009180e-01
2.59383589e-01 9.91397738e-01 -9.29923654e-01 -3.46433550e-01
-9.62667614e-02 2.60015458e-01 -8.32190335e-01 2.53049284e-01
-4.80016053e-01 4.31173682e-01 -5.99940181e-01 1.94757119e-01
4.87351656e-01 -7.40553677e-01 1.01402067e-01 2.60677785e-01
4.99626219e-01 2.34355509e-01 -1.03302550e+00 1.43234706e+00
-9.53231156e-02 4.21133935e-01 -2.77900428e-01 -9.94972110e-01
6.22728169e-01 1.31803565e-02 3.82265151e-01 2.22926840e-01
-1.94693264e-02 -2.07772613e-01 5.22591807e-02 -1.97901577e-02
3.70281726e-01 -5.50764263e-01 -1.22108392e-01 3.41747195e-01
4.23270017e-01 -3.35250288e-01 1.45123526e-01 4.98490334e-01
1.28704524e+00 1.36589140e-01 1.05879702e-01 -2.08299607e-01
-1.77212954e-01 -1.10781454e-01 5.23850262e-01 1.09482479e+00
3.29360925e-02 7.08355904e-01 7.71882415e-01 7.90527090e-02
-1.31146777e+00 -1.70917225e+00 -2.40526915e-01 9.95691717e-01
-6.93686754e-02 -3.53413373e-01 -6.75794661e-01 -9.54750717e-01
1.04912467e-01 8.43252420e-01 -8.43093872e-01 -2.85762161e-01
-5.77405393e-02 -8.49071443e-01 3.43784779e-01 6.88170910e-01
1.34256203e-02 -6.80741012e-01 -2.14760676e-01 2.16568023e-01
3.14141363e-01 -1.12484682e+00 -6.42446041e-01 6.77352369e-01
-1.23529005e+00 -9.61832345e-01 -7.75208354e-01 -6.12290144e-01
8.94899189e-01 1.08614936e-01 1.13632131e+00 -5.14242530e-01
-3.38522822e-01 7.38569856e-01 -4.68083620e-02 -4.23327804e-01
-4.47164655e-01 1.63271353e-01 -2.28222996e-01 -1.86423764e-01
1.90262735e-01 -5.68079352e-01 -4.53441978e-01 1.31716192e-01
-8.22289884e-01 1.84362933e-01 7.67953992e-01 8.05059969e-01
9.46505606e-01 -9.19844722e-04 6.33692071e-02 -1.23358762e+00
3.50147992e-01 -6.05351090e-01 -8.22647095e-01 3.11453849e-01
-6.01061642e-01 4.16772544e-01 4.32161450e-01 -6.04735911e-01
-1.09255385e+00 5.68306684e-01 2.51520753e-01 -5.10155737e-01
-1.63159028e-01 2.59471953e-01 -1.59563735e-01 3.23524475e-01
5.83362043e-01 1.76229358e-01 -5.56716472e-02 -3.19966197e-01
6.16400957e-01 1.21162526e-01 2.33746305e-01 -6.16482675e-01
1.12556994e+00 3.45962793e-01 4.38519269e-01 -8.17600667e-01
-1.21044827e+00 -4.10425305e-01 -7.19056606e-01 7.22950250e-02
9.99974787e-01 -1.05129123e+00 -4.68710065e-01 1.79755270e-01
-8.69318426e-01 -5.68302870e-01 -5.41765273e-01 4.65201586e-01
-8.30170989e-01 4.32398319e-01 -6.06389761e-01 -9.24919426e-01
-1.25193268e-01 -9.17158842e-01 1.24824226e+00 1.30158171e-01
-2.33464345e-01 -1.09149539e+00 2.34188929e-01 5.29546142e-01
-6.18400872e-02 2.34309789e-02 1.03560913e+00 -8.25010538e-01
-1.02744520e+00 -3.77857119e-01 -5.95085248e-02 5.76549470e-01
-1.98832825e-01 9.37499180e-02 -6.76850021e-01 -8.39791670e-02
-1.31427079e-01 -2.10335061e-01 9.61704314e-01 6.40607595e-01
1.08165836e+00 -3.36704314e-01 -3.73576254e-01 5.12891471e-01
1.18899441e+00 2.95184553e-02 3.88770700e-01 -3.91448617e-01
9.62574124e-01 4.01765853e-01 3.36571962e-01 2.64062494e-01
9.22322646e-02 1.81102023e-01 4.21013907e-02 2.56880432e-01
9.74351093e-02 -8.05680454e-01 3.17713857e-01 5.03719211e-01
1.52024940e-01 -5.30701756e-01 -8.26973259e-01 3.95886809e-01
-1.82008100e+00 -1.02327311e+00 -2.45658651e-01 2.16939163e+00
7.37876773e-01 1.75950915e-01 6.44029379e-02 -1.93063006e-01
5.30811191e-01 -1.91224799e-01 -7.17187881e-01 -8.59568864e-02
1.11606807e-01 2.97283441e-01 7.74129570e-01 6.14305675e-01
-1.09496844e+00 8.94760668e-01 7.08097315e+00 8.32925260e-01
-4.99070734e-01 1.15408517e-01 1.02479529e+00 -6.53010532e-02
-5.11888087e-01 1.10883161e-01 -1.28987145e+00 3.59632909e-01
9.32148218e-01 -1.45710364e-01 2.36587524e-01 1.31083882e+00
-2.55784005e-01 -1.19980343e-01 -1.46883738e+00 5.03610790e-01
-1.90692618e-01 -1.13292634e+00 4.77401130e-02 4.36272591e-01
1.04332149e+00 1.39456153e-01 4.83443558e-01 1.53091863e-01
1.15078306e+00 -1.18197429e+00 4.42504615e-01 1.52284071e-01
5.19334257e-01 -5.67733943e-01 2.71956295e-01 3.91461611e-01
-7.59149134e-01 9.09620076e-02 -5.29559314e-01 2.32292950e-01
2.26086855e-01 6.98118985e-01 -1.10320652e+00 -5.34641780e-02
4.01938736e-01 7.51387835e-01 -3.10484409e-01 1.02931595e+00
-3.54421407e-01 1.02962804e+00 -6.07703209e-01 3.74906249e-02
1.53087392e-01 -4.58784819e-01 4.88146394e-01 1.02938759e+00
4.39622730e-01 -2.79288869e-02 2.10225672e-01 9.43160176e-01
-1.37657955e-01 1.04734823e-01 -6.72562301e-01 -3.26506227e-01
5.24092875e-02 1.20788634e+00 -1.00592899e+00 -2.93799937e-01
-1.63587555e-01 9.65108395e-01 3.67584795e-01 4.16530281e-01
-7.21961141e-01 1.01831630e-01 5.44027746e-01 -3.08322851e-02
7.46504307e-01 -3.74830395e-01 -5.58972120e-01 -1.15356135e+00
-1.32097423e-01 -5.41151166e-01 3.36730093e-01 -5.48217118e-01
-1.65177095e+00 3.33083302e-01 2.31454223e-01 -8.93810511e-01
-2.67668337e-01 -6.23727202e-01 -3.92329127e-01 8.37192476e-01
-9.10405099e-01 -1.07510805e+00 1.36250883e-01 4.91317809e-01
4.27686155e-01 3.55479009e-02 7.02740610e-01 -2.13044092e-01
-5.47974288e-01 5.46915829e-01 2.89813578e-01 -9.74064469e-02
3.32363874e-01 -1.58906174e+00 2.91017771e-01 9.26573694e-01
6.53302431e-01 5.33158660e-01 8.82282376e-01 -7.47108459e-01
-1.20317578e+00 -1.06995463e+00 4.68315870e-01 -9.72999275e-01
9.66941297e-01 -5.05671144e-01 -6.70904577e-01 1.01404941e+00
-1.82882175e-01 5.19896261e-02 1.09512424e+00 4.35885131e-01
-5.19916296e-01 1.56449258e-01 -8.91860127e-01 6.18041635e-01
1.17161214e+00 -3.83992434e-01 -3.09263289e-01 5.43212533e-01
6.82174802e-01 -2.67454892e-01 -7.77550757e-01 2.87327081e-01
3.13284576e-01 -7.00662613e-01 8.10495794e-01 -1.05206430e+00
7.37495363e-01 1.23612490e-02 -3.40418518e-01 -1.10935390e+00
-5.29918194e-01 -8.31252396e-01 -3.53008628e-01 1.02591288e+00
7.97355533e-01 -2.89816767e-01 1.18529606e+00 1.04644406e+00
1.45646008e-02 -6.95946097e-01 -8.13118339e-01 -8.08249772e-01
2.95152009e-01 -5.14771640e-01 -1.21404096e-01 3.53243440e-01
-3.93541038e-01 8.53182971e-01 -4.63515222e-01 3.57217044e-01
9.73403931e-01 1.58336326e-01 8.17752540e-01 -1.38287902e+00
-9.53058660e-01 -1.78703368e-01 -6.51272416e-01 -1.60614121e+00
1.54157311e-01 -1.02125132e+00 3.19587976e-01 -1.16773677e+00
6.79697037e-01 -6.23941004e-01 -1.32635877e-01 1.90904543e-01
-4.04389441e-01 2.24158075e-02 2.68910319e-01 4.85205017e-02
-7.41429031e-01 6.24185979e-01 1.30542183e+00 -1.79568380e-02
-1.15883999e-01 4.18638766e-01 -8.13669682e-01 6.27865016e-01
7.37654090e-01 -6.67723536e-01 -8.59340191e-01 -1.89589381e-01
1.79285750e-01 4.91835475e-02 2.70406783e-01 -7.23923326e-01
-1.94143549e-01 -3.01205873e-01 5.20959258e-01 -4.38219041e-01
2.96951920e-01 -7.45920479e-01 2.18702361e-01 3.17502886e-01
-7.83949912e-01 -2.57665157e-01 -1.77323222e-01 1.33675754e+00
2.18982369e-01 -4.02310014e-01 8.47524822e-01 -8.08204859e-02
-1.07952669e-01 6.57492459e-01 -4.02125448e-01 4.92713094e-01
1.17085290e+00 -2.35861391e-02 2.54626237e-02 -4.68775362e-01
-9.88065243e-01 7.51228184e-02 3.08225691e-01 1.72940075e-01
3.82533073e-01 -1.00873101e+00 -4.57240194e-01 -7.28955911e-03
-1.75110906e-01 3.01081508e-01 2.00939596e-01 6.96179748e-01
-1.27238616e-01 2.06869602e-01 1.97506532e-01 -7.54484177e-01
-1.19120932e+00 6.25615478e-01 1.34733453e-01 -5.31925499e-01
-6.98697865e-01 1.30466068e+00 7.42268085e-01 -1.17555849e-01
2.88041234e-01 -3.05684179e-01 4.10596550e-01 -1.19630553e-01
2.35209644e-01 1.19104043e-01 -4.67921525e-01 -2.82712609e-01
1.63417771e-01 1.93322450e-01 -4.05085742e-01 -2.44297698e-01
1.35544336e+00 6.66337237e-02 2.26415291e-01 5.10736644e-01
1.36476469e+00 -4.97332700e-02 -1.80781555e+00 -1.46215230e-01
1.32546633e-01 -4.09924328e-01 -1.10741720e-01 -4.83788103e-01
-8.37903321e-01 1.12382615e+00 4.19464260e-01 1.61667302e-01
6.17970943e-01 3.97745848e-01 5.81382394e-01 4.27811503e-01
2.93879211e-01 -9.53415453e-01 2.27209479e-01 5.14527082e-01
6.30513012e-01 -1.18030655e+00 -2.90958565e-02 -6.06302679e-01
-7.16783166e-01 8.69519591e-01 4.04266119e-01 -4.17349756e-01
8.45646322e-01 4.50337827e-01 -4.10913020e-01 -4.11265567e-02
-8.64211857e-01 -1.46008104e-01 5.67686617e-01 6.66256487e-01
2.51673937e-01 2.53376842e-01 1.85714245e-01 4.60987747e-01
-3.31877947e-01 -3.25178176e-01 2.42904842e-01 4.72111076e-01
-5.91215909e-01 -1.14488757e+00 3.40303704e-02 8.67386162e-01
-4.92662996e-01 -2.04609752e-01 -2.81900883e-01 4.33585107e-01
-2.25392565e-01 6.24603808e-01 6.91765249e-02 -1.98027283e-01
-2.54897833e-01 1.16301112e-01 7.30441749e-01 -1.03394818e+00
-6.98786527e-02 2.40666538e-01 -2.49882430e-01 -2.12917969e-01
-9.30564031e-02 -8.55245173e-01 -9.65665221e-01 1.30475936e-02
-3.03046137e-01 4.59455922e-02 2.80485362e-01 8.55999947e-01
2.50600547e-01 1.76401392e-01 5.37309647e-01 -8.42530668e-01
-9.31726098e-01 -7.86909163e-01 -5.99837959e-01 3.42391104e-01
1.09437756e-01 -4.29834485e-01 -6.33774638e-01 3.18465263e-01] | [7.676203727722168, 4.004696369171143] |
3cfda1d9-d6bf-4006-9137-7c9572ad2803 | a-framework-for-csi-based-indoor-localization | 2205.08068 | null | https://arxiv.org/abs/2205.08068v1 | https://arxiv.org/pdf/2205.08068v1.pdf | A Framework for CSI-Based Indoor Localization with 1D Convolutional Neural Networks | Modern indoor localization techniques are essential to overcome the weak GPS coverage in indoor environments. Recently, considerable progress has been made in Channel State Information (CSI) based indoor localization with signal fingerprints. However, CSI signal patterns can be complicated in the large and highly dynamic indoor spaces with complex interiors, thus a solution for solving this issue is urgently needed to expand the applications of CSI to a broader indoor space. In this paper, we propose an end-to-end solution including data collection, pattern clustering, denoising, calibration and a lightweight one-dimensional convolutional neural network (1D CNN) model with CSI fingerprinting to tackle this problem. We have also created and plan to open source a CSI dataset with a large amount of data collected across complex indoor environments at Colorado State University. Experiments indicate that our approach achieves up to 68.5% improved performance (mean distance error) with minimal number of parameters, compared to the best-known deep machine learning and CSI-based indoor localization works. | ['Sudeep Pasricha', 'Liping Wang'] | 2022-05-17 | null | null | null | null | ['indoor-localization'] | ['computer-vision'] | [-8.03701673e-03 -6.22171640e-01 3.11104000e-01 -6.61129296e-01
-8.96748483e-01 -4.12342340e-01 7.03028813e-02 -1.55797794e-01
-2.95841485e-01 1.07535338e+00 2.62143523e-01 -6.03632808e-01
-4.09320980e-01 -8.97014618e-01 -9.23596025e-01 -7.00264513e-01
-4.41555113e-01 -1.25999562e-02 -6.38787672e-02 2.03711400e-03
3.58945094e-02 4.86898124e-01 -1.00636494e+00 -2.29801372e-01
8.32597494e-01 1.24672592e+00 3.01964372e-01 7.35784173e-01
1.31959319e-01 6.55598462e-01 -7.38708973e-01 1.30530477e-01
1.50764242e-01 8.15872289e-03 -3.11509371e-01 -5.32318175e-01
4.79941547e-01 -2.47698039e-01 -6.57458127e-01 6.70887589e-01
1.31545138e+00 1.22570626e-01 3.34470421e-02 -1.04400492e+00
-4.24948961e-01 3.72892946e-01 -8.41856673e-02 1.51936129e-01
6.83976412e-01 -5.17665684e-01 8.85968953e-02 -5.98471880e-01
5.33267893e-02 7.02258706e-01 1.77136469e+00 2.17492576e-03
-1.02494991e+00 -9.26530004e-01 -6.53785691e-02 -5.69694228e-02
-1.80819666e+00 -5.93624294e-01 7.36607194e-01 9.37554762e-02
9.15133953e-01 2.94420213e-01 4.72592622e-01 1.51269031e+00
2.40984157e-01 4.72769648e-01 8.85662436e-01 -2.24909216e-01
4.54505771e-01 -1.12950765e-01 -2.76869804e-01 3.92736971e-01
4.04324502e-01 5.39558418e-02 -3.96410286e-01 -2.29173407e-01
7.86847234e-01 4.16851431e-01 -2.73109347e-01 -3.87677908e-01
-1.17456281e+00 2.11943015e-01 9.61937428e-01 6.54966831e-01
-2.91060090e-01 6.92700148e-01 -1.41975328e-01 3.14372361e-01
1.94877982e-01 4.50443029e-01 -7.68022597e-01 -5.51069975e-01
-1.09834814e+00 1.97804525e-01 1.13297498e+00 1.43320501e+00
6.53536201e-01 2.97272834e-03 5.43072708e-02 7.86245644e-01
4.39812094e-01 9.05940175e-01 -2.34757233e-02 -1.13056326e+00
6.40336931e-01 -6.69904202e-02 4.73457873e-01 -1.51846242e+00
-7.93191969e-01 -1.18930078e+00 -1.29205287e+00 -6.58605933e-01
2.40218461e-01 -6.54050529e-01 -7.14267910e-01 1.62368870e+00
1.04160331e-01 6.80279672e-01 -1.07350066e-01 4.37669992e-01
7.56950617e-01 3.28575701e-01 -4.90408182e-01 3.73241276e-01
5.23883641e-01 -7.08281040e-01 -7.19535291e-01 -3.10882300e-01
5.70914328e-01 -7.77965248e-01 5.77676475e-01 3.01118284e-01
-5.10885656e-01 -6.60026193e-01 -1.37277973e+00 2.98471153e-01
-4.48147595e-01 1.45851597e-01 1.00230944e+00 1.30622447e+00
-1.41114962e+00 4.41119790e-01 -1.04200029e+00 -4.30871814e-01
3.78176719e-01 9.48421657e-01 -2.30462372e-01 -6.62643611e-01
-1.19200230e+00 2.91947216e-01 -1.92853168e-01 4.67283100e-01
-3.79909366e-01 -6.02316260e-01 -8.78439426e-01 -1.17931411e-01
-1.45405248e-01 -7.83398747e-01 1.17237544e+00 -2.29205534e-01
-1.32516706e+00 -1.59907803e-01 -4.67450112e-01 -3.83920074e-01
2.41141662e-01 -9.58513990e-02 -8.13833177e-01 -7.94276416e-01
4.91653144e-01 5.20868786e-03 -1.46445408e-02 -1.40984833e+00
-6.13985240e-01 -2.66383648e-01 -8.62777904e-02 -1.58472642e-01
-8.23837668e-02 -4.66800570e-01 -6.95123911e-01 -3.44482154e-01
7.80478954e-01 -1.18610764e+00 -6.52266860e-01 -3.47405046e-01
-4.03365582e-01 5.62541127e-01 4.68686759e-01 -4.91642296e-01
1.14550829e+00 -2.06502843e+00 -8.78876925e-01 6.66382909e-01
1.45911332e-03 -1.32703006e-01 3.13402534e-01 5.02667665e-01
4.35542077e-01 -4.60435301e-02 1.91784710e-01 -1.04250097e+00
1.92830443e-01 4.75744516e-01 7.37223327e-02 7.11141050e-01
-7.68467188e-01 6.03327632e-01 -1.04194963e+00 -7.60430247e-02
3.73865426e-01 7.72710860e-01 -6.81590855e-01 -1.19122274e-01
7.73577571e-01 8.88519764e-01 -7.38138378e-01 9.02481318e-01
1.20213246e+00 -3.04436237e-01 1.03300549e-01 -3.35664526e-02
-3.49356920e-01 4.83023822e-01 -1.57481265e+00 2.14146066e+00
-9.60266709e-01 7.41406858e-01 3.34400713e-01 -8.35975349e-01
8.35737944e-01 2.30646491e-01 6.08960271e-01 -9.58818257e-01
1.34478971e-01 5.22874832e-01 -4.04949337e-01 -2.43533343e-01
4.92685884e-01 4.80148166e-01 -4.21384990e-01 -1.33055568e-01
-1.28332600e-01 1.52505264e-01 -4.71681744e-01 -1.05431248e-02
1.74986100e+00 -5.75242005e-02 -9.18059349e-02 -1.77930981e-01
3.97439092e-01 -3.02457303e-01 5.99659860e-01 1.45470202e+00
-2.28031725e-01 4.57752228e-01 -4.77222621e-01 -4.31685686e-01
-4.74751472e-01 -1.08307612e+00 -3.83598417e-01 6.18243814e-01
4.43159342e-01 -3.19180876e-01 -3.77845317e-01 -2.22502891e-02
2.65578926e-01 2.24835813e-01 -3.76707494e-01 1.79403216e-01
-5.87328434e-01 -5.78458726e-01 9.74443793e-01 5.58752120e-01
1.19376647e+00 -3.31101239e-01 1.74711898e-01 6.25111103e-01
-2.96572536e-01 -1.28954959e+00 -1.05197452e-01 5.85637271e-01
-5.20114243e-01 -4.21359867e-01 -5.44030964e-01 -8.93729627e-01
6.20940566e-01 6.63482726e-01 9.91730511e-01 -4.84487712e-02
9.55554992e-02 4.66134340e-01 -1.83611736e-01 -3.37070614e-01
4.37874645e-01 3.89940798e-01 4.79511917e-01 -3.08701307e-01
1.30005166e-01 -1.05490875e+00 -7.41198897e-01 5.02082050e-01
-4.70889099e-02 -3.76398474e-01 6.34691298e-01 6.36327505e-01
6.37279868e-01 3.62443328e-01 3.50101978e-01 -3.99025738e-01
4.72621948e-01 -7.59123147e-01 -5.46341181e-01 -4.08546552e-02
-3.50943923e-01 -3.61163735e-01 6.11273944e-01 3.97939682e-02
-8.42740655e-01 4.27532047e-01 -7.49720156e-01 2.10893244e-01
-4.62014169e-01 4.52494800e-01 -4.34747458e-01 -1.02884495e+00
5.28123617e-01 2.75160700e-01 -7.98529506e-01 -4.86746520e-01
-4.70848307e-02 1.04899430e+00 7.21887946e-01 -7.12070644e-01
7.55057156e-01 5.50087035e-01 1.99036062e-01 -8.39365542e-01
-8.07986736e-01 -7.55322576e-01 -5.85667014e-01 7.05140978e-02
2.67366648e-01 -1.25456476e+00 -9.21675861e-01 4.08467889e-01
-9.82258141e-01 -4.36438471e-01 3.84977072e-01 4.74726975e-01
-2.17781395e-01 1.65681005e-01 -4.27850872e-01 -7.77467251e-01
-2.12725446e-01 -9.45808649e-01 1.00715840e+00 3.51429015e-01
5.60600236e-02 -1.08191645e+00 3.17942739e-01 2.79407859e-01
1.28494191e+00 4.74836051e-01 -2.54069448e-01 -2.14902341e-01
-9.23596978e-01 -5.89245021e-01 -2.22181510e-02 -5.58915101e-02
3.74160081e-01 -6.93299413e-01 -1.19327307e+00 -3.92689526e-01
5.68326004e-02 2.42635027e-01 3.61176670e-01 8.28071773e-01
1.36893308e+00 -1.05617315e-01 -8.65534306e-01 1.34842885e+00
1.45953476e+00 3.62244040e-01 6.26834273e-01 6.05429590e-01
6.80419207e-01 -3.95627141e-01 3.01831335e-01 4.10237730e-01
7.72990704e-01 7.64043450e-01 3.77886504e-01 -4.27676380e-01
1.79796144e-02 -4.26189184e-01 -1.58621311e-01 6.60222471e-01
9.69819650e-02 -4.92019832e-01 -9.39749002e-01 3.97107989e-01
-1.78721070e+00 -7.68407285e-01 -3.42032254e-01 2.02006936e+00
4.17657167e-01 1.48199037e-01 -5.61560750e-01 3.34328681e-01
2.87011117e-01 8.57628137e-03 -2.17269808e-01 1.32027522e-01
7.02041760e-03 1.67628631e-01 1.32709134e+00 5.07937014e-01
-1.48700619e+00 6.21143043e-01 6.16002703e+00 5.53415775e-01
-1.18006945e+00 7.40949214e-02 3.19157958e-01 3.83862376e-01
-2.84993406e-02 -3.70832682e-01 -8.06381524e-01 6.71386361e-01
1.05518293e+00 6.23436928e-01 5.88534057e-01 1.20720959e+00
6.59912229e-02 -2.89616853e-01 -7.51446843e-01 1.46221042e+00
-2.76967555e-01 -1.61582303e+00 -1.16788495e+00 2.59605020e-01
1.06740069e+00 5.81948638e-01 1.67213976e-01 3.28513324e-01
3.67350757e-01 -9.20439601e-01 3.89562994e-01 4.77815449e-01
7.48299837e-01 -7.91328430e-01 1.13501453e+00 3.29514056e-01
-1.61447906e+00 -1.44346491e-01 -4.49307263e-01 -5.62181711e-01
1.28247544e-01 8.97920251e-01 -9.20939088e-01 6.83172703e-01
1.01959586e+00 7.84036398e-01 -5.49139082e-01 1.70937681e+00
-8.59095156e-02 5.81960380e-01 -8.04975688e-01 -1.76932111e-01
1.87256545e-01 2.67727137e-01 -7.18347542e-03 1.18410730e+00
8.42741072e-01 -2.34051242e-01 3.26196432e-01 4.73390728e-01
3.86622809e-02 -6.33087516e-01 -9.39242244e-01 6.92610800e-01
1.07072210e+00 9.58076119e-01 -4.97329414e-01 5.08347303e-02
-2.84948468e-01 1.13008070e+00 -3.31096202e-01 5.60297787e-01
-7.61112034e-01 -6.34176672e-01 7.29351521e-01 -1.35344237e-01
3.44677627e-01 -9.66533661e-01 -5.13931513e-01 -7.88364887e-01
-1.01587296e-01 -1.65758654e-01 -3.71290058e-01 -3.76814485e-01
-1.08935475e+00 3.48325819e-01 -6.57650888e-01 -1.09068418e+00
-3.74437831e-02 -3.58769149e-01 -4.65860367e-01 8.12391818e-01
-1.54779661e+00 -9.95274782e-01 -8.15705121e-01 1.01876414e+00
-2.52434239e-02 2.34521940e-01 1.14249742e+00 1.28459704e+00
-4.78430361e-01 8.91820550e-01 1.05801332e+00 3.48165691e-01
5.80660105e-01 -1.26104820e+00 7.97746301e-01 8.76821458e-01
-4.95507382e-02 1.12589085e+00 7.21175671e-01 -5.37776291e-01
-1.84772611e+00 -1.17643702e+00 1.00021529e+00 -5.01785636e-01
2.74660617e-01 -7.65708148e-01 -1.45584252e-02 6.94652259e-01
-1.11846484e-01 2.91604072e-01 9.17876482e-01 4.25313652e-01
2.30555922e-01 -5.32918572e-01 -1.44436312e+00 3.80829543e-01
1.38867402e+00 -4.14667428e-01 2.71456003e-01 3.36423635e-01
6.02899909e-01 -9.19709265e-01 -9.27463830e-01 3.66968542e-01
6.65015101e-01 -9.32921469e-01 1.47808909e+00 5.82136512e-01
-5.54992080e-01 -6.07314527e-01 -6.91095591e-01 -1.12608945e+00
-4.82284397e-01 -6.23294473e-01 -4.65081297e-02 1.23180950e+00
2.41139367e-01 -7.49843419e-01 1.07529497e+00 5.10454774e-01
-5.36740422e-01 -4.78436291e-01 -1.29081333e+00 -7.46105790e-01
-6.52416468e-01 -9.54194129e-01 1.02228832e+00 8.65989864e-01
-6.13638282e-01 -4.31079641e-02 -6.96768999e-01 7.87792683e-01
7.61403978e-01 -2.75315970e-01 1.21226978e+00 -1.16419566e+00
-1.43518165e-01 2.28466421e-01 -4.91765857e-01 -1.58130229e+00
-3.44601601e-01 -3.37586224e-01 3.07871759e-01 -1.86798680e+00
-5.88349700e-01 -1.27544928e+00 -5.43674529e-01 3.06543440e-01
4.02722239e-01 2.46901870e-01 -4.80090648e-01 -1.16046637e-01
-1.10924959e+00 4.01654512e-01 8.35009634e-01 -1.92299455e-01
-2.52023131e-01 7.65861034e-01 -6.50265515e-01 4.56152827e-01
9.06414926e-01 -4.98096198e-01 -2.20744893e-01 -6.94016337e-01
2.13425636e-01 -2.67452765e-02 2.99755901e-01 -2.04827404e+00
7.86019385e-01 1.76897749e-01 1.21508741e+00 -7.82876372e-01
5.97183585e-01 -1.35168731e+00 6.16727233e-01 4.75914389e-01
3.05824399e-01 -3.01193912e-02 2.68923193e-01 6.32880032e-01
-5.08255139e-03 5.20074725e-01 1.64606363e-01 4.13589813e-02
-7.84981251e-01 3.27385247e-01 -2.54561037e-01 -5.35522759e-01
4.86240923e-01 -4.23674732e-01 -1.33004546e-01 -6.20287716e-01
-3.68042380e-01 1.43023282e-01 4.57973391e-01 1.11423358e-01
4.52343374e-01 -1.52982700e+00 1.71879500e-01 5.00864923e-01
-1.02817081e-01 6.24073111e-02 3.19023579e-01 8.55935037e-01
-6.82999671e-01 8.23683977e-01 8.50985125e-02 -6.05603993e-01
-8.03762376e-01 -4.36723009e-02 6.19381130e-01 -2.21982785e-02
-3.59049678e-01 1.24379110e+00 -7.30317056e-01 -1.16119337e+00
8.14978182e-01 -7.00334787e-01 3.26091737e-01 -7.21864939e-01
3.44774634e-01 4.24946219e-01 3.54405433e-01 -4.63043660e-01
-7.91693509e-01 6.70162797e-01 6.78262234e-01 2.84833521e-01
1.43601882e+00 -5.44976175e-01 1.38247341e-01 2.90457662e-02
1.28521895e+00 4.52756345e-01 -1.15752828e+00 -2.69950360e-01
1.87859610e-02 -6.33897901e-01 1.67505279e-01 -9.01235282e-01
-7.88892508e-01 4.66521561e-01 1.18145144e+00 4.56542335e-02
9.22749460e-01 -2.69951850e-01 1.06606877e+00 9.62099791e-01
1.31173718e+00 -8.33491266e-01 -3.75668555e-01 9.04572725e-01
2.84813941e-01 -1.34852886e+00 -2.14139313e-01 -3.48481424e-02
2.56112695e-01 7.75887609e-01 1.87086165e-01 -1.91403985e-01
1.04886019e+00 6.91566229e-01 2.34728772e-02 1.04330592e-02
2.61123538e-01 9.01236311e-02 -1.00989126e-01 1.05955029e+00
4.78099167e-01 1.33671626e-01 3.93110782e-01 8.93244922e-01
-4.70974416e-01 1.03213407e-01 -4.99575026e-02 1.25702572e+00
-3.75845939e-01 -1.24890399e+00 -6.48181915e-01 3.99948657e-01
-4.16166455e-01 -1.17284581e-01 1.74822614e-01 5.29749095e-01
5.56193352e-01 1.31275630e+00 -3.50746751e-01 -8.54248524e-01
3.61177862e-01 -6.03535414e-01 3.89575124e-01 -1.82997286e-01
-3.23155761e-01 -1.97130635e-01 1.20601580e-01 -9.08921361e-01
-2.36354560e-01 -3.45831037e-01 -1.06366086e+00 -5.61352015e-01
-1.70415923e-01 1.18017957e-01 1.32711244e+00 8.45800757e-01
7.80824721e-01 8.70783091e-01 6.72129989e-01 -1.24971449e+00
3.85978371e-02 -6.81310058e-01 -5.46579719e-01 -5.01174569e-01
5.91171026e-01 -6.70303047e-01 -2.07886472e-01 -6.45564973e-01] | [6.405378341674805, 0.8989027142524719] |
68fed536-2bd3-4670-8805-e964fe7b85e0 | sambert-improve-aspect-sentiment-triplet | null | null | https://openreview.net/forum?id=Z9vIuaFlIXx | https://openreview.net/pdf?id=Z9vIuaFlIXx | SAMBERT: Improve Aspect Sentiment Triplet Extraction by Segmenting the Attention Maps of BERT | Aspect Sentiment Triplet Extraction (ASTE) performs fine-grained sentiment analysis in a unified way through extracting sentiment triplets comprised of aspect terms, opinion spans, and their sentiment relations in sentences. The previous works show the adoption of BERT, which simply leverages its last layer output as the word representation, is beneficial for recognizing triplet elements. However, their methods limit the potential of pretrained knowledge in BERT, since the different layers can capture multi-level linguistic information existing in sentences, which are useful for ASTE as well. In this work, we explore to access the rich pretrained knowledge by fully leveraging its attention maps of different layers. To this end, we propose to Segment the Attention Maps of BERT (SAMBERT) by taking the merits of semantic segmentation, which can effectively discriminate the desired objects from others in an image. In this procedure, we can further reason over the knowledge of different levels in these attention maps to distinguish aspect terms, opinion spans and their sentiment relations from other parts, which results in a same-shape tagging matrix of word pairs for deriving sentiment triplets. Through the extensive experiments on four benchmarks, we demonstrate our method can achieve a new state of the art. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['aspect-sentiment-triplet-extraction'] | ['natural-language-processing'] | [-7.30220526e-02 -7.31985718e-02 -7.61641562e-02 -5.87546468e-01
-6.49104357e-01 -6.91315770e-01 4.25712615e-01 5.26503101e-02
-2.38463238e-01 3.37506205e-01 3.68450254e-01 -1.30319744e-01
1.09798744e-01 -8.02434385e-01 -5.12881458e-01 -7.47184694e-01
5.59670568e-01 1.83324918e-01 1.19525634e-01 -4.74363416e-01
3.12291980e-01 -1.11206481e-02 -1.30019855e+00 6.08360231e-01
8.37543190e-01 1.60527730e+00 -5.78663535e-02 1.48138344e-01
-7.56868839e-01 8.61771643e-01 -6.49921954e-01 -8.75107884e-01
3.83309722e-02 -3.33399236e-01 -6.18456244e-01 2.74463564e-01
1.98262155e-01 -1.59996897e-02 2.60255426e-01 1.29042065e+00
2.11258218e-01 -2.02641323e-01 6.95504963e-01 -9.90021765e-01
-9.08310235e-01 8.40466261e-01 -9.15158212e-01 3.48149613e-02
-3.65739241e-02 6.03416078e-02 1.54315233e+00 -1.02448690e+00
9.86234937e-03 1.30350590e+00 5.75838745e-01 5.49160913e-02
-5.14210999e-01 -6.42984807e-01 6.62436128e-01 1.57584891e-01
-1.06156754e+00 -1.17725000e-01 1.09476864e+00 -4.09379005e-01
7.97302485e-01 2.00516075e-01 8.95313442e-01 8.86159539e-01
3.10664214e-02 1.19912791e+00 1.34688747e+00 -1.48837849e-01
-1.34014457e-01 3.51519436e-01 6.67560160e-01 6.91156805e-01
1.91534400e-01 -5.54901004e-01 -6.11524999e-01 2.97678083e-01
4.32430804e-01 8.25525522e-02 -1.33143589e-01 -1.64013803e-01
-1.10074377e+00 6.82976484e-01 7.65016913e-01 3.74232501e-01
-4.25974309e-01 -1.80270836e-01 5.21524549e-01 9.35963839e-02
6.37765706e-01 4.28313136e-01 -8.13516557e-01 8.50820094e-02
-4.49899524e-01 -3.70150745e-01 5.51739395e-01 1.07034719e+00
1.12267756e+00 -1.68606997e-01 -5.11197269e-01 7.44259536e-01
5.79247177e-01 5.42510867e-01 6.70895576e-01 -2.81952918e-01
6.36745691e-01 1.41887677e+00 -2.29774833e-01 -1.07648206e+00
-3.14986020e-01 -6.61608040e-01 -7.20983565e-01 -2.52070487e-01
-1.60491154e-01 -2.66966641e-01 -1.03141534e+00 1.48644447e+00
3.51427525e-01 -4.62134033e-02 2.18422875e-01 8.76206279e-01
1.09619319e+00 3.80006760e-01 -2.44046748e-02 4.03696746e-02
1.89719963e+00 -1.31409550e+00 -6.56241417e-01 -5.91955006e-01
4.33516592e-01 -9.00840461e-01 1.26899433e+00 1.36127248e-01
-7.29954481e-01 -6.48794055e-01 -1.13162148e+00 -1.65719956e-01
-7.23603964e-01 3.03852141e-01 9.20959830e-01 6.19111180e-01
-7.14657009e-01 1.68890476e-01 -7.04116583e-01 -8.06808993e-02
7.07045496e-01 2.99481720e-01 6.99452311e-03 1.32116109e-01
-1.35181570e+00 5.92177689e-01 3.34816307e-01 6.09039724e-01
-3.51835102e-01 -4.18275416e-01 -1.04159343e+00 2.27436781e-01
4.29800928e-01 -8.34920764e-01 1.03574872e+00 -1.24433076e+00
-1.19610858e+00 8.75807643e-01 -1.85110420e-01 1.87071934e-02
1.27279907e-01 -2.36763820e-01 -3.78470480e-01 1.26375377e-01
4.68799531e-01 5.11179090e-01 7.87160218e-01 -1.19180095e+00
-7.44601905e-01 -6.47771776e-01 6.44386113e-01 4.00578767e-01
-8.97750497e-01 5.07159494e-02 -8.70860815e-01 -5.55859566e-01
7.99517035e-02 -7.22227335e-01 -1.94973439e-01 -4.35912549e-01
-6.09747112e-01 -2.27110654e-01 5.16596735e-01 -2.16890424e-01
1.14414036e+00 -2.09905505e+00 -1.35472253e-01 2.30042234e-01
2.24394813e-01 1.56845286e-01 -1.77766204e-01 7.90719911e-02
1.60914913e-01 3.17494601e-01 -2.96820194e-01 -4.25100565e-01
2.43942872e-01 1.71843410e-01 -3.36543143e-01 -4.83627338e-03
5.69130421e-01 1.33707035e+00 -8.79644394e-01 -6.61402047e-01
-3.36925946e-02 3.01877081e-01 -3.93231541e-01 1.18393950e-01
-2.11443692e-01 3.50293607e-01 -1.10221350e+00 6.84554994e-01
7.79092789e-01 -3.59043419e-01 -9.22381803e-02 -6.82666719e-01
1.04536407e-01 3.68605047e-01 -7.90424943e-01 1.46732950e+00
-6.60023451e-01 4.48176950e-01 -1.25726387e-01 -1.04766667e+00
1.02859855e+00 1.12559907e-01 3.50656122e-01 -6.98308885e-01
5.40688634e-01 1.35332614e-01 -1.19418152e-01 -5.62598586e-01
4.83864516e-01 -2.73410589e-01 -3.98561776e-01 4.28595304e-01
1.05667964e-01 -1.81219764e-02 2.84012169e-01 9.95715931e-02
6.16593540e-01 1.01679396e-02 1.17970645e-01 -3.31130624e-01
8.09729278e-01 -3.40361558e-02 5.00073433e-01 3.77141625e-01
-1.63790900e-02 5.69974720e-01 7.67193139e-01 -3.22049230e-01
-4.84792292e-01 -7.33857393e-01 -1.00609265e-01 9.75013912e-01
4.77915615e-01 -4.75213289e-01 -6.77777767e-01 -1.12751830e+00
-3.40444088e-01 2.32218429e-01 -9.42037404e-01 -3.05658549e-01
-4.14466977e-01 -1.00749207e+00 1.86396450e-01 7.53292322e-01
7.01750040e-01 -1.29388762e+00 -4.12651181e-01 -3.54004055e-02
-3.30334008e-01 -1.45363903e+00 -5.29104471e-01 2.78573275e-01
-5.44105768e-01 -9.24116611e-01 -7.09196150e-01 -9.15734529e-01
8.84304345e-01 3.55171204e-01 1.37505054e+00 1.23470940e-01
2.57181853e-01 1.18758015e-01 -6.35507405e-01 -5.73549807e-01
3.03615212e-01 3.55346680e-01 -4.17147487e-01 5.58460355e-01
7.59077787e-01 -4.05491054e-01 -7.82733142e-01 3.17100048e-01
-1.02947283e+00 1.24879986e-01 8.85077178e-01 7.76737094e-01
7.57342041e-01 4.67632860e-02 4.46048766e-01 -1.35974574e+00
6.78114057e-01 -4.31899518e-01 -2.18674481e-01 2.96934605e-01
-4.41710114e-01 -1.38111208e-02 7.20325708e-01 -9.13223550e-02
-1.19311929e+00 -2.81974286e-01 -3.50051135e-01 -2.13292584e-01
-6.83357613e-03 9.04524386e-01 -5.37920773e-01 3.37292373e-01
3.69739011e-02 2.76096165e-01 -4.81456608e-01 -2.90727586e-01
4.44667429e-01 8.10158610e-01 1.86959952e-01 -5.89957893e-01
5.39363742e-01 6.34088576e-01 -2.82883197e-01 -3.09305876e-01
-1.75500345e+00 -6.72977746e-01 -6.86834157e-01 4.16654609e-02
1.03359854e+00 -1.02074277e+00 -5.25963604e-01 6.20339513e-01
-1.03349042e+00 3.12278062e-01 -2.35476568e-01 1.30163148e-01
-1.65922359e-01 1.94870248e-01 -4.90341723e-01 -6.04844451e-01
-4.76659089e-01 -1.33545732e+00 1.51408529e+00 5.20984888e-01
6.87825978e-02 -1.10122085e+00 -1.14259005e-01 6.03512049e-01
1.50358930e-01 5.81040420e-02 7.99176812e-01 -7.78673649e-01
-5.44602692e-01 -6.17946163e-02 -5.13568580e-01 5.56663156e-01
3.38165164e-01 -1.52162194e-01 -1.15976882e+00 3.56417634e-02
3.31594229e-01 -3.32132399e-01 1.29623473e+00 1.49480209e-01
1.02365625e+00 -2.80955613e-01 -1.17950141e-02 5.52554488e-01
1.37732673e+00 1.31540466e-02 5.14976740e-01 4.69590396e-01
1.18139803e+00 7.60911822e-01 8.64640594e-01 1.86628893e-01
6.83377862e-01 2.93979108e-01 4.64264482e-01 -4.29654688e-01
1.69368535e-02 -3.89683396e-02 4.57350969e-01 1.29126000e+00
1.44098043e-01 -9.66517180e-02 -6.19504452e-01 7.75806606e-01
-1.91337752e+00 -4.85584348e-01 -6.70041442e-02 1.50397158e+00
8.05798829e-01 4.41484541e-01 -2.38820061e-01 -7.06894994e-02
6.07960522e-01 5.54827332e-01 -6.57571971e-01 -4.96172786e-01
-3.34298879e-01 1.52774811e-01 1.95687190e-01 2.92835245e-03
-1.16641831e+00 1.08207238e+00 4.66023111e+00 9.80774999e-01
-1.10679400e+00 -1.74641218e-02 8.37169468e-01 1.47592932e-01
-8.31437051e-01 -5.62002473e-02 -8.08376491e-01 4.50892836e-01
4.05083090e-01 1.45425647e-01 -3.63733806e-02 6.96187675e-01
-1.59401789e-01 2.08089158e-01 -7.66659141e-01 7.80835986e-01
3.07401955e-01 -8.72179925e-01 3.88960212e-01 -2.77384669e-01
9.91023421e-01 -4.44441326e-02 2.08819300e-01 4.98862177e-01
2.11017013e-01 -7.62995183e-01 7.34085739e-01 4.00832117e-01
4.29768592e-01 -7.51848578e-01 1.15299714e+00 1.11326486e-01
-1.60555828e+00 5.53211048e-02 -4.56734329e-01 -1.62426680e-02
1.11364484e-01 8.92896056e-01 -3.76108140e-01 9.55268919e-01
7.91891277e-01 1.18226683e+00 -7.19180822e-01 6.36227250e-01
-5.87831855e-01 5.27942300e-01 -6.44576503e-03 -2.25960910e-01
5.79217732e-01 -4.52942401e-01 1.40890643e-01 1.16030228e+00
9.56341550e-02 -4.91211228e-02 1.77387267e-01 7.30273783e-01
-1.63848698e-01 2.87763566e-01 -2.01144546e-01 -2.07327247e-01
-4.82456386e-02 1.70430613e+00 -1.03951025e+00 -4.59178537e-01
-6.94229782e-01 7.81213939e-01 4.34275508e-01 3.92703354e-01
-7.93201923e-01 -4.28094566e-01 6.45876229e-01 -4.01383579e-01
7.10241675e-01 1.05024986e-01 -7.40026176e-01 -1.42101014e+00
4.09433216e-01 -8.37108970e-01 2.67332673e-01 -8.17835927e-01
-1.54888093e+00 1.05319428e+00 -4.50726777e-01 -1.42734599e+00
1.39256030e-01 -8.59303415e-01 -6.75041616e-01 9.04094756e-01
-1.90130007e+00 -1.73077703e+00 -3.18981022e-01 5.88418782e-01
7.03055143e-01 2.29091849e-02 4.93849158e-01 2.59667486e-01
-7.52890527e-01 5.84718466e-01 -2.98535854e-01 4.54433054e-01
4.62572277e-01 -1.40601599e+00 3.50710779e-01 7.86294103e-01
4.04883444e-01 8.33085895e-01 1.92854211e-01 -3.84598523e-01
-1.12133884e+00 -1.10428584e+00 6.92849278e-01 -4.74647641e-01
9.84316468e-01 -2.97475874e-01 -7.30645299e-01 5.65949500e-01
3.23164463e-01 -2.51876384e-01 8.03767145e-01 4.62916553e-01
-5.10608435e-01 -2.67375439e-01 -5.89767933e-01 6.16863132e-01
8.87762547e-01 -7.70870090e-01 -7.00406551e-01 6.76892102e-02
9.39556181e-01 -2.06505433e-01 -7.11060047e-01 6.45400763e-01
6.02580428e-01 -1.12306833e+00 8.36389840e-01 -5.06029725e-01
9.76397038e-01 -4.23074961e-01 -2.96399575e-02 -1.34796643e+00
-3.81377600e-02 -1.30608911e-02 9.91948545e-02 1.75030923e+00
6.94781482e-01 -6.55246615e-01 5.84581435e-01 2.54925460e-01
-2.86564052e-01 -1.28218770e+00 -4.99385625e-01 -2.63676971e-01
-6.37745485e-03 -5.37989974e-01 1.13104522e+00 7.09776521e-01
-1.11300848e-01 9.69660342e-01 -9.30499732e-02 7.20101520e-02
1.11161485e-01 8.16675961e-01 4.21945304e-01 -1.06357193e+00
-1.32545456e-01 -7.00234175e-01 -3.37496310e-01 -1.38470232e+00
3.22194487e-01 -8.22437525e-01 3.25412042e-02 -1.83012605e+00
4.67396110e-01 -4.42584068e-01 -8.25271726e-01 5.53105712e-01
-8.05981576e-01 5.57438672e-01 1.76757693e-01 1.48648769e-01
-9.98009205e-01 8.47804129e-01 1.78510880e+00 -5.74665606e-01
2.27064863e-01 -3.53795141e-02 -1.43305147e+00 1.09911823e+00
6.54456317e-01 -3.35257053e-01 -5.16973972e-01 -7.40396678e-01
6.62638664e-01 -4.40252185e-01 1.48260500e-03 -6.02799833e-01
1.54113755e-01 1.13647975e-01 3.58280420e-01 -6.98331416e-01
2.61089474e-01 -9.88331199e-01 -5.14199853e-01 -5.63162602e-02
-7.06882179e-02 -4.40556519e-02 1.82453349e-01 3.93067658e-01
-7.44632840e-01 -2.36312971e-01 3.30416858e-01 -2.62720346e-01
-7.76005626e-01 5.03792703e-01 3.24747637e-02 2.25280717e-01
6.28448665e-01 -1.76411837e-01 -3.94435763e-01 -8.81861821e-02
-4.08498645e-01 5.44528008e-01 3.40603530e-01 5.54302633e-01
4.10356045e-01 -1.33809042e+00 -3.97473902e-01 2.78889298e-01
4.85319465e-01 2.33749107e-01 3.91793072e-01 9.42164421e-01
-2.85862815e-02 2.09338829e-01 -1.14413425e-01 -6.08489096e-01
-9.60313261e-01 5.61918020e-01 1.41300648e-01 -6.49232686e-01
-9.65979770e-02 8.97663832e-01 7.22435713e-01 -4.56640720e-01
-2.43170246e-01 -4.64828372e-01 -9.76129115e-01 4.89082843e-01
3.14900160e-01 -3.62877905e-01 1.77171856e-01 -8.25731814e-01
-3.30847800e-01 1.16093636e+00 -2.16607943e-01 9.72667709e-02
1.08016670e+00 -4.24575984e-01 -3.92626554e-01 5.58758080e-01
1.37226796e+00 3.46666902e-01 -1.03308189e+00 -2.68566668e-01
-3.87918204e-02 -2.62294024e-01 -1.04453363e-01 -7.12596655e-01
-1.58765757e+00 1.08698070e+00 9.36133787e-02 4.29403961e-01
1.48337257e+00 2.40181431e-01 1.06876898e+00 1.51845887e-01
1.32548690e-01 -7.69100606e-01 8.71656090e-02 6.85344100e-01
5.87297618e-01 -1.49955142e+00 -2.37782508e-01 -5.23207784e-01
-9.41856682e-01 8.26049566e-01 7.07957983e-01 -1.36950582e-01
7.43047893e-01 1.54616043e-01 5.65833867e-01 -5.85844636e-01
-4.49481994e-01 -7.04602957e-01 5.95932126e-01 3.07039678e-01
3.91684502e-01 8.99038017e-02 -2.81386346e-01 1.27875483e+00
-4.47584391e-01 -4.49671268e-01 1.99906737e-01 5.97688615e-01
-3.42378229e-01 -9.92854238e-01 -7.78857246e-02 6.33243263e-01
-6.14754021e-01 -5.71068585e-01 -6.05144382e-01 3.05214882e-01
3.45071375e-01 1.00825322e+00 -1.89688187e-02 -3.73230487e-01
5.14428854e-01 -2.42819607e-01 7.69730434e-02 -7.42965937e-01
-9.21006501e-01 1.66267648e-01 1.23026215e-01 -2.94481546e-01
-7.67030239e-01 -2.93518066e-01 -1.21245480e+00 2.62763649e-01
-4.78162020e-01 3.21387351e-01 4.85981524e-01 1.32044339e+00
3.12663168e-01 9.80918825e-01 9.03298259e-01 -4.12492096e-01
-2.02703312e-01 -9.00391996e-01 -5.77165484e-01 5.67897975e-01
3.63190681e-01 -6.45057082e-01 -1.59749761e-01 -6.18731268e-02] | [11.475707054138184, 6.592503547668457] |
51f47584-920b-4cee-ba50-b900fe9ad636 | scallop-a-language-for-neurosymbolic | 2304.04812 | null | https://arxiv.org/abs/2304.04812v1 | https://arxiv.org/pdf/2304.04812v1.pdf | Scallop: A Language for Neurosymbolic Programming | We present Scallop, a language which combines the benefits of deep learning and logical reasoning. Scallop enables users to write a wide range of neurosymbolic applications and train them in a data- and compute-efficient manner. It achieves these goals through three key features: 1) a flexible symbolic representation that is based on the relational data model; 2) a declarative logic programming language that is based on Datalog and supports recursion, aggregation, and negation; and 3) a framework for automatic and efficient differentiable reasoning that is based on the theory of provenance semirings. We evaluate Scallop on a suite of eight neurosymbolic applications from the literature. Our evaluation demonstrates that Scallop is capable of expressing algorithmic reasoning in diverse and challenging AI tasks, provides a succinct interface for machine learning programmers to integrate logical domain knowledge, and yields solutions that are comparable or superior to state-of-the-art models in terms of accuracy. Furthermore, Scallop's solutions outperform these models in aspects such as runtime and data efficiency, interpretability, and generalizability. | ['Mayur Naik', 'Jiani Huang', 'Ziyang Li'] | 2023-04-10 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [-4.40212965e-01 2.01000243e-01 -5.01723886e-01 -3.03622931e-01
-1.20499596e-01 -6.81834698e-01 8.49294722e-01 4.54393983e-01
-7.15590715e-02 6.64916873e-01 8.43149498e-02 -8.09384763e-01
-4.45376575e-01 -1.15757358e+00 -9.27938819e-01 -2.04570033e-02
-5.06428123e-01 8.34549546e-01 5.54008782e-01 -3.57068926e-01
2.92754352e-01 7.09068656e-01 -1.73867452e+00 6.18482649e-01
9.07123923e-01 1.24928617e+00 -6.10451102e-01 4.53262150e-01
-2.24239051e-01 1.92077529e+00 -3.86039436e-01 -4.64741379e-01
1.30357057e-01 7.17717633e-02 -1.24336481e+00 -7.66877592e-01
4.28602129e-01 -2.05245197e-01 -2.97938675e-01 6.72000289e-01
-3.47168408e-02 -2.89820790e-01 4.12238389e-01 -1.73273540e+00
-8.09384167e-01 9.56818104e-01 -1.46899549e-02 2.67787348e-03
3.10263187e-01 2.74481833e-01 1.18838322e+00 -4.06435162e-01
6.54128075e-01 1.38891864e+00 8.23032916e-01 4.57158387e-01
-1.43985069e+00 -5.95490873e-01 -1.12072960e-01 3.15170646e-01
-1.21786666e+00 -5.23153484e-01 1.90655008e-01 -5.47048867e-01
1.37293231e+00 2.48493105e-01 8.20779979e-01 4.55237091e-01
4.87319440e-01 7.50642717e-01 9.80535626e-01 -2.67228514e-01
6.63558483e-01 -8.55817944e-02 3.22003275e-01 1.31331396e+00
2.13070676e-01 1.13844022e-01 -9.75916207e-01 -3.80147845e-01
7.08788157e-01 -4.60555315e-01 3.94010067e-01 -5.41031897e-01
-1.17620790e+00 4.67476308e-01 4.46373016e-01 -7.51279201e-03
4.68764920e-03 9.00137186e-01 7.47996032e-01 3.74642223e-01
-3.44632603e-02 9.50883567e-01 -5.69632292e-01 -2.36064151e-01
-7.56222606e-01 5.90972900e-01 1.36098325e+00 9.76531684e-01
2.67206609e-01 1.69184342e-01 1.01898834e-01 2.78058439e-01
3.38119566e-01 3.63625973e-01 1.18215203e-01 -1.57000732e+00
1.19904682e-01 9.78762805e-01 -2.14003757e-01 -6.21506155e-01
-3.34036857e-01 -1.06552921e-01 -3.62829924e-01 7.72784650e-01
2.96035916e-01 8.57366547e-02 -5.81967294e-01 1.68203998e+00
5.80507480e-02 2.93132633e-01 5.94350219e-01 4.20855492e-01
1.02057469e+00 4.51579034e-01 1.33922100e-01 1.25472993e-01
1.23962665e+00 -6.32946312e-01 -3.28092843e-01 -1.45843685e-01
6.93210602e-01 1.63678713e-02 1.17116666e+00 8.62996519e-01
-1.51748502e+00 -1.00379199e-01 -1.33687961e+00 -5.71525931e-01
-5.03830254e-01 -3.60472232e-01 1.53136241e+00 5.80083191e-01
-1.16249323e+00 5.57488382e-01 -1.15078890e+00 9.72517356e-02
7.86670089e-01 5.37388623e-01 -1.62250444e-01 7.35658854e-02
-9.91947770e-01 1.15343308e+00 7.72009134e-01 -3.94122899e-01
-1.01714063e+00 -1.07348049e+00 -8.88469875e-01 1.85559466e-01
4.77664113e-01 -7.94197500e-01 1.64121282e+00 -6.14099383e-01
-1.40154207e+00 6.63783967e-01 2.21763909e-01 -9.36789513e-01
3.39036912e-01 -1.01084970e-01 -3.16733003e-01 1.87586322e-01
-1.56934276e-01 5.23210883e-01 2.54162133e-01 -7.10609436e-01
-6.02472186e-01 -3.09585869e-01 5.99960506e-01 -8.47153440e-02
-5.58773838e-02 8.99173543e-02 -1.88709721e-01 -2.29156315e-01
-1.10623166e-01 -6.39522731e-01 2.56352603e-01 4.86145526e-01
-1.12208888e-01 -4.87083167e-01 6.65967047e-01 -1.85221419e-01
8.47236812e-01 -2.05200243e+00 3.49993348e-01 2.28933692e-01
6.57408297e-01 1.49219096e-01 3.18545401e-01 3.47738922e-01
4.86335764e-03 3.65183055e-01 -1.93866417e-01 3.57083827e-01
4.01145458e-01 5.16858995e-01 -5.83744764e-01 1.87108532e-01
4.79210824e-01 1.01756895e+00 -1.08877218e+00 -7.86821485e-01
2.87609361e-02 -1.53559225e-03 -9.14459646e-01 3.98397520e-02
-1.05059803e+00 -1.71588242e-01 -2.62211353e-01 8.49856734e-01
1.37301266e-01 -3.46889734e-01 4.87229884e-01 1.32091060e-01
-1.54522538e-01 7.60346770e-01 -1.10144949e+00 1.83482432e+00
-5.58372319e-01 6.34476483e-01 -8.49732682e-02 -5.19892514e-01
6.29861891e-01 2.73409694e-01 2.37854019e-01 -5.26601672e-01
-1.08162358e-01 2.21641362e-01 -2.96424311e-02 -5.20756125e-01
3.19119275e-01 -1.66161194e-01 -2.91642487e-01 7.19168663e-01
9.22976732e-02 -7.46251106e-01 5.39442599e-01 4.09081489e-01
1.46383834e+00 7.02245235e-01 3.77897680e-01 -4.90254372e-01
3.63887995e-01 1.67426512e-01 6.33358181e-01 5.73731482e-01
3.41745406e-01 -2.82516390e-01 1.15746510e+00 -7.80106604e-01
-7.79496610e-01 -1.48782802e+00 -2.29407310e-01 1.34262025e+00
-2.16364767e-02 -1.00336349e+00 -4.60149854e-01 -2.27575749e-01
2.17763871e-01 1.09414220e+00 -3.71773392e-01 -1.95613965e-01
-3.68526548e-01 -2.68997639e-01 1.32444870e+00 9.93779063e-01
5.41398287e-01 -1.20504689e+00 -1.00425231e+00 -1.43075973e-01
8.37477371e-02 -8.77932489e-01 4.66961920e-01 2.90142804e-01
-9.42951381e-01 -1.34869444e+00 7.07437694e-01 -6.18526459e-01
2.19603941e-01 -5.87541699e-01 1.52238309e+00 3.46289545e-01
-2.54523218e-01 1.16397895e-01 6.29771575e-02 -7.86901772e-01
-7.21527100e-01 -1.38594389e-01 -7.35619711e-03 -8.98107886e-01
3.02227020e-01 -7.70944953e-01 3.03459138e-01 -2.25443557e-01
-1.06465983e+00 2.87742645e-01 2.01134160e-01 6.27248883e-01
2.76128262e-01 4.36539918e-01 3.42220753e-01 -9.58240986e-01
6.59587085e-01 -3.24586451e-01 -1.14742708e+00 5.66045225e-01
-7.88739681e-01 6.11806154e-01 9.05775309e-01 5.37340678e-02
-8.30532432e-01 -1.99230611e-01 8.64934698e-02 -3.29770260e-02
1.20961949e-01 6.93616986e-01 -2.20426563e-02 -7.66568631e-02
1.01288342e+00 -7.57646561e-02 1.83845699e-01 4.59024347e-02
6.66841745e-01 3.46958935e-01 9.71177697e-01 -1.52591991e+00
5.60192585e-01 3.26624095e-01 4.01398957e-01 -5.01927197e-01
-6.74209058e-01 3.94841075e-01 -4.92177933e-01 2.69991845e-01
4.06761378e-01 -7.67490983e-01 -1.41389096e+00 4.35682684e-01
-1.02936149e+00 -5.87121427e-01 -5.99873364e-01 5.09893000e-02
-8.64690483e-01 2.82148365e-02 -9.61520135e-01 -6.21128261e-01
-2.81853914e-01 -1.17218900e+00 6.43601716e-01 -8.81354883e-02
-4.73463297e-01 -9.59033370e-01 -1.27081618e-01 7.37203583e-02
1.24604680e-01 3.24128687e-01 1.65149546e+00 -6.53792977e-01
-8.36472154e-01 7.01752454e-02 -2.44882509e-01 3.80404949e-01
-3.58207017e-01 3.02106470e-01 -8.25136364e-01 1.93252012e-01
-4.01386261e-01 -9.39324021e-01 2.80712694e-01 -2.31878415e-01
1.42268622e+00 -4.98938024e-01 -2.55206991e-02 7.82834113e-01
1.20400321e+00 -1.95327550e-01 7.41804779e-01 3.42683703e-01
2.80062973e-01 2.96228200e-01 2.60312796e-01 2.65237927e-01
6.46283865e-01 1.35262132e-01 7.83523560e-01 4.63549793e-01
4.87541184e-02 -2.67540842e-01 2.45117918e-01 3.17740560e-01
-1.93376467e-01 3.50381911e-01 -1.39754808e+00 4.13910210e-01
-2.15072942e+00 -1.13473010e+00 3.80508713e-02 1.82647967e+00
1.36914647e+00 4.91005421e-01 1.60561860e-01 2.78019994e-01
-5.76816201e-02 -2.13345185e-01 -6.02662802e-01 -8.15184712e-01
1.41915679e-01 8.01891029e-01 3.25125128e-01 3.61084253e-01
-7.76096404e-01 1.09533906e+00 7.41837597e+00 2.69920319e-01
-9.64411795e-01 -1.27716169e-01 1.35264859e-01 -1.58286110e-01
-3.53253752e-01 -6.36825012e-03 -5.24537385e-01 1.02469221e-01
1.14996624e+00 -4.38713133e-01 1.05290210e+00 1.00265789e+00
-4.81568724e-01 2.46703494e-02 -1.79640079e+00 6.55112088e-01
-8.18446949e-02 -2.03002739e+00 -1.57158628e-01 -2.57960677e-01
3.92470330e-01 1.85249284e-01 -1.17565520e-01 4.10571784e-01
9.74138498e-01 -1.54865444e+00 1.09905267e+00 5.56946814e-01
6.83228016e-01 -8.07906806e-01 5.14414132e-01 3.09053779e-01
-8.15443814e-01 -4.30236369e-01 6.27222806e-02 -4.70015079e-01
-3.67003053e-01 1.63206905e-01 -8.58650148e-01 4.81847256e-01
7.80129731e-01 8.20865214e-01 -6.46417797e-01 8.25180888e-01
-7.47109711e-01 4.25361395e-01 -4.86840576e-01 -2.11069528e-02
-3.54651324e-02 2.45284334e-01 3.41214053e-03 1.15026379e+00
-1.79139808e-01 9.50943828e-02 3.70980278e-02 1.32595026e+00
-4.20301855e-02 -5.03356159e-01 -5.72336316e-01 -3.53786618e-01
7.81639516e-01 9.01175618e-01 -4.45734113e-01 -4.50606823e-01
-2.50402421e-01 5.48943691e-02 4.38052654e-01 3.98786850e-02
-8.52988780e-01 -5.00278652e-01 6.66268349e-01 7.13104233e-02
6.98146373e-02 -3.63330066e-01 -7.94195116e-01 -1.04275727e+00
5.15376776e-02 -1.25692320e+00 5.55610716e-01 -1.13653314e+00
-1.14945555e+00 4.81945187e-01 3.24664026e-01 -4.87925649e-01
-5.60161471e-01 -1.01604652e+00 -2.29086444e-01 6.35419369e-01
-1.23426962e+00 -1.46904504e+00 -1.83858842e-01 5.32411933e-01
-2.46965006e-01 -2.56790608e-01 1.09958839e+00 -3.93116400e-02
-1.47071719e-01 2.27226257e-01 -4.92491990e-01 1.18214853e-01
2.34092116e-01 -1.53340626e+00 4.99398261e-01 7.09339082e-01
-7.92733394e-03 1.09816289e+00 4.65333432e-01 -4.73723948e-01
-1.93391740e+00 -8.79455388e-01 6.49936199e-01 -7.49367297e-01
1.07667315e+00 -3.57673049e-01 -7.15950906e-01 1.27207088e+00
-8.51824507e-02 1.93821684e-01 5.56407571e-01 4.67765778e-01
-1.02465153e+00 -5.09034753e-01 -1.05870736e+00 7.48887062e-01
1.11835587e+00 -7.68980682e-01 -1.19729161e+00 4.80805755e-01
7.69415259e-01 -7.87907779e-01 -1.23589444e+00 3.33030283e-01
7.40092933e-01 -9.45238590e-01 1.00630426e+00 -9.98254955e-01
7.96892464e-01 -5.35714328e-01 -2.81963319e-01 -7.86603630e-01
-1.83876976e-01 -6.98579311e-01 -8.48095417e-01 8.65045190e-01
2.48741135e-01 -6.12310648e-01 4.84482020e-01 8.77170682e-01
-2.26448789e-01 -8.62738132e-01 -4.96649802e-01 -6.52678490e-01
2.06000209e-01 -9.21589792e-01 1.03355730e+00 6.45184159e-01
6.78480208e-01 3.60356867e-01 3.69062871e-01 1.81780741e-01
3.66606981e-01 3.46634269e-01 6.51655674e-01 -1.53372419e+00
-5.12819052e-01 -6.70773864e-01 -4.76654947e-01 -6.23547018e-01
6.54665589e-01 -1.40302253e+00 -1.64465770e-01 -1.62756264e+00
-1.62177682e-01 -7.25527406e-01 -1.91075355e-01 1.34514546e+00
7.95788705e-01 -4.19810675e-02 -4.99458015e-02 1.36948362e-01
-6.61504924e-01 5.40921912e-02 7.13561177e-01 -6.69028834e-02
-3.56261292e-03 -3.21529388e-01 -9.45448458e-01 1.08070147e+00
5.03267825e-01 -1.32978782e-01 -6.15991712e-01 -5.43081760e-01
8.03280354e-01 -2.51512136e-02 6.68251872e-01 -1.15417302e+00
5.44349074e-01 -6.08832181e-01 2.43086591e-01 -2.83704489e-01
3.51790816e-01 -6.65131569e-01 1.22119663e-02 5.32104313e-01
-5.64105153e-01 2.30942935e-01 5.87404132e-01 -4.84689362e-02
9.13387612e-02 -1.35586083e-01 6.90117359e-01 -2.80093312e-01
-9.57616448e-01 -1.55127436e-01 -2.54903346e-01 3.92670095e-01
9.84362543e-01 2.31897160e-01 -9.17143941e-01 2.86996327e-02
-1.61883101e-01 3.88169050e-01 7.48831093e-01 3.17164093e-01
4.96118754e-01 -1.05209243e+00 -2.87798345e-01 2.71542788e-01
1.84399635e-01 2.53155351e-01 -6.48479402e-01 5.76354384e-01
-1.24384427e+00 4.53520596e-01 -6.27997816e-01 -3.06918532e-01
-1.05690503e+00 6.38147652e-01 4.47410136e-01 -1.67746201e-01
-6.49406016e-01 5.69013059e-01 -3.32146972e-01 -8.19186568e-01
4.77336258e-01 -9.34281409e-01 4.02307540e-01 -4.88013506e-01
7.70693719e-01 3.03844452e-01 1.35720298e-01 2.15723976e-01
-7.12904096e-01 -9.05000195e-02 2.14249179e-01 -1.42081514e-01
1.55499959e+00 7.87253320e-01 -1.01721299e+00 6.49660051e-01
2.17670366e-01 -1.01516411e-01 -8.42476189e-01 -1.06628329e-01
2.99247324e-01 -6.18928764e-03 -1.68676358e-02 -1.30659127e+00
-5.99718988e-01 6.64993584e-01 -2.67589360e-01 1.76747739e-01
9.68288004e-01 8.58541057e-02 5.09502053e-01 1.06320393e+00
5.37391067e-01 -9.98552501e-01 5.52983880e-02 7.86648691e-01
7.23848104e-01 -3.74164730e-01 4.20335740e-01 -3.16945553e-01
-3.77516299e-01 1.31186700e+00 7.40261197e-01 -1.68721437e-01
3.94245118e-01 9.84029830e-01 -3.56083393e-01 -3.40644836e-01
-1.25071406e+00 2.74513990e-01 -4.30014767e-02 5.88178396e-01
4.32108074e-01 -8.57629254e-02 1.46814108e-01 4.90271032e-01
-5.54876626e-01 7.26394832e-01 5.07469118e-01 1.28833795e+00
-8.79056528e-02 -1.18428397e+00 -1.49816066e-01 3.50089818e-01
-2.63933539e-01 -2.96850115e-01 -4.63546693e-01 1.02632082e+00
3.68934661e-01 5.80375850e-01 1.08688988e-01 -1.63744643e-01
2.46421285e-02 2.15516746e-01 9.93659377e-01 -8.01312327e-01
-7.59657919e-01 -7.61011004e-01 5.11875749e-01 -7.09606946e-01
-1.20005600e-01 -3.61953050e-01 -1.91808653e+00 -8.42696548e-01
2.55796283e-01 1.43415123e-01 6.76743507e-01 1.04396069e+00
5.19951403e-01 6.47386491e-01 -3.16234589e-01 -2.28347331e-01
-7.14411318e-01 -3.27327073e-01 -4.75623608e-01 7.80928507e-02
1.50256291e-01 -5.18964231e-01 1.68717563e-01 1.03391260e-01] | [9.097552299499512, 7.161159992218018] |
2ee67c1c-87d6-4c34-9858-e46f4b09fe14 | billion-user-customer-lifetime-value | 2208.13358 | null | https://arxiv.org/abs/2208.13358v1 | https://arxiv.org/pdf/2208.13358v1.pdf | Billion-user Customer Lifetime Value Prediction: An Industrial-scale Solution from Kuaishou | Customer Life Time Value (LTV) is the expected total revenue that a single user can bring to a business. It is widely used in a variety of business scenarios to make operational decisions when acquiring new customers. Modeling LTV is a challenging problem, due to its complex and mutable data distribution. Existing approaches either directly learn from posterior feature distributions or leverage statistical models that make strong assumption on prior distributions, both of which fail to capture those mutable distributions. In this paper, we propose a complete set of industrial-level LTV modeling solutions. Specifically, we introduce an Order Dependency Monotonic Network (ODMN) that models the ordered dependencies between LTVs of different time spans, which greatly improves model performance. We further introduce a Multi Distribution Multi Experts (MDME) module based on the Divide-and-Conquer idea, which transforms the severely imbalanced distribution modeling problem into a series of relatively balanced sub-distribution modeling problems hence greatly reduces the modeling complexity. In addition, a novel evaluation metric Mutual Gini is introduced to better measure the distribution difference between the estimated value and the ground-truth label based on the Lorenz Curve. The ODMN framework has been successfully deployed in many business scenarios of Kuaishou, and achieved great performance. Extensive experiments on real-world industrial data demonstrate the superiority of the proposed methods compared to state-of-the-art baselines including ZILN and Two-Stage XGBoost models. | ['Yang song', 'Xiao Fang', 'Naijun Yang', 'Guangcui Shao', 'Kunpeng Li'] | 2022-08-29 | null | null | null | null | ['value-prediction'] | ['computer-code'] | [-3.01205486e-01 -3.45510066e-01 -7.33543694e-01 -8.39406490e-01
-5.88193774e-01 -3.03687990e-01 4.36083883e-01 6.20736554e-02
-4.18293625e-02 7.82981932e-01 -2.74512947e-01 -3.92996073e-01
-5.89751542e-01 -1.01846337e+00 -5.40885746e-01 -9.41737115e-01
7.80731589e-02 1.14425302e+00 -2.96003342e-01 -3.47811803e-02
1.09931290e-01 2.80184239e-01 -1.20161009e+00 7.94205815e-02
9.77410257e-01 1.44641376e+00 3.09740193e-02 4.99556921e-02
-1.85541287e-01 9.50688422e-01 -7.65492320e-01 -5.57452500e-01
2.56699830e-01 -1.03794159e-02 -2.09305167e-01 -8.17334279e-02
-2.57346064e-01 -2.99971014e-01 -1.63910583e-01 9.60329592e-01
3.74440312e-01 6.21702373e-02 1.11202848e+00 -1.92031002e+00
-4.57939982e-01 8.86100531e-01 -1.16828167e+00 2.11741641e-01
-3.79476935e-01 -1.70180887e-01 1.06204736e+00 -1.91398412e-01
-2.85439529e-02 1.28405094e+00 4.80322897e-01 -1.07819587e-01
-1.21598554e+00 -1.02761126e+00 5.30579746e-01 4.70509708e-01
-1.36540985e+00 3.92140508e-01 9.33710635e-01 -3.88352215e-01
8.62656474e-01 -4.10193093e-02 4.36499089e-01 1.05593073e+00
6.29057586e-01 1.05379581e+00 1.19686508e+00 2.78364867e-02
1.50320604e-01 3.80157888e-01 5.03999963e-02 -1.74044311e-01
3.28850567e-01 2.73834802e-02 -2.69793391e-01 -1.34574160e-01
6.48854852e-01 4.83847797e-01 1.25810280e-01 -4.14051473e-01
-6.80848658e-01 1.01525104e+00 2.91182876e-01 -4.86152731e-02
-6.36602283e-01 5.86293973e-02 2.74190336e-01 2.21776187e-01
7.33342469e-01 -6.78806752e-02 -7.90804625e-01 -1.08765386e-01
-8.37742329e-01 3.30224901e-01 7.81393826e-01 1.10790503e+00
5.47704995e-01 -5.23528345e-02 -2.63434798e-01 8.97350252e-01
5.08284152e-01 5.75222552e-01 2.39388436e-01 -5.38491726e-01
6.09926641e-01 4.94795799e-01 1.27472341e-01 -1.06917918e+00
-2.78880060e-01 -1.05241919e+00 -9.28961098e-01 -4.96282466e-02
3.41442466e-01 -3.62008554e-03 -7.38554239e-01 1.55341625e+00
2.35505566e-01 1.48318633e-01 -8.53072628e-02 6.48466170e-01
1.78934216e-01 6.35584354e-01 1.22536354e-01 -2.47008219e-01
1.02156734e+00 -7.88927078e-01 -7.51185358e-01 -7.62967765e-02
1.07046649e-01 -6.75354302e-01 6.22499585e-01 1.03735650e+00
-7.52024770e-01 -4.28676546e-01 -1.05914402e+00 5.20832121e-01
-3.00778836e-01 -6.83458820e-02 7.85545647e-01 6.06987357e-01
-3.02271783e-01 6.16988838e-01 -7.18055725e-01 3.62355977e-01
5.73276997e-01 3.08692455e-01 2.54849255e-01 -3.58599931e-01
-1.32694447e+00 7.90185332e-01 5.77931166e-01 1.89343378e-01
-9.18134570e-01 -7.23655641e-01 -5.59032321e-01 2.60992199e-01
6.22247994e-01 -3.82412523e-01 1.36278582e+00 -6.32947206e-01
-1.27871919e+00 1.55801490e-01 1.92718416e-01 -5.08518517e-01
6.88208401e-01 -5.27585968e-02 -6.64716780e-01 -3.67590666e-01
-7.34930336e-02 1.30250871e-01 7.28985488e-01 -1.32814443e+00
-1.11235988e+00 -5.13281763e-01 -1.74322635e-01 5.47428206e-02
-2.24461734e-01 -3.02701056e-01 -3.07948172e-01 -6.93774760e-01
-2.34111622e-02 -7.16524243e-01 -1.36727527e-01 -9.31207180e-01
-5.39189994e-01 -5.17962515e-01 6.75003588e-01 -7.31986642e-01
1.35891116e+00 -1.84958577e+00 -1.91791132e-01 6.72622025e-01
9.58068967e-02 -1.76310882e-01 1.60626635e-01 3.73093098e-01
-7.79360384e-02 -7.33704641e-02 8.83340538e-02 -1.49138346e-01
5.33281982e-01 3.90183836e-01 -2.82395989e-01 3.71799976e-01
-1.42005503e-01 7.57028282e-01 -6.97378755e-01 -1.67459339e-01
2.10720569e-01 3.98166835e-01 1.16079897e-02 1.17953844e-01
-2.65032023e-01 2.90928602e-01 -3.54365110e-01 9.64910984e-01
1.20978522e+00 -1.75812170e-01 2.47214347e-01 -2.69034922e-01
3.03897500e-01 -6.59170747e-03 -1.18850565e+00 1.08634353e+00
-6.19262815e-01 -9.48973149e-02 -1.50266960e-01 -1.50693834e+00
1.08152580e+00 3.09033692e-02 7.20522344e-01 -7.83722460e-01
5.48460543e-01 5.31412102e-02 1.76169164e-02 -4.97465059e-02
1.39574215e-01 -5.12454689e-01 -4.47945118e-01 3.04003000e-01
2.20985889e-01 4.73674238e-02 2.72460610e-01 -2.14357808e-01
7.16639638e-01 1.30408242e-01 1.98998809e-01 -1.93018109e-01
8.29824805e-02 -2.27746233e-01 1.12294912e+00 6.16733611e-01
-1.29691660e-01 2.53434688e-01 9.12961364e-01 -2.14207694e-01
-6.42328560e-01 -1.40537858e+00 -2.09545732e-01 8.09559166e-01
4.43166018e-01 -3.67072262e-02 -4.17374611e-01 -9.08005178e-01
5.61313868e-01 1.36272097e+00 -4.65620577e-01 -1.72369733e-01
-5.05946912e-02 -1.22972786e+00 7.74678588e-02 7.56904960e-01
4.43085760e-01 -6.28472149e-01 -5.22581562e-02 3.99283588e-01
-2.59496599e-01 -9.04595733e-01 -1.16540119e-02 3.78123254e-01
-6.76293552e-01 -9.12831485e-01 -4.69881743e-01 -1.97442278e-01
3.20527434e-01 1.49261439e-02 1.29990041e+00 -7.62184620e-01
-1.03677519e-01 2.41380464e-02 -2.10866287e-01 -1.08177423e+00
5.51584736e-02 1.60623431e-01 -1.15872314e-02 -8.47840402e-03
9.88004684e-01 -5.90392709e-01 -5.33971727e-01 7.29955852e-01
-7.57664979e-01 -2.90463716e-01 7.73784280e-01 8.04252028e-01
9.22365904e-01 1.07224321e+00 1.06235754e+00 -1.09634447e+00
7.63936937e-01 -1.03896451e+00 -7.10765123e-01 3.62469703e-01
-1.15011287e+00 -1.99940562e-01 4.90164548e-01 -5.06265759e-01
-1.33443880e+00 -5.09395421e-01 1.52198240e-01 -5.09556055e-01
-7.16501176e-02 7.61106253e-01 -5.49327970e-01 6.03831410e-01
-9.09939408e-02 -6.80332482e-02 -4.61530052e-02 -4.62836742e-01
2.36651450e-01 9.83773828e-01 4.06242222e-01 -6.63198352e-01
6.71457589e-01 1.46963686e-01 5.78913651e-03 -1.58265121e-02
-1.00743496e+00 -6.64639175e-01 -3.36972982e-01 -1.60192400e-01
4.15934592e-01 -1.07085145e+00 -8.52750659e-01 7.15685606e-01
-6.53540909e-01 -2.65776813e-01 -2.79394627e-01 6.85257077e-01
-6.35405183e-01 -1.79601997e-01 -4.16210473e-01 -9.48265254e-01
-1.38358459e-01 -8.90251756e-01 7.69327164e-01 3.66046041e-01
9.89629403e-02 -1.13601112e+00 -1.80150107e-01 5.77440679e-01
5.04588246e-01 4.87125486e-01 1.21463227e+00 -9.11108911e-01
-4.32474136e-01 -4.48220968e-01 -2.12082148e-01 8.64687562e-01
2.92599946e-01 -2.90006071e-01 -6.75255597e-01 -3.35240871e-01
1.79940358e-01 -1.63159460e-01 7.11486578e-01 7.27314889e-01
1.48237205e+00 4.61341515e-02 -4.34648544e-01 4.08557385e-01
1.54622066e+00 6.50455356e-01 5.67995012e-01 3.65574658e-01
5.20444453e-01 7.45144904e-01 1.17292118e+00 8.79461884e-01
7.32929528e-01 3.90342027e-01 7.24040270e-01 -1.07175156e-01
6.02579713e-01 -2.77689338e-01 6.07583001e-02 7.42746890e-01
-3.39853279e-02 -6.88758373e-01 -6.73782706e-01 4.90572691e-01
-2.14578891e+00 -8.58318090e-01 4.72101085e-02 2.16446424e+00
5.35008788e-01 4.71803427e-01 1.87454134e-01 5.10767959e-02
7.11070120e-01 -6.30630478e-02 -9.17248428e-01 -2.41727948e-01
7.67855942e-02 1.64790414e-02 7.58442342e-01 -5.26315309e-02
-1.01795518e+00 3.27813357e-01 5.48021936e+00 1.47781205e+00
-9.35353220e-01 8.86363983e-02 1.09884834e+00 -3.41426760e-01
-2.75793582e-01 -2.62816131e-01 -9.90009189e-01 8.07270706e-01
8.48838747e-01 -3.68605226e-01 4.09179688e-01 1.08175611e+00
1.64735913e-01 -4.19468060e-02 -1.01769030e+00 1.13048303e+00
1.32019758e-01 -7.40256190e-01 -2.33038038e-01 4.27554756e-01
8.44878674e-01 3.81821319e-02 2.14366734e-01 8.14638019e-01
6.62665546e-01 -1.06869805e+00 5.75254858e-01 7.81458020e-01
4.53442127e-01 -1.58061028e+00 1.34524846e+00 5.42341530e-01
-1.04969382e+00 -4.80717093e-01 -4.25547630e-01 2.91236669e-01
5.28448462e-01 1.09152353e+00 -8.02309036e-01 1.07698870e+00
8.38593543e-01 5.29399812e-01 -3.61069962e-02 9.59243894e-01
-3.11789036e-01 6.92568362e-01 -2.90796787e-01 1.27839595e-01
5.11824526e-02 -5.04288793e-01 9.05714482e-02 7.18894720e-01
5.04584372e-01 -3.93396884e-01 3.20570320e-01 8.04501235e-01
-1.29604578e-01 -8.19909154e-04 -1.89592034e-01 2.61002388e-02
4.63810831e-01 1.47145593e+00 -5.87597251e-01 -2.19925866e-01
-5.42316914e-01 4.10755277e-01 -2.03443225e-03 4.08948332e-01
-1.13358998e+00 -4.63298202e-01 5.49219906e-01 2.70352870e-01
5.91035724e-01 1.10480957e-01 -1.92752838e-01 -9.21825230e-01
8.94788839e-03 -7.91505575e-01 4.31049705e-01 -3.67386550e-01
-2.03902292e+00 3.00561875e-01 4.37275320e-01 -1.36312509e+00
-3.96292120e-01 -6.35502696e-01 -5.28255284e-01 8.86611819e-01
-1.81844866e+00 -1.24697661e+00 -2.56119817e-01 2.95164585e-01
3.62674624e-01 -3.56080353e-01 2.81779855e-01 4.20915812e-01
-6.39990270e-01 5.17088771e-01 4.80755091e-01 -2.33036458e-01
6.41723812e-01 -1.43911231e+00 -7.30790347e-02 3.29472929e-01
-2.55041420e-01 4.86581415e-01 5.64729333e-01 -6.56472862e-01
-7.42353320e-01 -1.33637083e+00 4.86819446e-01 -1.80058822e-01
5.73555589e-01 -3.66051137e-01 -7.79738247e-01 8.00111473e-01
1.71296194e-01 -2.39993498e-01 1.03068292e+00 3.09780091e-01
-6.56164140e-02 -7.12474942e-01 -1.29818261e+00 -2.01791823e-01
4.83432204e-01 5.80649264e-02 -3.61630678e-01 1.85217112e-01
3.83383542e-01 -1.22573368e-01 -1.27236938e+00 7.03004420e-01
4.76972371e-01 -9.03801858e-01 8.90820324e-01 -3.49231035e-01
2.46407971e-01 -1.78053811e-01 -2.79867023e-01 -1.57081699e+00
-3.95648926e-01 -3.13761413e-01 -2.48370677e-01 1.75437534e+00
2.57259488e-01 -8.53379250e-01 5.68952560e-01 5.31714261e-01
2.63706714e-01 -1.11055684e+00 -8.02658737e-01 -1.03632450e+00
4.15668935e-02 -5.20163059e-01 1.26452541e+00 9.25625324e-01
-1.82758838e-01 2.64738262e-01 -4.69751596e-01 -8.78899023e-02
8.72856796e-01 3.44272286e-01 7.32799768e-01 -1.49130976e+00
-5.15413821e-01 -3.82211983e-01 -2.68332064e-01 -9.37589347e-01
2.65010387e-01 -7.23376870e-01 7.40859937e-03 -1.50421596e+00
4.10874009e-01 -6.17671430e-01 -1.13578844e+00 1.88364759e-01
-3.65188345e-02 -2.22174883e-01 -1.22321313e-02 1.41150178e-02
-5.74770510e-01 8.18572700e-01 9.82100725e-01 -3.03948283e-01
-3.08455098e-02 5.67641973e-01 -1.01203716e+00 8.12394381e-01
9.74956751e-01 -6.07945204e-01 -8.73129606e-01 -4.71517853e-02
2.38920644e-01 5.73314205e-02 4.94962856e-02 -5.92420638e-01
-1.34569123e-01 -4.60223079e-01 6.19641960e-01 -1.05234241e+00
1.27201930e-01 -9.18638468e-01 3.25307995e-01 -3.42646427e-02
9.24432948e-02 1.11668475e-01 -1.60159431e-02 7.85647452e-01
-5.13453305e-01 -1.56512186e-01 5.12365639e-01 1.11279570e-01
-3.87357801e-01 5.51255822e-01 2.79144607e-02 -2.23225057e-01
1.26113141e+00 1.26222312e-01 -3.24637890e-01 -5.39136648e-01
-5.12475252e-01 7.09129393e-01 -5.42587116e-02 4.40387607e-01
2.01027051e-01 -1.44748831e+00 -6.75652325e-01 -9.92311835e-02
1.17789447e-01 2.25892380e-01 4.69573498e-01 9.10474896e-01
-1.55039221e-01 2.38227472e-01 -1.06293075e-01 -5.51418126e-01
-7.08221674e-01 4.58758920e-01 6.22825474e-02 -1.02817512e+00
5.40838055e-02 6.11844420e-01 3.29371691e-01 -6.42052412e-01
9.53837633e-02 -3.48873824e-01 -3.55498940e-01 3.74384403e-01
2.48919621e-01 7.13007390e-01 9.03900489e-02 -4.57590669e-01
-2.48190567e-01 2.21047848e-01 -3.65503877e-01 1.61548853e-01
1.62197506e+00 -3.16150784e-01 -2.16286127e-02 5.24939775e-01
9.29002941e-01 -3.62840682e-01 -1.43480027e+00 -2.77730793e-01
-1.00430332e-01 -5.69898069e-01 2.64262468e-01 -1.17988157e+00
-1.28957272e+00 9.42542791e-01 4.73262697e-01 3.23635042e-01
1.30652332e+00 -1.04934022e-01 8.87350559e-01 -1.77248359e-01
5.08590281e-01 -1.50692439e+00 -6.41616136e-02 1.13799840e-01
5.55651128e-01 -1.22952092e+00 6.64662868e-02 -2.65610695e-01
-8.76550496e-01 8.87833178e-01 6.84900105e-01 -5.01348339e-02
8.22524250e-01 3.55352521e-01 5.39081916e-02 5.43473335e-03
-7.04204202e-01 1.80435315e-01 1.11532435e-01 6.59561992e-01
2.89284391e-03 4.68799025e-01 -1.54385298e-01 1.30732024e+00
-2.34377310e-01 9.19212624e-02 8.89560282e-02 7.82263517e-01
-1.50431190e-02 -1.25640476e+00 -1.12246715e-01 8.32909763e-01
-7.43904412e-01 2.45876551e-01 3.07707518e-01 1.03998649e+00
3.85142982e-01 1.06961012e+00 2.44718045e-01 -5.33334196e-01
5.72515786e-01 -2.70696640e-01 3.99810463e-01 -2.86915779e-01
-1.76818907e-01 5.53978443e-01 -1.26162201e-01 -2.68779218e-01
-2.35769972e-01 -8.42357934e-01 -1.09395981e+00 -2.77241528e-01
-5.59868753e-01 7.65281171e-02 7.89567709e-01 8.72895837e-01
1.99680299e-01 9.82048273e-01 9.40849125e-01 -5.64288199e-01
-9.06238019e-01 -1.21399891e+00 -1.28159642e+00 2.75909036e-01
-1.10350415e-01 -1.16031790e+00 -3.11450243e-01 -2.65809476e-01] | [9.72372055053711, 5.252973556518555] |
b5b022f7-4cc5-4387-b79d-db9d33176708 | a-proposal-of-automatic-error-correction-in | 2112.01846 | null | https://arxiv.org/abs/2112.01846v1 | https://arxiv.org/pdf/2112.01846v1.pdf | A Proposal of Automatic Error Correction in Text | The great amount of information that can be stored in electronic media is growing up daily. Many of them is got mainly by typing, such as the huge of information obtained from web 2.0 sites; or scaned and processing by an Optical Character Recognition software, like the texts of libraries and goverment offices. Both processes introduce error in texts, so it is difficult to use the data for other purposes than just to read it, i.e. the processing of those texts by other applications like e-learning, learning of languages, electronic tutorials, data minning, information retrieval and even more specialized systems such as tiflologic software, specifically blinded people-oriented applications like automatic reading, where the text would be error free as possible in order to make easier the text to speech task, and so on. In this paper it is showed an application of automatic recognition and correction of ortographic errors in electronic texts. This task is composed of three stages: a) error detection; b) candidate corrections generation; and c) correction -selection of the best candidate. The proposal is based in part of speech text categorization, word similarity, word diccionaries, statistical measures, morphologic analisys and n-grams based language model of Spanish. | ['Carlos R. Jaimez-González', 'Wulfrano A. Luna-Ramírez'] | 2021-09-24 | null | null | null | null | ['word-similarity', 'text-categorization'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.07564628e-01 -1.92357705e-03 1.46448195e-01 -1.74081624e-01
-5.50105274e-01 -4.99788791e-01 5.39212167e-01 9.03734744e-01
-9.00376976e-01 6.72561765e-01 1.36522830e-01 -8.84471059e-01
-2.39339635e-01 -8.47149551e-01 -1.58226550e-01 -2.21970931e-01
5.06217897e-01 8.29849184e-01 4.76972014e-01 -4.04270977e-01
8.82328510e-01 6.17487907e-01 -1.91547918e+00 1.79522172e-01
1.17921329e+00 6.76724434e-01 4.37495828e-01 6.79296196e-01
-8.74614537e-01 3.86959910e-01 -5.93051195e-01 -3.28638524e-01
-1.31921038e-01 -5.05187511e-01 -1.00084865e+00 1.18823379e-01
-2.11084053e-01 -1.64389983e-01 2.24754214e-01 1.36712921e+00
3.38098347e-01 1.16923265e-01 8.31627786e-01 -5.41192949e-01
-4.68290567e-01 7.29490221e-01 -2.79956549e-01 2.27247417e-01
7.20253170e-01 -3.14873904e-01 4.68863904e-01 -8.59457076e-01
6.17353320e-01 1.31239891e+00 4.01864558e-01 3.16326350e-01
-6.68105006e-01 -2.82683730e-01 -5.00867367e-01 2.76619554e-01
-1.18378925e+00 -3.43921751e-01 2.22456872e-01 -6.75448358e-01
8.62676501e-01 5.52003145e-01 3.43870074e-01 6.07297063e-01
2.51079202e-01 4.13517088e-01 1.00940800e+00 -1.11242735e+00
1.10160917e-01 6.67931795e-01 6.53562307e-01 5.72634697e-01
5.59877396e-01 -4.33346331e-01 -2.88457721e-01 1.55102134e-01
1.53962925e-01 6.30274341e-02 -2.99334526e-01 5.10235786e-01
-9.73205924e-01 5.32886147e-01 -2.75618106e-01 1.11080933e+00
-5.24503231e-01 -7.69091308e-01 4.75217193e-01 4.59764928e-01
2.73948610e-01 3.21680993e-01 -4.87819791e-01 -3.33413780e-01
-1.07438624e+00 -1.46535665e-01 7.26895094e-01 9.31783319e-01
5.13979435e-01 -2.56696999e-01 2.22815648e-01 7.96918809e-01
5.16249657e-01 5.61958015e-01 1.41005981e+00 7.56593421e-02
6.75289512e-01 1.06847858e+00 -7.60008618e-02 -1.03509009e+00
-3.68993014e-01 1.82645217e-01 -6.41217589e-01 2.20442891e-01
6.59214079e-01 1.67538691e-02 -7.00933754e-01 9.41779673e-01
2.12869376e-01 -8.45586002e-01 -1.03592500e-01 5.13149083e-01
9.49247599e-01 1.01586616e+00 3.34177054e-02 -5.84131896e-01
1.65598130e+00 -4.57198262e-01 -1.16210115e+00 6.82617053e-02
7.63859749e-01 -1.47749782e+00 9.22515988e-01 7.51220942e-01
-1.05236328e+00 -5.13916433e-01 -7.96013176e-01 -7.22831041e-02
-8.34541917e-01 5.34930825e-01 2.48139855e-02 7.99222350e-01
-7.37881660e-01 9.14428234e-01 -5.21325231e-01 -8.94630432e-01
-8.22066888e-02 4.93061095e-02 -3.78204584e-01 1.63609490e-01
-9.86915410e-01 1.10228968e+00 6.87942863e-01 8.21215287e-03
1.11217499e-01 -2.60138154e-01 -6.67078495e-01 1.07839853e-01
1.89972728e-01 -1.47076130e-01 1.00843334e+00 -1.22837579e+00
-1.19692194e+00 1.15400147e+00 -1.68756396e-01 -1.30847007e-01
6.87613308e-01 -2.70368587e-02 -7.66616702e-01 2.73649860e-02
1.06209993e-01 -2.37687752e-01 7.02590168e-01 -5.66044927e-01
-7.86440194e-01 -7.97540665e-01 -6.41972661e-01 1.24184586e-01
-3.17579240e-01 5.21864235e-01 -2.84361541e-01 -6.47844136e-01
2.80215293e-01 -4.56954032e-01 1.53033406e-01 -1.21811457e-01
-4.41230118e-01 -5.53948343e-01 5.62734663e-01 -1.52612543e+00
1.55600572e+00 -2.09247994e+00 -1.08017623e-01 5.67084730e-01
3.69412638e-02 7.69451857e-01 4.26374674e-01 7.25028157e-01
-7.55895823e-02 3.45776737e-01 -1.08690066e-02 6.46220967e-02
-1.22788679e-02 -1.91775218e-01 1.14372529e-01 1.43074840e-01
-3.56011271e-01 4.88643274e-02 -7.15099394e-01 -8.36818516e-01
1.73575789e-01 3.96833681e-02 -1.08625852e-01 -3.62408049e-02
1.34184346e-01 1.71563271e-02 -6.26903117e-01 4.82317388e-01
3.60021859e-01 9.21336040e-02 4.74872366e-02 2.49175467e-02
-6.21474504e-01 4.61555123e-01 -1.28353894e+00 1.14367223e+00
-4.80808109e-01 7.33014941e-01 -2.86964953e-01 -9.81441975e-01
1.21740913e+00 4.33387637e-01 1.38355374e-01 -5.96899450e-01
6.13632977e-01 7.76462078e-01 -1.19917713e-01 -8.28073025e-01
8.29434276e-01 3.04863364e-01 3.39863569e-01 5.50519228e-01
-1.12939067e-01 -7.72409467e-03 6.41605139e-01 9.64631364e-02
5.75402081e-01 -2.32512221e-01 6.45836234e-01 -1.48785561e-01
9.81244147e-01 1.10113956e-01 -1.28345653e-01 2.72896647e-01
2.11695090e-01 2.31646001e-01 3.12786669e-01 -5.11480384e-02
-1.31728745e+00 -3.69828880e-01 -5.36610544e-01 7.46394455e-01
-2.12380663e-01 -2.38049239e-01 -1.05255914e+00 -5.01450062e-01
-1.23540826e-01 8.37666690e-01 -7.03597516e-02 1.22242093e-01
-2.81039089e-01 -2.12681741e-01 2.30609119e-01 -9.83075425e-02
4.29469079e-01 -1.44404221e+00 -4.90869701e-01 3.02551150e-01
-7.54519403e-02 -4.73423034e-01 -3.09584141e-01 1.67948931e-01
-9.41550553e-01 -1.04408932e+00 -6.40744150e-01 -8.25402737e-01
6.62116706e-01 1.75521523e-01 6.19193375e-01 5.04805684e-01
-3.59227091e-01 8.67317021e-02 -9.05181050e-01 -6.10266030e-01
-1.00694871e+00 7.76308030e-02 9.51545015e-02 -1.47987291e-01
7.63798773e-01 -2.14852765e-01 5.02425209e-02 4.48016003e-02
-9.33973670e-01 -2.49040201e-01 7.52276540e-01 5.96136510e-01
2.70052582e-01 9.29491874e-03 4.32033330e-01 -8.78577113e-01
9.21056807e-01 -3.73386860e-01 -6.33852482e-01 4.04656559e-01
-8.98765683e-01 5.46533149e-03 7.63761878e-01 -1.34336948e-01
-1.02747357e+00 -3.09190273e-01 -7.07520127e-01 3.56058091e-01
-4.56041843e-01 4.47627634e-01 -1.71032593e-01 -2.50774231e-02
8.38324487e-01 4.96284276e-01 2.00196907e-01 -7.67634332e-01
-8.80199075e-02 1.60959947e+00 2.84641325e-01 5.21252230e-02
6.47240281e-01 -3.14227372e-01 -4.72886592e-01 -1.34975863e+00
-1.40084848e-01 -7.70962715e-01 -6.57912791e-01 -3.34198862e-01
1.12631011e+00 -1.67239502e-01 -5.57474971e-01 4.53628808e-01
-1.28802180e+00 2.28160411e-01 -1.67134896e-01 6.96992159e-01
-1.23847993e-02 7.42634058e-01 -5.27164876e-01 -9.18785930e-01
-4.22476381e-01 -9.11481380e-01 6.47182703e-01 4.19370145e-01
-3.92460942e-01 -8.58280241e-01 1.05039531e-03 4.96660441e-01
1.49869621e-01 -4.19983596e-01 1.15737736e+00 -1.20120716e+00
-1.08164586e-01 -5.08829892e-01 -4.29131314e-02 6.47531748e-01
-4.44306433e-02 2.88122565e-01 -5.94850302e-01 7.88616538e-02
5.78992106e-02 -1.07829213e-01 5.32723248e-01 6.79066852e-02
1.08958232e+00 -4.16630507e-01 -2.59262145e-01 6.93654492e-02
1.20368731e+00 7.62439132e-01 1.02147400e+00 4.50604588e-01
3.35008770e-01 6.06828570e-01 6.66842341e-01 5.11570513e-01
3.78750339e-02 6.18848920e-01 2.21515913e-02 2.34391436e-01
6.15199059e-02 -2.22056165e-01 1.61226586e-01 1.45002186e+00
-1.72729954e-01 -2.86260277e-01 -9.92060244e-01 4.30022717e-01
-1.57360017e+00 -1.13565230e+00 -8.09933722e-01 2.56286097e+00
7.37920702e-01 1.24104835e-01 -5.05452193e-02 8.11442614e-01
1.02145743e+00 -3.91665816e-01 2.41630077e-01 -7.63416529e-01
2.22172990e-01 1.59761041e-01 4.69620019e-01 5.81420302e-01
-7.31812775e-01 5.30089140e-01 4.88177967e+00 1.12993932e+00
-1.05592072e+00 1.06732048e-01 3.92642349e-01 5.26791394e-01
-2.88216263e-01 -3.28217328e-01 -8.95155549e-01 8.42998624e-01
7.52870321e-01 -7.04070032e-02 2.40435719e-01 6.50795162e-01
4.21726644e-01 -6.47472799e-01 -6.62779212e-01 1.16814280e+00
3.26009482e-01 -9.68280375e-01 1.28006756e-01 -7.43909404e-02
3.69202316e-01 -4.41836983e-01 -6.08316123e-01 1.23639971e-01
-1.82453781e-01 -6.83332205e-01 9.01814878e-01 6.38772845e-01
4.57315505e-01 -4.84312773e-01 9.96434629e-01 4.95922327e-01
-6.65976524e-01 1.09821908e-01 -5.76919913e-01 1.06829844e-01
3.39937694e-02 7.19068885e-01 -6.18640423e-01 6.29663765e-01
4.53897625e-01 1.66296795e-01 -7.71612942e-01 1.44275355e+00
-1.52688930e-02 5.34205616e-01 -1.69527471e-01 -7.29852974e-01
1.65500659e-02 -4.66812938e-01 6.34318173e-01 1.36100817e+00
7.49736667e-01 2.32853934e-01 -1.16102912e-01 3.98887992e-01
2.18689263e-01 1.07406950e+00 -5.07379651e-01 -1.17199905e-01
5.75892985e-01 1.18481874e+00 -9.82126594e-01 -4.20806974e-01
-2.09653854e-01 9.04276431e-01 -1.18546836e-01 -5.69019429e-02
-2.64644295e-01 -1.13568568e+00 -4.86600623e-02 2.61318803e-01
5.10251075e-02 -8.51235837e-02 -3.07923675e-01 -1.09993958e+00
1.04505040e-01 -1.20084536e+00 1.70889527e-01 -7.22689509e-01
-9.90071177e-01 6.61006033e-01 -2.52982199e-01 -1.16474307e+00
-3.88750345e-01 -9.70291913e-01 -4.61365223e-01 1.18039513e+00
-7.97116876e-01 -3.62135291e-01 -1.55905172e-01 5.98289192e-01
6.91811144e-01 -5.30660689e-01 6.26194477e-01 6.81913495e-01
-4.85072404e-01 3.51754189e-01 5.17620504e-01 1.92672908e-01
8.39830995e-01 -1.17953992e+00 -2.51559585e-01 9.13346648e-01
2.17404008e-01 3.83090407e-01 6.09859228e-01 -8.53076696e-01
-9.47735846e-01 -3.57984722e-01 1.84743524e+00 -1.84872910e-01
7.13088870e-01 -4.68218252e-02 -1.00904322e+00 1.15248144e-01
6.05048686e-02 -6.75371766e-01 3.24617028e-01 -1.55633196e-01
1.20608754e-01 -1.74039379e-01 -1.06815493e+00 4.22850877e-01
4.50737000e-01 -3.11242461e-01 -8.99063468e-01 6.68324649e-01
2.18740419e-01 -1.78227916e-01 -5.52647889e-01 -5.27051151e-01
3.50313753e-01 -9.04650450e-01 4.92720693e-01 -6.20427907e-01
5.06363988e-01 -2.79926807e-01 2.25372583e-01 -1.26185298e+00
1.38930127e-01 -4.88008440e-01 8.48853171e-01 1.45424628e+00
7.49592483e-01 -6.62360489e-01 2.67572105e-01 3.19436729e-01
-3.30168754e-01 -3.16408984e-02 -8.32049489e-01 -4.85011518e-01
-3.00409555e-01 -3.47541749e-01 3.03066820e-01 1.12048054e+00
5.36386251e-01 1.69001207e-01 7.81308115e-02 -1.25264466e-01
2.13973001e-01 -4.53022391e-01 5.12080133e-01 -1.46894455e+00
-3.58470052e-01 -7.60923624e-01 -5.66343307e-01 -6.10662401e-01
-2.00498864e-01 -7.56669879e-01 9.97157991e-02 -1.64619434e+00
-2.20093042e-01 -3.25542241e-01 5.51189005e-01 3.42439115e-01
-2.41974089e-02 -1.68392360e-01 2.26046979e-01 3.12624246e-01
-1.99708313e-01 2.12966148e-02 9.57964242e-01 7.54051954e-02
-2.64134109e-01 2.29134664e-01 -2.91293442e-01 8.14256847e-01
5.73669434e-01 -4.93020922e-01 -9.32364687e-02 -1.33393601e-01
5.09247363e-01 2.20496222e-01 -1.87077567e-01 -8.89757752e-01
2.73315668e-01 -5.68590462e-02 2.82005966e-01 -5.21444261e-01
-4.89133358e-01 -9.89714384e-01 -3.88343856e-02 5.76688528e-01
-2.23971426e-01 4.63865519e-01 -1.25876591e-01 1.66117355e-01
-3.01550478e-01 -1.38955641e+00 5.88876426e-01 -1.72892123e-01
-4.71386254e-01 -2.42960453e-01 -1.04599333e+00 -1.27919734e-01
9.40030217e-01 -4.85003740e-01 -2.85421014e-01 -3.71222675e-01
-6.31390154e-01 -2.39619821e-01 2.61474937e-01 8.77937824e-02
1.96776807e-01 -8.91375899e-01 -6.10048532e-01 1.53467759e-01
6.09119646e-02 -4.41021264e-01 1.40884668e-01 9.32755589e-01
-1.21094108e+00 4.12160069e-01 -3.99936229e-01 -1.69049073e-02
-1.64699626e+00 5.06802857e-01 4.90049049e-02 -2.95488179e-01
-4.14136499e-01 3.07543248e-01 -4.52902555e-01 -2.87316263e-01
4.01623070e-01 -4.02405679e-01 -1.03052115e+00 4.34421331e-01
9.78195488e-01 5.94107330e-01 6.81088150e-01 -5.81034482e-01
-1.66437194e-01 5.45983851e-01 -6.05459372e-03 -1.98821247e-01
9.44254816e-01 -2.05900356e-01 -5.07700920e-01 5.61320603e-01
8.47094417e-01 4.07972395e-01 -1.31790757e-01 -3.00371088e-02
2.97849894e-01 -4.01639760e-01 -5.85256070e-02 -8.55775237e-01
-4.92983520e-01 1.01220369e+00 5.59662163e-01 8.30207407e-01
9.06293571e-01 -3.60437721e-01 4.90486681e-01 7.78728843e-01
3.98946583e-01 -1.51957798e+00 -4.40800399e-01 6.91161692e-01
6.63453221e-01 -1.17963159e+00 -9.87835154e-02 -1.81556016e-01
-5.01821578e-01 1.77346373e+00 1.05907142e-01 3.58669758e-01
6.98235512e-01 -2.56899390e-02 -2.89660152e-02 3.77981924e-02
-1.86469734e-01 -3.89792085e-01 4.94878262e-01 4.63796109e-01
8.56867969e-01 -9.33755115e-02 -1.46712863e+00 5.82627594e-01
-1.91314518e-01 -6.76391274e-02 8.06390941e-01 7.13710010e-01
-9.62830782e-01 -1.29823339e+00 -9.38270032e-01 7.86588669e-01
-7.82515407e-01 -1.76281959e-01 -3.97167891e-01 6.66654766e-01
1.34154409e-01 9.87494409e-01 1.27307773e-01 -1.36429220e-01
3.08154374e-01 3.85837018e-01 2.75717378e-01 -4.47944999e-01
-8.46932948e-01 -4.69834171e-02 3.96104604e-01 -9.25515369e-02
-1.55994684e-01 -8.71684849e-01 -1.07723403e+00 -5.53087592e-01
-4.98510778e-01 5.47313929e-01 1.31855392e+00 1.27784216e+00
-3.77494156e-01 2.24840999e-01 3.74750435e-01 -5.00819206e-01
-6.43451571e-01 -1.18442595e+00 -7.26058602e-01 4.94020015e-01
-2.30017215e-01 -2.03428492e-01 -3.61889869e-01 2.66960621e-01] | [10.812272071838379, 10.520987510681152] |
5d28d779-ce98-497f-abb2-df5e6cc10014 | training-free-object-counting-with-prompts | 2307.00038 | null | https://arxiv.org/abs/2307.00038v1 | https://arxiv.org/pdf/2307.00038v1.pdf | Training-free Object Counting with Prompts | This paper tackles the problem of object counting in images. Existing approaches rely on extensive training data with point annotations for each object, making data collection labor-intensive and time-consuming. To overcome this, we propose a training-free object counter that treats the counting task as a segmentation problem. Our approach leverages the Segment Anything Model (SAM), known for its high-quality masks and zero-shot segmentation capability. However, the vanilla mask generation method of SAM lacks class-specific information in the masks, resulting in inferior counting accuracy. To overcome this limitation, we introduce a prior-guided mask generation method that incorporates three types of priors into the segmentation process, enhancing efficiency and accuracy. Additionally, we tackle the issue of counting objects specified through free-form text by proposing a two-stage approach that combines reference object selection and prior-guided mask generation. Extensive experiments on standard datasets demonstrate the competitive performance of our training-free counter compared to learning-based approaches. This paper presents a promising solution for counting objects in various scenarios without the need for extensive data collection and model training. Code is available at https://github.com/shizenglin/training-free-object-counter. | ['Mengmi Zhang', 'Ying Sun', 'Zenglin Shi'] | 2023-06-30 | null | null | null | null | ['zero-shot-segmentation', 'object-counting'] | ['computer-vision', 'computer-vision'] | [ 4.51677561e-01 -1.95766971e-01 -9.04011875e-02 -4.57364410e-01
-8.43482137e-01 -5.06431103e-01 6.21412456e-01 2.38285527e-01
-8.10490012e-01 5.49823642e-01 -3.63109350e-01 -2.65492409e-01
2.92259783e-01 -8.42999756e-01 -5.58133543e-01 -4.65218872e-01
5.15615821e-01 6.02293909e-01 7.51953840e-01 3.59478682e-01
5.09262264e-01 4.35616761e-01 -1.51368296e+00 -1.71957891e-02
9.92603183e-01 7.80406594e-01 5.93567610e-01 7.55348623e-01
-5.31181335e-01 6.50613368e-01 -6.78735197e-01 -5.69267750e-01
4.33742166e-01 -3.44391257e-01 -6.52726889e-01 3.10191989e-01
4.59706455e-01 -5.06499112e-01 6.90801516e-02 9.15055573e-01
3.75487179e-01 8.12056512e-02 7.05655456e-01 -1.10327792e+00
-1.43010005e-01 3.62512738e-01 -9.44976211e-01 2.59512007e-01
-4.69433628e-02 2.48811141e-01 6.66752219e-01 -9.23730135e-01
1.65718287e-01 8.50022972e-01 7.05394745e-01 6.32283866e-01
-1.26996291e+00 -6.71259701e-01 -2.50524469e-02 -1.59262463e-01
-1.53098392e+00 -3.97361308e-01 5.25986612e-01 -6.72356665e-01
5.84311068e-01 2.97126532e-01 7.23672152e-01 6.31295502e-01
-3.67373288e-01 1.02423096e+00 1.03336871e+00 -7.12487280e-01
2.96051711e-01 1.35599598e-01 3.29262942e-01 7.85371482e-01
6.08037889e-01 -1.38594657e-01 -2.50011325e-01 -1.02536418e-01
1.03019047e+00 -1.43874986e-02 1.55183170e-02 -2.90172577e-01
-1.00774157e+00 7.92947948e-01 2.96590447e-01 2.24428147e-01
-2.25980788e-01 1.63528502e-01 2.65733361e-01 -4.95206594e-01
6.22861683e-01 1.85557917e-01 -2.40517035e-01 9.11075547e-02
-1.71329248e+00 2.25714996e-01 6.63340449e-01 1.17983902e+00
6.93205178e-01 -1.90334365e-01 -7.08707273e-01 7.70751953e-01
2.32859567e-01 2.78299093e-01 1.73183024e-01 -9.20476973e-01
5.68248630e-01 7.55571663e-01 4.28630769e-01 -7.77075529e-01
-4.02797967e-01 -3.65210593e-01 -5.53096712e-01 8.64178911e-02
7.32048273e-01 7.62355104e-02 -1.21481764e+00 1.31763363e+00
7.09765255e-01 4.11499947e-01 -4.73760486e-01 8.17744434e-01
7.16315985e-01 3.48821104e-01 2.94784039e-01 -2.53045131e-02
1.51831198e+00 -1.13216233e+00 -5.85398614e-01 -2.67747551e-01
4.28146660e-01 -7.26304710e-01 1.12257063e+00 2.81143904e-01
-1.24419427e+00 -4.91657495e-01 -1.04288459e+00 -1.78621784e-01
-2.52334744e-01 3.75059515e-01 7.91412950e-01 9.92088795e-01
-5.33142745e-01 4.29939568e-01 -1.04013228e+00 -2.13671654e-01
9.46731508e-01 4.27053690e-01 1.34794176e-01 5.05713113e-02
-4.36446935e-01 7.18327165e-01 7.04512537e-01 -6.15748651e-02
-4.77916479e-01 -6.29136801e-01 -7.56867290e-01 1.19441021e-02
7.99348533e-01 -6.95356727e-01 1.57829654e+00 -5.96520960e-01
-1.39303839e+00 8.74484003e-01 -3.94626617e-01 -3.22755784e-01
8.91821384e-01 -2.21783996e-01 2.38670126e-01 2.99778014e-01
3.63655388e-01 7.23703325e-01 8.44783366e-01 -1.49478829e+00
-6.68559670e-01 -1.25054434e-01 3.54607888e-02 -7.77963102e-02
-1.72362268e-01 1.01211376e-01 -9.32850361e-01 -7.02711761e-01
1.45198649e-03 -5.02612352e-01 -4.71079320e-01 -6.95247203e-02
-5.40689349e-01 -9.70631614e-02 6.19861662e-01 -5.33210278e-01
1.37882698e+00 -1.94173825e+00 -4.81337994e-01 1.27799213e-01
1.59500420e-01 5.38602114e-01 2.72723157e-02 6.24156743e-02
5.31257510e-01 1.03810988e-01 -6.88709974e-01 -9.14736271e-01
-9.26813930e-02 3.00102234e-01 2.70089284e-02 3.93543243e-01
4.72095788e-01 8.21625233e-01 -1.02114093e+00 -1.22916520e+00
5.39543748e-01 4.61926609e-01 -5.68856537e-01 2.04130664e-01
-3.62355560e-01 7.25194693e-01 -4.37073052e-01 8.35770667e-01
9.47982073e-01 -4.06034291e-01 -6.78592473e-02 -4.85664271e-02
-3.17636162e-01 -1.00623176e-01 -1.42839074e+00 1.57709336e+00
-3.82899135e-01 -8.82160268e-04 9.13424417e-03 -7.00518668e-01
6.66903913e-01 6.13266639e-02 3.70758325e-01 -4.47932333e-01
5.00506938e-01 3.84326071e-01 -2.76388794e-01 -2.61388570e-01
8.40134561e-01 -1.18708769e-02 3.23351361e-02 3.75962228e-01
3.31890360e-02 -5.51508129e-01 7.43820190e-01 2.42918432e-01
9.37176168e-01 4.31987524e-01 4.98848319e-01 -8.74926969e-02
4.30664927e-01 1.60818264e-01 6.23075306e-01 1.06465173e+00
-3.04036260e-01 1.00483382e+00 1.47435412e-01 -1.01502649e-01
-9.84769821e-01 -6.86032176e-01 -4.05777812e-01 8.51346314e-01
2.40498319e-01 -2.72548914e-01 -1.20603096e+00 -6.43737733e-01
-1.45941332e-01 7.29428053e-01 -4.94352192e-01 4.40503448e-01
-6.69111550e-01 -8.23672891e-01 4.63809222e-01 7.12589562e-01
6.48920476e-01 -8.90178323e-01 -1.04146278e+00 2.86914319e-01
-3.30024481e-01 -1.33583879e+00 -6.26818895e-01 1.63386896e-01
-9.42914903e-01 -1.20205462e+00 -9.16708410e-01 -3.90331894e-01
9.10292327e-01 3.37368727e-01 1.13742757e+00 4.74517167e-01
-5.67443907e-01 2.86066413e-01 -2.61304259e-01 -6.69236720e-01
-1.57062650e-01 1.36650085e-01 -4.38562840e-01 -6.93107620e-02
5.18175781e-01 -8.06941092e-02 -7.50601888e-01 4.46585923e-01
-1.10842431e+00 2.08286375e-01 5.06369233e-01 6.25370502e-01
7.13991344e-01 -1.60641849e-01 3.88286680e-01 -1.03144336e+00
2.46315017e-01 -3.43091398e-01 -1.05581367e+00 1.67704761e-01
-2.00202495e-01 -1.69187546e-01 3.56884599e-01 -4.51805025e-01
-1.13734090e+00 5.80721676e-01 5.16791157e-02 -2.62340933e-01
-1.65338933e-01 -2.91038323e-02 -1.35120779e-01 -8.13461021e-02
5.20090938e-01 4.59463820e-02 -3.78697991e-01 -5.65548360e-01
3.12050909e-01 5.29914975e-01 6.57144606e-01 -5.74859858e-01
7.86550939e-01 7.80294955e-01 -1.27858385e-01 -6.97309911e-01
-1.19080698e+00 -9.60644186e-01 -1.00045228e+00 -1.42687812e-01
1.03858483e+00 -6.28793716e-01 -4.68020469e-01 5.07116914e-01
-1.30733919e+00 -4.42495048e-01 -3.93838346e-01 3.87605369e-01
-6.13251507e-01 5.28398335e-01 -4.83161539e-01 -1.09034896e+00
-2.76281357e-01 -9.58268583e-01 1.26712823e+00 4.20515567e-01
-1.37699619e-01 -6.98649943e-01 -1.05777748e-01 6.02707446e-01
1.92881867e-01 2.93454736e-01 3.93842936e-01 -6.07090116e-01
-9.59660172e-01 -3.89689535e-01 -5.86749673e-01 3.03370297e-01
-1.62656263e-01 -3.24268788e-02 -1.12415135e+00 -2.50750706e-02
-5.46100270e-03 -1.65176451e-01 9.64683533e-01 5.61083376e-01
1.46718681e+00 3.59439403e-02 -4.48593378e-01 5.34493983e-01
1.61697161e+00 -9.87411290e-02 4.07285273e-01 2.92220026e-01
6.92119241e-01 3.89762074e-01 7.58167863e-01 6.06523454e-01
4.30363387e-01 6.11962736e-01 2.47390464e-01 -2.00962827e-01
-2.12644234e-01 -1.91751376e-01 -3.62604499e-01 4.61307764e-01
-2.99337208e-01 -3.91315639e-01 -9.98650908e-01 7.84664869e-01
-1.90608144e+00 -8.02557111e-01 -5.21093249e-01 2.20622206e+00
8.70359898e-01 1.94226190e-01 3.45254928e-01 4.31485116e-01
8.24227214e-01 -1.44951388e-01 -3.00185859e-01 2.74834465e-02
2.16307223e-01 4.57063347e-01 7.90643871e-01 5.01204610e-01
-1.33530915e+00 1.02460921e+00 5.48199034e+00 1.02057242e+00
-5.73866546e-01 4.56536204e-01 4.97676343e-01 -7.13230297e-02
2.58850515e-01 7.22229406e-02 -1.18512845e+00 5.54797471e-01
5.23916781e-01 1.54561251e-01 4.39429134e-02 7.10283518e-01
2.82208771e-01 -8.14855099e-01 -9.27639723e-01 9.28658783e-01
1.36298249e-02 -1.09950256e+00 -1.95410907e-01 -9.20800567e-02
6.50791168e-01 -1.63991868e-01 -3.81808519e-01 2.08320558e-01
2.76664734e-01 -8.04702401e-01 1.02873600e+00 5.48258007e-01
7.80944228e-01 -4.57803547e-01 6.87200367e-01 6.67510748e-01
-1.26890397e+00 1.00392126e-01 -4.53246027e-01 -6.73879310e-02
3.61408055e-01 8.91607106e-01 -8.28267992e-01 5.19631147e-01
4.18983281e-01 8.25053081e-02 -6.24766827e-01 1.58248544e+00
-2.23055899e-01 7.07822144e-01 -5.82047939e-01 1.93593621e-01
1.13101810e-01 -4.85963747e-02 3.51116985e-01 1.64393270e+00
1.92250043e-01 2.68444568e-01 5.13800323e-01 1.08057618e+00
-3.26054208e-02 1.24057718e-01 -2.54636854e-01 1.77487254e-01
5.57531178e-01 1.35992813e+00 -1.57117534e+00 -5.23174942e-01
-3.98466468e-01 8.76863897e-01 2.65427262e-01 4.83432934e-02
-1.00682008e+00 -2.34518513e-01 -2.09212825e-01 3.47625524e-01
5.18974662e-01 -3.27945650e-01 -8.03190708e-01 -1.05349636e+00
1.05813578e-01 -3.82735610e-01 3.12404513e-01 -3.15265149e-01
-1.20244455e+00 2.39007860e-01 4.47044790e-01 -1.03498042e+00
1.09092869e-01 -4.65887874e-01 -6.66222453e-01 6.58398032e-01
-1.70671380e+00 -1.36050713e+00 -5.15809655e-01 3.46123487e-01
7.77954459e-01 3.54613751e-01 4.12160456e-01 4.07508820e-01
-5.58852017e-01 5.16146302e-01 -1.62610993e-01 2.44563103e-01
3.11991900e-01 -1.28886890e+00 3.95392299e-01 8.95987391e-01
8.86047184e-02 5.28822243e-01 4.18271840e-01 -6.61155283e-01
-6.88901961e-01 -1.20709085e+00 7.70568013e-01 -6.71344876e-01
2.81730622e-01 -5.70549011e-01 -8.92528176e-01 4.61670786e-01
-1.96145058e-01 2.04546615e-01 5.77206314e-01 -3.11547428e-01
8.72650489e-05 3.53828311e-01 -1.32418811e+00 3.64593863e-01
9.53989208e-01 1.51440699e-03 -4.71712291e-01 3.41558963e-01
3.48458320e-01 -4.68303680e-01 -3.02891731e-01 3.50108147e-01
3.46640140e-01 -1.01302743e+00 8.92363131e-01 6.63064271e-02
2.76367068e-01 -4.84724730e-01 8.09375569e-02 -5.99200368e-01
-1.32187769e-01 -2.77792424e-01 -2.08976567e-01 1.35156989e+00
1.49608672e-01 -3.52322906e-01 8.51935506e-01 6.61720335e-01
-5.18712997e-02 -5.31513453e-01 -8.13290298e-01 -9.62743282e-01
-4.22270484e-02 -6.09493971e-01 6.07818067e-01 6.82903767e-01
-4.45466727e-01 1.05734318e-01 -1.08732902e-01 1.78407356e-01
7.38546133e-01 -3.47635932e-02 1.01625216e+00 -1.15322316e+00
-3.29293847e-01 -4.55201894e-01 -1.22448683e-01 -1.13364089e+00
-1.18137039e-01 -7.65251875e-01 3.26549411e-01 -1.73295712e+00
4.87427920e-01 -6.88603461e-01 1.43689826e-01 4.67881560e-01
-5.54531932e-01 5.87288499e-01 4.20351982e-01 2.86316305e-01
-7.69860983e-01 1.81180537e-01 1.17878342e+00 7.01344535e-02
-2.05736473e-01 2.18271792e-01 -2.83880591e-01 1.07709992e+00
9.84456122e-01 -8.02891016e-01 -2.33019769e-01 -5.37081838e-01
-1.63273811e-01 -2.19019875e-01 5.60269296e-01 -1.22891331e+00
2.45035917e-01 -2.03143030e-01 3.19295555e-01 -9.30783093e-01
4.16514069e-01 -7.54332006e-01 -1.45161971e-01 3.86073560e-01
1.48741305e-01 -4.31200981e-01 2.02905118e-01 6.11617506e-01
1.50509581e-01 -8.91729295e-01 1.04772806e+00 -4.80806351e-01
-4.46857274e-01 2.15481490e-01 -9.23324898e-02 8.68108496e-02
1.13986647e+00 -6.88801527e-01 -5.59728071e-02 1.53977841e-01
-4.82530832e-01 2.61249840e-01 5.58771908e-01 -7.92447031e-02
2.95098692e-01 -8.57578397e-01 -6.76737368e-01 1.79217070e-01
4.18749638e-02 5.55074632e-01 1.34757962e-02 9.13227558e-01
-5.47840476e-01 3.01250398e-01 2.77584702e-01 -8.12958002e-01
-1.15319932e+00 5.22127092e-01 1.19735613e-01 -5.29852450e-01
-6.49780154e-01 9.81687605e-01 1.45127639e-01 -3.83288592e-01
1.28011256e-01 -4.82453287e-01 -8.92280862e-02 -2.44637560e-02
4.42190468e-01 5.82951844e-01 8.07046965e-02 -3.84270221e-01
-2.23554939e-01 5.21493793e-01 -1.26510458e-02 -1.58333153e-01
1.01208615e+00 8.24059173e-03 1.83033109e-01 4.37414736e-01
5.30434012e-01 3.79055813e-02 -1.41604316e+00 -2.36257955e-01
3.58997613e-01 -9.26274717e-01 3.98598202e-02 -7.13267922e-01
-8.59972537e-01 7.93068409e-01 3.12612027e-01 1.28971055e-01
8.83081675e-01 -1.10398866e-01 8.42789531e-01 4.39297445e-02
4.85991150e-01 -1.18314040e+00 1.03807561e-01 4.01697040e-01
4.73361522e-01 -1.51203465e+00 1.92221731e-01 -7.46915400e-01
-3.53230089e-01 8.18318188e-01 6.63638473e-01 -1.81473233e-02
3.81260097e-01 4.68447268e-01 1.55535210e-02 -2.40813509e-01
-1.37216225e-01 -5.20388305e-01 2.22067147e-01 4.37968194e-01
3.80885959e-01 2.86964756e-02 -6.19085729e-01 6.18995130e-01
-5.76364994e-03 3.95234257e-01 4.45930302e-01 1.40214872e+00
-4.85752493e-01 -1.08346391e+00 -6.66894138e-01 6.28412902e-01
-5.64232886e-01 -1.93767264e-01 -1.44089207e-01 7.39458859e-01
3.75208557e-01 1.01128864e+00 1.47935763e-01 3.11518520e-01
1.75288409e-01 8.21767524e-02 6.58853829e-01 -1.04352081e+00
-5.40898502e-01 2.15524286e-01 -1.66355506e-01 -3.49595755e-01
-6.96539819e-01 -5.97792089e-01 -1.24290502e+00 -1.68707967e-01
-1.01493967e+00 -1.35864452e-01 7.09588647e-01 8.69207203e-01
1.45168360e-02 5.12704253e-01 1.83199450e-01 -1.24275303e+00
-4.65990126e-01 -9.38315749e-01 -4.11304176e-01 2.56849945e-01
8.98092762e-02 -8.40427518e-01 -6.64296672e-02 2.65292138e-01] | [9.141847610473633, 0.4487902820110321] |
0048fd10-e125-4879-923e-f31873399ba2 | demn-distilled-exposition-enhanced-matching | 1901.02252 | null | http://arxiv.org/abs/1901.02252v1 | http://arxiv.org/pdf/1901.02252v1.pdf | DEMN: Distilled-Exposition Enhanced Matching Network for Story Comprehension | This paper proposes a Distilled-Exposition Enhanced Matching Network (DEMN)
for story-cloze test, which is still a challenging task in story comprehension.
We divide a complete story into three narrative segments: an
\textit{exposition}, a \textit{climax}, and an \textit{ending}. The model
consists of three modules: input module, matching module, and distillation
module. The input module provides semantic representations for the three
segments and then feeds them into the other two modules. The matching module
collects interaction features between the ending and the climax. The
distillation module distills the crucial semantic information in the exposition
and infuses it into the matching module in two different ways. We evaluate our
single and ensemble model on ROCStories Corpus \cite{Mostafazadeh2016ACA},
achieving an accuracy of 80.1\% and 81.2\% on the test set respectively. The
experimental results demonstrate that our DEMN model achieves a
state-of-the-art performance. | ['Shan Jiang', 'Haiou Zhang', 'Dong Yu', 'Chunhua Liu'] | 2019-01-08 | null | https://aclanthology.org/Y18-1045 | https://aclanthology.org/Y18-1045.pdf | paclic-2018-12 | ['cloze-test'] | ['natural-language-processing'] | [ 2.19909638e-01 8.31519142e-02 -4.22063656e-02 -2.45731652e-01
-9.02928710e-01 -5.23787856e-01 8.76564264e-01 2.32997507e-01
-3.20992798e-01 4.98198032e-01 8.43096495e-01 -1.50394499e-01
-2.87924036e-02 -8.81104529e-01 -5.07573128e-01 -4.50305551e-01
1.71525165e-01 4.62953508e-01 3.62652034e-01 -5.19653976e-01
2.12657526e-01 -5.11490166e-01 -1.41226518e+00 7.95478582e-01
9.48220491e-01 8.83351326e-01 2.67710865e-01 6.74717546e-01
-1.66060373e-01 1.15775681e+00 -7.56980002e-01 -5.90350509e-01
-2.76181996e-01 -6.98503673e-01 -9.19184625e-01 -2.15638056e-01
8.07316788e-03 -5.51242709e-01 -7.16848552e-01 7.21366763e-01
6.29819512e-01 2.96280861e-01 7.25948453e-01 -1.15878665e+00
-6.38361454e-01 1.33328795e+00 -5.03154933e-01 4.68601465e-01
7.84316957e-01 8.53185430e-02 1.10265815e+00 -8.37558925e-01
7.57032454e-01 9.53388155e-01 4.09972161e-01 4.32359576e-01
-6.87914193e-01 -8.27838957e-01 3.46036404e-01 5.33748269e-01
-1.17440820e+00 -2.93995976e-01 9.77472067e-01 -6.92295790e-01
6.89568758e-01 3.64685118e-01 5.93732834e-01 1.40163827e+00
-4.23689932e-01 1.44822645e+00 8.98604989e-01 -1.17133766e-01
4.96155024e-02 -7.92777259e-03 6.03876710e-01 3.43095303e-01
-2.69942492e-01 -1.53059021e-01 -7.60602951e-01 2.96688974e-01
5.38216233e-01 -3.76905113e-01 -5.29044032e-01 2.18247294e-01
-1.12307692e+00 9.79114115e-01 4.42062348e-01 3.66849542e-01
-1.16872616e-01 -5.01616560e-02 6.78765178e-01 1.30100176e-01
3.87241036e-01 1.62712961e-01 1.20888054e-01 -5.36084652e-01
-9.64689314e-01 7.43363440e-01 6.16145670e-01 1.16867757e+00
6.62566498e-02 -2.62943685e-01 -6.23720586e-01 1.09939992e+00
3.03566623e-02 8.37828964e-02 4.64436233e-01 -3.91465366e-01
1.25985432e+00 6.58363938e-01 4.54839766e-02 -1.12803149e+00
-4.67126191e-01 -6.81675494e-01 -8.83718252e-01 -2.69666672e-01
1.87454119e-01 -1.19555980e-01 -3.71673554e-01 1.91781628e+00
1.58808693e-01 1.52784824e-01 1.19447365e-01 1.10188937e+00
1.86028850e+00 7.27773428e-01 2.09537894e-01 -7.27965161e-02
1.62996745e+00 -1.46764970e+00 -9.18284476e-01 -3.58545005e-01
7.62415826e-01 -6.56783402e-01 1.36918318e+00 3.97901952e-01
-1.37687063e+00 -6.28992975e-01 -1.63947511e+00 -4.22795534e-01
-2.71902114e-01 2.81520635e-01 4.75819826e-01 9.03029144e-02
-5.34946740e-01 5.71208715e-01 -3.88023198e-01 -3.36931527e-01
5.13565898e-01 -2.08557665e-01 -1.50547504e-01 -6.27231747e-02
-1.55414808e+00 7.29771197e-01 7.24169612e-01 -7.41561279e-02
-1.00561249e+00 -5.70653617e-01 -9.74318326e-01 1.90848373e-02
3.36706012e-01 -5.98428130e-01 1.43794227e+00 -5.12837827e-01
-1.29992306e+00 1.11420727e+00 6.91363886e-02 -9.37714353e-02
8.02944183e-01 -5.74014068e-01 -4.87206459e-01 -3.84839438e-02
1.80345684e-01 2.25220188e-01 2.31287643e-01 -9.51780081e-01
-5.98375797e-01 -9.49736312e-02 3.64020765e-01 5.64673305e-01
-1.33431345e-01 1.54748857e-01 -6.63279951e-01 -1.02286422e+00
1.75415680e-01 -3.45245510e-01 4.63045627e-01 -6.29694462e-01
-6.78047836e-01 -4.19420332e-01 6.77834153e-01 -1.00387347e+00
1.90479171e+00 -2.06314325e+00 3.90439987e-01 -3.80136520e-01
3.69545728e-01 1.14257969e-02 -3.05065289e-02 8.29252779e-01
-2.71175921e-01 -9.51571167e-02 -1.41694114e-01 -4.81234163e-01
1.47257924e-01 -2.85068750e-01 -3.46713185e-01 1.71359926e-01
-2.29633987e-01 1.03059292e+00 -1.15096986e+00 -3.65913332e-01
4.69754077e-02 7.47984797e-02 -4.25764740e-01 3.83858323e-01
-4.53031719e-01 2.46091977e-01 -4.87636030e-01 2.58102268e-01
5.59158087e-01 -1.65909067e-01 -2.57550150e-01 5.84084205e-02
-1.53632820e-01 7.20766425e-01 -1.14412761e+00 2.26100445e+00
3.11072506e-02 7.13077128e-01 -3.11839461e-01 -5.82859099e-01
8.95523012e-01 4.47352767e-01 5.14902081e-03 -6.44430161e-01
7.11493790e-01 -2.05615699e-01 -3.16354632e-01 -1.01853669e+00
7.69412756e-01 -1.01045765e-01 -6.02956533e-01 7.79671729e-01
-8.95256235e-04 -2.90027767e-01 3.84514809e-01 5.19948363e-01
1.15824258e+00 3.19815993e-01 1.64903298e-01 -1.21932462e-01
4.73337293e-01 -9.07960832e-02 2.72256494e-01 5.72174191e-01
-2.36332566e-02 7.44793653e-01 8.01711321e-01 -1.05565764e-01
-9.36886668e-01 -1.06774402e+00 2.96191305e-01 1.19952309e+00
3.91470343e-01 -8.73094857e-01 -9.67849791e-01 -6.28169835e-01
-4.64011073e-01 1.30434370e+00 -7.69351780e-01 -2.72631913e-01
-5.45120180e-01 -6.46457672e-01 6.09843850e-01 7.65396595e-01
1.31880987e+00 -1.01373351e+00 -5.79247415e-01 1.20081589e-01
-8.96613836e-01 -9.32213783e-01 -4.34093326e-01 -2.08093226e-01
-2.77955353e-01 -1.01315868e+00 -2.90441364e-01 -7.05926180e-01
1.21029630e-01 1.32413104e-01 1.13561761e+00 1.62368491e-01
2.36348659e-01 -3.50971729e-01 -7.15076506e-01 -3.42666179e-01
-8.86629671e-02 1.92764133e-01 -5.57515264e-01 -1.88069910e-01
4.87710655e-01 -8.77342165e-01 -5.22005260e-01 2.54207909e-01
-7.37080872e-01 9.31309760e-01 1.62800148e-01 6.77665114e-01
2.06066906e-01 6.21050298e-02 4.02211815e-01 -9.64673221e-01
9.53951538e-01 -8.97808790e-01 1.28667310e-01 3.70269790e-02
5.59187159e-02 -2.71086484e-01 4.76463914e-01 -4.57733274e-01
-1.02931345e+00 -5.17128587e-01 -2.51307070e-01 3.56755629e-02
-1.56356376e-02 8.46444607e-01 -6.66771829e-01 1.07658780e+00
6.61986172e-01 2.34752804e-01 -5.98901391e-01 -5.51021934e-01
4.37817603e-01 8.52464378e-01 1.19012105e+00 -6.54354274e-01
7.85010338e-01 7.86015168e-02 -7.03665018e-01 -3.48245233e-01
-1.27659023e+00 -3.47419918e-01 -5.00946403e-01 -7.02339411e-01
9.81708884e-01 -1.02046335e+00 -5.80553114e-01 6.70095682e-01
-1.19348466e+00 -2.79921353e-01 -3.11348528e-01 2.41905466e-01
-6.77869022e-01 -1.05606772e-01 -5.88598192e-01 -4.22291309e-01
-2.76050627e-01 -7.28122354e-01 8.99506211e-01 5.12205124e-01
-6.60482407e-01 -7.51655757e-01 9.18721035e-02 7.71648765e-01
-5.67890257e-02 8.23376179e-01 1.05945110e+00 -9.09717500e-01
-3.43251973e-01 -1.54579192e-01 -3.08341950e-01 -1.71685085e-01
-3.55652571e-01 -2.34270170e-01 -1.17745018e+00 -4.84456606e-02
5.40006757e-02 -6.58251762e-01 9.04350817e-01 -5.59035800e-02
9.34182405e-01 -2.41849005e-01 -2.90439367e-01 6.00636899e-01
1.16137493e+00 3.27626288e-01 9.08587933e-01 5.72189450e-01
4.70957428e-01 3.90055001e-01 5.46807408e-01 4.24158245e-01
8.51311982e-01 5.72251856e-01 3.97958726e-01 1.37455329e-01
-2.10251227e-01 -8.13236296e-01 2.41261721e-01 1.08227694e+00
-6.05986565e-02 -6.88099563e-01 -9.57245886e-01 6.25415862e-01
-2.19713068e+00 -1.29783022e+00 -4.41733450e-01 1.73592341e+00
9.81909633e-01 3.51041138e-01 1.30796328e-01 5.54258108e-01
5.61537743e-01 6.66702032e-01 -3.63728344e-01 -2.59243399e-01
-3.05650622e-01 -8.66119191e-02 -3.38188350e-01 4.61864442e-01
-1.25806141e+00 9.98433530e-01 5.28226423e+00 1.07330644e+00
-5.37325025e-01 3.10496449e-01 5.62116385e-01 -2.60950595e-01
-3.58363241e-01 -1.51268080e-01 -5.38588822e-01 6.02284729e-01
5.65203428e-01 -3.54129821e-01 2.28199482e-01 5.83547771e-01
2.85992563e-01 -4.76041228e-01 -1.33012557e+00 1.00276148e+00
4.28528041e-01 -1.28120589e+00 9.51842517e-02 -5.88274956e-01
5.55552363e-01 -4.88464028e-01 -1.05076492e-01 7.10344493e-01
2.95590729e-01 -1.20102155e+00 1.19366574e+00 5.89836776e-01
8.42976153e-01 -9.41549838e-01 6.76657796e-01 6.92223132e-01
-1.33816206e+00 9.74208936e-02 1.50751159e-01 -4.57380831e-01
4.09533650e-01 2.29779467e-01 -3.51434976e-01 7.61738122e-01
6.45115197e-01 9.24033821e-01 -5.14695168e-01 9.24580872e-01
-8.70408893e-01 6.00469053e-01 3.59969772e-02 -2.74703622e-01
3.54418248e-01 -1.30060673e-01 7.77673006e-01 1.19727576e+00
-6.59508109e-02 6.57562375e-01 9.76363420e-02 9.59613442e-01
-2.37031221e-01 2.08917096e-01 -3.19872975e-01 -3.92210744e-02
6.26886725e-01 9.74268496e-01 -5.75623155e-01 -4.30657387e-01
-9.48314071e-02 1.13616097e+00 4.98561531e-01 2.45015189e-01
-1.23669040e+00 -7.29814947e-01 2.11396486e-01 1.03845112e-01
4.02701274e-02 1.42736420e-01 -5.40624321e-01 -1.06443679e+00
2.43635073e-01 -7.33786523e-01 6.76753044e-01 -1.27758908e+00
-9.29052949e-01 6.58162892e-01 4.30802524e-01 -1.16565526e+00
-8.77161101e-02 -9.12833437e-02 -1.11919332e+00 8.85078490e-01
-9.54710186e-01 -1.33600628e+00 -8.83347750e-01 4.18375462e-01
9.77995276e-01 -2.10257433e-02 7.17696071e-01 3.03003788e-01
-7.75419712e-01 7.41494775e-01 6.15076609e-02 4.97061342e-01
2.94164330e-01 -1.18650174e+00 3.31141263e-01 8.17000747e-01
-2.80023992e-01 2.55281091e-01 8.61832261e-01 -7.15406597e-01
-6.38329208e-01 -9.49441195e-01 1.14408672e+00 -5.40664256e-01
5.95976532e-01 -6.71231091e-01 -7.33793736e-01 7.56733477e-01
3.95542979e-01 -9.70058620e-01 6.90654695e-01 6.21827133e-02
-4.75080907e-01 3.33988249e-01 -8.99314284e-01 7.99265325e-01
1.31859815e+00 -3.47904861e-01 -1.18418849e+00 3.55834603e-01
9.26328659e-01 -8.45282435e-01 -5.35468280e-01 2.41044521e-01
3.91644061e-01 -9.74276245e-01 7.41665125e-01 -7.00649083e-01
1.28233957e+00 -1.25564188e-01 -1.38265505e-01 -1.32764268e+00
-3.99790674e-01 -6.31272137e-01 -1.23115204e-01 1.63428783e+00
3.24656665e-01 6.17715865e-02 5.29926121e-01 4.41626072e-01
-3.80901754e-01 -8.02583218e-01 -6.72308564e-01 -3.84047836e-01
2.68713593e-01 -8.26492667e-01 8.04715455e-01 9.71622348e-01
5.32271445e-01 8.39454532e-01 -2.55988479e-01 -2.09979743e-01
3.92197013e-01 4.53946926e-02 7.11427033e-01 -9.52737331e-01
-2.54147083e-01 -7.27426231e-01 2.43255887e-02 -1.40281177e+00
-1.46377578e-01 -1.14617527e+00 -1.26709700e-01 -1.83949542e+00
7.13696003e-01 -7.25866705e-02 6.18202286e-03 2.24325597e-01
-3.91904265e-01 -1.46946609e-01 2.26200372e-01 4.96097915e-02
-7.23681808e-01 6.90067351e-01 1.18701303e+00 -2.50434786e-01
-1.84113234e-01 -2.06829235e-01 -8.08493137e-01 8.33680809e-01
7.07864881e-01 -1.25630319e-01 -8.07519257e-01 -6.61301255e-01
2.23476574e-01 2.53501892e-01 3.23711604e-01 -1.19213045e+00
3.05916578e-01 1.39706746e-01 2.43182123e-01 -1.11471903e+00
3.36251467e-01 -1.70497209e-01 2.15798914e-01 -5.28735407e-02
-6.41182542e-01 1.35748118e-01 2.08564490e-01 2.81268209e-01
-1.98313192e-01 -3.05641353e-01 5.89253426e-01 2.47253869e-02
-4.38820124e-01 -1.02677764e-02 -1.81676850e-01 4.65578824e-01
1.07396913e+00 -2.38970980e-01 -7.45033026e-01 -6.57017171e-01
-8.85617256e-01 6.10928357e-01 5.00434823e-03 8.34349215e-01
6.39746547e-01 -1.63864553e+00 -1.11091661e+00 1.03418246e-01
2.09737763e-01 2.95571089e-01 6.02619529e-01 6.08808935e-01
-4.97328788e-01 2.03520179e-01 1.03456356e-01 -2.70715535e-01
-1.29639864e+00 3.89685482e-01 1.77762091e-01 -5.20211279e-01
-6.68224514e-01 1.31375539e+00 1.68292403e-01 -1.44400599e-03
5.28049409e-01 -1.40251935e-01 -6.95969522e-01 2.92774171e-01
8.93773019e-01 4.52924639e-01 -2.47541294e-01 -7.91810811e-01
-3.16067003e-02 4.01642591e-01 -2.34943271e-01 -3.47765476e-01
1.38645828e+00 -9.57704261e-02 3.60252298e-02 7.27952242e-01
9.91289139e-01 -3.36911947e-01 -9.95442390e-01 -3.15662593e-01
3.07699423e-02 -2.97924846e-01 -1.43918857e-01 -1.21155727e+00
-7.47912765e-01 9.23794150e-01 -1.47801563e-01 8.56867209e-02
1.14795053e+00 1.61990359e-01 9.36016858e-01 -8.09296295e-02
2.09756806e-01 -1.18465805e+00 2.06278101e-01 7.49328017e-01
1.43930292e+00 -6.63724959e-01 -2.43711974e-02 -3.32901865e-01
-8.43211710e-01 7.82538712e-01 8.36486161e-01 -7.99286962e-02
3.58603060e-01 1.22704014e-01 -3.23150009e-01 -4.08235431e-01
-8.75537157e-01 1.03360727e-01 2.52393007e-01 2.67806053e-01
3.20241958e-01 5.13099395e-02 -6.13178074e-01 1.67473340e+00
-8.95281255e-01 -8.03691745e-02 3.36461067e-01 8.38288188e-01
-4.77770299e-01 -5.17131507e-01 -1.38830304e-01 2.33837396e-01
-8.46458599e-02 -1.52073428e-01 -6.51353657e-01 1.11906421e+00
4.21218663e-01 1.11000001e+00 1.49044707e-01 -9.18079436e-01
7.16171205e-01 2.31567785e-01 2.04281762e-01 -4.74220455e-01
-9.67665195e-01 -5.44917099e-02 5.54822683e-01 -4.58282471e-01
-7.49130100e-02 -7.82001138e-01 -1.30333424e+00 -6.38985276e-01
-2.32150272e-01 3.02150398e-01 1.37876675e-01 1.09104276e+00
-1.50636449e-01 9.93544638e-01 3.72498095e-01 -4.49326724e-01
-6.84275944e-03 -1.40446413e+00 -3.87319773e-01 6.20769978e-01
-7.36800814e-03 -7.23328471e-01 -2.08433479e-01 1.01556264e-01] | [11.360431671142578, 8.842357635498047] |
959d71c3-9939-4c0d-9af5-abcf41e2598e | a-tale-of-a-probe-and-a-parser | 2005.01641 | null | https://arxiv.org/abs/2005.01641v2 | https://arxiv.org/pdf/2005.01641v2.pdf | A Tale of a Probe and a Parser | Measuring what linguistic information is encoded in neural models of language has become popular in NLP. Researchers approach this enterprise by training "probes" - supervised models designed to extract linguistic structure from another model's output. One such probe is the structural probe (Hewitt and Manning, 2019), designed to quantify the extent to which syntactic information is encoded in contextualised word representations. The structural probe has a novel design, unattested in the parsing literature, the precise benefit of which is not immediately obvious. To explore whether syntactic probes would do better to make use of existing techniques, we compare the structural probe to a more traditional parser with an identical lightweight parameterisation. The parser outperforms structural probe on UUAS in seven of nine analysed languages, often by a substantial amount (e.g. by 11.1 points in English). Under a second less common metric, however, there is the opposite trend - the structural probe outperforms the parser. This begs the question: which metric should we prefer? | ['Josef Valvoda', 'Rowan Hall Maudslay', 'Adina Williams', 'Ryan Cotterell', 'Tiago Pimentel'] | 2020-05-04 | a-tale-of-a-probe-and-a-parser-1 | https://aclanthology.org/2020.acl-main.659 | https://aclanthology.org/2020.acl-main.659.pdf | acl-2020-6 | ['contextualised-word-representations'] | ['natural-language-processing'] | [ 2.72150099e-01 4.76143926e-01 -2.47011185e-01 -5.93713403e-01
-6.50102317e-01 -1.02609229e+00 7.50972569e-01 2.47198209e-01
-6.95926726e-01 6.86900258e-01 6.37617230e-01 -7.81335354e-01
-1.24534622e-01 -6.22775555e-01 -3.63459527e-01 -6.28137589e-01
2.35311352e-02 1.07695773e-01 1.54943630e-01 -2.39339903e-01
5.23008823e-01 4.15281355e-01 -1.12475514e+00 3.05365622e-01
3.49935174e-01 6.39195919e-01 3.23486835e-01 3.70636642e-01
-7.67329693e-01 8.23796391e-01 -4.54560876e-01 -4.50234324e-01
6.16330840e-02 -5.65413773e-01 -1.10608017e+00 -3.52896392e-01
1.79896012e-01 1.42716587e-01 3.26219767e-01 1.09463310e+00
3.32747608e-01 -9.22131017e-02 3.84307265e-01 -4.35963482e-01
-4.64772642e-01 9.63231921e-01 -2.47687146e-01 6.61630750e-01
4.85388219e-01 -1.06973037e-01 1.30317318e+00 -6.20410860e-01
8.63881946e-01 1.54516780e+00 6.24349594e-01 4.45865035e-01
-1.60157704e+00 -3.49765271e-01 1.39361054e-01 -4.01316673e-01
-1.04001117e+00 -6.58675611e-01 6.74235523e-01 -5.38279831e-01
1.12249720e+00 2.65388489e-01 4.51058932e-02 1.08085883e+00
2.73618877e-01 4.73106176e-01 1.73873127e+00 -7.14669466e-01
2.77164102e-01 3.71078551e-01 3.50078732e-01 5.74994624e-01
3.61611813e-01 2.38700464e-01 -4.16860491e-01 -1.36933088e-01
3.45187068e-01 -7.50143826e-01 -1.60749152e-01 -1.15658730e-01
-9.40880954e-01 1.00123107e+00 2.24311188e-01 9.45893943e-01
-1.14220850e-01 -1.21464774e-01 7.06201494e-01 6.11113489e-01
4.82575268e-01 7.33511925e-01 -9.23784792e-01 -3.27920943e-01
-8.00922751e-01 1.64012879e-01 9.10860538e-01 4.55606848e-01
6.67816103e-01 2.06050240e-02 2.82191336e-01 7.80853391e-01
5.58261514e-01 -3.40058357e-02 7.85828054e-01 -8.25974762e-01
5.97570121e-01 5.36538243e-01 -3.13347504e-02 -8.00311804e-01
-5.20288348e-01 -2.35219374e-01 -1.21807486e-01 2.77968869e-02
8.23532999e-01 -2.97973126e-01 -5.15932977e-01 1.96279287e+00
1.49787456e-01 -4.76528168e-01 7.23795146e-02 6.21415496e-01
6.41474962e-01 4.97678727e-01 7.51270235e-01 -2.40630955e-01
1.41449594e+00 -4.45701599e-01 -3.93099993e-01 -7.35740364e-01
1.05776608e+00 -8.65083694e-01 1.37329447e+00 1.44185737e-01
-9.33677137e-01 -3.54038239e-01 -8.89796674e-01 -1.18994609e-01
-5.20392835e-01 -1.66706249e-01 7.21339166e-01 1.10975099e+00
-1.16970956e+00 6.28094435e-01 -6.37087047e-01 -5.56268394e-01
-6.05268441e-02 2.30128646e-01 -5.27036309e-01 2.16524273e-01
-1.06369364e+00 1.23934054e+00 6.40931964e-01 -2.76340425e-01
-1.82082448e-02 -2.28622884e-01 -1.04773641e+00 1.07037928e-02
3.47039610e-01 -2.01304689e-01 1.33728409e+00 -1.32133687e+00
-1.58341515e+00 1.43770850e+00 -3.90379608e-01 -3.75569612e-01
1.54775977e-01 4.39890288e-02 -3.95084500e-01 -1.45721272e-01
2.25113079e-01 5.07176101e-01 2.96421915e-01 -1.11703420e+00
-4.03258353e-01 -5.09104550e-01 1.92716241e-01 -8.89308676e-02
1.43952817e-01 6.39620125e-01 1.70386717e-01 -6.33616567e-01
2.03221753e-01 -7.41270840e-01 -1.58156604e-01 -4.73898292e-01
-2.08160773e-01 -6.19489729e-01 2.14944229e-01 -6.55648530e-01
1.28720450e+00 -2.22267604e+00 -1.08631171e-01 2.67489627e-02
1.65933482e-02 3.03568691e-01 -1.93988264e-01 5.62620819e-01
-4.55533952e-01 5.49750626e-01 -3.98276120e-01 -1.70096144e-01
3.18046361e-01 4.89843130e-01 -3.46207857e-01 3.38468581e-01
5.10132551e-01 1.00278735e+00 -7.93400884e-01 -4.78088588e-01
-9.20221210e-02 2.40457326e-01 -3.55826378e-01 -1.30180836e-01
-7.65237510e-02 4.02957797e-01 -4.93948787e-01 5.67118943e-01
2.22785354e-01 3.94557826e-02 4.20573026e-01 3.22514504e-01
-5.05145967e-01 1.03967559e+00 -9.12303150e-01 1.49919415e+00
-2.54964799e-01 6.80767298e-01 3.57741602e-02 -1.23299813e+00
1.06614971e+00 3.16411257e-01 -2.82859474e-01 -9.15695310e-01
3.35226089e-01 4.39889193e-01 4.73657370e-01 -6.68278515e-01
4.55674917e-01 -5.85197508e-01 -3.33599091e-01 3.17275077e-01
4.13559154e-02 2.38447174e-01 1.47251159e-01 -1.75523058e-01
1.27803302e+00 3.97551090e-01 8.45917046e-01 -8.57702315e-01
5.05673170e-01 1.41831739e-02 6.47839427e-01 6.69868827e-01
-3.26272339e-01 2.46372044e-01 9.00029719e-01 -3.49235743e-01
-8.21083605e-01 -1.02186763e+00 -3.70469958e-01 1.42363000e+00
-5.20954192e-01 -3.80266070e-01 -8.79646957e-01 -7.22124279e-01
-1.80945113e-01 9.62484121e-01 -6.07754886e-01 1.08654298e-01
-7.69226670e-01 -5.62098801e-01 5.48265159e-01 5.42815506e-01
9.12002847e-02 -1.51310742e+00 -9.26489055e-01 3.43304098e-01
-3.63641046e-02 -8.91620874e-01 1.47243589e-01 7.15058804e-01
-1.14202142e+00 -6.12121463e-01 -3.42548698e-01 -8.95594180e-01
4.25446361e-01 -2.56208688e-01 1.28699589e+00 -5.09755276e-02
1.39127940e-01 6.11391142e-02 -4.64387476e-01 -3.38630706e-01
-7.08909094e-01 2.20908985e-01 -2.82470107e-01 -4.66552228e-01
9.11042988e-01 -4.72354621e-01 -8.81191343e-02 -1.89891040e-01
-8.78833294e-01 -4.38722283e-01 8.44601929e-01 5.66770494e-01
1.82484418e-01 -5.11616230e-01 6.45393729e-01 -1.32850397e+00
8.65295708e-01 -6.72847688e-01 -3.39373618e-01 -5.27370423e-02
-4.50329125e-01 4.94787633e-01 6.77605569e-01 -2.26363167e-01
-8.90147507e-01 -7.83630461e-02 -5.98614097e-01 4.84113514e-01
-5.41831315e-01 8.58447492e-01 -3.60685766e-01 2.50234734e-02
8.19683135e-01 -1.28013462e-01 -2.23915279e-01 -7.92630851e-01
2.76464611e-01 5.43860614e-01 3.08169305e-01 -7.48949528e-01
4.86201912e-01 -2.60217283e-02 -2.50738472e-01 -8.85582983e-01
-8.47343206e-01 -4.14440274e-01 -6.04626298e-01 3.39753240e-01
7.94727802e-01 -4.66541260e-01 -2.81885743e-01 -1.54863536e-01
-1.28751516e+00 -4.08898205e-01 -3.43572378e-01 3.81207883e-01
-5.42196929e-01 4.76581395e-01 -6.54240012e-01 -7.33135700e-01
4.55126725e-02 -9.85839963e-01 7.27768004e-01 -1.28763869e-01
-8.76573741e-01 -1.09512937e+00 2.73847729e-01 -8.54652077e-02
4.45866495e-01 2.87176013e-01 1.35612619e+00 -1.17281508e+00
2.22345576e-01 -9.50860307e-02 -2.42729187e-01 3.24123442e-01
1.49148121e-01 -3.89248043e-01 -1.05032372e+00 -3.11298054e-02
5.33616364e-01 -1.93973929e-01 6.39972270e-01 5.12939952e-02
6.44909739e-01 -3.74158174e-01 -4.64222096e-02 1.19599991e-01
1.60341072e+00 3.68081927e-01 5.68955779e-01 6.60136342e-01
1.42862692e-01 1.13238442e+00 1.58122107e-01 -1.88951358e-01
1.89911410e-01 3.53664607e-01 -9.09057334e-02 3.46604884e-01
7.64595866e-02 -3.54393750e-01 5.77736557e-01 9.78169322e-01
1.76254407e-01 -2.49426395e-01 -1.12922835e+00 5.99880099e-01
-1.30520689e+00 -6.21852815e-01 -2.60715205e-02 2.01277995e+00
7.70674169e-01 7.02039838e-01 -4.14599553e-02 2.91256517e-01
4.08860505e-01 4.72615868e-01 3.43679115e-02 -1.11650467e+00
-6.64455444e-03 5.10026872e-01 3.83328348e-01 6.05880082e-01
-8.63659739e-01 1.05201399e+00 6.64025402e+00 4.06161338e-01
-1.05715442e+00 7.26462230e-02 5.97224057e-01 4.92837965e-01
-3.47885787e-01 3.38732153e-01 -7.27204978e-01 5.99907875e-01
1.57784653e+00 -7.64966244e-03 7.91839138e-02 6.98098838e-01
2.43477792e-01 -3.66033375e-01 -1.23827422e+00 4.75660145e-01
-8.03075135e-02 -8.89879525e-01 -2.45599493e-01 1.60602391e-01
1.53832480e-01 1.82932943e-01 -2.74821758e-01 3.63460422e-01
2.84129143e-01 -1.06937301e+00 6.98843777e-01 2.62967587e-01
5.57079732e-01 -5.07327318e-01 7.04938114e-01 5.87649882e-01
-9.29619491e-01 1.10707351e-03 -4.01229143e-01 -4.47917461e-01
2.06189871e-01 3.25922102e-01 -5.71299553e-01 1.89235151e-01
5.33099592e-01 7.38309100e-02 -7.80028403e-01 6.94491506e-01
-4.99238104e-01 1.13913763e+00 -3.39049727e-01 -1.68833822e-01
6.91641569e-01 -1.88122764e-01 5.22733808e-01 1.70609009e+00
3.83550711e-02 -4.44413833e-02 -2.12874129e-01 7.43151188e-01
1.96913511e-01 4.06015426e-01 -8.69774818e-01 -2.60656416e-01
4.12674576e-01 1.07194507e+00 -8.70608151e-01 -1.75326437e-01
-6.47267342e-01 5.60126305e-01 5.17662048e-01 1.13689259e-01
-2.66329855e-01 -3.76453668e-01 2.53159046e-01 2.54167318e-01
2.97605604e-01 -2.62054771e-01 -3.24016839e-01 -6.94977164e-01
2.14243174e-01 -8.12059999e-01 1.34453058e-01 -6.08473301e-01
-1.27961421e+00 7.19390750e-01 7.58271515e-02 -5.25501430e-01
-5.31184375e-01 -1.01530421e+00 -6.32056177e-01 1.05253673e+00
-1.38305461e+00 -7.84995496e-01 3.48751038e-01 1.21930189e-01
3.62011462e-01 1.00214988e-01 1.13359082e+00 5.10125468e-03
-3.68836105e-01 4.11857605e-01 -1.00118764e-01 3.49445164e-01
5.25479376e-01 -1.43666697e+00 7.78377354e-01 7.11506248e-01
3.67309421e-01 8.97159219e-01 8.27761352e-01 -4.07587111e-01
-9.24485981e-01 -6.13333821e-01 1.77753961e+00 -6.32812381e-01
9.66090143e-01 -2.25679949e-01 -1.13155782e+00 8.54435623e-01
3.05830866e-01 -1.23415418e-01 7.68554330e-01 2.67549127e-01
-5.63092649e-01 4.22465950e-01 -1.16113305e+00 5.05154371e-01
9.34045911e-01 -7.32747376e-01 -1.43870926e+00 -1.19387269e-01
7.01598704e-01 8.31440017e-02 -8.40662599e-01 2.19307721e-01
4.84888792e-01 -1.09482384e+00 5.94617307e-01 -7.20410228e-01
2.98972756e-01 1.25624668e-02 -3.43360454e-01 -1.15668452e+00
-4.54783827e-01 -3.82885754e-01 5.72060049e-01 1.50825548e+00
6.73875034e-01 -1.09963846e+00 6.74680293e-01 5.65181613e-01
-6.04281016e-02 -7.27367580e-01 -1.14181471e+00 -7.77858615e-01
6.49789929e-01 -5.27831376e-01 3.72241318e-01 1.06699502e+00
2.82333732e-01 5.96836329e-01 3.58389616e-01 -2.60136157e-01
1.54807687e-01 -3.07067215e-01 2.44211406e-01 -1.34053528e+00
-2.87068963e-01 -6.09392107e-01 -5.79268813e-01 -9.32000160e-01
4.35509950e-01 -1.07243860e+00 6.50921017e-02 -1.27124619e+00
-8.11198354e-02 -2.72667259e-01 -3.46561760e-01 3.47261369e-01
2.50770170e-02 3.71132651e-03 2.93180287e-01 8.56452882e-02
-1.86625883e-01 -4.78594191e-02 6.53250456e-01 3.94438058e-01
-1.30651236e-01 -3.47187072e-01 -1.04812062e+00 1.16308343e+00
1.11383975e+00 -7.46724427e-01 -1.55859768e-01 -5.60636699e-01
4.42010224e-01 -7.32313842e-02 2.87023455e-01 -6.84442043e-01
-3.94226834e-02 -4.91760485e-02 2.15680718e-01 9.66885500e-03
-8.29636008e-02 -5.34793317e-01 -8.86301845e-02 3.42538744e-01
-4.89783227e-01 5.50102830e-01 4.05605942e-01 2.17092484e-01
-2.00070366e-01 -8.60548675e-01 6.58321917e-01 -6.09349191e-01
-8.34313691e-01 -4.42050219e-01 -6.48234785e-01 3.90010774e-01
5.84745765e-01 -3.50576639e-01 -1.42907560e-01 1.70483306e-01
-7.80963361e-01 -3.59874576e-01 3.62337232e-01 3.71096641e-01
1.20291978e-01 -9.54702854e-01 -4.79634523e-01 1.47265404e-01
1.47473440e-01 -3.70094299e-01 -3.78485560e-01 7.55896866e-01
-4.20160025e-01 6.56318307e-01 1.19347729e-01 -2.12272316e-01
-8.03602278e-01 5.39685786e-01 2.16310561e-01 -4.00189221e-01
-5.22158265e-01 8.69581878e-01 2.35660866e-01 -5.60162187e-01
-1.32364482e-01 -5.47412395e-01 -3.36458594e-01 2.57747382e-01
4.12244111e-01 -1.25596058e-02 3.83953713e-02 -8.96794975e-01
-4.33924764e-01 3.77642006e-01 -1.43107027e-01 -2.82553822e-01
1.20188487e+00 -1.14714898e-01 -2.56508738e-01 1.03657675e+00
1.24118876e+00 3.41422766e-01 -8.41103435e-01 -3.18775028e-01
1.03456032e+00 -2.25074857e-01 -1.79015413e-01 -9.24326837e-01
-4.60167140e-01 9.01598752e-01 3.56367826e-01 5.80095947e-01
6.44418895e-01 2.44147286e-01 4.68952209e-01 2.43069232e-01
4.65740323e-01 -9.60017800e-01 -6.06672943e-01 6.16489291e-01
7.97394097e-01 -1.01415837e+00 -2.59546787e-01 -1.45015329e-01
-4.96825367e-01 9.98407364e-01 1.26663759e-01 -3.74225140e-01
5.45591831e-01 3.65941346e-01 1.41041487e-01 -2.84093946e-01
-5.78884363e-01 -1.85324326e-01 3.98731083e-02 3.27117026e-01
9.57892120e-01 1.95839450e-01 -8.84807825e-01 7.18489885e-01
-5.62241852e-01 -2.49827221e-01 2.18828917e-01 1.03249633e+00
-6.74744368e-01 -1.33747327e+00 -2.93895513e-01 4.20743525e-01
-8.68481100e-01 -2.55840600e-01 -7.01462150e-01 1.06116629e+00
8.94387588e-02 1.01008427e+00 -3.35202068e-02 -1.19727634e-01
3.38485986e-01 6.36022151e-01 2.55129367e-01 -8.22105050e-01
-7.55253196e-01 1.02958918e-01 4.66357917e-01 -6.29026115e-01
-5.68634868e-01 -6.82135880e-01 -1.14159167e+00 -7.00851008e-02
-2.04767630e-01 2.65567362e-01 6.62961543e-01 1.31031394e+00
-3.38586047e-02 1.20595448e-01 1.76406562e-01 -6.26290500e-01
-7.32712686e-01 -1.10938275e+00 -4.29198205e-01 2.09734485e-01
1.73822194e-01 -5.25176942e-01 -6.33032441e-01 -1.55490890e-01] | [10.517135620117188, 9.481322288513184] |
1cb1bd23-8969-450f-bd52-8596bfe43727 | semi-supervised-classification-with-graph | 1609.02907 | null | http://arxiv.org/abs/1609.02907v4 | http://arxiv.org/pdf/1609.02907v4.pdf | Semi-Supervised Classification with Graph Convolutional Networks | We present a scalable approach for semi-supervised learning on
graph-structured data that is based on an efficient variant of convolutional
neural networks which operate directly on graphs. We motivate the choice of our
convolutional architecture via a localized first-order approximation of
spectral graph convolutions. Our model scales linearly in the number of graph
edges and learns hidden layer representations that encode both local graph
structure and features of nodes. In a number of experiments on citation
networks and on a knowledge graph dataset we demonstrate that our approach
outperforms related methods by a significant margin. | ['Max Welling', 'Thomas N. Kipf'] | 2016-09-09 | null | null | null | null | ['node-classification-on-non-homophilic', 'graph-regression'] | ['graphs', 'graphs'] | [-9.09897238e-02 6.89871609e-01 -4.10701245e-01 -3.73881310e-01
-1.27521604e-01 -5.33955216e-01 6.78766906e-01 5.79975486e-01
-6.08203560e-02 3.59659165e-01 2.57143080e-01 -7.32616842e-01
-2.57839829e-01 -1.31869984e+00 -9.86867547e-01 -5.11790253e-03
-1.02426302e+00 4.46664572e-01 1.21716321e-01 -1.81297019e-01
8.71083736e-02 5.70879817e-01 -9.16011512e-01 3.38461280e-01
4.71525997e-01 7.56215572e-01 -2.57196784e-01 9.96054530e-01
-3.74849379e-01 1.41407156e+00 -2.24273503e-01 -5.09658396e-01
2.16890872e-01 -2.43465811e-01 -1.26485384e+00 1.95420474e-01
6.68923259e-01 -9.34595615e-02 -9.80375051e-01 1.09380007e+00
6.32617995e-02 1.16962098e-01 6.50567114e-01 -1.18427896e+00
-1.20421994e+00 9.63129699e-01 -2.00252846e-01 5.21422148e-01
2.40595177e-01 -1.58346519e-01 1.61616218e+00 -6.17043197e-01
8.43318582e-01 1.18133223e+00 1.15470755e+00 -8.10550805e-03
-1.47126293e+00 -2.17639819e-01 2.54261374e-01 8.09512958e-02
-1.26138866e+00 -1.58031777e-01 8.25669169e-01 -4.96950567e-01
1.39187181e+00 -2.50836253e-01 7.41572499e-01 6.61108673e-01
9.05661669e-04 4.36696261e-01 6.43326700e-01 -4.85434085e-01
4.72971126e-02 -4.44057524e-01 4.57569391e-01 1.62052822e+00
4.90957975e-01 -5.00410572e-02 -3.17646652e-01 -3.95358771e-01
1.08278573e+00 -1.59548968e-01 1.41659910e-02 -5.29874563e-01
-9.91165161e-01 1.13137269e+00 1.16559732e+00 3.06453437e-01
-3.04219306e-01 9.26651776e-01 4.22492951e-01 7.05349267e-01
6.47555172e-01 3.51516455e-01 -3.47504675e-01 4.01628673e-01
-7.73810983e-01 -2.15303048e-01 1.18644464e+00 1.10081863e+00
1.09604657e+00 1.01408072e-01 1.57989487e-01 3.66328984e-01
2.10426837e-01 5.42876720e-02 -4.71655168e-02 -7.35159993e-01
2.27171019e-01 9.71615195e-01 -4.50666070e-01 -1.02730381e+00
-6.32075965e-01 -5.84975064e-01 -8.13914835e-01 -1.58682908e-03
2.38852426e-01 -1.00627914e-01 -1.11664402e+00 1.29892361e+00
-1.70690164e-01 4.37130988e-01 -6.69796616e-02 4.92483735e-01
1.36178923e+00 2.43947193e-01 -9.56393853e-02 1.78373694e-01
8.32288265e-01 -1.06359482e+00 -5.77579916e-01 -4.95531559e-02
9.96333539e-01 -2.40684539e-01 7.43236423e-01 -1.77055076e-01
-1.01577735e+00 -2.69412547e-01 -9.64179337e-01 -1.62062466e-01
-7.64572680e-01 -2.98502725e-02 1.35059488e+00 3.23622942e-01
-1.79034734e+00 1.06238401e+00 -7.31878757e-01 -5.24461269e-01
7.16103733e-01 4.99124706e-01 -6.73568606e-01 7.02502355e-02
-1.05136025e+00 5.20785391e-01 6.35092258e-01 -5.23477942e-02
-6.91128671e-01 -4.37963635e-01 -1.15860367e+00 5.60691476e-01
1.33879825e-01 -6.91740096e-01 1.02886760e+00 -1.02984667e+00
-1.18701422e+00 9.57008362e-01 1.55113995e-01 -8.95398319e-01
-2.48901937e-02 3.22123408e-01 -3.49493980e-01 3.47193569e-01
-6.08670488e-02 1.85958087e-01 6.76748753e-01 -8.97321641e-01
-9.50995088e-02 -2.87707373e-02 2.85742223e-01 -1.46636710e-01
-5.42427778e-01 -7.77373388e-02 -2.75183439e-01 -5.00754178e-01
-1.22081153e-02 -7.15516508e-01 -5.09777367e-01 -1.49208814e-01
-6.00533068e-01 -4.90815014e-01 6.40528738e-01 -5.03357172e-01
1.11012101e+00 -1.79265928e+00 1.13411807e-01 7.53653288e-01
1.00080693e+00 -6.62841722e-02 -3.57663751e-01 8.40780973e-01
-3.37788820e-01 3.74071121e-01 -9.01740193e-02 -1.19824834e-01
-3.55728343e-02 2.24354014e-01 -4.71074842e-02 5.00872731e-01
4.25400048e-01 1.68053031e+00 -1.13067138e+00 -3.97596210e-01
-6.65387586e-02 3.16002250e-01 -3.81676257e-01 1.27955601e-01
-3.28608960e-01 -2.71602094e-01 -2.74511665e-01 5.91728270e-01
3.68573695e-01 -1.32020938e+00 6.30835652e-01 1.99308991e-02
3.11490506e-01 4.87678587e-01 -8.03555071e-01 1.71815062e+00
-1.12243302e-01 7.65111446e-01 1.75941885e-01 -1.43535864e+00
6.39315307e-01 1.41013339e-01 3.11018258e-01 -4.56810415e-01
5.01661282e-03 -9.98350512e-03 -4.69969027e-02 -6.58330619e-02
1.07794680e-01 2.89387107e-01 1.06974952e-01 7.35280156e-01
6.83816254e-01 2.09545135e-01 3.23838085e-01 9.32152092e-01
1.80978107e+00 -2.55853087e-01 2.89069206e-01 -6.38634443e-01
9.27091464e-02 -5.97344227e-02 -2.33697549e-01 9.09186065e-01
7.27823600e-02 1.09953351e-01 9.90621626e-01 -9.68328953e-01
-8.91165018e-01 -1.06591308e+00 2.56826460e-01 1.14699650e+00
-3.87567133e-01 -1.05225325e+00 -5.55015385e-01 -1.04953909e+00
2.51482338e-01 3.85994054e-02 -9.78328466e-01 -8.87726899e-03
-4.78210181e-01 -3.32305312e-01 5.41380048e-01 7.50644565e-01
1.88361079e-01 -1.09040093e+00 1.37513384e-01 1.84205934e-01
5.21604419e-01 -1.28801131e+00 -2.94522047e-01 5.02996147e-01
-1.02512670e+00 -1.63588095e+00 -2.45123982e-01 -1.29521954e+00
8.75808239e-01 3.22885573e-01 1.56972253e+00 7.21656084e-01
-2.52083480e-01 3.93941820e-01 -1.79991931e-01 -8.90126750e-02
-3.00999582e-01 5.52636206e-01 -4.53098923e-01 -2.32806101e-01
2.22729340e-01 -8.43761504e-01 -1.71583921e-01 -6.10738873e-01
-6.07152879e-01 -1.05053678e-01 6.09767795e-01 7.87871480e-01
2.23168597e-01 2.11106509e-01 3.76154333e-01 -1.42772019e+00
1.07817090e+00 -6.95218921e-01 -8.23996067e-01 2.55709022e-01
-6.53729737e-01 4.82789397e-01 7.10451663e-01 -5.02726063e-02
-3.04457784e-01 8.66392553e-02 1.82374597e-01 -4.88747358e-01
1.90831393e-01 1.15427458e+00 4.89632219e-01 -6.95996344e-01
6.92497134e-01 -7.02781156e-02 -2.96763470e-03 -2.79947549e-01
8.39632750e-01 1.80785075e-01 4.03600693e-01 -2.74709165e-01
1.00984395e+00 5.44905961e-01 4.60140914e-01 -8.34856033e-01
-7.39853740e-01 -2.61954904e-01 -8.19053233e-01 -8.16177651e-02
6.47699118e-01 -8.91330779e-01 -8.23202431e-01 -2.13028095e-03
-9.97504473e-01 -7.05612063e-01 -1.18585855e-01 2.49713555e-01
-2.95952976e-01 4.65089053e-01 -1.17759001e+00 -3.06125283e-01
-3.30021858e-01 -4.71565098e-01 8.34225833e-01 -1.64327353e-01
6.31522462e-02 -1.61663830e+00 1.91106066e-01 -2.14929178e-01
5.02636969e-01 3.92551988e-01 1.22169423e+00 -8.89696360e-01
-8.09733570e-01 -2.52743512e-01 -7.01138794e-01 6.44529983e-02
-1.10672461e-02 2.04701070e-03 -5.70433378e-01 -4.59521413e-01
-9.63296056e-01 -5.09233713e-01 1.36533296e+00 4.21261609e-01
1.17159224e+00 -4.83177125e-01 -3.81114602e-01 8.78339112e-01
1.44107223e+00 -6.78503513e-01 3.63536835e-01 -5.78432977e-02
1.20873237e+00 2.26545691e-01 -4.54855740e-01 -1.34116700e-02
4.48225200e-01 -1.71089813e-01 5.59433997e-01 -4.14819986e-01
-3.50737005e-01 -4.65908408e-01 -7.15712905e-02 8.47766042e-01
-3.37339967e-01 -1.51952490e-01 -1.03760695e+00 7.07717776e-01
-2.06265068e+00 -8.94553304e-01 -4.30598706e-01 1.69385660e+00
3.56039435e-01 2.60297447e-01 2.57053614e-01 -1.91669956e-01
8.01721811e-01 3.57650787e-01 -2.62079179e-01 -5.71056783e-01
-1.74517244e-01 8.54920864e-01 9.26057816e-01 7.07972348e-01
-1.22532487e+00 1.29611540e+00 8.56262016e+00 2.73619056e-01
-6.87191010e-01 -6.69515505e-02 2.98944056e-01 2.38534987e-01
-5.45573413e-01 2.67675519e-01 1.19038403e-01 -9.38699692e-02
1.25778818e+00 -2.43624657e-01 9.39483166e-01 8.74783933e-01
-5.68581820e-01 6.98159456e-01 -1.12969637e+00 7.58683801e-01
-1.43854082e-01 -2.05664659e+00 4.14453074e-02 3.29321265e-01
9.01002228e-01 6.67330980e-01 -2.59246498e-01 1.22584134e-01
1.17655587e+00 -1.62690365e+00 1.41623333e-01 3.98591727e-01
7.84122348e-01 -7.64041901e-01 4.33947921e-01 -1.03371412e-01
-1.57952261e+00 -9.06832423e-03 -4.51085061e-01 -5.14742136e-01
-3.53598714e-01 6.19853556e-01 -1.06156552e+00 5.50663292e-01
6.14572763e-01 1.12298632e+00 -7.57642746e-01 5.92727959e-01
-6.34865105e-01 8.43247950e-01 -1.12440996e-01 -5.74914478e-02
5.32237351e-01 -1.07567646e-01 1.46310017e-01 1.56604028e+00
-1.15113325e-01 1.04889214e-01 3.81632954e-01 1.06500018e+00
-8.37460458e-01 6.73077255e-02 -1.08009195e+00 -8.52237999e-01
2.99404562e-01 1.38809371e+00 -9.08109903e-01 -4.95406598e-01
-7.57873416e-01 8.40921462e-01 1.28945017e+00 7.41337240e-01
-3.31351608e-01 -7.57927477e-01 2.68188894e-01 1.49968892e-01
7.00268924e-01 -5.93006909e-01 1.42199785e-01 -1.27638400e+00
-1.54836372e-01 -3.42049271e-01 8.43547881e-01 -5.06811023e-01
-1.56688559e+00 6.05495036e-01 -3.99935126e-01 -3.69762093e-01
-1.85533136e-01 -8.25888574e-01 -8.16120386e-01 8.62503767e-01
-1.58982027e+00 -1.35582781e+00 -2.85169661e-01 8.09826255e-01
-3.61036509e-01 -1.44152239e-01 1.03877187e+00 -8.82760659e-02
-1.14960305e-01 3.53700310e-01 4.69910493e-03 8.47340763e-01
1.32094041e-01 -1.68815136e+00 1.28502369e+00 7.21265018e-01
6.57616079e-01 7.10832536e-01 1.84091032e-01 -7.25626230e-01
-1.55691612e+00 -1.14342010e+00 9.87601817e-01 -1.53772861e-01
1.15218019e+00 -5.81162632e-01 -9.49119687e-01 1.45482075e+00
4.34973359e-01 8.41111422e-01 5.49562871e-01 6.85593188e-01
-7.79608965e-01 3.89232576e-01 -7.94831753e-01 2.84740955e-01
1.52349007e+00 -1.05602431e+00 -2.81717867e-01 7.28373706e-01
7.55465031e-01 -1.45936131e-01 -1.14159381e+00 2.34612212e-01
1.84025019e-01 -7.19778419e-01 8.92269194e-01 -1.24372494e+00
2.43861318e-01 3.60135809e-02 2.11501792e-01 -1.42440176e+00
-9.82069016e-01 -8.94559860e-01 -7.95362413e-01 3.45323503e-01
3.56376708e-01 -6.34297132e-01 1.09168136e+00 -1.93216689e-02
8.50503054e-03 -5.77315986e-01 -5.87379277e-01 -6.57324612e-01
1.00659899e-01 -4.68008816e-02 4.30644304e-01 1.26148391e+00
3.96533191e-01 5.94678521e-01 7.25562274e-02 2.45098755e-01
8.77698004e-01 4.06654716e-01 6.03305519e-01 -1.59076226e+00
-1.81623742e-01 -5.89215398e-01 -9.53045607e-01 -7.66535282e-01
7.03742504e-01 -1.52808177e+00 -5.55145144e-01 -2.06677294e+00
2.70519823e-01 4.18630764e-02 -4.99553531e-01 7.03271031e-01
1.89127084e-02 3.30822766e-01 -3.24440271e-01 -8.20645243e-02
-9.39087689e-01 1.27386913e-01 7.70072937e-01 -2.96878576e-01
3.33995782e-02 -3.86236131e-01 -6.09027326e-01 5.32471001e-01
6.53625488e-01 -2.68956155e-01 -4.18786228e-01 -4.10129488e-01
6.38196290e-01 -1.16996579e-01 5.99171162e-01 -6.25473380e-01
3.73591453e-01 7.47937784e-02 4.04979974e-01 -3.37368757e-01
-2.36451134e-01 -3.28007102e-01 -1.58097014e-01 5.21409512e-01
-5.41537225e-01 5.15979044e-02 5.49611747e-02 8.44795883e-01
-7.91538879e-02 3.04287374e-01 4.46043938e-01 -4.84355003e-01
-7.08361447e-01 6.54175818e-01 -1.02651648e-01 1.28151059e-01
5.31471193e-01 1.40317723e-01 -4.79937166e-01 -7.90309370e-01
-9.35116351e-01 2.14964733e-01 4.60520148e-01 1.98729396e-01
6.18319333e-01 -1.47190750e+00 -7.25324392e-01 9.10140350e-02
2.79701203e-01 -2.56750911e-01 -3.84723306e-01 4.59026515e-01
-7.75015652e-01 5.35814226e-01 4.76160683e-02 -1.32773578e-01
-9.10630107e-01 8.28146696e-01 3.85249466e-01 -3.78126413e-01
-8.26196611e-01 8.77424121e-01 -1.78394094e-01 -4.92663652e-01
2.07237363e-01 -4.62762058e-01 8.48768204e-02 -4.29338336e-01
2.88285315e-01 1.40357390e-01 4.73724604e-02 -5.17381549e-01
-4.18675035e-01 7.96196088e-02 4.30918708e-02 3.94154996e-01
1.50120461e+00 3.14427793e-01 -5.58248997e-01 2.56723791e-01
1.50207162e+00 -1.61311433e-01 -1.04509425e+00 -5.67675412e-01
3.50224108e-01 -5.88264130e-02 3.02678555e-01 -4.30974990e-01
-1.12430084e+00 6.93654239e-01 -1.01394303e-01 7.44386971e-01
6.97482228e-01 4.97126251e-01 5.67915022e-01 9.00816321e-01
1.21863462e-01 -8.86997283e-01 2.61629134e-01 6.67416811e-01
5.72705388e-01 -1.22807372e+00 2.55742162e-01 -6.51099324e-01
5.43384030e-02 1.53598583e+00 1.75298601e-01 -8.35481584e-01
1.07243717e+00 2.26692602e-01 -3.21538776e-01 -1.13718569e+00
-7.97425032e-01 -5.69924295e-01 5.59646547e-01 7.10557163e-01
6.68235540e-01 3.03746611e-01 6.02256581e-02 1.56522572e-01
-4.52189967e-02 -2.24978045e-01 4.43779409e-01 8.34424675e-01
-4.87289220e-01 -9.09895599e-01 3.42302531e-01 7.76815355e-01
-3.78679603e-01 -5.17051697e-01 -1.05312705e+00 8.25211346e-01
-5.32531798e-01 8.17282081e-01 6.76787868e-02 -4.54997629e-01
-4.19484861e-02 -5.53606823e-03 7.05209970e-01 -8.79483819e-01
-4.47015971e-01 -3.86250049e-01 3.31807375e-01 -7.76166558e-01
-4.82129246e-01 -8.17045867e-02 -1.41858912e+00 -6.13075554e-01
-2.75175989e-01 2.21944064e-01 3.95646781e-01 6.54227674e-01
4.88949805e-01 7.43939221e-01 3.65578681e-01 -6.61495686e-01
-3.30751628e-01 -9.56593275e-01 -8.69189203e-01 5.09691536e-01
3.79080385e-01 -2.30883807e-01 -3.52640003e-01 -1.63162485e-01] | [6.916823387145996, 6.2176313400268555] |
b3ce9c43-d03d-4401-b761-aa8d02e74e61 | learning-from-label-relationships-in-human | 2207.05577 | null | https://arxiv.org/abs/2207.05577v2 | https://arxiv.org/pdf/2207.05577v2.pdf | Learning from Label Relationships in Human Affect | Human affect and mental state estimation in an automated manner, face a number of difficulties, including learning from labels with poor or no temporal resolution, learning from few datasets with little data (often due to confidentiality constraints) and, (very) long, in-the-wild videos. For these reasons, deep learning methodologies tend to overfit, that is, arrive at latent representations with poor generalisation performance on the final regression task. To overcome this, in this work, we introduce two complementary contributions. First, we introduce a novel relational loss for multilabel regression and ordinal problems that regularises learning and leads to better generalisation. The proposed loss uses label vector inter-relational information to learn better latent representations by aligning batch label distances to the distances in the latent feature space. Second, we utilise a two-stage attention architecture that estimates a target for each clip by using features from the neighbouring clips as temporal context. We evaluate the proposed methodology on both continuous affect and schizophrenia severity estimation problems, as there are methodological and contextual parallels between the two. Experimental results demonstrate that the proposed methodology outperforms all baselines. In the domain of schizophrenia, the proposed methodology outperforms previous state-of-the-art by a large margin, achieving a PCC of up to 78% performance close to that of human experts (85%) and much higher than previous works (uplift of up to 40%). In the case of affect recognition, we outperform previous vision-based methods in terms of CCC on both the OMG and the AMIGOS datasets. Specifically for AMIGOS, we outperform previous SoTA CCC for both arousal and valence by 9% and 13% respectively, and in the OMG dataset we outperform previous vision works by up to 5% for both arousal and valence. | ['Ioannis Patras', 'Niki Maria Foteinopoulou'] | 2022-07-12 | null | null | null | null | ['continuous-affect-estimation'] | ['computer-vision'] | [ 2.52716333e-01 2.31601045e-01 -1.61393762e-01 -5.84102035e-01
-1.03202927e+00 -3.01285833e-01 6.48632407e-01 2.84502417e-01
-6.91672444e-01 7.49097645e-01 3.48844528e-01 6.52063012e-01
-3.65478337e-01 -9.98872593e-02 -1.35336801e-01 -8.91121328e-01
-1.58801958e-01 4.51712519e-01 -3.32407802e-01 5.28095327e-02
4.79494072e-02 3.47409457e-01 -1.51159871e+00 6.10527456e-01
5.98070920e-01 1.29169917e+00 -2.45938316e-01 3.77139986e-01
3.95390570e-01 9.70194042e-01 -3.40971828e-01 -6.72571301e-01
-9.47270077e-03 -2.52671570e-01 -7.99995720e-01 1.70491248e-01
5.74911118e-01 -2.75948465e-01 1.07724527e-02 8.22096229e-01
7.32897162e-01 -3.96189615e-02 8.98013115e-01 -1.20792282e+00
-4.45199817e-01 1.49175555e-01 -7.83197224e-01 7.08385482e-02
4.50366169e-01 8.10064301e-02 1.08517814e+00 -8.69562805e-01
5.91177583e-01 1.36061895e+00 8.08755815e-01 6.00105882e-01
-1.46196592e+00 -5.72847247e-01 2.96597093e-01 3.42634588e-01
-1.25156379e+00 -6.44856215e-01 5.24063051e-01 -7.51934886e-01
1.06355095e+00 2.07353890e-01 6.41066790e-01 1.52110493e+00
-3.59745920e-02 7.56483674e-01 1.55008507e+00 -2.23338082e-01
2.11433649e-01 3.24218929e-01 4.56741042e-02 4.85761940e-01
-3.34637552e-01 -1.19530171e-01 -4.79489386e-01 -1.18456513e-01
4.81840849e-01 -2.05618590e-01 -2.48017028e-01 -2.86328137e-01
-1.21820748e+00 1.00009549e+00 3.47029448e-01 3.83222282e-01
-6.76746666e-01 2.43799478e-01 7.55940676e-01 3.18824798e-01
8.29176784e-01 5.05687058e-01 -3.18887085e-01 -3.46329004e-01
-1.27267218e+00 2.10872769e-01 5.72936237e-01 2.08841234e-01
3.20985109e-01 -4.62084413e-02 -5.69211245e-01 1.10380054e+00
1.24273136e-01 2.34400347e-01 5.83160639e-01 -1.10459387e+00
3.58696252e-01 3.53706867e-01 -1.25953062e-02 -1.11740363e+00
-7.96381056e-01 -4.60065961e-01 -9.92158234e-01 3.66811454e-01
4.04075354e-01 -8.17246884e-02 -7.84528077e-01 2.09329700e+00
1.49066895e-01 5.51217608e-02 -1.52426526e-01 9.47571993e-01
6.68167591e-01 4.47476298e-01 2.13480473e-01 -9.13804054e-01
1.39511347e+00 -9.05392468e-01 -9.45318103e-01 -3.04181665e-01
4.37354684e-01 -5.30265093e-01 8.40973139e-01 8.84336174e-01
-1.35840440e+00 -5.14431059e-01 -8.83299768e-01 9.18796808e-02
-2.56052911e-01 3.43172103e-01 5.99469125e-01 4.35965151e-01
-1.35974884e+00 5.97476244e-01 -6.78123832e-01 -4.94068384e-01
6.58901513e-01 4.47209835e-01 -5.49542606e-01 -6.28787745e-03
-1.18073976e+00 1.07480049e+00 2.07417563e-01 2.04749569e-01
-6.90111578e-01 -7.83397079e-01 -7.23578870e-01 -1.30102500e-01
1.24459542e-01 -7.20436811e-01 1.03270268e+00 -1.44013488e+00
-1.50689042e+00 1.42588389e+00 1.40710995e-01 -5.98634005e-01
7.83706427e-01 -4.17720199e-01 -2.80731291e-01 4.70166653e-01
-2.77400762e-02 1.01890683e+00 1.08641279e+00 -8.64949465e-01
-2.89824098e-01 -5.25653899e-01 3.50726508e-02 3.00250560e-01
-3.62705976e-01 1.05432548e-01 -3.57230246e-01 -5.32293797e-01
-2.32128367e-01 -1.01847494e+00 -1.49781704e-01 1.58951402e-01
-6.16332144e-02 -2.44630381e-01 3.59101683e-01 -6.80191100e-01
1.17230666e+00 -2.16226339e+00 5.63657343e-01 1.80625934e-02
3.64434481e-01 2.97518581e-01 -8.49831551e-02 2.16321990e-01
-5.71237862e-01 -2.83846498e-01 -2.21673176e-01 -7.75414824e-01
1.27392530e-01 5.84940650e-02 -6.91546574e-02 8.60146403e-01
2.69079745e-01 7.12694764e-01 -7.56392479e-01 -4.77678329e-01
1.89608246e-01 5.81836820e-01 -4.88315940e-01 2.02866718e-01
1.55405581e-01 4.87881899e-01 7.06302002e-02 4.04130071e-01
3.81930858e-01 7.66080469e-02 -1.43365301e-02 -1.31179169e-01
-5.53174764e-02 3.95411812e-02 -9.97017026e-01 1.76163137e+00
-5.31646132e-01 6.53362095e-01 5.75560778e-02 -1.36989379e+00
8.55944872e-01 7.47455299e-01 7.28718340e-01 -6.43030345e-01
6.49499893e-02 1.23721302e-01 -1.15007497e-01 -8.01400721e-01
-1.80585444e-01 -3.97166908e-01 -1.37906998e-01 1.06283657e-01
3.14320654e-01 1.49727408e-02 3.98132168e-02 5.70059791e-02
9.45547104e-01 8.78611803e-02 5.28302073e-01 -1.05676942e-01
6.27427459e-01 -5.90055287e-01 5.47295034e-01 4.49284345e-01
-5.16903758e-01 7.01545715e-01 1.14180398e+00 -4.17044431e-01
-8.44694078e-01 -7.55997360e-01 -3.22723657e-01 1.01250648e+00
-5.52118599e-01 -2.59185314e-01 -7.94587612e-01 -6.17684484e-01
-2.46967882e-01 5.17968535e-01 -1.06793869e+00 -2.59488523e-01
-2.75382340e-01 -8.86575103e-01 4.48449463e-01 5.47585130e-01
1.45416543e-01 -1.31559753e+00 -6.27879620e-01 1.95295811e-01
-2.10752413e-01 -1.27660763e+00 8.89670849e-02 1.93598107e-01
-7.17571735e-01 -6.08255744e-01 -8.85587573e-01 -3.49002510e-01
2.46245578e-01 -4.55375612e-01 1.11377621e+00 -3.79894823e-01
-4.71178293e-01 5.46663940e-01 -2.29428589e-01 -3.64957005e-01
6.98085278e-02 -1.18374750e-01 1.62004784e-01 4.21052426e-01
3.39824975e-01 -7.25815535e-01 -6.77224636e-01 -1.32504493e-01
-7.23011613e-01 4.25381996e-02 7.70341635e-01 8.36575091e-01
4.28431869e-01 -4.78562444e-01 9.11117435e-01 -7.91359365e-01
4.42051589e-01 -6.63385212e-01 -2.18921319e-01 -8.36300924e-02
-7.03941464e-01 -2.60493279e-01 2.59506941e-01 -4.52043653e-01
-8.92204463e-01 1.71046615e-01 -6.98996857e-02 -5.85311532e-01
-4.70170408e-01 4.54655856e-01 1.73557520e-01 3.66887361e-01
7.06602991e-01 -2.91222751e-01 7.39304796e-02 -2.58125246e-01
3.75904530e-01 6.00204527e-01 4.59082931e-01 -6.49962053e-02
3.01158309e-01 6.13231063e-01 3.50795984e-02 -4.95463490e-01
-1.20063877e+00 -5.61653495e-01 -8.51259887e-01 -4.19054836e-01
1.12461567e+00 -1.13647103e+00 -6.06371462e-01 5.13488412e-01
-9.62105989e-01 -1.85632512e-01 -1.58360064e-01 6.77541554e-01
-9.96067882e-01 1.75635457e-01 -7.45696247e-01 -1.10785735e+00
-4.32975024e-01 -1.01049936e+00 1.22660697e+00 -1.85658470e-01
-7.74513721e-01 -9.47007298e-01 4.03153449e-01 5.15724301e-01
1.85246989e-01 8.11617613e-01 7.83900917e-01 -5.69655001e-01
2.20552146e-01 -2.59498328e-01 -2.91091859e-01 4.49442089e-01
-1.35562807e-01 -1.97222382e-01 -1.47165191e+00 -3.72675329e-01
2.84623474e-01 -8.61159742e-01 1.05564439e+00 6.16319895e-01
9.54625189e-01 -4.59848009e-02 -3.76631431e-02 3.73239368e-01
1.29286289e+00 7.82016257e-04 8.52704406e-01 3.57317895e-01
4.59068924e-01 1.11422217e+00 7.61450410e-01 6.75433636e-01
1.16228312e-01 9.81982768e-01 6.19694829e-01 -3.92177850e-01
-2.97773182e-02 5.49170494e-01 5.64920723e-01 6.68625057e-01
-2.44769573e-01 2.29825824e-01 -6.28877938e-01 8.05972278e-01
-2.16506481e+00 -1.07866931e+00 -1.87696502e-01 2.08263874e+00
9.17246044e-01 -2.36924528e-03 3.93107831e-01 2.92838901e-01
4.39414352e-01 2.09915146e-01 -3.41502815e-01 -8.25912118e-01
8.82677585e-02 2.71025985e-01 -3.28659192e-02 3.18148941e-01
-1.31025636e+00 6.53692245e-01 5.61785650e+00 6.99464917e-01
-1.16437507e+00 2.84993589e-01 8.76252294e-01 -7.02042758e-01
2.84021884e-01 -4.73217160e-01 -2.83512294e-01 3.18991542e-01
1.09215844e+00 4.99083281e-01 4.07470375e-01 6.25399232e-01
5.04287362e-01 5.10032959e-02 -1.32645822e+00 1.44543898e+00
3.66762042e-01 -6.90604031e-01 -5.28009057e-01 -6.34623319e-02
7.48822927e-01 -1.73746109e-01 4.24572587e-01 4.13845152e-01
-2.66764641e-01 -1.39150822e+00 5.95289588e-01 8.87310326e-01
8.25766802e-01 -8.82649541e-01 9.54614520e-01 7.03980774e-02
-6.44262612e-01 -2.15532377e-01 -2.52956033e-01 -2.90235817e-01
-1.01818413e-01 7.35582530e-01 -5.86849689e-01 3.85997415e-01
7.15459526e-01 9.00405109e-01 -4.51048315e-01 9.08922851e-01
-2.46404529e-01 4.38876659e-01 3.05526853e-02 3.28186154e-01
6.18031621e-01 -8.76424238e-02 2.68119991e-01 1.55827725e+00
1.40389100e-01 3.28268223e-02 -2.43244860e-02 7.65271783e-01
1.13119006e-01 4.05749053e-01 -5.06011605e-01 1.67706072e-01
-1.21305481e-01 1.66864133e+00 -3.99555057e-01 -2.60823995e-01
-4.42924470e-01 1.10832262e+00 6.52536213e-01 2.45265216e-01
-9.86248851e-01 -5.75872883e-02 5.80145776e-01 -1.43162847e-01
3.40199441e-01 3.61005187e-01 -1.35264710e-01 -9.15493786e-01
4.55547459e-02 -9.00318265e-01 5.43270826e-01 -8.88607264e-01
-1.30087709e+00 7.28111386e-01 -4.25937138e-02 -1.20258367e+00
-5.72901666e-01 -4.65360641e-01 -3.41694891e-01 6.80752337e-01
-1.41647542e+00 -1.30790043e+00 -2.10955188e-01 5.95408618e-01
4.30702478e-01 -6.03228062e-03 1.03741229e+00 4.53218877e-01
-5.01396894e-01 4.52073842e-01 -1.72683537e-01 -1.96591273e-01
1.18882096e+00 -1.38444746e+00 -3.28593701e-01 4.27347422e-01
1.46848917e-01 1.26893595e-01 6.64985955e-01 -8.70115682e-02
-7.01926589e-01 -1.03071654e+00 8.96785498e-01 -5.10138512e-01
6.18136108e-01 -4.18135613e-01 -8.67899776e-01 5.30421734e-01
3.72258991e-01 -1.19275190e-02 7.60157168e-01 3.16586912e-01
-4.75685447e-01 -1.67864010e-01 -1.22894561e+00 4.50699180e-01
7.85744548e-01 -5.92902958e-01 -5.73159933e-01 4.54514444e-01
2.65058279e-01 -1.10210471e-01 -1.06174576e+00 3.32231373e-01
7.41135120e-01 -1.27899325e+00 8.29303265e-01 -5.06250381e-01
6.97327614e-01 5.86629957e-02 -1.08790308e-01 -1.38629520e+00
-5.39471149e-01 -4.24817115e-01 2.44690590e-02 1.35106051e+00
2.48173848e-01 -3.48438472e-01 3.86592031e-01 5.37548304e-01
2.89345123e-02 -1.07506478e+00 -1.07201147e+00 -2.50878990e-01
-1.92618128e-02 -6.43716812e-01 -1.01310723e-01 1.16712046e+00
2.44635925e-01 6.00213826e-01 -7.47080266e-01 -1.12292938e-01
4.05144095e-01 -3.08712963e-02 2.65977085e-01 -1.32493806e+00
-2.09101900e-01 -5.84422052e-01 -7.39524007e-01 -2.47028634e-01
3.72598529e-01 -7.24489868e-01 -1.28131568e-01 -1.41294277e+00
4.33780104e-01 4.79839034e-02 -6.09657586e-01 5.61194479e-01
-4.43624705e-02 6.84151590e-01 1.36017203e-01 9.23785046e-02
-7.93873966e-01 4.64747131e-01 8.14941227e-01 -9.56069678e-02
-8.30171630e-02 -1.93619370e-01 -6.38233662e-01 1.02164423e+00
5.50756156e-01 -4.51913893e-01 -3.76070410e-01 -3.23704593e-02
3.36803764e-01 1.42654985e-01 3.59301716e-01 -1.07519889e+00
-7.71685913e-02 1.55647054e-01 4.88339484e-01 -2.85756379e-01
8.16627324e-01 -7.68840909e-01 -1.73445400e-02 4.02362019e-01
-6.35955870e-01 -3.73235613e-01 9.86912400e-02 4.01772976e-01
-2.28903651e-01 -1.89023003e-01 1.06276178e+00 -1.46489799e-01
-4.35766071e-01 1.54970482e-01 -3.58367652e-01 -3.86375636e-02
1.00109708e+00 -2.02742685e-02 -1.73825640e-04 -7.33368337e-01
-1.14490902e+00 4.70485315e-02 -3.86243174e-03 4.32055414e-01
3.80978584e-01 -1.44120300e+00 -8.99901092e-01 -1.22410595e-01
2.70999193e-01 -5.08635223e-01 5.27943254e-01 1.69890213e+00
4.18621004e-02 4.31434512e-01 -4.28636044e-01 -7.75740445e-01
-1.48627520e+00 6.01792395e-01 2.83563644e-01 -4.94719416e-01
-3.64808142e-01 8.24159801e-01 5.17639935e-01 -2.56045610e-01
3.76566976e-01 -3.30198973e-01 -6.42483115e-01 8.88966560e-01
7.01806009e-01 3.83598596e-01 2.37706020e-01 -7.91317821e-01
-3.47288609e-01 6.37216508e-01 -1.51388749e-01 -1.70767888e-01
1.63855088e+00 -7.73544982e-02 -2.30385497e-01 7.70283878e-01
1.45319462e+00 -2.91325420e-01 -1.22718060e+00 -7.53427222e-02
8.29379633e-03 -2.43934721e-01 2.29370117e-01 -9.40280497e-01
-1.08154809e+00 9.55645204e-01 1.08363605e+00 1.32379293e-01
1.32412410e+00 8.03770944e-02 2.78980583e-01 4.75685559e-02
-1.84140652e-01 -1.36585188e+00 3.98173690e-01 1.44321680e-01
1.15800238e+00 -1.29182839e+00 5.72263338e-02 -4.88698855e-02
-1.07598460e+00 1.01346707e+00 3.99250656e-01 -1.70306310e-01
2.09413186e-01 -5.62963542e-03 1.88703790e-01 -5.12205482e-01
-1.06146371e+00 -2.50327826e-01 5.47526121e-01 6.15057528e-01
6.43611073e-01 -1.62563119e-02 -3.98564935e-01 5.90219259e-01
5.70102446e-02 5.46982735e-02 3.74673396e-01 3.27339530e-01
-1.26413360e-01 -7.63564408e-01 -3.10023606e-01 4.56119746e-01
-9.78528976e-01 1.02581523e-01 -3.17961365e-01 6.87316537e-01
3.17228794e-01 7.46742845e-01 1.28648683e-01 6.08971603e-02
4.08683956e-01 2.99642354e-01 4.42588747e-01 -4.53519732e-01
-6.51459575e-01 4.02830303e-01 3.71691376e-01 -9.50448096e-01
-9.37512100e-01 -1.02278960e+00 -8.14707100e-01 2.33092412e-01
-6.41151890e-02 -3.07813615e-01 4.75438684e-01 1.07964885e+00
7.33617842e-02 5.53955019e-01 6.11456573e-01 -9.26879287e-01
-4.07520324e-01 -1.15015244e+00 -6.50905192e-01 7.12647855e-01
3.64002973e-01 -8.08006585e-01 -3.86744291e-01 2.11297702e-02] | [13.598389625549316, 1.9669139385223389] |
2d653c66-656a-47f6-9282-18c9bc330ca9 | calibrated-one-class-classification-for | 2207.12201 | null | https://arxiv.org/abs/2207.12201v1 | https://arxiv.org/pdf/2207.12201v1.pdf | Calibrated One-class Classification for Unsupervised Time Series Anomaly Detection | Unsupervised time series anomaly detection is instrumental in monitoring and alarming potential faults of target systems in various domains. Current state-of-the-art time series anomaly detectors mainly focus on devising advanced neural network structures and new reconstruction/prediction learning objectives to learn data normality (normal patterns and behaviors) as accurately as possible. However, these one-class learning methods can be deceived by unknown anomalies in the training data (i.e., anomaly contamination). Further, their normality learning also lacks knowledge about the anomalies of interest. Consequently, they often learn a biased, inaccurate normality boundary. This paper proposes a novel one-class learning approach, named calibrated one-class classification, to tackle this problem. Our one-class classifier is calibrated in two ways: (1) by adaptively penalizing uncertain predictions, which helps eliminate the impact of anomaly contamination while accentuating the predictions that the one-class model is confident in, and (2) by discriminating the normal samples from native anomaly examples that are generated to simulate genuine time series abnormal behaviors on the basis of original data. These two calibrations result in contamination-tolerant, anomaly-informed one-class learning, yielding a significantly improved normality modeling. Extensive experiments on six real-world datasets show that our model substantially outperforms twelve state-of-the-art competitors and obtains 6% - 31% F1 score improvement. The source code is available at \url{https://github.com/xuhongzuo/couta}. | ['Guansong Pang', 'Yongjun Wang', 'Qing Liao', 'Songlei Jian', 'Yijie Wang', 'Hongzuo Xu'] | 2022-07-25 | null | null | null | null | ['one-class-classifier', 'one-class-classification'] | ['methodology', 'miscellaneous'] | [ 1.90140590e-01 -8.22066516e-02 -8.01955238e-02 -4.10107940e-01
-6.66089296e-01 -2.80591339e-01 4.03705508e-01 3.39729339e-01
3.91243733e-02 3.92499924e-01 -2.39976615e-01 -6.33026242e-01
-3.08792014e-02 -7.28073657e-01 -7.40412593e-01 -9.01857734e-01
-3.02998155e-01 4.10589695e-01 8.84535015e-02 -1.28272578e-01
2.18939215e-01 4.76425081e-01 -1.61359954e+00 1.40647829e-01
1.23629284e+00 1.39306629e+00 -5.93589723e-01 4.23314750e-01
-2.34044835e-01 6.80448651e-01 -8.18555176e-01 -1.20347016e-01
3.88955861e-01 -4.33462054e-01 -9.86383483e-02 6.88012317e-02
1.82050094e-01 -3.06847066e-01 -2.23727658e-01 1.34896207e+00
4.90305945e-02 2.00005658e-02 5.03888011e-01 -1.64196920e+00
-4.57958251e-01 3.84299934e-01 -6.30580127e-01 5.40779471e-01
-7.05232844e-02 2.44842485e-01 8.63213658e-01 -8.30014288e-01
-1.09543830e-01 7.70944715e-01 6.83752954e-01 5.88691354e-01
-9.35312808e-01 -9.39203203e-01 4.75975871e-01 2.19055548e-01
-1.12362492e+00 -3.51974130e-01 1.01851499e+00 -3.88106197e-01
7.49934256e-01 4.20924067e-01 4.39768374e-01 1.35951078e+00
3.73990238e-01 7.56905079e-01 6.29338086e-01 -1.07972763e-01
4.94347572e-01 -1.42637715e-01 8.37959871e-02 9.68961045e-02
5.95030427e-01 1.28974780e-01 -1.89342290e-01 -5.20092607e-01
4.62364346e-01 4.88455385e-01 -2.80914366e-01 -2.52338141e-01
-9.75480437e-01 6.10405743e-01 2.06242234e-01 3.38518083e-01
-7.21361518e-01 -3.65361094e-01 7.02602506e-01 5.99817991e-01
7.33256996e-01 4.00051802e-01 -6.68301046e-01 -1.23348460e-01
-6.35994792e-01 1.40411362e-01 6.45084321e-01 6.23982072e-01
3.81818920e-01 9.28609610e-01 -2.00513024e-02 7.15155840e-01
1.52343482e-01 5.69849730e-01 8.97941053e-01 -2.38243297e-01
3.63884211e-01 7.45036006e-01 -1.91315398e-01 -8.31885397e-01
-1.97825715e-01 -8.79797757e-01 -1.13763797e+00 1.97624058e-01
4.01115716e-01 -1.20756932e-01 -1.11399090e+00 1.53770101e+00
3.17590714e-01 6.64179802e-01 2.10668206e-01 5.71315169e-01
1.26306176e-01 7.96077847e-01 -5.23223132e-02 -3.44712555e-01
6.78662062e-01 -5.54200947e-01 -6.88321173e-01 -2.27259815e-01
6.45890653e-01 -3.82744372e-01 9.68448400e-01 7.88018227e-01
-5.16517699e-01 -3.26036543e-01 -1.03114736e+00 9.10532415e-01
-3.63737702e-01 -8.00147206e-02 4.02926266e-01 5.61482012e-01
-4.44599390e-01 5.64701676e-01 -1.15813470e+00 -8.62212926e-02
5.04243433e-01 1.11065395e-01 -2.86527544e-01 8.71991739e-02
-1.10544837e+00 4.28638846e-01 4.88983363e-01 7.56707266e-02
-9.63691890e-01 -6.69520557e-01 -6.88930154e-01 1.63468793e-02
4.51867282e-01 1.34130359e-01 1.23335230e+00 -1.19596255e+00
-9.30476487e-01 4.23538953e-01 5.78951053e-02 -7.37971604e-01
6.49415255e-01 -3.89198005e-01 -1.12600327e+00 -2.54978567e-01
1.05453342e-01 -2.72168875e-01 1.20641541e+00 -1.05942106e+00
-8.03960085e-01 -4.66039807e-01 -7.16182888e-01 -2.36703515e-01
-6.31612599e-01 -2.69014508e-01 4.32872437e-02 -1.07831347e+00
5.11483908e-01 -5.63357174e-01 -2.37399027e-01 -8.56612623e-02
-6.64077282e-01 -1.62558407e-01 1.33698928e+00 -6.98320866e-01
1.32483947e+00 -2.25216055e+00 -6.50573969e-01 7.38037407e-01
2.51470834e-01 4.62914854e-01 8.72652531e-02 1.87080085e-01
-5.92170954e-01 -1.18565314e-01 -4.85853791e-01 3.45502980e-02
-1.95079610e-01 3.34385246e-01 -9.22791839e-01 6.74023390e-01
4.15223926e-01 4.08999264e-01 -8.19031537e-01 1.88529044e-01
1.72074109e-01 9.61001441e-02 -3.31325889e-01 4.48637217e-01
-2.14484543e-01 6.28376365e-01 -4.82177526e-01 1.14524937e+00
5.45034587e-01 1.12866595e-01 -1.21258020e-01 1.41039804e-01
1.04930401e-01 -3.67108136e-02 -1.15164578e+00 7.98119247e-01
5.76590858e-02 3.57312918e-01 -4.37459409e-01 -1.65343392e+00
1.26282871e+00 3.37602705e-01 6.92484558e-01 -6.98990226e-01
9.26368758e-02 5.54381549e-01 1.62012517e-01 -3.42464715e-01
6.04178868e-02 1.60614341e-01 -7.82713816e-02 3.66093516e-01
-1.24469444e-01 4.79967594e-01 -2.77727004e-02 -2.49994144e-01
1.28739989e+00 -1.62736863e-01 2.57569045e-01 -1.76372062e-02
5.62382877e-01 -2.56413072e-01 1.15032375e+00 7.93501794e-01
-4.20895457e-01 3.96814346e-01 6.98360622e-01 -6.99452162e-01
-1.02949131e+00 -1.36756074e+00 -1.59392163e-01 8.20089877e-01
-2.77053624e-01 4.54137512e-02 -3.01688105e-01 -8.78796518e-01
1.58048391e-01 1.12906086e+00 -5.18356025e-01 -6.96281254e-01
-5.78070760e-01 -8.07135999e-01 7.53215790e-01 7.07337737e-01
2.28758395e-01 -1.00868344e+00 -4.33802933e-01 5.83249964e-02
5.41712493e-02 -7.40169346e-01 -1.44971296e-01 4.19243008e-01
-1.09420240e+00 -1.19755113e+00 -4.36051250e-01 -3.17441672e-01
8.29964459e-01 -1.35531574e-01 9.10533130e-01 -5.18649779e-02
-1.02032259e-01 1.31701931e-01 -4.64821815e-01 -9.03152645e-01
-4.41735715e-01 -3.62147808e-01 6.41318440e-01 4.05985564e-01
7.94190049e-01 -7.61673212e-01 -2.70141065e-01 3.21348906e-01
-1.03106296e+00 -6.91626847e-01 6.31967306e-01 9.59499359e-01
8.68591726e-01 4.65261579e-01 1.07245088e+00 -1.04725611e+00
5.45720875e-01 -1.03390491e+00 -7.39869595e-01 5.01258001e-02
-1.01065707e+00 -1.95069134e-01 1.03846931e+00 -7.08331943e-01
-8.57031345e-01 -2.24463224e-01 -2.03537747e-01 -1.04922581e+00
-5.88830173e-01 4.34529811e-01 -3.09029788e-01 2.20840156e-01
7.78917253e-01 6.16391361e-01 1.00716874e-01 -4.83957142e-01
-2.42471278e-01 5.71076512e-01 8.82638752e-01 -5.76242387e-01
1.06859577e+00 2.64516234e-01 -2.41989285e-01 -7.75803804e-01
-7.29705334e-01 -5.71358263e-01 -4.08766270e-01 -2.18934357e-01
1.42908812e-01 -8.56388867e-01 -4.99378107e-02 9.77557242e-01
-5.94882965e-01 -1.13886958e-02 -6.11678600e-01 4.63795096e-01
-2.29866683e-01 3.99121344e-01 -3.71338487e-01 -1.06516552e+00
-4.22437489e-01 -6.42148733e-01 5.91288924e-01 1.49740562e-01
-1.59834698e-01 -9.43602443e-01 3.27242836e-02 -1.85222149e-01
5.69638312e-01 7.00901091e-01 9.57669139e-01 -1.73580015e+00
-1.34209454e-01 -8.99999797e-01 1.87634587e-01 8.12491417e-01
3.72043133e-01 2.10705355e-01 -1.06248927e+00 -5.06156802e-01
3.96961451e-01 -4.56687659e-02 5.16942739e-01 2.44529247e-01
1.55902076e+00 -5.84625065e-01 -1.40807658e-01 5.39703548e-01
1.11635399e+00 5.90375006e-01 6.77083910e-01 4.34584767e-01
7.19702125e-01 2.53235072e-01 7.62773752e-01 7.64510214e-01
-1.51176795e-01 9.99971777e-02 7.71706104e-01 6.81842864e-02
6.31560862e-01 -1.80177778e-01 4.46396858e-01 7.48232126e-01
2.75871485e-01 -2.58107990e-01 -1.18263197e+00 7.13525414e-01
-1.91173673e+00 -9.23750341e-01 -7.80733079e-02 2.45481634e+00
5.17046154e-01 5.94085813e-01 2.26947457e-01 6.78868532e-01
8.23056400e-01 2.42661908e-02 -1.15177345e+00 -1.17209405e-01
-2.74890363e-01 2.82182377e-02 1.72684088e-01 -2.24391669e-01
-1.39553821e+00 4.01470631e-01 4.72598410e+00 6.12131476e-01
-1.32601166e+00 -2.50779450e-01 8.42999518e-01 -1.17781097e-02
-1.68144144e-02 -2.66215205e-01 -4.30296898e-01 5.83259761e-01
1.13226521e+00 -3.36444080e-01 -8.69075060e-02 1.27005768e+00
1.01387210e-01 4.44768757e-01 -1.09707797e+00 8.42342019e-01
6.05674833e-03 -7.69414842e-01 2.28186965e-01 -3.44000980e-02
6.21881962e-01 1.42575160e-01 1.12865351e-01 5.41056097e-01
-1.87821805e-01 -7.47641683e-01 6.91763997e-01 5.66616356e-01
5.31350315e-01 -9.14841592e-01 1.09371495e+00 4.23568666e-01
-9.80437160e-01 -4.41386312e-01 -1.01182200e-01 6.46856474e-03
-2.01813146e-01 1.00680804e+00 -5.05071521e-01 5.30760527e-01
8.78199160e-01 8.68999004e-01 -4.75716770e-01 1.20330727e+00
-5.58765493e-02 1.17478311e+00 -2.83455253e-01 3.22767586e-01
1.62036896e-01 -9.14803147e-02 9.56796885e-01 7.19307423e-01
4.50695813e-01 -1.41394541e-01 2.73659140e-01 7.14297771e-01
1.62503272e-01 5.08630136e-03 -6.09482050e-01 -2.83691376e-01
4.62046802e-01 8.31941068e-01 -5.34929395e-01 -1.82160065e-01
-4.64563102e-01 5.48955560e-01 -4.36761267e-02 4.04982030e-01
-7.84962952e-01 -7.12660372e-01 8.30415189e-01 8.87673721e-02
8.91174152e-02 1.88865438e-01 -3.59961689e-01 -1.31568456e+00
4.46574688e-01 -1.05337131e+00 7.27723241e-01 -2.13822663e-01
-1.80328608e+00 7.12238610e-01 -1.93331629e-01 -1.76922882e+00
-4.79368508e-01 -6.86742961e-01 -1.12228560e+00 6.06525719e-01
-1.21030915e+00 -6.51265979e-01 -3.56721133e-01 6.34280741e-01
4.75894690e-01 -5.56738138e-01 6.95178270e-01 4.15105045e-01
-8.78732145e-01 8.05500746e-01 3.44387084e-01 4.53704774e-01
6.84831321e-01 -1.22759557e+00 6.04008436e-01 1.41787243e+00
-1.02228180e-01 2.06374183e-01 7.09267795e-01 -7.46102691e-01
-9.71021295e-01 -1.57422459e+00 4.01830792e-01 -2.15288401e-01
8.96900058e-01 -5.84202744e-02 -1.72976649e+00 7.16959417e-01
-6.09280109e-01 3.49741876e-01 7.49436200e-01 -9.32416245e-02
-5.19780099e-01 -2.58895487e-01 -1.25601435e+00 4.26699966e-01
6.21230781e-01 -3.64368528e-01 -5.95681965e-01 2.75837123e-01
3.47193182e-01 -3.76557857e-01 -5.74471116e-01 6.99239969e-01
1.98354527e-01 -9.77205157e-01 7.25729525e-01 -8.92055035e-01
1.35755584e-01 -2.70590723e-01 -2.72934198e-01 -1.28856254e+00
-9.85868648e-02 -2.99344778e-01 -8.05062234e-01 1.11619270e+00
2.40374595e-01 -1.28682041e+00 6.83361232e-01 3.64220113e-01
-5.24697602e-01 -1.04803073e+00 -9.03216183e-01 -1.10944450e+00
-1.19194612e-02 -7.64486253e-01 8.08171809e-01 1.29529321e+00
-3.14653039e-01 -3.32400858e-01 -2.77315438e-01 4.97985601e-01
7.88084924e-01 -1.89390749e-01 6.08184397e-01 -1.51907527e+00
4.74453084e-02 -5.54841638e-01 -6.98716164e-01 -3.92666101e-01
1.90187186e-01 -5.14149964e-01 6.98954612e-02 -7.29729235e-01
-4.78263557e-01 -3.91660661e-01 -8.53993714e-01 7.52640903e-01
-2.20807940e-01 -1.02176452e-02 -5.36274195e-01 4.10878420e-01
-3.68683517e-01 6.84210002e-01 4.68686730e-01 -5.99267483e-02
-2.26920485e-01 6.00213826e-01 -6.39723718e-01 8.55877638e-01
1.24830365e+00 -4.35890108e-01 -4.49016303e-01 2.64115119e-03
-2.17334256e-01 -2.02203482e-01 4.40688699e-01 -1.29516160e+00
1.24050342e-01 -2.63324559e-01 6.06793165e-01 -6.71642125e-01
-1.44833088e-01 -8.23763609e-01 -8.85824412e-02 5.68221271e-01
-1.05955757e-01 3.38245392e-01 3.21870446e-01 9.84781325e-01
-4.57504779e-01 1.06260262e-01 8.51832092e-01 1.38692170e-01
-7.96570420e-01 5.43112040e-01 -4.26567614e-01 2.13274941e-01
1.13135886e+00 -2.01169670e-01 -2.79898256e-01 -4.43357170e-01
-4.63787913e-01 3.25545847e-01 2.29118913e-01 7.14129806e-01
7.57721841e-01 -1.35068631e+00 -7.20751584e-01 8.24046552e-01
5.23222148e-01 5.94911501e-02 2.43108794e-01 7.27547824e-01
-2.48895109e-01 2.61596322e-01 -2.12024570e-01 -8.28076243e-01
-8.62060964e-01 4.36991870e-01 5.05243242e-01 -6.08103573e-02
-7.33058453e-01 6.35677457e-01 6.11409321e-02 -4.75291789e-01
4.53134686e-01 -2.01123700e-01 -1.70919560e-02 -3.15588474e-01
6.01045907e-01 4.34513330e-01 4.14803684e-01 -4.60841954e-01
-3.09819013e-01 9.53923017e-02 -3.73104930e-01 5.24442136e-01
1.17963266e+00 3.25438857e-01 8.93836189e-03 7.28979945e-01
7.93502092e-01 -2.06711784e-01 -1.20788729e+00 -4.13720757e-01
3.49848479e-01 -5.01736104e-01 -2.38587081e-01 -6.22189760e-01
-1.13934600e+00 6.29992902e-01 7.79499114e-01 4.77171749e-01
1.53751802e+00 -3.70888650e-01 7.95520067e-01 5.32027185e-01
1.12629339e-01 -8.76086056e-01 1.30361512e-01 5.26040018e-01
6.40894949e-01 -1.32981229e+00 -3.40834796e-01 1.25803456e-01
-6.15289211e-01 1.07150567e+00 1.15662539e+00 -2.42426604e-01
6.92203581e-01 1.60910413e-01 3.03135484e-01 -2.32667029e-02
-8.70765865e-01 3.58633667e-01 3.39859694e-01 6.50908351e-01
1.18153669e-01 6.98247105e-02 2.63943762e-01 9.33905721e-01
-3.53483111e-02 -5.30591726e-01 3.71051699e-01 9.89263415e-01
-4.64652210e-01 -6.98710203e-01 -5.15926659e-01 1.06331086e+00
-6.11555874e-01 2.18548968e-01 -2.13461459e-01 6.19187891e-01
-1.03588074e-01 7.68601656e-01 3.70201498e-01 -4.81939465e-01
5.00163019e-01 3.59187961e-01 -4.34913993e-01 -2.51131356e-01
-2.16873825e-01 6.19367845e-02 -3.16230595e-01 -5.73032022e-01
2.99168974e-01 -7.56893754e-01 -1.22047949e+00 -1.22287959e-01
-3.09998155e-01 3.74306172e-01 3.09998542e-01 9.15743589e-01
2.75751740e-01 7.99014986e-01 9.69352543e-01 -5.90688765e-01
-1.08410561e+00 -8.85882437e-01 -6.78258121e-01 5.29222608e-01
6.72006130e-01 -5.54570794e-01 -8.43626142e-01 -1.91359878e-01] | [7.572902679443359, 2.3890087604522705] |
75894d3c-9f54-4b73-acd2-553623668400 | zero-shot-sketch-image-hashing | 1803.02284 | null | http://arxiv.org/abs/1803.02284v1 | http://arxiv.org/pdf/1803.02284v1.pdf | Zero-Shot Sketch-Image Hashing | Recent studies show that large-scale sketch-based image retrieval (SBIR) can
be efficiently tackled by cross-modal binary representation learning methods,
where Hamming distance matching significantly speeds up the process of
similarity search. Providing training and test data subjected to a fixed set of
pre-defined categories, the cutting-edge SBIR and cross-modal hashing works
obtain acceptable retrieval performance. However, most of the existing methods
fail when the categories of query sketches have never been seen during
training. In this paper, the above problem is briefed as a novel but realistic
zero-shot SBIR hashing task. We elaborate the challenges of this special task
and accordingly propose a zero-shot sketch-image hashing (ZSIH) model. An
end-to-end three-network architecture is built, two of which are treated as the
binary encoders. The third network mitigates the sketch-image heterogeneity and
enhances the semantic relations among data by utilizing the Kronecker fusion
layer and graph convolution, respectively. As an important part of ZSIH, we
formulate a generative hashing scheme in reconstructing semantic knowledge
representations for zero-shot retrieval. To the best of our knowledge, ZSIH is
the first zero-shot hashing work suitable for SBIR and cross-modal search.
Comprehensive experiments are conducted on two extended datasets, i.e., Sketchy
and TU-Berlin with a novel zero-shot train-test split. The proposed model
remarkably outperforms related works. | ['Yuming Shen', 'Li Liu', 'Fumin Shen', 'Ling Shao'] | 2018-03-06 | zero-shot-sketch-image-hashing-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Shen_Zero-Shot_Sketch-Image_Hashing_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Shen_Zero-Shot_Sketch-Image_Hashing_CVPR_2018_paper.pdf | cvpr-2018-6 | ['sketch-based-image-retrieval'] | ['computer-vision'] | [-2.43283948e-03 -4.43516642e-01 -5.01759648e-01 -2.64794558e-01
-1.16480970e+00 -3.54373842e-01 7.56807327e-01 -7.32162893e-02
-3.23749095e-01 3.35639983e-01 -9.69337821e-02 3.14502977e-02
-4.90960687e-01 -1.00363684e+00 -6.76484108e-01 -6.64264977e-01
1.14272512e-01 5.59831262e-01 2.51948774e-01 -3.02575350e-01
4.52622712e-01 2.91298896e-01 -1.82619798e+00 2.75615126e-01
2.37340763e-01 1.18371415e+00 2.66695857e-01 6.02424204e-01
-1.07459821e-01 5.33216953e-01 -3.81739616e-01 -7.28178144e-01
2.33867750e-01 -7.29287639e-02 -5.47747374e-01 -2.67007321e-01
6.22924507e-01 -6.79560959e-01 -1.02453148e+00 8.12003613e-01
9.09156203e-01 3.39117020e-01 6.68010890e-01 -1.61376023e+00
-1.39551628e+00 4.62531149e-01 -2.43213937e-01 -1.19333282e-01
4.67624009e-01 -1.65016636e-01 1.39511323e+00 -1.38033211e+00
6.96413517e-01 1.31643343e+00 6.28013730e-01 5.29082656e-01
-9.29101646e-01 -8.54409158e-01 -3.22472870e-01 5.32307088e-01
-2.11046338e+00 -2.73883939e-01 1.11480987e+00 -8.30548480e-02
7.07949340e-01 2.58664917e-02 3.81216854e-01 1.09729183e+00
-1.27108172e-01 1.02416158e+00 5.33966601e-01 -1.60734922e-01
3.11898977e-01 9.29811299e-02 4.34449464e-02 8.53558302e-01
7.42137060e-02 7.83542469e-02 -7.67457604e-01 -4.12891358e-01
7.08742142e-01 6.62258744e-01 -8.81514475e-02 -6.94261730e-01
-1.08488786e+00 1.16029072e+00 7.61572480e-01 4.49647039e-01
-1.37753729e-02 4.22303140e-01 5.02722621e-01 4.67910409e-01
3.42364497e-02 -5.34123741e-02 2.40366563e-01 3.09464246e-01
-1.22677588e+00 2.34564081e-01 7.06457078e-01 1.18701029e+00
9.62046504e-01 -2.46324003e-01 -3.25425357e-01 1.11869025e+00
3.35879624e-01 6.73459947e-01 5.72523415e-01 -5.24038494e-01
1.92576334e-01 3.12627882e-01 -2.64203727e-01 -1.24710023e+00
3.48180592e-01 -5.75411990e-02 -9.23489034e-01 -3.25883508e-01
-1.84418604e-01 7.47007310e-01 -9.90867138e-01 1.58244658e+00
4.57503833e-02 3.78754973e-01 2.95371488e-02 1.06149542e+00
1.27263319e+00 7.56098509e-01 -5.52570708e-02 2.43473232e-01
1.38242722e+00 -9.76258576e-01 -6.42918825e-01 2.54206210e-01
1.16705813e-01 -7.76120842e-01 1.05243981e+00 2.10544970e-02
-1.10498238e+00 -7.69057333e-01 -1.26468182e+00 -4.80422467e-01
-9.28805828e-01 -1.16901482e-02 6.68149650e-01 4.47110236e-01
-1.22022486e+00 5.32471001e-01 -3.82568300e-01 -3.19922894e-01
3.51390839e-01 3.02085966e-01 -5.04550934e-01 -5.44489145e-01
-1.53810608e+00 5.06230175e-01 2.27553874e-01 -1.62292212e-01
-1.09229279e+00 -6.09282851e-01 -1.08016801e+00 2.70234764e-01
1.57159671e-01 -5.76958716e-01 9.41011727e-01 -2.02545971e-01
-1.01452613e+00 9.36801612e-01 4.41821702e-02 -2.39234775e-01
3.48740593e-02 2.83549875e-01 -3.56878728e-01 4.80661035e-01
2.01283157e-01 9.52527642e-01 1.13253260e+00 -1.23965418e+00
-2.37722665e-01 -4.42987144e-01 4.90527041e-02 -3.62816714e-02
-6.23937309e-01 -2.60252327e-01 -9.49559867e-01 -8.05609763e-01
-3.27803604e-02 -8.21029127e-01 2.84391165e-01 4.24310118e-01
-2.61150990e-02 -4.52726603e-01 1.11815357e+00 -3.67076963e-01
1.40829194e+00 -2.37015152e+00 1.56275854e-01 1.51953384e-01
1.05030060e-01 2.71370232e-01 -3.26872945e-01 8.76507819e-01
1.76333427e-01 -8.02287646e-03 -1.97111964e-01 -6.04172349e-01
5.44080377e-01 4.18589383e-01 -7.36131966e-01 4.16794062e-01
1.15912884e-01 1.25500131e+00 -8.62579226e-01 -9.11363959e-01
2.99647242e-01 9.59433436e-01 -4.42024291e-01 3.41040671e-01
1.61498100e-01 -3.82085443e-01 -1.86310962e-01 1.12364101e+00
6.39566123e-01 -4.64108467e-01 -1.27046674e-01 -3.11475039e-01
2.76361525e-01 -9.60725993e-02 -1.01302910e+00 2.31019425e+00
-4.31810915e-01 3.34524721e-01 -3.03342253e-01 -9.37699914e-01
1.05329263e+00 3.14640015e-01 3.37707520e-01 -9.45452869e-01
-1.43706918e-01 3.64090711e-01 -6.59869432e-01 -5.75948656e-02
7.36526608e-01 -2.85286725e-01 -4.50000376e-01 4.65518206e-01
5.14226019e-01 -1.54406384e-01 1.25183254e-01 5.57396829e-01
8.70900273e-01 -2.39410818e-01 1.18583448e-01 1.49099603e-01
5.09811580e-01 -3.42309296e-01 4.80169579e-02 8.72511268e-01
-1.32457554e-01 8.20393205e-01 1.51207624e-02 -5.39416969e-01
-1.17489755e+00 -1.31596112e+00 -1.97310343e-01 1.11622334e+00
6.57414556e-01 -5.30888379e-01 -1.59659594e-01 -4.75022763e-01
4.01350945e-01 2.50594109e-01 -6.69016540e-01 -5.91179073e-01
-1.59465194e-01 -2.13473320e-01 8.52672637e-01 3.97465646e-01
5.29973090e-01 -1.11828816e+00 -3.36096734e-01 -1.96940135e-02
-9.29006487e-02 -8.19358468e-01 -7.47977555e-01 -1.28222048e-01
-4.31702137e-01 -8.77526641e-01 -1.09734249e+00 -1.10975850e+00
2.86178857e-01 8.83987129e-01 9.96704698e-01 2.89091706e-01
-8.47968400e-01 7.52058387e-01 -4.72993970e-01 2.37586096e-01
6.90424368e-02 2.65265778e-02 -1.76959425e-01 -2.69717760e-02
4.60247338e-01 -6.28080308e-01 -9.18946147e-01 4.61245060e-01
-1.25867033e+00 -4.64857191e-01 7.73834646e-01 1.36057353e+00
6.89696014e-01 -8.56943801e-02 5.19086957e-01 -2.09761307e-01
5.70259750e-01 -5.81527591e-01 -2.64748782e-01 6.57131374e-01
-6.22910321e-01 2.07186177e-01 4.82388437e-01 -5.78765273e-01
-5.83686769e-01 -1.43828496e-01 1.01988169e-03 -1.16473603e+00
2.17131943e-01 2.83115000e-01 6.93087354e-02 -4.93574142e-01
1.87601283e-01 6.61819458e-01 2.55086999e-02 -3.29004198e-01
5.21740079e-01 7.73642361e-01 5.37860572e-01 -6.06704175e-01
9.54987705e-01 5.58192492e-01 -3.67515795e-02 -7.36924469e-01
-5.44916391e-01 -8.59206140e-01 -3.65398794e-01 3.35971229e-02
6.51487648e-01 -1.01825368e+00 -7.81366229e-01 2.15628177e-01
-1.02101183e+00 1.35172278e-01 -2.09572583e-01 1.02151893e-01
-6.22978151e-01 6.71802640e-01 -6.35140657e-01 -8.68353069e-01
-7.58486092e-01 -1.13389599e+00 1.80011344e+00 3.98345925e-02
3.31446499e-01 -8.17944705e-01 2.44343996e-01 1.68973163e-01
5.19546986e-01 -2.56131083e-01 8.68716240e-01 -5.44909000e-01
-7.77113497e-01 -4.51128185e-01 -4.96271610e-01 5.43232597e-02
-1.83128446e-01 -2.65653342e-01 -1.03506839e+00 -6.42189205e-01
-4.61585373e-01 -8.57524872e-01 1.11010778e+00 -1.64334297e-01
1.22581279e+00 -1.75591946e-01 -2.21191511e-01 6.20873928e-01
1.74964547e+00 8.32965970e-02 8.14896166e-01 5.55447210e-03
5.01008868e-01 1.03623316e-01 6.80730402e-01 6.57153308e-01
6.16108775e-01 8.93233299e-01 2.98897833e-01 4.30533811e-02
-4.12860245e-01 -7.62472570e-01 1.02717303e-01 8.73514473e-01
4.40079778e-01 -2.11735517e-01 -5.82567155e-01 7.80020237e-01
-1.75055242e+00 -1.22332013e+00 6.64453328e-01 2.25231266e+00
6.96361840e-01 -2.94317812e-01 -1.17748313e-01 2.65087664e-01
7.58912981e-01 4.82108712e-01 -3.84399712e-01 -8.26508626e-02
-3.97712290e-02 4.96813864e-01 1.52090624e-01 3.33004534e-01
-1.18240833e+00 9.21496928e-01 5.14495611e+00 1.28200626e+00
-1.01330900e+00 2.08730847e-01 1.87274024e-01 3.01931854e-02
-3.67083639e-01 4.25712094e-02 -7.05752909e-01 5.20082414e-01
5.71979105e-01 -8.68694410e-02 7.01694310e-01 9.96061802e-01
-8.94167006e-01 3.47268492e-01 -1.06694913e+00 1.54528391e+00
6.35045111e-01 -1.26184106e+00 4.98925954e-01 -1.66050196e-01
3.35617840e-01 -2.08219200e-01 3.17472249e-01 7.02929854e-01
-1.17327981e-01 -9.30544376e-01 5.87600112e-01 5.36503732e-01
1.42027020e+00 -8.23288679e-01 4.94076788e-01 -6.03490695e-02
-1.70514035e+00 -1.37134984e-01 -7.18619049e-01 5.24747133e-01
-8.21685977e-03 7.42006600e-02 -5.64707935e-01 6.68956339e-01
5.52914083e-01 7.69677222e-01 -5.04025996e-01 9.01670933e-01
1.82397798e-01 -1.33917212e-01 -1.43672153e-01 -3.07269879e-02
3.84408504e-01 1.22157842e-01 2.10745066e-01 1.15116167e+00
4.53329474e-01 3.91007848e-02 6.33966327e-02 8.73639584e-01
-1.35423556e-01 -1.07875571e-01 -1.01225102e+00 -2.17121378e-01
6.29029870e-01 1.28852522e+00 -3.74756068e-01 -4.75826591e-01
-4.42908823e-01 1.33119047e+00 3.18876117e-01 2.00305507e-01
-8.88329148e-01 -8.79012525e-01 5.60444295e-01 -2.13790148e-01
6.94280565e-01 -5.94270490e-02 2.72793829e-01 -1.32164681e+00
5.40824933e-03 -4.98868883e-01 6.39195204e-01 -8.25183630e-01
-1.58126271e+00 4.74587798e-01 -7.84341469e-02 -1.25338304e+00
-1.51887640e-01 -4.33530420e-01 -3.24579149e-01 4.95756805e-01
-1.86750269e+00 -1.50631034e+00 -4.64605838e-01 9.86977518e-01
5.52365482e-01 -2.82712340e-01 1.07951784e+00 8.71306002e-01
-1.77440315e-01 1.11958528e+00 5.90384156e-02 1.59768105e-01
8.42572093e-01 -9.03137386e-01 2.95958996e-01 3.11894566e-01
4.15740818e-01 9.00604069e-01 1.41029939e-01 -5.68943024e-01
-2.04731393e+00 -9.66393113e-01 9.67991710e-01 7.25531429e-02
7.92836547e-01 -5.77016592e-01 -9.39651787e-01 1.94299847e-01
-9.32433754e-02 4.47898656e-01 6.44957423e-01 -3.33212644e-01
-1.01846588e+00 -2.36334071e-01 -1.11945605e+00 3.11047643e-01
1.14865637e+00 -1.35579574e+00 -6.35879874e-01 2.88272321e-01
7.55077243e-01 -3.40397619e-02 -1.06968033e+00 3.36002529e-01
9.94804561e-01 -7.36000419e-01 1.62168956e+00 -3.89402151e-01
4.04695511e-01 -2.48241425e-01 -6.68293357e-01 -6.86981022e-01
-2.81129062e-01 -2.68986344e-01 -2.51463145e-01 1.08556926e+00
-1.80830508e-01 -2.40223065e-01 5.46240330e-01 1.69454023e-01
1.38669804e-01 -6.95015132e-01 -1.00377214e+00 -8.49545121e-01
-2.68002659e-01 -6.10954016e-02 6.21240914e-01 8.63556921e-01
4.64005070e-03 2.09190145e-01 -6.66569710e-01 -1.86068825e-02
9.43561673e-01 5.72571039e-01 6.36496305e-01 -1.10232353e+00
-2.68062085e-01 -3.61258209e-01 -8.07719767e-01 -9.58220840e-01
3.16511482e-01 -1.12693739e+00 3.47234029e-03 -1.25574434e+00
5.29240668e-01 -5.32163858e-01 -4.90714192e-01 4.85402167e-01
-4.25331928e-02 6.90815151e-01 3.37544680e-01 4.55971032e-01
-1.04939711e+00 9.56538200e-01 8.86748850e-01 -5.53552508e-01
3.90641928e-01 -4.14117247e-01 -3.25553119e-01 -1.74374357e-01
4.50196147e-01 -2.06804395e-01 -6.66576684e-01 -2.80820042e-01
1.05860919e-01 2.67650276e-01 8.10061634e-01 -7.73132205e-01
7.59912074e-01 2.90497750e-01 1.61350802e-01 -8.73414338e-01
7.71358430e-01 -8.90527546e-01 1.70561418e-01 3.49404305e-01
-4.39569026e-01 -5.22114225e-02 -1.75442263e-01 8.12780738e-01
-5.94979465e-01 -1.27081543e-01 7.05765188e-01 -1.02044486e-01
-9.14622843e-01 7.38150597e-01 2.97850251e-01 -1.24566756e-01
8.60607028e-01 -1.78011760e-01 -4.51605707e-01 -4.69438404e-01
-4.11665827e-01 8.54734033e-02 4.97618377e-01 5.28020740e-01
1.27302575e+00 -1.84726346e+00 -3.61877173e-01 4.30948257e-01
6.07541978e-01 -3.36035758e-01 5.06539404e-01 4.33056384e-01
-7.89835379e-02 7.19597459e-01 -3.34581077e-01 -4.10850137e-01
-1.02675104e+00 1.16417253e+00 -8.79118145e-02 -1.30088359e-01
-5.21082163e-01 8.76212001e-01 1.23806469e-01 -3.56022209e-01
5.90794861e-01 3.42979491e-01 1.71658859e-01 1.03138767e-01
6.89128876e-01 2.66107529e-01 3.76779772e-02 -6.13141596e-01
-3.86190951e-01 8.34937394e-01 -1.50191635e-01 -4.19468619e-02
1.21510422e+00 -2.87147257e-02 -1.25218809e-01 4.00489241e-01
1.81132066e+00 -6.20918989e-01 -8.02738905e-01 -5.64803898e-01
-2.48466805e-01 -7.01897740e-01 -4.63050120e-02 -3.91078502e-01
-1.09842622e+00 1.05821431e+00 6.22169733e-01 -1.76878721e-02
9.57304895e-01 2.04212502e-01 1.28043473e+00 6.56457603e-01
7.31594861e-01 -7.73434401e-01 4.60531890e-01 2.15393201e-01
9.23068941e-01 -1.33929133e+00 -1.71945691e-01 5.50180748e-02
-4.67037529e-01 1.01051497e+00 2.98343718e-01 -3.06974053e-01
6.00888491e-01 7.18855113e-03 -5.51929772e-01 -3.52616072e-01
-7.59697616e-01 -2.64889926e-01 4.91362333e-01 3.40034455e-01
-1.13720223e-01 -1.02625087e-01 -2.03915238e-01 5.12021720e-01
1.70859203e-01 2.48796027e-02 -2.33478054e-01 1.13888097e+00
-3.50782484e-01 -1.03164113e+00 -2.24746287e-01 1.55810788e-01
-6.17969595e-03 -3.50799203e-01 -3.90494525e-01 6.58271194e-01
-1.26393214e-01 5.34264326e-01 7.37808794e-02 -6.00523233e-01
1.17636338e-01 1.55660376e-01 4.40425128e-01 -2.56941170e-01
-3.59567583e-01 -2.84301132e-01 -5.41205764e-01 -5.83680987e-01
-1.81726560e-01 -1.09206818e-01 -9.93203223e-01 -3.95235747e-01
-3.29363197e-01 2.26767719e-01 8.13780308e-01 3.72680515e-01
5.25814593e-01 -1.75462347e-02 8.02577138e-01 -9.15149629e-01
-7.69513369e-01 -5.42945087e-01 -7.35450268e-01 5.60438693e-01
3.45498025e-01 -1.05915630e+00 -4.53753710e-01 -2.75980055e-01] | [11.55553913116455, 0.7607250213623047] |
9ab57deb-1058-4880-8f97-f8fc70c8def3 | recolornerf-layer-decomposed-radiance-field | 2301.07958 | null | https://arxiv.org/abs/2301.07958v2 | https://arxiv.org/pdf/2301.07958v2.pdf | RecolorNeRF: Layer Decomposed Radiance Fields for Efficient Color Editing of 3D Scenes | Radiance fields have gradually become a main representation of media. Although its appearance editing has been studied, how to achieve view-consistent recoloring in an efficient manner is still under explored. We present RecolorNeRF, a novel user-friendly color editing approach for the neural radiance fields. Our key idea is to decompose the scene into a set of pure-colored layers, forming a palette. By this means, color manipulation can be conducted by altering the color components of the palette directly. To support efficient palette-based editing, the color of each layer needs to be as representative as possible. In the end, the problem is formulated as an optimization problem, where the layers and their blending weights are jointly optimized with the NeRF itself. Extensive experiments show that our jointly-optimized layer decomposition can be used against multiple backbones and produce photo-realistic recolored novel-view renderings. We demonstrate that RecolorNeRF outperforms baseline methods both quantitatively and qualitatively for color editing even in complex real-world scenes. | ['Qi Dou', 'Xiaoguang Han', 'Yuehao Wang', 'Bingchen Gong'] | 2023-01-19 | null | null | null | null | ['color-manipulation'] | ['computer-vision'] | [ 4.14851993e-01 -2.63607681e-01 2.79453188e-01 -4.29990351e-01
-3.38174999e-01 -6.51810527e-01 4.29777473e-01 -1.96545154e-01
-2.44377121e-01 4.70427901e-01 7.47246742e-02 -9.74904522e-02
2.58799672e-01 -8.33969951e-01 -9.16344166e-01 -7.39178360e-01
4.03943777e-01 -1.44236200e-02 7.47064352e-02 -2.43196458e-01
1.46499664e-01 7.50456989e-01 -1.50211358e+00 4.09304172e-01
1.13055372e+00 7.00047791e-01 3.48406225e-01 8.52406800e-01
-2.11526617e-01 6.83212876e-01 -5.17673075e-01 -2.73717046e-01
4.99361426e-01 -4.46538359e-01 -3.85315657e-01 3.10973912e-01
9.32731509e-01 -4.34560806e-01 -7.70612583e-02 1.13241506e+00
3.86023462e-01 4.33204442e-01 6.17289662e-01 -1.01075137e+00
-8.49747419e-01 3.48985702e-01 -1.18119347e+00 -3.21978331e-01
1.03422634e-01 -1.66090876e-01 8.58345628e-01 -1.02727044e+00
7.09669948e-01 1.34137356e+00 3.36598516e-01 4.92603540e-01
-1.51402307e+00 -6.97625935e-01 6.98677659e-01 -9.77317318e-02
-1.26035404e+00 -2.33914018e-01 1.06682169e+00 -2.25943461e-01
5.49739957e-01 5.76842964e-01 1.00825298e+00 7.11129665e-01
7.73571059e-02 6.67528808e-01 1.28295493e+00 -5.05346954e-01
3.06242198e-01 8.24902803e-02 -1.32542580e-01 8.47622991e-01
2.02851012e-01 -1.75690129e-01 -6.21711493e-01 -6.42624423e-02
9.29286301e-01 -3.59262004e-02 -5.65196991e-01 -6.49629712e-01
-1.12373102e+00 4.10866469e-01 7.62452781e-01 -2.13176757e-01
-3.10431629e-01 5.34406245e-01 -8.33206773e-02 6.30041361e-02
8.58035445e-01 4.89489615e-01 -2.91515857e-01 3.07447910e-01
-8.19982052e-01 2.96879828e-01 4.33175921e-01 8.61592114e-01
9.31317329e-01 1.87670261e-01 -2.47092158e-01 1.25393617e+00
1.54706180e-01 4.60283428e-01 -3.35938364e-01 -1.35069108e+00
2.91076660e-01 5.20203948e-01 3.11000437e-01 -9.93468642e-01
-2.03748062e-01 -2.39665404e-01 -1.00228691e+00 9.15800393e-01
2.99227655e-01 -2.34503850e-01 -1.04101193e+00 1.79651427e+00
5.06127715e-01 1.77940086e-01 -2.66950577e-01 9.72709239e-01
5.53776026e-01 1.05269372e+00 1.64170936e-02 2.29547381e-01
1.31748962e+00 -1.13309991e+00 -5.12775302e-01 -2.06944272e-01
3.78957316e-02 -8.26628566e-01 1.28326392e+00 5.85087478e-01
-1.25036824e+00 -3.17126513e-01 -1.12508523e+00 -4.75315154e-01
-3.26831281e-01 1.79418966e-01 8.44738781e-01 5.40043652e-01
-1.08311522e+00 3.55502248e-01 -4.90638584e-01 2.46115681e-02
3.74754697e-01 -8.03209171e-02 -2.54246503e-01 -3.32251281e-01
-8.51949930e-01 5.80976188e-01 1.53008655e-01 2.81219184e-01
-6.32368147e-01 -1.01149666e+00 -7.31884122e-01 1.02302507e-01
2.78286785e-01 -9.24970567e-01 9.75194871e-01 -1.12438416e+00
-1.66660905e+00 9.02325451e-01 -1.82189360e-01 2.41725087e-01
6.52476132e-01 -1.69927239e-01 -2.20055550e-01 -3.04679666e-02
-3.36771935e-01 8.00520241e-01 1.03145969e+00 -1.91131437e+00
-6.87934995e-01 -1.12925977e-01 4.11490351e-01 5.63500524e-01
-4.06407565e-01 -2.30069146e-01 -1.18552971e+00 -9.34100628e-01
1.61371410e-01 -9.34724033e-01 -2.76222855e-01 7.37771809e-01
-5.50164521e-01 3.06518346e-01 5.43306053e-01 -6.73869729e-01
1.08019245e+00 -2.17323613e+00 2.75541425e-01 4.19636995e-01
5.63899696e-01 -2.64724083e-02 -3.24495971e-01 2.13012815e-01
-1.52613223e-01 -6.66617528e-02 -3.90413433e-01 -5.74628353e-01
-6.11139908e-02 1.48099661e-02 -2.21333742e-01 4.11725312e-01
-1.99072771e-02 6.30418301e-01 -9.69883978e-01 -1.94465995e-01
3.39298964e-01 9.63182390e-01 -8.63169968e-01 1.71196997e-01
-3.83684129e-01 3.04976046e-01 -1.82621866e-01 4.04049546e-01
1.15126252e+00 -1.25335068e-01 1.08102337e-01 -6.12099111e-01
-2.90172547e-01 -1.04771383e-01 -1.35756135e+00 1.79861879e+00
-6.34398818e-01 8.68425965e-01 3.01964730e-01 -2.24815160e-01
7.46248186e-01 -3.36852312e-01 3.60393226e-01 -8.07633519e-01
-1.62707850e-01 -1.94137305e-01 -4.44008946e-01 6.32703826e-02
9.77901101e-01 2.22765848e-01 1.99619979e-01 6.03408158e-01
-6.15164876e-01 -2.04448268e-01 1.62029594e-01 3.08025777e-01
5.63632488e-01 4.51226532e-01 -1.53822929e-01 -1.62848189e-01
1.50696278e-01 -2.11636081e-01 5.40574610e-01 6.55788541e-01
3.86088401e-01 1.04982078e+00 3.93507838e-01 -2.87411153e-01
-1.05338240e+00 -1.12491930e+00 2.04686392e-02 1.23484874e+00
3.48524213e-01 -3.38997990e-01 -9.78637218e-01 -3.34224999e-01
-5.29330596e-02 7.19634175e-01 -8.10357630e-01 6.41925484e-02
-5.66877425e-01 -8.44592392e-01 3.34308669e-02 3.08395147e-01
5.03171384e-01 -7.38270938e-01 -7.47846782e-01 4.68613431e-02
-6.01629261e-03 -7.63104975e-01 -7.96180367e-01 9.54094008e-02
-5.28329909e-01 -8.95966291e-01 -1.24235320e+00 -5.80027103e-01
1.11196208e+00 8.87816131e-01 1.22300053e+00 1.68725953e-01
-3.14107686e-01 3.00872445e-01 -1.12564825e-01 -2.45046899e-01
-2.48558745e-01 -2.57275462e-01 -4.83137101e-01 3.13271463e-01
-4.66881990e-01 -6.38421595e-01 -1.02593136e+00 1.02284424e-01
-1.18267143e+00 8.30549896e-01 3.12607110e-01 4.07070041e-01
8.39870691e-01 -1.89002454e-01 -9.90906805e-02 -1.45781362e+00
5.60578525e-01 6.44416958e-02 -8.28445554e-01 6.54054880e-01
-3.95503521e-01 1.68067619e-01 6.82037234e-01 -3.83972168e-01
-1.50220966e+00 -8.06243625e-03 2.65821777e-02 -3.84668708e-01
1.36571169e-01 9.76310596e-02 -3.11923653e-01 -3.63488317e-01
4.16493475e-01 -2.51093637e-02 -4.60209310e-01 -6.10590637e-01
9.05469358e-01 2.86586106e-01 5.67098856e-01 -7.07826853e-01
9.43567216e-01 7.02685833e-01 -2.42922325e-02 -7.30582058e-01
-6.41479075e-01 -9.49042886e-02 -3.24989498e-01 -5.60255468e-01
7.39835501e-01 -9.58528638e-01 -5.66959143e-01 7.68303573e-01
-1.20686948e+00 -6.75309539e-01 -1.94944531e-01 -7.60605633e-02
-2.66381592e-01 2.06194609e-01 -4.36879873e-01 -6.25971735e-01
-9.05593187e-02 -1.01765740e+00 1.01707244e+00 4.35328782e-01
1.87273428e-01 -7.67247558e-01 1.83186725e-01 -2.90879654e-03
3.40220898e-01 5.04824579e-01 1.20008922e+00 5.90764165e-01
-7.95099914e-01 2.14606389e-01 -6.95426226e-01 2.30250210e-01
2.67667562e-01 5.10840118e-01 -1.13635767e+00 -2.76297837e-01
-6.59051716e-01 -7.17716338e-03 1.05321348e+00 3.23566258e-01
1.60105526e+00 -8.66784900e-02 -9.96563733e-02 1.18295574e+00
1.51285744e+00 1.89368740e-01 7.72357702e-01 2.49475762e-01
1.13313699e+00 4.84278768e-01 1.98663443e-01 4.42110032e-01
4.07781243e-01 6.71051383e-01 3.98826599e-01 -7.19573498e-01
-5.02323925e-01 -3.23600203e-01 8.71770009e-02 6.02875292e-01
-1.93341330e-01 -4.68341291e-01 -4.51715946e-01 5.15343342e-03
-1.60370493e+00 -8.18963110e-01 -2.60730330e-02 2.17101240e+00
8.09481263e-01 -2.05015451e-01 -1.60951510e-01 -2.18654573e-01
6.83875740e-01 5.86670399e-01 -7.54636824e-01 -4.31127161e-01
-2.93845206e-01 -5.45492396e-03 5.22434711e-01 5.61563253e-01
-8.60070944e-01 8.22203636e-01 6.35637331e+00 5.50380230e-01
-1.30853617e+00 -1.35983229e-01 9.16626096e-01 -2.41115883e-01
-1.02857471e+00 -7.01544806e-02 -4.04819280e-01 2.54421085e-01
-3.72063108e-02 2.10386217e-01 9.74817872e-01 4.79753822e-01
2.89926678e-01 -3.01126212e-01 -7.40653574e-01 1.17688429e+00
1.09414093e-01 -1.43804717e+00 3.21101129e-01 -1.40191227e-01
1.18454742e+00 -3.24641854e-01 3.27583700e-01 -4.12234366e-02
4.44596738e-01 -6.94023013e-01 1.04651320e+00 7.54047632e-01
1.09782302e+00 -8.74050021e-01 -3.97718936e-01 -1.27667561e-01
-1.31588447e+00 2.35359862e-01 -4.51664209e-01 4.00319159e-01
1.96171388e-01 7.90399790e-01 -1.97959900e-01 4.54769343e-01
7.07501173e-01 6.64129317e-01 -5.82733214e-01 1.11612880e+00
-2.19143197e-01 2.26180926e-01 -2.60415345e-01 2.17389420e-01
7.86586255e-02 -6.25153363e-01 3.14483792e-01 1.27654755e+00
2.11101204e-01 1.67771176e-01 4.93639242e-03 1.12633824e+00
-3.70064348e-01 -1.89517494e-02 -1.86030537e-01 1.40838176e-01
3.26766551e-01 1.49853075e+00 -7.39146948e-01 -1.46324441e-01
-3.99471790e-01 1.52560997e+00 6.59789741e-01 9.11743224e-01
-9.17538643e-01 -4.75335747e-01 9.26676333e-01 6.18539155e-02
1.87438384e-01 -1.89617679e-01 -2.33110696e-01 -1.21295881e+00
1.46663806e-03 -7.57803857e-01 1.88194178e-02 -1.35577714e+00
-1.04600132e+00 7.22878635e-01 -2.26558205e-02 -1.16852832e+00
3.73091221e-01 -6.89278960e-01 -6.75448060e-01 1.00634646e+00
-1.89486432e+00 -1.13308120e+00 -6.62837148e-01 5.62509596e-01
2.93240905e-01 3.32310736e-01 6.01379514e-01 4.03673679e-01
-7.46773899e-01 4.84413713e-01 3.66120964e-01 -1.29733056e-01
8.22979629e-01 -1.35055745e+00 5.48029184e-01 1.01947045e+00
1.60096809e-01 5.23828924e-01 5.72718561e-01 -3.85806143e-01
-1.41627169e+00 -1.14642882e+00 1.65886894e-01 9.66089368e-02
8.42945501e-02 -6.20397270e-01 -9.37555552e-01 2.80947894e-01
3.32392484e-01 7.46677145e-02 3.61736059e-01 -5.83971404e-02
-7.25950420e-01 -3.89484972e-01 -8.46604526e-01 1.08506787e+00
1.17219067e+00 -4.58740592e-01 1.79726154e-01 1.93440527e-01
5.61273634e-01 -7.19385743e-01 -4.29319739e-01 4.86412235e-02
6.56591594e-01 -1.16692030e+00 1.16396844e+00 -3.61108154e-01
3.27457368e-01 -6.79844856e-01 -3.25209945e-02 -1.65485024e+00
-4.54812795e-01 -7.53225625e-01 3.99366990e-02 1.14777458e+00
3.59140635e-01 -4.31183219e-01 5.92881322e-01 8.19234908e-01
-1.07823588e-01 -5.79246521e-01 -3.45764875e-01 -1.75097659e-01
-1.19712174e-01 -3.90318811e-01 7.86433399e-01 8.75649452e-01
-6.80808604e-01 4.24140040e-03 -6.06195748e-01 1.35335281e-01
8.02788913e-01 3.69549632e-01 8.17810059e-01 -1.02298093e+00
-3.39876801e-01 -5.25145888e-01 2.61702240e-01 -1.20591235e+00
-9.02847424e-02 -6.70522988e-01 2.18231827e-01 -1.89895129e+00
3.43882829e-01 -7.74476707e-01 -2.80469924e-01 4.07382101e-01
-4.19728070e-01 4.39663440e-01 5.27573109e-01 -1.47580877e-01
-3.87231857e-01 5.43899596e-01 1.79722512e+00 -3.33670139e-01
-2.21786410e-01 -3.60669345e-01 -9.17578280e-01 6.66109622e-01
6.63897991e-01 -2.34179512e-01 -6.02284968e-01 -8.74351203e-01
6.81537688e-01 -2.70882726e-01 2.53844202e-01 -8.05470884e-01
7.74260089e-02 -4.82899278e-01 6.27906084e-01 -5.40322900e-01
5.31925857e-01 -8.39630604e-01 5.13584375e-01 -8.73608962e-02
-3.43652934e-01 2.04369202e-01 1.80863753e-01 5.69183111e-01
1.54329404e-01 6.10292256e-02 1.00191975e+00 -7.59968311e-02
-8.34827125e-01 4.65304852e-01 -2.19754636e-01 3.54311466e-02
7.96439707e-01 -2.73566693e-01 -3.89763385e-01 -3.63965750e-01
-2.70016521e-01 6.87270239e-02 9.29320574e-01 3.32792610e-01
6.56252205e-01 -1.35438764e+00 -6.83793545e-01 2.30832547e-01
1.53274968e-01 1.72600821e-01 6.17506683e-01 2.76449680e-01
-7.98658669e-01 -3.04873914e-01 -2.63727129e-01 -3.14470679e-01
-1.19281006e+00 2.64099479e-01 4.62558001e-01 9.89244599e-03
-8.79369795e-01 9.89705563e-01 7.26739943e-01 -2.46805936e-01
4.18418229e-01 -5.32472730e-01 -5.02160527e-02 -1.00494400e-01
7.30079710e-01 3.30640972e-01 3.71054886e-03 -3.63928378e-01
-3.39297839e-02 7.86196887e-01 -9.33347121e-02 -2.42229342e-01
1.49972486e+00 -2.54440844e-01 -2.98891693e-01 3.24219793e-01
1.19162178e+00 4.07782584e-01 -1.84589982e+00 -9.86925513e-02
-8.19181263e-01 -8.00764561e-01 3.05573165e-01 -8.33048046e-01
-1.66682315e+00 6.63010359e-01 5.70468187e-01 -4.24750447e-02
1.35965252e+00 -5.47957599e-01 6.67682946e-01 2.34106645e-01
2.85127848e-01 -1.04184353e+00 -1.01491921e-02 2.98356712e-01
1.05360079e+00 -7.40063190e-01 1.39985025e-01 -6.81916893e-01
-5.62927127e-01 9.77687955e-01 6.75509334e-01 -1.18200101e-01
5.87326527e-01 4.44397897e-01 2.62807816e-01 -6.05331138e-02
-4.74408060e-01 8.94450247e-02 6.47606313e-01 5.62400699e-01
5.43033421e-01 1.14253618e-01 2.11343482e-01 -6.23702537e-03
6.98047206e-02 -3.18899632e-01 4.28190470e-01 6.88904047e-01
-2.70781189e-01 -1.10729575e+00 -4.96192962e-01 2.50886172e-01
-2.57648621e-03 -1.70944542e-01 -3.41727853e-01 4.45928901e-01
1.14786528e-01 5.61628044e-01 9.04082507e-02 -1.53665468e-01
5.25535941e-01 -3.97199243e-01 6.57580972e-01 -4.22965437e-01
-4.90297854e-01 1.92248404e-01 -4.78724837e-02 -7.69305110e-01
-3.76715541e-01 -3.09930056e-01 -1.07671642e+00 -5.19843459e-01
-4.65403944e-02 -2.54019827e-01 6.81550503e-01 3.75205189e-01
2.90919155e-01 8.53901684e-01 7.23724723e-01 -1.25210774e+00
1.43650800e-01 -2.80001581e-01 -7.77736604e-01 5.50035238e-01
3.15985978e-01 -5.97444057e-01 -1.02217235e-01 1.26321629e-01] | [11.410377502441406, -1.0391401052474976] |
7c5d02b6-57b0-4773-b059-21c2bcc93e7f | 4d-isip-4d-implicit-surface-interest-point | 1705.03634 | null | http://arxiv.org/abs/1705.03634v2 | http://arxiv.org/pdf/1705.03634v2.pdf | 4d isip: 4d implicit surface interest point detection | In this paper, we propose a new method to detect 4D spatiotemporal interest
points though an implicit surface, we refer to as the 4D-ISIP. We use a 3D
volume which has a truncated signed distance function(TSDF) for every voxel to
represent our 3D object model. The TSDF represents the distance between the
spatial points and object surface points which is an implicit surface
representation. Our novelty is to detect the points where the local
neighborhood has significant variations along both spatial and temporal
directions. We established a system to acquire 3D human motion dataset using
only one Kinect. Experimental results show that our method can detect 4D-ISIP
for different human actions. | ['Changlin Xiao', 'Shirui Li', 'Alper Yilmaz', 'Hua Li'] | 2017-05-10 | null | null | null | null | ['interest-point-detection'] | ['computer-vision'] | [-2.35547125e-01 -4.37227875e-01 -7.92821869e-02 -1.21882074e-01
-1.91954821e-01 -2.30334446e-01 3.32921207e-01 -2.38490701e-01
-4.87022221e-01 2.74649829e-01 2.41914764e-02 3.00339282e-01
2.22925097e-01 -7.97980011e-01 -5.85815370e-01 -2.34675094e-01
-3.30119193e-01 2.97131956e-01 1.24923229e+00 -7.39679709e-02
5.31066358e-01 9.01164889e-01 -1.32300806e+00 5.17101511e-02
4.75663155e-01 9.78729606e-01 2.87625231e-02 3.69322538e-01
-1.50044829e-01 1.10459529e-01 -5.80434561e-01 4.02904689e-01
4.57105845e-01 -2.29582429e-01 -5.19974172e-01 5.66524193e-02
1.64002627e-01 -4.30302173e-01 -3.19256365e-01 8.80697131e-01
3.55604589e-01 3.25187951e-01 6.64794505e-01 -1.33563435e+00
-3.49696040e-01 -3.37805003e-01 -1.02678835e+00 3.14361513e-01
7.61011600e-01 5.95242679e-02 2.06998721e-01 -1.22080159e+00
7.87582040e-01 1.33166993e+00 8.49324882e-01 5.64487278e-01
-7.63918340e-01 -3.67533624e-01 -8.09822902e-02 3.51581395e-01
-1.67704213e+00 -1.02045320e-01 1.08223116e+00 -5.51277518e-01
9.20083225e-01 3.79136115e-01 8.98073018e-01 4.81838077e-01
5.65973878e-01 7.99853981e-01 1.03664124e+00 -2.75570482e-01
3.93781930e-01 -4.06529754e-01 3.15298766e-01 7.07995951e-01
5.31894155e-02 8.53968635e-02 -5.30739784e-01 -3.35664243e-01
1.64825392e+00 2.51568139e-01 -2.61306942e-01 -7.75193036e-01
-1.30907381e+00 3.81739348e-01 3.73075634e-01 3.46804351e-01
-3.23115289e-01 1.65920720e-01 1.20391533e-01 -1.99514106e-01
3.88143718e-01 -3.68289977e-01 -2.73841918e-01 -3.47217858e-01
-5.03925025e-01 4.26785797e-01 4.74219680e-01 8.54669750e-01
6.75577223e-01 -3.15344542e-01 -2.40491293e-02 5.36955535e-01
4.34525132e-01 5.97434938e-01 5.18552244e-01 -1.20117664e+00
3.86488169e-01 1.01828599e+00 4.02790219e-01 -1.23661220e+00
-3.32032472e-01 3.89988661e-01 -4.01892126e-01 5.46569467e-01
2.82622218e-01 2.77344584e-01 -6.59838915e-01 7.99448490e-01
1.00469887e+00 5.22907615e-01 -4.74531442e-01 1.30413890e+00
8.63583446e-01 5.23855388e-01 -3.53878975e-01 -5.34703918e-02
1.14859498e+00 -5.54681480e-01 -7.37028837e-01 2.28934456e-02
3.78373116e-01 -1.83559567e-01 1.11048615e+00 1.10056549e-01
-1.07782602e+00 -5.71332753e-01 -9.00009334e-01 -2.99585052e-02
-1.17740713e-01 -2.43242413e-01 -3.18208598e-02 2.48956829e-01
-7.43409872e-01 5.44898927e-01 -1.10321021e+00 -3.53196502e-01
1.74694970e-01 1.32047728e-01 -4.95449811e-01 3.72838676e-01
-9.68771279e-01 8.22416961e-01 7.77933151e-02 1.71931148e-01
-4.93707240e-01 -2.29907066e-01 -8.79242837e-01 -4.77630496e-01
1.84448898e-01 -4.05069500e-01 9.48267639e-01 -3.14730078e-01
-1.56245828e+00 1.21074259e+00 -5.73460281e-01 -2.40877755e-02
5.44507146e-01 -2.48379931e-01 -1.99662611e-01 2.41128847e-01
3.50432903e-01 6.62534982e-02 6.30275249e-01 -1.28040588e+00
-3.98828238e-01 -9.11778688e-01 -3.38665098e-01 3.88657808e-01
2.02707306e-01 1.65073588e-01 -4.57587898e-01 -6.31881237e-01
9.12505090e-01 -8.62364233e-01 -2.39985600e-01 3.60300571e-01
-2.63448000e-01 -5.40070772e-01 1.26929247e+00 -4.79013801e-01
1.11935377e+00 -2.19865489e+00 3.67760770e-02 3.69650424e-01
1.79857954e-01 1.90750748e-01 3.57420236e-01 5.45990616e-02
4.01363194e-01 4.36601378e-02 -5.45772851e-01 4.06935401e-02
-2.92450666e-01 2.32856527e-01 3.52101438e-02 7.88902640e-01
-6.75576702e-02 8.36465418e-01 -9.60743427e-01 -6.89684212e-01
3.00689071e-01 4.10689056e-01 -1.11759424e-01 1.55899242e-01
2.19233528e-01 5.20391822e-01 -9.65317547e-01 8.45234811e-01
8.61334562e-01 1.79957107e-01 -4.72425491e-01 1.09887971e-02
-5.20597816e-01 -7.23472014e-02 -1.43239713e+00 1.82982683e+00
3.58619273e-01 4.08173263e-01 1.07539073e-01 -6.77513182e-01
1.32100844e+00 1.18139721e-01 7.29344070e-01 -6.63788915e-01
7.92403445e-02 2.11524874e-01 -7.58571565e-01 -8.25055480e-01
2.98185974e-01 -8.93535614e-02 6.85445890e-02 1.73715577e-01
-5.62732756e-01 -1.85841188e-01 -5.65837801e-01 -2.94009209e-01
1.12812686e+00 3.39414954e-01 3.99187565e-01 -3.21371496e-01
5.87005615e-01 2.11248755e-01 7.55126059e-01 3.25248659e-01
-8.46796274e-01 8.91397476e-01 3.13684970e-01 -5.04576206e-01
-7.34131694e-01 -1.41222227e+00 -3.06102991e-01 2.46383533e-01
9.28874373e-01 -8.66538361e-02 -6.01891398e-01 -6.74246192e-01
2.68498152e-01 1.93455786e-01 -6.65853918e-01 1.66151986e-01
-1.07275808e+00 6.05698042e-02 2.54810840e-01 7.16158271e-01
6.58864975e-01 -1.01819515e+00 -1.27222347e+00 8.63530952e-03
-2.43046298e-03 -9.13156927e-01 -8.15919697e-01 -3.88896704e-01
-1.46962178e+00 -1.22324121e+00 -9.52702045e-01 -7.30674803e-01
5.61066449e-01 4.21897948e-01 7.51183987e-01 -1.50631964e-01
-3.32871854e-01 7.62582719e-01 -4.18508351e-01 -2.34889239e-01
1.81377143e-01 -6.93919539e-01 9.97829363e-02 -1.37436308e-03
6.83624864e-01 -4.41985577e-01 -8.47718596e-01 7.66476631e-01
-4.54192787e-01 -2.74167508e-01 -9.43434164e-02 1.78384051e-01
1.08224881e+00 -1.56514063e-01 -1.09421283e-01 7.76159540e-02
4.74257827e-01 -2.82422811e-01 -5.28572381e-01 8.36557671e-02
4.29227836e-02 -2.35126629e-01 -2.96641707e-01 -3.95506263e-01
-6.71579480e-01 1.77786723e-01 2.76722252e-01 -7.81070173e-01
-3.35629463e-01 1.76763967e-01 -9.99838635e-02 -1.31106406e-01
5.25254428e-01 3.15537930e-01 2.56880790e-01 -6.82597816e-01
1.01556949e-01 7.51990259e-01 5.76918542e-01 -4.04580712e-01
6.59418464e-01 1.09740329e+00 2.24320009e-01 -1.08293974e+00
-2.26990297e-01 -7.70769298e-01 -1.42797995e+00 -6.28455281e-01
1.00636876e+00 -6.37709141e-01 -8.18211317e-01 6.24196887e-01
-1.49646926e+00 -2.76009828e-01 -4.85399872e-01 6.58594728e-01
-6.23832881e-01 7.10789859e-01 -5.02092719e-01 -9.81789589e-01
-2.44070172e-01 -8.69359434e-01 1.32558811e+00 5.76296076e-02
-3.15134943e-01 -9.72609460e-01 5.23474216e-01 6.36633188e-02
-1.31119072e-01 7.89536536e-01 1.92531303e-01 3.15015428e-02
-4.41436857e-01 -2.86325097e-01 1.59646511e-01 5.61592728e-02
3.59715670e-01 -7.08968341e-02 -5.44366181e-01 3.15637812e-02
4.89947081e-01 2.49027282e-01 3.27741116e-01 7.21891880e-01
1.10170519e+00 1.46242991e-01 -6.21873438e-01 4.70852762e-01
1.11118543e+00 5.29825211e-01 7.96495557e-01 6.37595117e-01
6.73574626e-01 5.39225578e-01 1.08927774e+00 5.90370476e-01
3.00176322e-01 9.57660139e-01 3.46108794e-01 1.44980282e-01
1.12233728e-01 -2.65332818e-01 4.69981343e-01 6.78175747e-01
-5.21882117e-01 4.35790479e-01 -1.04406548e+00 6.05620265e-01
-1.88797331e+00 -9.03351784e-01 -8.18341613e-01 2.37848258e+00
6.52823448e-01 2.66381307e-04 2.79398203e-01 2.06465438e-01
7.93623924e-01 2.56618159e-03 -6.15020752e-01 -1.15695909e-01
1.90972820e-01 1.25852108e-01 3.50863516e-01 5.37331879e-01
-9.38485980e-01 6.88130379e-01 6.83549595e+00 5.06977081e-01
-1.01360226e+00 1.52878910e-01 -9.14259329e-02 -1.51188925e-01
1.58263389e-02 -2.03514189e-01 -7.14442194e-01 5.03662765e-01
2.58960098e-01 -1.67868823e-01 -1.98769554e-01 8.55838239e-01
7.33481169e-01 -3.61112237e-01 -8.09775352e-01 1.12417316e+00
8.48601293e-03 -9.05670702e-01 -9.10310969e-02 8.39282051e-02
4.31229979e-01 -2.13111028e-01 -2.45673701e-01 -1.13024347e-01
-3.96241397e-01 -8.56692612e-01 7.52684772e-01 6.79765701e-01
8.45711768e-01 -4.61448401e-01 2.77720928e-01 6.99181199e-01
-1.75209272e+00 4.61402059e-01 -4.17784303e-01 -2.51080126e-01
3.03641856e-01 3.61517996e-01 -6.75995409e-01 2.66201317e-01
9.70422447e-01 8.40240955e-01 -2.01086581e-01 1.28720772e+00
1.27896955e-02 2.16413602e-01 -3.90824944e-01 -1.26678109e-01
5.80372289e-02 -5.04512966e-01 9.72186863e-01 7.92667568e-01
5.97859204e-01 6.78967357e-01 1.24382183e-01 1.01680982e+00
5.46719193e-01 -3.76055688e-02 -9.32858229e-01 6.62354827e-01
4.54257488e-01 6.69718325e-01 -8.57224524e-01 -2.32053876e-01
-1.17642440e-01 1.24978256e+00 -8.19527432e-02 2.82802939e-01
-8.43574524e-01 -4.56211448e-01 4.95929867e-01 6.59383893e-01
5.59776351e-02 -8.05706799e-01 -4.79735672e-01 -1.15121484e+00
4.88554686e-01 -1.84020489e-01 1.84256658e-01 -8.87900054e-01
-1.10958612e+00 7.84991980e-02 2.82223076e-01 -1.57342827e+00
3.66542727e-01 -5.77961445e-01 -8.25015485e-01 9.38861132e-01
-9.23452675e-01 -5.85051298e-01 -6.22177899e-01 1.11330724e+00
6.01583779e-01 3.05680335e-01 4.03729349e-01 -1.10059902e-01
-2.79053822e-02 -4.41552512e-02 -1.28137767e-01 2.62606710e-01
2.85591781e-01 -1.02005053e+00 5.94119191e-01 4.60815430e-01
-1.68507919e-01 5.62149465e-01 3.01142246e-01 -1.15888107e+00
-1.48278177e+00 -7.80368090e-01 6.96875691e-01 -9.91840959e-01
1.76149145e-01 -2.21989438e-01 -1.07542670e+00 7.80817688e-01
-4.58697796e-01 3.97657424e-01 2.49369234e-01 -4.83924478e-01
2.19249263e-01 9.42482129e-02 -1.52689087e+00 3.84716004e-01
1.42626047e+00 -3.41397613e-01 -1.06313467e+00 -7.33350683e-03
5.83498538e-01 -7.47242749e-01 -8.89841855e-01 7.26295233e-01
5.81493676e-01 -1.01681960e+00 1.16198158e+00 -6.47719577e-02
-2.51522988e-01 -6.60807073e-01 -1.39910772e-01 -8.60956967e-01
-3.06684107e-01 -3.03836972e-01 -4.34532344e-01 6.22930825e-01
-3.69591087e-01 -5.93037426e-01 1.11227977e+00 6.90599620e-01
-6.98458925e-02 -8.70613277e-01 -1.56716788e+00 -1.21732068e+00
-1.16119795e-01 -4.67650443e-01 5.05431116e-01 8.31228673e-01
2.35077396e-01 -2.66598105e-01 8.11095610e-02 2.68907458e-01
9.58864272e-01 -6.77476227e-02 7.32469618e-01 -1.25229633e+00
4.41176683e-01 -2.57718682e-01 -8.76961768e-01 -1.52969253e+00
-7.49751329e-02 -5.47605693e-01 2.52836719e-02 -1.63005257e+00
-6.56288043e-02 -5.39555252e-01 -1.34468660e-01 -2.71879844e-02
6.56647682e-02 5.30217998e-02 -1.19085133e-01 5.74826002e-01
-4.56097007e-01 8.54701996e-01 1.55827105e+00 6.30822852e-02
-5.70849001e-01 -1.32893380e-02 3.94480854e-01 8.75906944e-01
5.84954143e-01 -3.55391204e-01 -1.22969098e-01 -3.15832794e-01
-4.55640793e-01 1.24209903e-01 7.22863078e-01 -1.13943696e+00
2.12273777e-01 -5.58339596e-01 3.44532639e-01 -1.29808819e+00
5.59357166e-01 -9.63191986e-01 2.17061847e-01 7.30522513e-01
1.30910456e-01 -5.84420487e-02 2.54482683e-02 6.14223123e-01
-1.59651965e-01 -1.36832044e-01 5.90899765e-01 -4.59469974e-01
-9.72650409e-01 5.07266283e-01 -7.36485571e-02 -1.05414286e-01
1.65849400e+00 -9.34375048e-01 3.26865762e-01 1.67664856e-01
-7.11932659e-01 1.12281948e-01 7.61069000e-01 3.42597038e-01
1.49012613e+00 -1.74887288e+00 -4.35771465e-01 2.66728103e-01
-4.02876996e-02 3.46595466e-01 -5.31680584e-02 7.40572155e-01
-8.45706880e-01 3.08685362e-01 -2.12082684e-01 -1.10242426e+00
-1.43805003e+00 3.10336411e-01 5.72884798e-01 2.78367132e-01
-1.09929979e+00 7.93050528e-01 7.74108618e-02 -3.98430735e-01
1.68978140e-01 -6.56194210e-01 -2.61165410e-01 -3.26732665e-01
4.53888029e-01 9.40374017e-01 -2.53004819e-01 -9.80768323e-01
-7.40638733e-01 1.34025228e+00 7.92070210e-01 -3.24940532e-01
9.50476348e-01 -6.77377731e-02 -2.77301043e-01 9.70937610e-01
1.29688251e+00 -7.70879835e-02 -1.46971333e+00 -6.63099512e-02
9.57849026e-02 -1.00059199e+00 -2.51633584e-01 -1.30569622e-01
-8.55648458e-01 8.08432937e-01 9.60543394e-01 1.35407120e-01
6.63767874e-01 1.62545905e-01 1.12813663e+00 -4.02714089e-02
1.04350817e+00 -1.07085478e+00 1.90471098e-01 5.96908510e-01
1.15979385e+00 -1.04954898e+00 -4.79926132e-02 -5.89452028e-01
-5.81759512e-01 1.26775360e+00 6.43866420e-01 -4.99382526e-01
1.01143527e+00 1.05258517e-01 -5.21308295e-02 -6.09593391e-01
6.18239976e-02 1.08811662e-01 3.87681603e-01 7.85182238e-01
1.13070421e-01 1.75868049e-01 -4.73007262e-01 3.69265288e-01
1.78608373e-01 3.39575887e-01 1.05366923e-01 1.44968390e+00
-8.11296821e-01 -6.49493992e-01 -8.16843927e-01 -4.66657691e-02
-4.73914072e-02 7.35922396e-01 -3.24518919e-01 7.86092699e-01
1.65203124e-01 7.20776916e-01 4.01114970e-01 -5.49825788e-01
9.71147180e-01 -1.10327773e-01 5.84332049e-01 -4.96539265e-01
-1.89952955e-01 -5.10604493e-02 -4.77458715e-01 -1.11549270e+00
-4.78847265e-01 -7.71699488e-01 -1.99757195e+00 -5.90961389e-02
2.59167934e-03 2.74820700e-02 6.74982548e-01 5.70127010e-01
3.36619645e-01 -2.20781907e-01 6.51989639e-01 -1.16085613e+00
-3.11487347e-01 -6.32766604e-01 -8.77862394e-01 4.18907762e-01
3.48041952e-01 -1.01282978e+00 -3.33000839e-01 8.15920383e-02] | [7.967260360717773, 0.1220424622297287] |
69c078de-fb71-44de-9029-9010ee07959a | stochastic-rising-bandits | 2212.03798 | null | https://arxiv.org/abs/2212.03798v1 | https://arxiv.org/pdf/2212.03798v1.pdf | Stochastic Rising Bandits | This paper is in the field of stochastic Multi-Armed Bandits (MABs), i.e., those sequential selection techniques able to learn online using only the feedback given by the chosen option (a.k.a. arm). We study a particular case of the rested and restless bandits in which the arms' expected payoff is monotonically non-decreasing. This characteristic allows designing specifically crafted algorithms that exploit the regularity of the payoffs to provide tight regret bounds. We design an algorithm for the rested case (R-ed-UCB) and one for the restless case (R-less-UCB), providing a regret bound depending on the properties of the instance and, under certain circumstances, of $\widetilde{\mathcal{O}}(T^{\frac{2}{3}})$. We empirically compare our algorithms with state-of-the-art methods for non-stationary MABs over several synthetically generated tasks and an online model selection problem for a real-world dataset. Finally, using synthetic and real-world data, we illustrate the effectiveness of the proposed approaches compared with state-of-the-art algorithms for the non-stationary bandits. | ['Marcello Restelli', 'Matteo Pirola', 'Francesco Trovò', 'Alberto Maria Metelli'] | 2022-12-07 | null | null | null | null | ['multi-armed-bandits'] | ['miscellaneous'] | [ 3.57688099e-01 8.56489502e-03 -6.54689670e-01 -2.97124356e-01
-1.07262182e+00 -6.66201234e-01 9.58723798e-02 -2.97427773e-02
-5.41117966e-01 1.41125727e+00 -3.21631104e-01 -7.36527681e-01
-9.14380729e-01 -6.27460957e-01 -1.10665834e+00 -9.66592848e-01
-2.56758422e-01 8.33324373e-01 2.05754414e-02 9.95471925e-02
2.02996820e-01 3.19955558e-01 -1.32735026e+00 1.37371942e-01
9.42620754e-01 1.56890380e+00 8.31634775e-02 5.51444352e-01
-1.29811585e-01 4.03562963e-01 -3.52766812e-01 -6.35101497e-01
6.57559693e-01 -6.10156834e-01 -7.88101971e-01 2.44746834e-01
1.81846693e-02 -4.30417210e-02 -2.42773201e-02 9.68293786e-01
3.55129302e-01 3.95375341e-01 5.71357250e-01 -1.06883430e+00
-1.09687693e-01 8.53400469e-01 -1.01170003e+00 5.31177104e-01
-3.80600467e-02 -4.23621647e-02 1.18089604e+00 -1.89794041e-02
3.20207655e-01 1.08236444e+00 1.87751442e-01 5.40068150e-01
-1.36895084e+00 -7.55987644e-01 6.00045204e-01 4.76343185e-03
-7.26377845e-01 -4.20028865e-01 5.59960544e-01 -2.54807562e-01
3.25271130e-01 7.55300641e-01 7.10809469e-01 1.01618946e+00
-3.35998684e-01 1.13473082e+00 1.29420233e+00 -5.94041407e-01
5.35874963e-01 5.61660491e-02 3.04305702e-01 3.78463119e-01
5.95029116e-01 2.22921491e-01 -5.60905576e-01 -4.99726087e-01
6.61863208e-01 4.47174311e-02 -1.34366870e-01 -5.61124265e-01
-9.22399938e-01 9.18424368e-01 -6.07582927e-02 -1.67832762e-01
-7.20285654e-01 3.41307461e-01 2.16867805e-01 5.78664422e-01
6.78342462e-01 1.21020079e-01 -6.65232360e-01 -9.86364633e-02
-9.62076843e-01 3.49162668e-01 7.26973116e-01 1.08304131e+00
4.94484663e-01 -8.77116695e-02 -6.44208014e-01 8.22441459e-01
-1.04903705e-01 2.99296051e-01 2.27670982e-01 -6.85027063e-01
9.44463253e-01 1.26094446e-01 9.96417046e-01 -1.38535231e-01
-3.05120379e-01 -9.50563788e-01 -5.13523519e-01 -1.30372792e-01
6.47111714e-01 -5.11648715e-01 -6.98071778e-01 1.82295763e+00
5.04048526e-01 2.62489710e-02 -5.34817316e-02 7.60162950e-01
-1.76304653e-01 5.36201358e-01 -2.77400404e-01 -8.67023408e-01
8.30350399e-01 -9.78756368e-01 -4.59649622e-01 -3.02332997e-01
2.88844198e-01 -4.11067665e-01 9.97449696e-01 6.02286458e-01
-1.25250530e+00 1.05812013e-01 -8.09132338e-01 8.50264490e-01
2.04787761e-01 2.43810471e-02 8.13045025e-01 9.73107398e-01
-4.90664482e-01 4.42411095e-01 -7.09603786e-01 -1.30155059e-02
4.50294584e-01 4.33935761e-01 1.24517433e-01 5.68992533e-02
-8.28525424e-01 3.51678938e-01 3.31733465e-01 1.53253645e-01
-9.51479673e-01 -4.58913147e-01 -1.36275291e-01 2.10783761e-02
9.86475170e-01 -6.27373338e-01 1.50096428e+00 -1.31832492e+00
-1.59967065e+00 6.45131588e-01 -1.80513427e-01 -9.09707248e-01
1.16498280e+00 -1.83794394e-01 -5.95866777e-02 -6.95045814e-02
-2.86258180e-02 -2.55062729e-01 8.15923870e-01 -1.10846663e+00
-1.08960342e+00 -5.10355175e-01 3.99877757e-01 6.83514327e-02
-3.26592177e-01 -2.45284550e-02 -3.52729037e-02 -6.59821153e-01
-2.63558924e-02 -1.10282087e+00 -2.92531163e-01 -4.86583292e-01
-6.13556504e-01 -5.69738634e-02 1.96708202e-01 -2.77237803e-01
1.44154799e+00 -1.86663747e+00 1.12121433e-01 3.71088356e-01
-6.35532022e-01 -5.94842322e-02 7.06682876e-02 5.50140619e-01
5.87178655e-02 1.85616314e-01 3.54117751e-02 -2.22780898e-01
7.72597045e-02 -3.24281980e-03 -2.36662030e-01 7.31520057e-01
-6.07901871e-01 4.82304841e-01 -6.30080044e-01 -2.19554767e-01
-2.92593658e-01 -6.41927063e-01 -4.03025657e-01 3.86974365e-02
-6.38589978e-01 3.33171278e-01 -9.29097116e-01 5.17034292e-01
4.15958047e-01 -3.93657416e-01 1.95248812e-01 4.89894062e-01
-6.06677756e-02 -2.19114497e-02 -1.56378758e+00 1.20809650e+00
-5.95851183e-01 -2.88934726e-02 2.73182780e-01 -1.24613750e+00
4.95081544e-01 1.66512668e-01 3.35156709e-01 -3.20983380e-01
8.14243108e-02 3.58652502e-01 -9.04894061e-03 -2.09959596e-01
-9.43294391e-02 -3.09162408e-01 1.79944828e-01 7.60299265e-01
-2.23821476e-01 3.92718792e-01 3.97369802e-01 -2.11888924e-01
1.01596045e+00 -6.17488325e-02 3.03753734e-01 -4.53720391e-01
2.71204859e-01 -2.07000077e-01 6.11738205e-01 1.40954852e+00
7.58398138e-03 1.70002341e-01 6.27699554e-01 -5.34580469e-01
-7.80217648e-01 -7.87302732e-01 2.45328676e-02 1.50460374e+00
1.23114616e-01 4.34953690e-01 -5.57869136e-01 -6.22578323e-01
4.71812040e-01 1.01908338e+00 -8.95772040e-01 6.90014735e-02
-1.11349173e-01 -9.31318700e-01 -9.26731154e-02 3.47551927e-02
3.98975819e-01 -9.03141558e-01 -9.63583648e-01 3.86921614e-01
-5.40672876e-02 -7.72141397e-01 -4.73120660e-01 5.20371377e-01
-9.73210156e-01 -8.59773874e-01 -7.41540074e-01 -1.95865169e-01
5.40531158e-01 2.45972350e-01 8.48107576e-01 -3.19584101e-01
2.78414823e-02 2.03575492e-01 -3.53258401e-01 -8.02890539e-01
1.06434755e-01 2.93581873e-01 -3.53953056e-02 6.83577776e-01
-1.46077707e-01 -3.76137376e-01 -8.56691480e-01 4.74074066e-01
-8.20075393e-01 1.89964533e-01 5.77257812e-01 1.08124495e+00
7.45751262e-01 1.94342304e-02 8.15586686e-01 -1.17865133e+00
6.25859141e-01 -4.37726527e-01 -1.15342486e+00 7.18640745e-01
-4.75913197e-01 3.01372141e-01 6.25784338e-01 -8.12827468e-01
-1.05498159e+00 2.30321780e-01 5.83698630e-01 -2.73783833e-01
2.41948813e-01 4.87635285e-01 1.61480427e-01 2.36967187e-02
4.35613662e-01 4.41092670e-01 -2.17093199e-01 -5.43023705e-01
1.34954257e-02 7.04766393e-01 -8.70404299e-03 -1.03218091e+00
4.00266677e-01 5.71568489e-01 1.76067755e-01 -3.98192614e-01
-1.30266297e+00 -3.20490390e-01 1.37477890e-01 -1.02395996e-01
1.19034484e-01 -5.23783088e-01 -1.12979472e+00 2.37210944e-01
-6.86479151e-01 -4.36135858e-01 -4.05887127e-01 5.76373160e-01
-1.16331935e+00 -9.86036938e-03 -2.77540296e-01 -1.88131690e+00
-3.49961698e-01 -7.80019701e-01 5.31957746e-01 3.77181351e-01
1.95064172e-01 -5.91496348e-01 -1.05098598e-01 6.46583617e-01
2.09072188e-01 2.58907050e-01 9.62709904e-01 -8.88124228e-01
-5.88404894e-01 -2.91801006e-01 1.23522684e-01 1.57694772e-01
8.46378878e-03 -3.47049445e-01 -4.73205388e-01 -5.22378862e-01
-5.21744080e-02 -4.79227811e-01 9.44234133e-01 8.65262628e-01
1.38925064e+00 -7.33520389e-01 -4.00930911e-01 4.00971860e-01
1.32424676e+00 6.32346809e-01 1.74752921e-01 7.37083018e-01
-2.38999382e-01 3.62586111e-01 9.00518060e-01 1.07641757e+00
-2.14489564e-01 5.97169995e-01 6.53352737e-01 2.40033120e-01
7.81808794e-01 -1.23081967e-01 -2.73536295e-02 -2.10408047e-01
-1.85046196e-01 -5.51633298e-01 -5.64760685e-01 7.71624982e-01
-2.34887314e+00 -9.78617787e-01 3.14324439e-01 2.75027800e+00
1.14664316e+00 4.35954243e-01 6.13179147e-01 -9.97927971e-03
1.04807723e+00 7.47698620e-02 -1.06519508e+00 -4.87200707e-01
-1.69541433e-01 2.44393766e-01 1.00260186e+00 3.23216468e-01
-1.01677108e+00 5.59357584e-01 5.27970266e+00 1.26051259e+00
-7.84935057e-01 2.28281930e-01 1.16784835e+00 -7.55421579e-01
-1.26257151e-01 -1.55425165e-02 -7.58716524e-01 6.96437478e-01
8.29948008e-01 -4.13975477e-01 9.16292429e-01 9.83899176e-01
3.24845076e-01 -3.91319513e-01 -1.07682312e+00 9.78695273e-01
-5.02013922e-01 -1.36663783e+00 -2.11461112e-01 1.04424305e-01
1.07617819e+00 -2.41236210e-01 2.12686583e-01 5.85304154e-03
6.40846074e-01 -6.71468794e-01 1.07437098e+00 4.91841525e-01
5.78847766e-01 -9.70835209e-01 5.90319216e-01 7.97598302e-01
-6.42518997e-01 -6.92779779e-01 -1.99650064e-01 3.71283144e-02
-1.89152372e-03 7.07460761e-01 -3.67443562e-01 7.72293508e-01
6.05834007e-01 -7.79879615e-02 1.16487905e-01 1.37247276e+00
9.16996226e-02 7.79352784e-01 -5.83364904e-01 -5.30981541e-01
4.49651629e-01 -3.92145872e-01 5.89350343e-01 8.88964951e-01
3.81456912e-01 1.92772463e-01 5.26059605e-02 4.27564174e-01
-8.27963725e-02 3.34261805e-01 -5.59939966e-02 -5.73796667e-02
4.37700629e-01 8.12273204e-01 -7.37204432e-01 -1.70413509e-01
-1.21230319e-01 6.65132165e-01 3.24217349e-01 4.96259511e-01
-8.69656086e-01 -1.46742880e-01 5.59032619e-01 1.56956956e-01
7.66323924e-01 1.70892417e-01 -3.41928154e-01 -8.43096495e-01
2.86946803e-01 -6.29027843e-01 8.13864768e-01 -2.69773602e-01
-1.37897968e+00 2.18678609e-01 2.08471417e-01 -9.53426957e-01
-2.37278774e-01 -2.48697296e-01 -2.98204333e-01 7.58953810e-01
-1.39905989e+00 -6.90999448e-01 2.55336553e-01 5.76192796e-01
5.62344074e-01 -1.69280916e-01 3.68021339e-01 -6.78598136e-02
-6.54162169e-01 6.41313434e-01 8.42068911e-01 -4.61494207e-01
1.44021779e-01 -9.90762353e-01 -7.56391361e-02 6.00293100e-01
4.27036546e-02 3.47553402e-01 9.42964792e-01 -3.04188907e-01
-1.47215223e+00 -8.04565549e-01 1.81994513e-01 1.43365815e-01
7.89749086e-01 -3.30009431e-01 -3.33624221e-02 7.02359080e-01
-2.47000322e-01 1.41350806e-01 5.42862117e-01 1.34544462e-01
1.65079776e-02 -5.78395128e-01 -1.26891255e+00 4.78558838e-01
1.31580973e+00 1.07598066e-01 4.96605299e-02 5.75919509e-01
5.72149634e-01 -4.67812151e-01 -3.28378767e-01 2.87281692e-01
8.16354513e-01 -1.03168011e+00 7.04305708e-01 -1.21429205e+00
1.46397129e-02 2.48684302e-01 -1.85947999e-01 -1.24573863e+00
-1.63340434e-01 -1.17300391e+00 -1.78441927e-01 8.54806125e-01
7.61632383e-01 -9.27149117e-01 1.04479051e+00 5.80655575e-01
3.54553372e-01 -9.51090157e-01 -1.55266881e+00 -1.13682306e+00
-1.25541668e-02 -2.11170450e-01 6.54395759e-01 4.94202912e-01
-3.17663223e-01 -8.82750824e-02 -8.26756239e-01 -1.23841465e-01
7.34591842e-01 7.88729668e-01 6.12100959e-01 -1.15031481e+00
-9.10466194e-01 -2.66235203e-01 2.10457027e-01 -1.05003607e+00
7.10871667e-02 -2.77104169e-01 -3.26144814e-01 -1.14045429e+00
3.70808333e-01 -6.19269848e-01 -8.12866688e-01 3.72193515e-01
5.03309667e-02 -4.98705864e-01 2.40716785e-01 1.57049522e-02
-7.88529456e-01 2.98438996e-01 1.10087359e+00 -8.25549513e-02
-4.06919569e-01 9.68670368e-01 -7.39433169e-01 3.10206264e-01
8.37147772e-01 -7.13307321e-01 -5.22365689e-01 -1.73554897e-01
3.58504325e-01 7.78740406e-01 -1.97187271e-02 -5.29102743e-01
2.25099195e-02 -9.41714108e-01 2.30687615e-02 -6.11068904e-01
3.04536283e-01 -8.54294598e-01 4.36435521e-01 6.93596959e-01
-7.44208992e-01 -2.19279468e-01 -6.90610558e-02 9.96253014e-01
2.86435932e-01 -4.91711378e-01 7.85529196e-01 -3.74776453e-01
1.39834270e-01 3.58411968e-01 -1.21585265e-01 1.70341834e-01
1.33140254e+00 -1.70689389e-01 -2.82811850e-01 -7.45019197e-01
-6.84981763e-01 3.20788741e-01 -4.11319919e-02 -1.68708146e-01
3.60137001e-02 -9.20700848e-01 -6.45171165e-01 -1.30835414e-01
-7.97066763e-02 -1.61410257e-01 3.19271266e-01 8.47995400e-01
-7.81992152e-02 6.26819670e-01 -2.67521832e-02 -1.81588620e-01
-1.07496738e+00 5.90385258e-01 3.60947669e-01 -6.83754504e-01
4.88530770e-02 8.78360391e-01 -8.16352814e-02 1.53670669e-01
4.46398079e-01 -1.70680478e-01 3.54913384e-01 1.19915105e-01
2.14645147e-01 6.57024264e-01 1.04904279e-01 1.22852854e-01
-1.63191035e-01 -3.57082449e-02 -1.75625309e-01 -4.96730715e-01
1.56803536e+00 -7.18535185e-02 9.27379131e-02 5.46679735e-01
6.11254990e-01 -1.30021214e-01 -1.33245826e+00 -3.63359094e-01
1.18012354e-01 -6.94809258e-01 -7.10132420e-02 -1.00995052e+00
-1.03787076e+00 3.69506657e-01 5.55579782e-01 6.57915056e-01
1.12484252e+00 -2.10202709e-01 5.00192642e-01 4.18657988e-01
1.02118528e+00 -1.21938932e+00 -3.48806053e-01 9.73060876e-02
6.23960137e-01 -9.94564176e-01 8.16897526e-02 -6.17185272e-02
-6.78554773e-01 8.93567860e-01 1.24287687e-01 -5.50458543e-02
5.64413607e-01 -2.06472784e-01 -5.77383816e-01 3.64869475e-01
-9.71880913e-01 -3.12398881e-01 -6.92364722e-02 5.62722124e-02
2.20841059e-04 3.72294307e-01 -8.25074911e-01 8.14090133e-01
-5.87789565e-02 1.35712147e-01 2.58869052e-01 1.10878789e+00
-3.50048900e-01 -1.01865983e+00 -4.24049169e-01 9.27485704e-01
-8.91230583e-01 1.61669269e-01 -3.42418522e-01 7.13136554e-01
-4.52526957e-01 1.02883923e+00 -2.12677896e-01 1.43605873e-01
3.31186801e-01 1.23652086e-01 6.13628685e-01 -3.48559201e-01
-5.35715878e-01 2.80438304e-01 2.81443208e-01 -1.44877031e-01
-3.88889462e-01 -8.13634872e-01 -6.03794456e-01 -1.05107792e-01
-6.55042470e-01 4.93479967e-01 5.96267402e-01 8.30540776e-01
2.48678848e-01 1.18076228e-01 1.00732577e+00 -5.44929683e-01
-1.36475468e+00 -8.33086431e-01 -1.09256589e+00 2.42103606e-01
3.96876663e-01 -7.93451905e-01 -3.92155975e-01 -4.47716802e-01] | [4.537148475646973, 3.2817187309265137] |
ec94d39b-cea1-43dc-aa02-8afea571bbdb | composition-based-heterogeneous-graph-multi | null | null | https://aclanthology.org/2022.coling-1.594 | https://aclanthology.org/2022.coling-1.594.pdf | Composition-based Heterogeneous Graph Multi-channel Attention Network for Multi-aspect Multi-sentiment Classification | Aspect-based sentiment analysis (ABSA) has drawn more and more attention because of its extensive applications. However, towards the sentence carried with more than one aspect, most existing works generate an aspect-specific sentence representation for each aspect term to predict sentiment polarity, which neglects the sentiment relationship among aspect terms. Besides, most current ABSA methods focus on sentences containing only one aspect term or multiple aspect terms with the same sentiment polarity, which makes ABSA degenerate into sentence-level sentiment analysis. In this paper, to deal with this problem, we construct a heterogeneous graph to model inter-aspect relationships and aspect-context relationships simultaneously and propose a novel Composition-based Heterogeneous Graph Multi-channel Attention Network (CHGMAN) to encode the constructed heterogeneous graph. Meanwhile, we conduct extensive experiments on three datasets: MAMSATSA, Rest14, and Laptop14, experimental results show the effectiveness of our method. | ['Yangyong Zhu', 'Yao Zhang', 'Hongrun Ren', 'Xiaosu Wang', 'Zhongchen Miao', 'Jian Gao', 'Yun Xiong', 'Hao Niu'] | null | null | null | null | coling-2022-10 | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [ 8.96900892e-02 -1.05754361e-01 -2.05296740e-01 -5.36662400e-01
-2.49035671e-01 -4.04788703e-01 3.30669165e-01 2.65287519e-01
-7.66274938e-03 3.13427806e-01 4.88716751e-01 -4.09683406e-01
-6.23653270e-03 -1.16558790e+00 -5.55437565e-01 -4.19863492e-01
5.12851775e-01 1.06070764e-01 3.51101495e-02 -6.24684155e-01
1.28973499e-01 -3.25840026e-01 -6.65867507e-01 3.58963847e-01
8.51967990e-01 9.50257003e-01 1.04767710e-01 1.22169875e-01
-6.53152287e-01 7.27042019e-01 -6.12206757e-01 -8.51522267e-01
-1.98848739e-01 -6.52379394e-01 -6.76259398e-01 3.72718155e-01
-2.19301745e-01 2.45094016e-01 -1.84520736e-01 1.20394969e+00
3.22274029e-01 -2.09660068e-01 3.06249976e-01 -1.11717093e+00
-8.63963366e-01 9.22756374e-01 -9.62390959e-01 7.54040405e-02
2.65439779e-01 2.93258540e-02 1.46903932e+00 -5.30333340e-01
3.48855019e-01 1.28891730e+00 6.42710209e-01 1.41722947e-01
-6.99263096e-01 -4.85640079e-01 1.00643611e+00 2.91497290e-01
-9.38771009e-01 2.23679036e-01 1.24107337e+00 5.31627424e-03
9.76632357e-01 2.97042400e-01 1.12940109e+00 8.27041805e-01
3.19194019e-01 9.89709556e-01 1.08310270e+00 2.43837517e-02
-1.07229158e-01 4.78277989e-02 4.75176305e-01 4.64087278e-01
5.08809268e-01 -8.07505906e-01 -2.18197018e-01 1.41019195e-01
1.05392449e-01 -7.07174325e-03 -1.34279773e-01 -1.75468534e-01
-1.05479789e+00 8.08656037e-01 3.73794377e-01 3.12553167e-01
-2.82162577e-01 -1.51993901e-01 7.25074172e-01 3.58831942e-01
6.38459265e-01 2.36595333e-01 -5.65410495e-01 3.59085831e-03
-2.64686018e-01 1.49618134e-01 7.86253452e-01 1.09709752e+00
8.51965189e-01 1.25740364e-01 -2.29061559e-01 7.82493412e-01
5.10791659e-01 6.03435516e-01 5.66715658e-01 -1.86833680e-01
7.87061870e-01 1.36962688e+00 -4.83314872e-01 -1.69385099e+00
-6.11027300e-01 -7.79313207e-01 -9.16663647e-01 -4.18915153e-01
-3.87700915e-01 -3.11937362e-01 -5.91483593e-01 1.63831890e+00
3.83734614e-01 -1.34434089e-01 2.15647906e-01 9.90480483e-01
1.09554803e+00 7.48064637e-01 -6.92662410e-03 -3.12646061e-01
1.63117349e+00 -1.36502779e+00 -8.53006721e-01 -5.81827939e-01
6.14553154e-01 -7.93445051e-01 1.16515422e+00 1.78577736e-01
-7.79733121e-01 -2.64165401e-01 -1.17131031e+00 -4.20652442e-02
-4.82397825e-01 -1.56542167e-01 1.03686810e+00 4.64470029e-01
-6.12701356e-01 -1.23023227e-01 -4.14328814e-01 -9.39363837e-02
3.48839581e-01 2.84192383e-01 -1.70953825e-01 -1.94462791e-01
-1.41155195e+00 4.25125033e-01 2.43945375e-01 4.99867588e-01
-2.40553364e-01 -2.88817734e-01 -1.13204348e+00 2.00742692e-01
6.78962111e-01 -8.55690897e-01 9.16976511e-01 -1.13664532e+00
-1.18694007e+00 5.06128073e-01 -3.33315104e-01 -7.38155842e-02
-1.00362785e-01 2.85172146e-02 -7.64503181e-01 -1.81307152e-01
1.78444296e-01 2.08051968e-02 4.26010907e-01 -1.05572343e+00
-4.10608858e-01 -6.81375742e-01 6.72087550e-01 6.31377816e-01
-6.34205759e-01 -5.95574826e-03 -8.34153473e-01 -6.93221867e-01
7.14725852e-02 -6.75862491e-01 -3.57476562e-01 -7.42324412e-01
-7.92695880e-01 -1.56421870e-01 6.88734055e-01 -3.62750441e-01
1.54049265e+00 -1.94940627e+00 2.57807255e-01 1.83246970e-01
1.50157705e-01 1.47059441e-01 -4.27566946e-01 3.75082135e-01
-1.60250831e-02 1.47504628e-01 -4.01785403e-01 -2.19454363e-01
-5.89961298e-02 1.15207545e-01 -1.82278052e-01 -8.89347866e-02
2.04832464e-01 9.83111382e-01 -9.14853990e-01 -7.11185217e-01
-1.59403369e-01 3.51434737e-01 -6.81913197e-01 -1.42616713e-02
-4.29059207e-01 2.70768851e-01 -9.45632756e-01 7.86898494e-01
8.92506540e-01 -4.53833103e-01 2.42348000e-01 -3.89561921e-01
1.97097450e-01 2.74790943e-01 -8.16237509e-01 1.58902049e+00
-7.79758751e-01 3.00425977e-01 -4.75870706e-02 -8.91116321e-01
9.53920782e-01 4.63213362e-02 4.26340640e-01 -7.30452776e-01
3.46582621e-01 -2.76936349e-02 1.10858522e-01 -5.21432579e-01
5.91302991e-01 -2.97404587e-01 -2.91825652e-01 3.62590641e-01
-7.76915699e-02 -1.57428309e-01 4.48428661e-01 4.80573595e-01
9.11176026e-01 -9.41846892e-02 1.38309434e-01 -3.72752436e-02
9.56346333e-01 1.37802307e-02 6.60103559e-01 9.40873325e-02
2.76645683e-02 6.97971463e-01 1.02277160e+00 -3.50999326e-01
-5.65794408e-01 -3.15814704e-01 3.10299009e-01 6.51762664e-01
6.84677720e-01 -9.66679335e-01 -5.79236746e-01 -8.54548216e-01
-2.83978611e-01 4.73275661e-01 -5.29482603e-01 -3.43269587e-01
-4.19648945e-01 -1.21024799e+00 -7.15028793e-02 3.29338521e-01
8.44393432e-01 -1.20791745e+00 3.12062263e-01 1.15544647e-01
-4.90833104e-01 -1.25115073e+00 -7.42627800e-01 -3.21738034e-01
-6.03628218e-01 -1.13372469e+00 -4.75973845e-01 -8.69455516e-01
6.24187052e-01 4.52190340e-01 1.15021646e+00 1.30505458e-01
2.70730495e-01 1.44376069e-01 -5.91714203e-01 -3.56375486e-01
1.06542863e-01 4.69090521e-01 -5.17946124e-01 3.07377696e-01
6.36024415e-01 -6.01684928e-01 -6.40734196e-01 5.27208783e-02
-9.72782791e-01 2.03061700e-01 7.60969758e-01 6.79024339e-01
5.69349110e-01 2.79401183e-01 6.39631748e-01 -1.36443758e+00
1.03549814e+00 -7.12811291e-01 -2.58377522e-01 2.44807824e-01
-7.86238790e-01 -3.73512894e-01 8.95488024e-01 -1.28364921e-01
-1.05695605e+00 -4.39152896e-01 -3.78368318e-01 -1.23592354e-01
2.18922988e-01 1.33293498e+00 -7.34557688e-01 3.27869385e-01
-2.27772951e-01 4.30217713e-01 -2.32557654e-01 -5.69614097e-02
1.57062411e-01 6.02726698e-01 1.63581781e-02 -2.21575305e-01
5.81176579e-01 2.84461945e-01 8.45858082e-03 -5.44220984e-01
-1.21293879e+00 -4.13442880e-01 -3.47139500e-02 -1.34224951e-01
7.60626078e-01 -1.03500271e+00 -7.37547755e-01 6.51846468e-01
-1.17356586e+00 2.03125522e-01 8.57416689e-02 3.56748551e-01
-1.09203391e-01 4.66768175e-01 -5.39391518e-01 -5.75175285e-01
-6.47141099e-01 -1.32494414e+00 1.00074100e+00 4.47852224e-01
3.62027660e-02 -9.62651432e-01 5.54159433e-02 6.28778934e-01
3.64413738e-01 7.99617618e-02 1.10214567e+00 -5.33022881e-01
-5.51591754e-01 -2.51781911e-01 -3.17998111e-01 2.18929768e-01
2.77354091e-01 -2.11608574e-01 -6.06040776e-01 -3.56847830e-02
1.51978001e-01 -7.25236759e-02 8.16533089e-01 1.96837991e-01
1.12312174e+00 -3.59742284e-01 -1.39580250e-01 5.04823446e-01
1.38323641e+00 1.49723321e-01 4.54175711e-01 5.64978659e-01
1.15821528e+00 4.47091192e-01 6.90104187e-01 2.05098853e-01
8.62874687e-01 4.40073609e-01 5.77892542e-01 -2.24232394e-02
5.83369173e-02 -3.07156712e-01 2.99492449e-01 1.74271524e+00
1.13400415e-01 -4.61653173e-01 -4.65832591e-01 7.39322484e-01
-1.80753672e+00 -6.27081037e-01 -4.87061620e-01 1.46025229e+00
6.41742826e-01 4.01235551e-01 -8.71388614e-02 3.49931256e-03
6.12518370e-01 7.52501786e-01 -5.45462370e-01 -5.23799479e-01
-5.45957386e-01 -3.25135469e-01 3.55786690e-03 3.06654721e-01
-1.00444531e+00 9.33474958e-01 4.61618805e+00 9.56945539e-01
-1.00094748e+00 7.18942732e-02 8.56356561e-01 1.62463054e-01
-1.14478266e+00 1.85998619e-01 -5.00371873e-01 7.50389457e-01
2.42201537e-01 -3.45932931e-01 2.80513257e-01 9.63206828e-01
1.11295395e-02 2.39321515e-01 -2.61927783e-01 6.97706163e-01
3.03025156e-01 -9.56751287e-01 4.46864814e-01 -2.31116593e-01
8.41753304e-01 -1.05363742e-01 1.26362666e-01 4.98096943e-01
-7.33436048e-02 -6.24622464e-01 3.53472859e-01 1.49010763e-01
3.67360204e-01 -1.08657885e+00 1.12235665e+00 3.33478115e-02
-1.58226204e+00 2.05281526e-01 -4.22665864e-01 -1.52640820e-01
1.80520564e-01 8.74233902e-01 -1.59663573e-01 1.12372148e+00
5.27122200e-01 1.16693091e+00 -6.53476179e-01 7.47301400e-01
-4.49001163e-01 6.09367013e-01 3.69211771e-02 -4.52519476e-01
5.43231428e-01 -6.53813779e-01 6.94392860e-01 9.77316678e-01
1.16805673e-01 8.60877186e-02 1.02600560e-01 6.37885451e-01
-2.58478969e-01 6.71890318e-01 -6.22077644e-01 -3.00211996e-01
-1.76309850e-02 1.64760613e+00 -7.55948842e-01 -2.18783572e-01
-9.44439113e-01 8.58013570e-01 2.17319712e-01 3.32204133e-01
-7.06098497e-01 -6.65685713e-01 4.90137041e-01 -2.92752981e-01
3.12542140e-01 1.50466934e-04 -5.03658175e-01 -1.60999322e+00
4.31422383e-01 -9.93507445e-01 3.68036389e-01 -8.21113765e-01
-1.31559670e+00 8.73773694e-01 -4.02968585e-01 -1.21323073e+00
2.27660507e-01 -3.58795166e-01 -1.10197842e+00 7.18913555e-01
-1.77895784e+00 -1.51044309e+00 -3.24429244e-01 4.57402825e-01
5.45690119e-01 -1.15979105e-01 4.66297120e-01 3.34100813e-01
-7.23377228e-01 4.59491968e-01 -3.93961191e-01 1.68643847e-01
3.45690548e-01 -1.13080394e+00 5.04129052e-01 7.78641164e-01
-6.06413521e-02 7.57962465e-01 5.46897650e-01 -6.72621548e-01
-1.65794528e+00 -1.15586412e+00 1.06363893e+00 -1.04575887e-01
8.38190496e-01 -2.99241006e-01 -8.47690582e-01 7.92882621e-01
6.08138621e-01 -3.47916812e-01 6.12408698e-01 6.11731470e-01
-2.85544038e-01 -3.48027587e-01 -5.36853015e-01 8.07899594e-01
9.50872660e-01 -4.03032780e-01 -3.91220987e-01 3.50419283e-01
1.28167415e+00 -3.34670126e-01 -5.99513352e-01 7.02084124e-01
2.80494571e-01 -8.51480186e-01 5.96068084e-01 -4.82479602e-01
9.31696773e-01 -5.18650830e-01 7.45171160e-02 -1.49226201e+00
-6.56485707e-02 -1.55961767e-01 4.82367352e-02 1.52265942e+00
6.18227303e-01 -7.15209305e-01 6.86811924e-01 1.63700357e-01
-2.46252954e-01 -1.22780502e+00 -4.83870059e-01 -3.36200148e-01
-1.46748364e-01 -4.97062385e-01 1.19989860e+00 1.00376129e+00
2.32846856e-01 1.10345578e+00 -3.19200426e-01 1.17655396e-01
1.17081143e-01 9.37113583e-01 6.80945754e-01 -7.05467343e-01
-4.38098818e-01 -5.96471667e-01 -3.94929498e-01 -9.91837978e-01
3.57487559e-01 -8.96691084e-01 -2.08640397e-01 -1.79158294e+00
5.11639833e-01 -3.40320140e-01 -3.94859344e-01 3.12060297e-01
-6.64332092e-01 1.15957044e-01 9.40677300e-02 -2.42921427e-01
-1.01839745e+00 9.33399618e-01 1.73413980e+00 -6.29089296e-01
1.07811317e-01 1.22893136e-02 -1.30935550e+00 7.19340146e-01
6.82037413e-01 -2.39141658e-01 -7.21979558e-01 -6.50996447e-01
9.67809916e-01 1.58080440e-02 -1.30926549e-01 -6.76056504e-01
1.22992374e-01 -1.98087618e-01 1.92619823e-02 -6.78297162e-01
1.98993295e-01 -8.25425208e-01 -2.40691066e-01 2.70670921e-01
-2.11078823e-01 2.46231243e-01 5.69236763e-02 8.12693417e-01
-7.49839306e-01 -8.57710838e-02 1.12376325e-01 -1.37254536e-01
-6.75408781e-01 5.68423390e-01 -2.00418547e-01 3.05441231e-01
6.21495008e-01 2.37879947e-01 -3.58961701e-01 -5.15663922e-01
-3.35238665e-01 6.48001492e-01 2.93712288e-01 5.93710423e-01
5.64872503e-01 -1.40546906e+00 -6.31108046e-01 9.19317678e-02
3.19699228e-01 4.16631699e-01 6.79533958e-01 8.48095119e-01
-2.73606956e-01 3.22624832e-01 1.80115104e-01 -2.74528325e-01
-1.12122703e+00 6.96794510e-01 1.04825325e-01 -6.19666040e-01
-3.61404002e-01 6.27678931e-01 4.29427207e-01 -7.30295718e-01
-4.11626697e-01 -1.47045627e-01 -7.39332795e-01 2.28508133e-02
2.30223462e-01 -2.07873404e-01 3.23618539e-02 -8.36296976e-01
-3.31208378e-01 8.24933052e-01 -1.67311862e-01 1.93065107e-01
1.14789820e+00 -3.21419686e-01 -5.32939374e-01 5.10913610e-01
1.22056508e+00 3.27076882e-01 -4.48367894e-01 -9.61074084e-02
-3.54787499e-01 -3.08709890e-01 2.63748970e-02 -5.70826948e-01
-1.58622348e+00 8.95636976e-01 -2.26478294e-01 3.78103346e-01
1.26925528e+00 -1.43928215e-01 1.16703022e+00 3.59022766e-01
3.13713849e-01 -8.69861186e-01 3.44471354e-03 7.11677372e-01
7.24171102e-01 -1.31895304e+00 5.47519885e-02 -6.73611999e-01
-9.55293715e-01 7.82715678e-01 9.31309521e-01 -5.35836816e-02
7.00110018e-01 -6.21834509e-02 1.40859827e-01 -4.50085789e-01
-6.89941466e-01 -1.27702728e-01 1.88315317e-01 2.23223135e-01
4.11409050e-01 1.16331503e-01 -7.48047531e-01 1.24066162e+00
-2.31164053e-01 -2.66723454e-01 5.08603394e-01 7.31963277e-01
-9.81643423e-03 -1.15680218e+00 1.79172188e-01 5.19568682e-01
-5.90257347e-01 -4.52270806e-01 -4.11970854e-01 5.96372128e-01
-6.17797486e-02 1.13088822e+00 -1.50459856e-01 -6.65702343e-01
4.55562413e-01 -2.56762415e-01 -1.37271415e-02 -4.79860902e-01
-7.43936896e-01 2.23792478e-01 2.73838937e-01 -2.78043598e-01
-5.58616042e-01 -5.10123610e-01 -1.31088722e+00 -1.86362579e-01
-4.44365948e-01 3.57095718e-01 5.54364741e-01 1.12790012e+00
4.89307880e-01 9.98815238e-01 8.97990942e-01 2.69741155e-02
2.05851600e-01 -1.07496238e+00 -6.78097129e-01 4.55916703e-01
1.43831819e-02 -2.03802079e-01 -1.83863908e-01 -4.77743685e-01] | [11.50041389465332, 6.602731704711914] |
3c16d982-0ba7-4f1a-a3bd-1a2d68811d60 | a-novel-attention-model-for-salient-structure | 2201.06174 | null | https://arxiv.org/abs/2201.06174v1 | https://arxiv.org/pdf/2201.06174v1.pdf | A novel attention model for salient structure detection in seismic volumes | A new approach to seismic interpretation is proposed to leverage visual perception and human visual system modeling. Specifically, a saliency detection algorithm based on a novel attention model is proposed for identifying subsurface structures within seismic data volumes. The algorithm employs 3D-FFT and a multi-dimensional spectral projection, which decomposes local spectra into three distinct components, each depicting variations along different dimensions of the data. Subsequently, a novel directional center-surround attention model is proposed to incorporate directional comparisons around each voxel for saliency detection within each projected dimension. Next, the resulting saliency maps along each dimension are combined adaptively to yield a consolidated saliency map, which highlights various structures characterized by subtle variations and relative motion with respect to their neighboring sections. A priori information about the seismic data can be either embedded into the proposed attention model in the directional comparisons, or incorporated into the algorithm by specifying a template when combining saliency maps adaptively. Experimental results on two real seismic datasets from the North Sea, Netherlands and Great South Basin, New Zealand demonstrate the effectiveness of the proposed algorithm for detecting salient seismic structures of different natures and appearances in one shot, which differs significantly from traditional seismic interpretation algorithms. The results further demonstrate that the proposed method outperforms comparable state-of-the-art saliency detection algorithms for natural images and videos, which are inadequate for seismic imaging data. | ['Ghassan AlRegib', 'Haibin Di', 'Zhiling Long', 'Muhammad Amir Shafiq'] | 2022-01-17 | null | null | null | null | ['seismic-imaging', 'seismic-interpretation'] | ['miscellaneous', 'miscellaneous'] | [ 6.01489425e-01 7.98114110e-03 3.74184310e-01 4.16007861e-02
-5.87424278e-01 -2.72459269e-01 3.57715338e-01 3.68460000e-01
-3.87778014e-01 2.39061207e-01 5.78890085e-01 -2.56742239e-02
-2.66374588e-01 -4.58080471e-01 -4.68520612e-01 -8.93685937e-01
-3.72221440e-01 3.29409912e-03 1.12015760e+00 -3.92435431e-01
9.86078501e-01 5.40616035e-01 -1.76545680e+00 2.93094754e-01
7.02176869e-01 7.53261447e-01 1.01238096e+00 7.05252767e-01
-2.47330056e-03 5.07947981e-01 -3.93559575e-01 3.86059523e-01
8.01611170e-02 -2.77646750e-01 -6.79666758e-01 2.72946328e-01
2.18236968e-01 -2.19804198e-01 -8.98166075e-02 1.20789862e+00
5.35501480e-01 5.01059532e-01 6.22156739e-01 -8.23315442e-01
-6.37593269e-01 3.71195793e-01 -1.08222294e+00 9.83318686e-01
3.65234464e-01 3.20736989e-02 1.05878627e+00 -1.46355593e+00
4.46475416e-01 1.29562378e+00 5.87172329e-01 1.98797360e-02
-1.11758101e+00 -2.01154754e-01 8.21679905e-02 4.05142307e-01
-1.07089424e+00 -2.29977891e-01 1.43742383e+00 -5.27767301e-01
8.85506809e-01 5.46057642e-01 6.45635188e-01 2.43927628e-01
2.22042203e-01 8.20831418e-01 7.86095798e-01 -4.35342878e-01
3.36270213e-01 -2.16229826e-01 3.19112130e-02 4.83486444e-01
2.25510970e-01 -5.36186025e-02 -9.65865493e-01 -1.19553521e-01
9.24334586e-01 -1.71530306e-01 -4.94963229e-01 -5.92706382e-01
-1.66082001e+00 7.67196536e-01 6.81165099e-01 2.67937273e-01
-6.94043398e-01 -8.69084373e-02 4.08290416e-01 -3.62284392e-01
5.19800127e-01 5.29400826e-01 -2.15941947e-02 3.46581250e-01
-1.22701812e+00 3.44408482e-01 8.12488347e-02 4.32340831e-01
8.01210999e-01 4.82051998e-01 -7.63115585e-02 6.59391224e-01
4.59166914e-01 6.85994744e-01 5.69635153e-01 -7.65973449e-01
4.56552863e-01 3.89876008e-01 2.24272788e-01 -1.58900976e+00
-6.55155480e-01 -1.90887764e-01 -4.67530608e-01 3.69188637e-01
6.98452070e-02 1.24495603e-01 -9.84381318e-01 1.32223105e+00
4.58747327e-01 4.16771024e-01 1.31611392e-01 1.46877503e+00
8.52616251e-01 5.93293846e-01 -8.00286457e-02 -3.06971334e-02
1.47860217e+00 -8.20283830e-01 -6.24228477e-01 -4.80487823e-01
-1.15811196e-03 -7.24887729e-01 1.12743568e+00 7.27295410e-03
-1.20279324e+00 -5.16203940e-01 -1.21288764e+00 1.09755509e-01
-8.56159031e-02 -8.73271227e-02 1.45775050e-01 2.33345643e-01
-1.18954897e+00 2.38071814e-01 -7.71046460e-01 -1.38502106e-01
6.99433237e-02 -2.41667964e-02 1.02818318e-01 4.54472721e-01
-1.26751876e+00 7.50879645e-01 4.40799326e-01 2.27616072e-01
-8.66688609e-01 -6.55111313e-01 -1.09269583e+00 2.44706973e-01
2.89591789e-01 -2.58795470e-01 1.03365648e+00 -9.53245521e-01
-9.48693156e-01 5.10679960e-01 -3.69434863e-01 -4.23328578e-01
2.15359718e-01 -1.16897136e-01 -4.02159393e-01 9.02121425e-01
4.96774375e-01 4.89986688e-01 1.15423787e+00 -1.58643651e+00
-7.83125460e-01 -5.12076579e-02 -2.05620289e-01 5.94341099e-01
9.53094289e-03 1.97642058e-01 -1.64207384e-01 -1.05471301e+00
7.06826091e-01 -4.26880836e-01 -3.44867319e-01 -2.69617379e-01
-3.61855865e-01 3.86978835e-01 1.10292721e+00 -9.77199376e-01
1.19267535e+00 -2.27177072e+00 2.65343100e-01 4.07195508e-01
3.07804435e-01 -1.41072217e-02 5.43183498e-02 1.47737443e-01
-2.76084542e-01 -1.71841875e-01 -7.91037738e-01 3.00037619e-02
-3.64218682e-01 -2.45678946e-01 -4.47127998e-01 5.09553969e-01
4.45031464e-01 4.81115550e-01 -1.23700809e+00 -5.54199994e-01
3.83087844e-01 2.80699372e-01 -3.31222892e-01 -5.04139476e-02
1.89394817e-01 4.44463193e-01 -5.08642435e-01 7.44199872e-01
8.46987963e-01 -8.58560391e-03 -3.15057158e-01 -1.31764144e-01
-5.40009677e-01 7.27085620e-02 -1.41626513e+00 1.29094446e+00
2.45662034e-01 7.84481645e-01 9.74904895e-02 -9.98925924e-01
1.04216921e+00 2.68295676e-01 3.50290626e-01 -5.54689705e-01
-2.17099637e-01 3.55797172e-01 2.81030629e-02 -7.23338187e-01
1.16273260e+00 -1.91446841e-01 2.95077078e-02 3.28706205e-01
-2.82559514e-01 -1.73986912e-01 1.68226078e-01 1.89504728e-01
6.34319246e-01 -1.56346396e-01 2.00645074e-01 -7.77359486e-01
6.59923196e-01 1.75137788e-01 4.02452230e-01 7.06706047e-01
-2.55206347e-01 1.06525540e+00 2.63675660e-01 -4.71343011e-01
-1.14781618e+00 -1.10975063e+00 -2.55051404e-02 1.06648970e+00
6.72075331e-01 1.84712112e-01 -8.16373467e-01 -1.01482794e-01
-2.04137683e-01 6.55668080e-01 -8.54969680e-01 -3.38802449e-02
-6.79643571e-01 -6.09891772e-01 7.78690130e-02 5.48699617e-01
5.64557076e-01 -1.49202549e+00 -1.74002075e+00 3.26756805e-01
-3.81802559e-01 -6.69759035e-01 -6.25272691e-01 7.57274032e-02
-9.26806390e-01 -9.16377902e-01 -1.09845734e+00 -1.02560031e+00
8.16132665e-01 1.03679895e+00 7.34064758e-01 4.30465341e-02
-2.36055806e-01 4.33233827e-01 -3.45491379e-01 -3.59326065e-01
-2.62197871e-02 -3.84679765e-01 -1.06303193e-01 3.52847844e-01
-1.11261113e-02 -3.54330987e-01 -7.83777952e-01 1.56313270e-01
-9.25545394e-01 2.78946728e-01 4.93506998e-01 8.63799274e-01
3.50689411e-01 3.90249342e-02 4.57971513e-01 -3.66444737e-01
4.96539474e-01 -6.71198905e-01 -2.97604829e-01 6.10839725e-02
5.33318706e-03 5.82035407e-02 1.36377871e-01 -2.79804647e-01
-1.25045252e+00 -8.77467319e-02 4.06282425e-01 -3.90540689e-01
7.53825763e-03 8.64523590e-01 1.78321481e-01 2.13017967e-02
5.21642029e-01 6.91261470e-01 -1.86640397e-01 -2.84509122e-01
6.26214743e-02 4.71372515e-01 1.03797853e+00 -2.28348017e-01
7.84008145e-01 6.91949248e-01 -1.76810324e-01 -1.23512816e+00
-4.16936040e-01 -6.90553248e-01 -7.73758054e-01 -5.10680258e-01
9.90648985e-01 -7.13607550e-01 -6.76546767e-02 5.29821396e-01
-1.05491889e+00 1.42150417e-01 -1.90975785e-01 4.45326507e-01
-5.14019847e-01 6.84231102e-01 -3.32549274e-01 -1.08683431e+00
-2.89281398e-01 -1.25419402e+00 1.30338645e+00 4.93044853e-01
-2.53579736e-01 -8.50214005e-01 -1.35177255e-01 -4.42068726e-02
3.72275859e-01 2.96190828e-01 7.98358679e-01 -4.08982307e-01
-4.34280187e-01 1.93957135e-01 -2.34659001e-01 -2.23937035e-01
2.86584377e-01 -8.31544213e-03 -8.60610962e-01 -2.46063575e-01
2.20788315e-01 2.89655924e-01 1.04500020e+00 7.65509486e-01
6.31045341e-01 3.21123973e-02 -2.67734140e-01 2.97888666e-01
1.21230114e+00 4.33449149e-01 4.36955392e-01 8.85558486e-01
7.39100039e-01 7.99209595e-01 8.95812631e-01 6.68946922e-01
1.72747701e-01 3.59853834e-01 6.51435912e-01 -5.60220838e-01
3.99467573e-02 8.75124559e-02 9.38778222e-02 5.48015058e-01
-1.73562422e-01 1.28686070e-01 -1.14618993e+00 1.14450443e+00
-1.66091740e+00 -1.15924907e+00 -4.10917252e-01 2.18957829e+00
4.19105142e-01 9.52747762e-02 5.16014285e-02 4.40530121e-01
1.03443778e+00 5.80228508e-01 -5.28091669e-01 -1.30776614e-01
-4.12506700e-01 -3.09422184e-02 4.41758633e-01 4.67350125e-01
-1.27980590e+00 7.16252744e-01 6.62822819e+00 4.81199741e-01
-1.28271067e+00 -9.17561278e-02 3.73940766e-01 2.26714104e-01
-7.03296185e-01 1.54838562e-02 -2.36486420e-01 5.41844964e-01
4.36520278e-01 -3.33803505e-01 2.99295262e-02 5.61443031e-01
6.38875723e-01 -6.03138506e-01 -3.98089945e-01 7.48141766e-01
2.21703619e-01 -1.39129019e+00 1.10047720e-01 -4.36412156e-01
6.98791623e-01 -1.86766893e-01 2.84253895e-01 -4.85408247e-01
-1.58613563e-01 -5.51535606e-01 1.42452860e+00 5.85471928e-01
2.52167314e-01 -9.11143005e-01 5.32972455e-01 5.63179851e-02
-1.49862945e+00 -1.94666341e-01 -2.74084210e-01 -6.53450415e-02
3.95411819e-01 2.73395777e-01 -8.99737239e-01 4.93075311e-01
1.10694921e+00 6.39766157e-01 -6.15802407e-01 1.27046537e+00
-6.92424104e-02 4.98009473e-01 -1.82009608e-01 3.34390700e-02
6.13983512e-01 -1.37309074e-01 1.26016116e+00 1.29708922e+00
5.58456063e-01 1.82780787e-01 6.06243312e-02 9.30409968e-01
7.04867125e-01 -8.24542157e-03 -3.63430113e-01 4.23821747e-01
5.71849763e-01 1.00269616e+00 -1.31176066e+00 -5.81482589e-01
-5.10429323e-01 6.43185794e-01 -1.94961742e-01 6.01261914e-01
-7.86506295e-01 -4.82747763e-01 2.94510186e-01 2.66162902e-01
7.96531320e-01 -2.27104306e-01 -6.55351043e-01 -7.00828195e-01
-2.86013912e-03 -5.44092834e-01 4.88734603e-01 -1.19073749e+00
-6.36963367e-01 4.89176571e-01 2.71017998e-01 -1.59935832e+00
6.43074214e-02 -6.23172373e-02 -8.66114795e-01 1.09626663e+00
-1.59210765e+00 -8.11174929e-01 -3.26443106e-01 4.86991495e-01
1.01287675e+00 -4.12588529e-02 2.28257895e-01 -2.86570549e-01
-2.35251307e-01 -3.01566739e-02 1.50045350e-01 -9.33536738e-02
2.68514186e-01 -1.21217704e+00 5.61318755e-01 1.40841770e+00
-5.83187118e-02 3.40253323e-01 1.15328014e+00 -9.27714825e-01
-1.00876427e+00 -8.16655397e-01 6.19840264e-01 8.95575210e-02
7.72214293e-01 5.80530874e-02 -1.52480459e+00 3.42707038e-01
2.31724650e-01 -2.34988168e-01 3.18102986e-01 -6.49005175e-01
1.80168957e-01 4.76831436e-01 -1.07975638e+00 6.79701805e-01
4.30167705e-01 -4.67634648e-01 -1.22603261e+00 -8.23846683e-02
7.48752058e-01 -5.77459335e-01 -3.30223203e-01 3.04402858e-01
2.30098903e-01 -1.00237453e+00 1.18629968e+00 -4.34603076e-03
4.64750350e-01 -8.24682951e-01 -1.46644026e-01 -1.19455051e+00
-6.20640099e-01 -4.33735758e-01 1.44992426e-01 6.54684246e-01
2.47203678e-01 -4.92661625e-01 5.39395928e-01 3.03560138e-01
-6.20214820e-01 -2.05666080e-01 -1.04043579e+00 -3.88626963e-01
-5.52998185e-01 -3.13283145e-01 4.35803443e-01 8.01822543e-01
9.84996110e-02 -6.54559284e-02 -2.81564593e-01 8.12022030e-01
8.96648228e-01 2.79374689e-01 1.81200132e-01 -1.04215991e+00
-6.35370389e-02 -5.34947157e-01 -5.64257145e-01 -8.77832294e-01
-3.28947246e-01 -4.76927519e-01 4.28353995e-01 -1.40607750e+00
2.77819455e-01 8.79829675e-02 -5.46452105e-01 2.68115163e-01
-6.29514337e-01 2.60663897e-01 1.30340502e-01 4.22682732e-01
-1.15637250e-01 6.38273716e-01 1.24771035e+00 -2.60416299e-01
-2.81517714e-01 -3.14012498e-01 -5.54007709e-01 9.19807911e-01
5.62051296e-01 -3.09811950e-01 -4.09593076e-01 -4.74706203e-01
-2.95127630e-01 2.53027380e-01 6.36858940e-01 -1.14795840e+00
3.73410195e-01 -2.53490925e-01 2.65593350e-01 -1.13811707e+00
1.35365829e-01 -5.44495344e-01 -4.03254963e-02 5.90530992e-01
-2.12578103e-01 1.62557721e-01 3.54466558e-01 6.68178260e-01
-4.54746366e-01 -2.89333731e-01 8.93905938e-01 -5.36219999e-02
-1.46575558e+00 -3.00260276e-01 -5.58878064e-01 -6.31895959e-02
9.80713487e-01 -8.12276483e-01 8.99336785e-02 -2.30700240e-01
-7.73959637e-01 1.38552219e-01 3.56814444e-01 3.69773626e-01
1.26709843e+00 -1.26294363e+00 -8.71723592e-01 4.26375449e-01
3.33288289e-03 4.32997607e-02 7.33278513e-01 1.03343761e+00
-7.97964931e-01 1.64452851e-01 -4.05852884e-01 -9.04585361e-01
-1.26766253e+00 5.49863219e-01 1.31671757e-01 2.11034477e-01
-9.04435933e-01 8.68428648e-01 6.69737875e-01 1.91403434e-01
-1.54154062e-01 -5.78082740e-01 -6.91884339e-01 1.50651142e-01
8.33076775e-01 4.13817734e-01 -1.35988608e-01 -1.12648094e+00
-4.81429994e-01 8.37097108e-01 2.61176199e-01 -4.09448355e-01
1.45382094e+00 -4.62740868e-01 1.02146104e-01 6.30551219e-01
6.35803878e-01 -1.35923699e-01 -1.70187199e+00 -4.36022878e-01
2.70315468e-01 -7.81732917e-01 1.74583465e-01 -2.38206655e-01
-1.10597599e+00 7.88529634e-01 5.39456189e-01 3.24021876e-01
1.38015497e+00 6.22841455e-02 5.16192913e-01 -1.74578980e-01
2.23266020e-01 -1.10704899e+00 3.14666063e-01 4.23421919e-01
1.29516447e+00 -1.04224372e+00 7.37105384e-02 -3.71499181e-01
-9.63500082e-01 1.18011534e+00 2.19475538e-01 -3.11800867e-01
3.39498699e-01 1.22110412e-01 -4.90036607e-03 -5.02540112e-01
-2.37612650e-01 -1.08440444e-01 5.54128885e-01 6.09826803e-01
9.01027992e-02 -1.16975062e-01 -9.71308500e-02 3.68184477e-01
-7.69351572e-02 -7.19940126e-01 7.23326325e-01 1.21969700e+00
-1.15878880e+00 -1.10718250e-01 -1.12011349e+00 1.45053551e-01
-1.34028152e-01 -3.42018872e-01 -5.34423776e-02 6.23419344e-01
-1.53671414e-01 7.53943861e-01 3.49098265e-01 -2.59644330e-01
3.52302909e-01 -2.90841937e-01 -1.62607417e-01 -4.96897578e-01
-3.84007841e-01 5.64432681e-01 -2.07862183e-01 -2.75074750e-01
-6.19899750e-01 -1.26626170e+00 -1.71109962e+00 3.40760171e-01
-2.10292205e-01 2.44150832e-01 3.32181811e-01 8.21950912e-01
8.80774781e-02 7.78540909e-01 6.64384425e-01 -1.78834987e+00
-1.24607302e-01 -9.70150828e-01 -7.84053028e-01 4.06609535e-01
7.02858150e-01 -1.09654105e+00 -5.96888125e-01 3.91997486e-01] | [9.676641464233398, -0.47169220447540283] |
2f03ef12-2da7-4c19-aa41-bd9b0bd63ad9 | adaptive-perception-transformer-for-temporal | 2208.11908 | null | https://arxiv.org/abs/2208.11908v2 | https://arxiv.org/pdf/2208.11908v2.pdf | Adaptive Perception Transformer for Temporal Action Localization | Temporal action localization aims to predict the boundary and category of each action instance in untrimmed long videos. Most of previous methods based on anchors or proposals neglect the global-local context interaction in entire video sequences. Besides, their multi-stage designs cannot generate action boundaries and categories straightforwardly. To address the above issues, this paper proposes a end-to-end model, called Adaptive Perception transformer (AdaPerFormer for short). Specifically, AdaPerFormer explores a dual-branch attention mechanism. One branch takes care of the global perception attention, which can model entire video sequences and aggregate global relevant contexts. While the other branch concentrates on the local convolutional shift to aggregate intra-frame and inter-frame information through our bidirectional shift operation. The end-to-end nature produces the boundaries and categories of video actions without extra steps. Extensive experiments together with ablation studies are provided to reveal the effectiveness of our design. Our method obtains competitive performance on the THUMOS14 and ActivityNet-1.3 dataset. | ['Hongfa Wang', 'Weibo Gu', 'Tianjin Zhang', 'Yizheng Ouyang'] | 2022-08-25 | null | null | null | null | ['action-localization'] | ['computer-vision'] | [ 4.75940883e-01 8.13455284e-02 -4.78734344e-01 -4.16310549e-01
-5.00226200e-01 -3.97692531e-01 6.48904502e-01 -4.27957922e-01
-3.25315148e-01 4.29595947e-01 7.72847176e-01 -9.40471217e-02
1.61198333e-01 -2.85885334e-01 -7.05014586e-01 -5.75842381e-01
-1.63315579e-01 -2.74035156e-01 7.24413991e-01 -5.11137843e-02
3.67639273e-01 2.64391042e-02 -1.41597593e+00 9.04007792e-01
6.20291591e-01 1.25483406e+00 4.73733515e-01 7.01294780e-01
6.17182106e-02 1.31975818e+00 -3.54885966e-01 -1.83471501e-01
2.43454382e-01 -6.30054057e-01 -7.29364634e-01 3.95254850e-01
5.71986914e-01 -7.63094127e-01 -5.71288049e-01 8.21488917e-01
4.24313068e-01 1.63053423e-01 1.52385548e-01 -1.37233150e+00
-7.18431473e-01 7.01329529e-01 -5.57533801e-01 5.70509613e-01
5.01010954e-01 5.03877103e-01 1.04560924e+00 -1.06076276e+00
4.54408377e-01 1.31105042e+00 4.58809495e-01 5.09108305e-01
-7.66335487e-01 -5.71316600e-01 7.65307844e-01 7.11129189e-01
-1.14446974e+00 -5.36766350e-01 7.10010350e-01 -4.11656916e-01
9.82673407e-01 -1.83269754e-03 6.95411742e-01 1.44123781e+00
3.65500391e-01 1.15834153e+00 7.47125566e-01 -6.50726259e-02
1.58022910e-01 -6.16593540e-01 -1.13768287e-01 5.44572711e-01
-2.33424768e-01 1.34968489e-01 -8.33121121e-01 2.18060762e-01
1.04093254e+00 2.81416506e-01 -4.75151956e-01 -2.44606495e-01
-1.74367869e+00 2.90673584e-01 5.11888921e-01 2.70976454e-01
-5.83415926e-01 5.79872429e-01 4.49347496e-01 1.29928723e-01
1.88516751e-01 1.04892932e-01 -4.47870284e-01 -4.32329357e-01
-8.13641608e-01 8.10577124e-02 1.41447544e-01 1.25416255e+00
4.37096000e-01 1.27431098e-02 -6.53744519e-01 4.86503154e-01
2.18234792e-01 8.89209360e-02 5.86097777e-01 -1.36186576e+00
6.30111337e-01 5.20256460e-01 2.28532419e-01 -8.95614684e-01
-2.28645921e-01 -4.47714657e-01 -5.14747024e-01 -6.32720068e-02
2.95065761e-01 -9.75825638e-02 -1.06126881e+00 1.78935921e+00
2.41550073e-01 7.16692865e-01 -5.34740388e-02 1.20685315e+00
6.90353930e-01 5.97294509e-01 4.25941974e-01 -1.91607356e-01
1.38670421e+00 -1.52788019e+00 -9.37945247e-01 -4.66404587e-01
3.91504407e-01 -5.77867031e-01 1.15724862e+00 2.85661399e-01
-1.16691840e+00 -9.76988256e-01 -8.79268587e-01 -2.50668943e-01
-6.36164099e-02 2.52248228e-01 5.66728950e-01 -6.61376640e-02
-1.01196384e+00 3.20854068e-01 -9.42290068e-01 -2.82945991e-01
5.67463279e-01 1.80786133e-01 -2.71948665e-01 2.32098680e-02
-1.06261408e+00 4.14738774e-01 3.67321581e-01 2.16674700e-01
-1.28902721e+00 -4.62020278e-01 -7.76320219e-01 1.98919877e-01
8.53974640e-01 -5.93568921e-01 1.44404137e+00 -1.51720476e+00
-1.35697472e+00 3.71010035e-01 -2.71093190e-01 -4.65237677e-01
4.50930893e-01 -5.64155221e-01 -2.33662263e-01 3.91198546e-01
2.65593022e-01 9.96427059e-01 8.14469635e-01 -8.93043458e-01
-1.13455939e+00 -3.23196277e-02 4.37662661e-01 4.04062450e-01
-2.42232308e-01 4.06674072e-02 -7.66317129e-01 -1.01939201e+00
1.52605787e-01 -7.29062259e-01 -2.50395089e-01 4.24165204e-02
-1.30595222e-01 -2.78500319e-01 8.39141548e-01 -6.34280920e-01
1.31331778e+00 -2.28017473e+00 2.37431511e-01 -4.13775325e-01
1.67314693e-01 2.50781346e-02 -3.39559227e-01 2.79046595e-01
-2.34216705e-01 -1.00066692e-01 -3.61203663e-02 -2.01308504e-01
1.36850439e-02 1.47203416e-01 -5.11818290e-01 2.88673908e-01
3.08911145e-01 9.93400514e-01 -1.18359542e+00 -5.42233765e-01
2.44810611e-01 3.18711489e-01 -9.37060893e-01 3.18496078e-01
-2.58584619e-01 4.66702938e-01 -5.71247935e-01 9.60048199e-01
2.49924928e-01 -1.42922148e-01 1.99172795e-01 -4.13405031e-01
-7.50493109e-02 2.80937016e-01 -8.07981312e-01 2.13615513e+00
-1.89672679e-01 6.52329981e-01 -1.38364315e-01 -8.68317783e-01
4.11238581e-01 4.20431942e-01 5.71090579e-01 -7.26027727e-01
4.46179397e-02 -2.54039299e-02 1.28824696e-01 -6.80777371e-01
5.58057547e-01 1.41954422e-01 -1.50087401e-01 6.11278415e-02
1.89795896e-01 8.66272628e-01 1.91920862e-01 1.47636041e-01
1.36802423e+00 8.88309538e-01 3.53939265e-01 -1.07231036e-01
4.74204957e-01 -1.17271088e-01 1.01191342e+00 6.24886096e-01
-7.39649355e-01 6.60172939e-01 6.26885593e-01 -5.25284350e-01
-7.06262827e-01 -9.61288750e-01 5.64058363e-01 1.55792475e+00
5.21260977e-01 -5.29263973e-01 -7.55206645e-01 -9.98486102e-01
-4.09863144e-01 5.86428523e-01 -7.32345283e-01 -1.84285536e-01
-7.59171128e-01 -1.80777729e-01 2.74614960e-01 9.87891614e-01
9.13882971e-01 -1.39325523e+00 -9.02604342e-01 3.81358624e-01
-7.86045849e-01 -1.35886002e+00 -9.14400101e-01 -2.40112275e-01
-7.39021182e-01 -1.23907065e+00 -5.88958681e-01 -6.56227350e-01
4.48654056e-01 4.89465833e-01 1.00056756e+00 -1.28464326e-01
1.68710768e-01 3.17796290e-01 -7.43669510e-01 1.15227893e-01
5.12391739e-02 -2.17960939e-01 -1.04320914e-01 4.39999104e-01
5.12113690e-01 -6.30873322e-01 -1.13306618e+00 7.08743334e-01
-6.73645258e-01 3.07638258e-01 8.78502309e-01 6.69361532e-01
6.45878494e-01 -1.60861865e-01 5.37996411e-01 -3.81406605e-01
1.78063333e-01 -5.78994989e-01 -1.30256832e-01 3.24232072e-01
-1.07052051e-01 -2.73263544e-01 5.62417150e-01 -5.74504614e-01
-1.15977728e+00 2.28402406e-01 9.95158404e-02 -8.37762952e-01
-2.92803168e-01 2.89821059e-01 -4.86586273e-01 3.97771716e-01
3.61863106e-01 4.39014107e-01 -4.16803151e-01 -3.82232010e-01
2.95893252e-01 5.14600635e-01 8.93367231e-01 -3.95007044e-01
2.03801706e-01 5.44881582e-01 -3.82217050e-01 -4.15043443e-01
-8.05512369e-01 -4.16373253e-01 -6.68389022e-01 -5.78825116e-01
1.14947820e+00 -1.17625105e+00 -6.73343122e-01 5.01138031e-01
-1.10430205e+00 -6.78826332e-01 -1.42788634e-01 4.74462748e-01
-8.83987844e-01 5.22880852e-01 -4.87272382e-01 -5.66204250e-01
-4.88055684e-02 -1.20623815e+00 1.18743908e+00 1.54363811e-01
-2.16078669e-01 -5.61136246e-01 -3.79962027e-01 4.49037641e-01
1.87665880e-01 1.40721276e-01 3.01447302e-01 -3.49659175e-01
-9.93063748e-01 3.95209193e-02 -2.00061351e-01 2.16517627e-01
1.43896684e-01 -1.64806232e-01 -7.53571451e-01 -2.31848523e-01
-1.22075535e-01 -2.14718997e-01 9.22536075e-01 4.06599998e-01
1.55494225e+00 -3.34077120e-01 -2.59381890e-01 5.78742087e-01
1.15924621e+00 4.77230191e-01 9.06162977e-01 2.04315856e-01
8.32588732e-01 3.61841321e-01 1.12742209e+00 3.94657612e-01
4.79006499e-01 8.50018263e-01 6.74508870e-01 2.32715234e-02
-3.99401277e-01 -4.50992733e-01 8.20179045e-01 4.56363887e-01
-3.46354514e-01 -4.33602393e-01 -6.39775693e-01 7.27441072e-01
-2.19683123e+00 -1.35477245e+00 2.67168254e-01 1.83756912e+00
4.86981690e-01 3.45205367e-01 2.16038004e-01 -1.73239306e-01
7.35802650e-01 4.03429121e-01 -6.15852654e-01 -1.14868648e-01
1.55054340e-02 -2.69624054e-01 3.19959730e-01 2.67986119e-01
-1.25392091e+00 1.09759092e+00 5.76508713e+00 9.05260563e-01
-8.66889596e-01 2.63755202e-01 6.61981106e-01 -2.62794048e-01
1.05215847e-01 1.80453703e-01 -5.85502744e-01 7.06091404e-01
5.20713747e-01 2.39180848e-01 3.36326241e-01 8.20418358e-01
4.62600082e-01 -2.00492606e-01 -1.26613998e+00 1.01571286e+00
5.68880327e-02 -1.17853498e+00 1.33006081e-01 8.79212096e-03
6.29965246e-01 -1.13664411e-01 -1.18937679e-01 4.11655247e-01
7.16781095e-02 -7.16834664e-01 1.30261970e+00 7.42611110e-01
7.52139032e-01 -5.58733642e-01 5.92454910e-01 1.89635068e-01
-1.50410807e+00 -6.05884373e-01 -1.30817831e-01 -2.91462153e-01
3.97224873e-01 -6.82365336e-03 -3.74249429e-01 3.29819739e-01
7.96544373e-01 1.16008914e+00 -5.47306895e-01 9.63596225e-01
-3.87491703e-01 8.19147706e-01 3.77722010e-02 2.26817399e-01
5.70699215e-01 7.57621080e-02 4.96887147e-01 1.21275711e+00
4.10755426e-01 3.42148304e-01 2.96760023e-01 5.57578802e-01
4.95258570e-02 -3.93325806e-01 -2.61418104e-01 -5.27312309e-02
4.08356309e-01 1.06995690e+00 -6.68914914e-01 -5.93214571e-01
-5.98171830e-01 1.17757010e+00 8.18369985e-02 6.16558492e-01
-1.31817853e+00 -2.76568502e-01 7.55022824e-01 1.71199277e-01
6.68471694e-01 -1.87963113e-01 -1.87553927e-01 -1.19333303e+00
1.66685343e-01 -1.03302228e+00 5.91906846e-01 -1.06610680e+00
-8.55709314e-01 4.65069473e-01 9.96119156e-02 -1.56050622e+00
-9.66726542e-02 -3.20251316e-01 -7.21341252e-01 4.08597678e-01
-1.27926993e+00 -1.18510818e+00 -5.38454831e-01 6.27114594e-01
1.24787211e+00 -1.12002246e-01 2.55077869e-01 2.95551330e-01
-6.26049399e-01 5.32680213e-01 -4.14103687e-01 2.67542064e-01
6.53786063e-01 -1.04035807e+00 3.34852636e-01 1.27353442e+00
-1.12771809e-01 4.18985248e-01 5.63839734e-01 -6.41996145e-01
-1.04292047e+00 -1.14987314e+00 8.33145738e-01 -4.48556334e-01
6.20510519e-01 -2.45957866e-01 -6.84265137e-01 8.77171278e-01
4.46795791e-01 1.03554919e-01 4.77208912e-01 -2.46349007e-01
-2.83044487e-01 -1.64246276e-01 -6.47298634e-01 9.00542200e-01
1.76811767e+00 -3.41993839e-01 -6.70355916e-01 1.05896592e-01
8.59196365e-01 -3.02591652e-01 -7.00432420e-01 5.36614120e-01
6.18361533e-01 -1.33071518e+00 9.38929021e-01 -6.63295865e-01
1.00089025e+00 -6.04215324e-01 -2.88594604e-01 -8.42841446e-01
-6.94768965e-01 -8.06796670e-01 -4.39396113e-01 1.11517847e+00
6.60553621e-03 -2.23294079e-01 5.99144280e-01 3.31598550e-01
-6.36180699e-01 -9.88042533e-01 -8.24565411e-01 -5.07954597e-01
-4.43374068e-01 -4.57267165e-01 5.97669542e-01 6.96860135e-01
1.36267738e-02 2.54603237e-01 -6.56315625e-01 9.92549211e-02
4.38917220e-01 2.66563535e-01 7.10127652e-01 -4.03391093e-01
-3.99110824e-01 -5.04309893e-01 -5.44425845e-01 -1.68620813e+00
-5.82020879e-02 -3.48527789e-01 2.74750501e-01 -1.34737599e+00
2.66850919e-01 1.70964655e-02 -6.31574392e-01 7.59061277e-01
-3.66452038e-01 3.19997102e-01 2.98128575e-01 2.13996336e-01
-1.27049208e+00 7.51207471e-01 1.49338937e+00 5.11595921e-04
-2.85443142e-02 -3.05009574e-01 -6.62188530e-01 7.87167966e-01
7.80784667e-01 -1.64829314e-01 -8.06675315e-01 -5.69392443e-01
-2.26692274e-01 1.51116744e-01 6.99338794e-01 -1.24661505e+00
1.54409572e-01 -5.72708666e-01 3.68612915e-01 -7.21718490e-01
4.07041013e-01 -7.26826847e-01 -7.40842149e-02 3.47321838e-01
-5.34352481e-01 1.36945739e-01 -1.50962636e-01 8.69076610e-01
-3.32161129e-01 3.41117799e-01 5.76905251e-01 -1.69765174e-01
-1.46606386e+00 4.51635569e-01 -4.84462529e-01 -6.00767210e-02
1.25300658e+00 -4.98471588e-01 -2.84172386e-01 -4.55556005e-01
-8.18427801e-01 3.89274091e-01 3.92076969e-01 7.33863294e-01
7.34600127e-01 -1.55623436e+00 -4.62071836e-01 9.42938849e-02
6.55339360e-02 -2.93211751e-02 5.13946414e-01 1.05377340e+00
-3.42738956e-01 4.41229999e-01 -4.04308140e-01 -6.53717756e-01
-1.23231792e+00 7.55006075e-01 4.75625873e-01 -2.38050465e-02
-5.22063375e-01 9.27356660e-01 8.25510144e-01 3.25114787e-01
4.29315358e-01 -4.77784872e-01 -2.30552569e-01 -1.02322847e-01
6.82239652e-01 1.89478993e-01 -5.58241367e-01 -8.13924968e-01
-4.25944686e-01 3.57631892e-01 3.38871256e-02 3.99749354e-02
1.00901914e+00 -4.18718606e-01 1.54061750e-01 2.92014509e-01
8.89223874e-01 -3.96973163e-01 -2.14199781e+00 -1.83041140e-01
-1.62178963e-01 -8.95095825e-01 -1.36917263e-01 -8.55011284e-01
-1.12675154e+00 8.26882541e-01 4.49281961e-01 -6.57022372e-02
1.60540688e+00 -5.02115451e-02 8.35926831e-01 1.36310145e-01
3.75573993e-01 -1.17059386e+00 5.57151914e-01 4.08684224e-01
1.09145010e+00 -1.05812430e+00 -1.86596692e-01 -4.27132279e-01
-8.26024592e-01 8.64489019e-01 1.13758659e+00 1.09430272e-02
3.44112962e-01 -1.09613530e-01 -8.47344287e-04 3.13207041e-03
-1.11934769e+00 -2.96835274e-01 1.69095203e-01 5.59336483e-01
2.55253285e-01 -2.01711729e-01 -3.40554684e-01 8.15391243e-01
3.72124076e-01 2.25140229e-01 3.86575133e-01 1.04213011e+00
-5.06542563e-01 -8.06252241e-01 -1.74626932e-01 2.42538825e-02
-5.19386530e-01 -2.46361271e-02 -3.40171993e-01 5.60787261e-01
5.79469144e-01 1.00149798e+00 2.25459367e-01 -6.88914597e-01
2.20715195e-01 -4.72632349e-02 2.43461534e-01 -3.64282489e-01
-3.70794952e-01 4.35593992e-01 8.34648535e-02 -1.24317241e+00
-7.65520692e-01 -6.56543195e-01 -1.27134478e+00 -4.61298637e-02
-3.69340591e-02 -1.84385762e-01 3.71556841e-02 9.11085427e-01
6.74373567e-01 7.82873213e-01 5.74536502e-01 -1.13248956e+00
-3.04884076e-01 -1.06106877e+00 -2.21175924e-01 5.52688539e-01
3.84618700e-01 -8.66293311e-01 -4.75292839e-02 4.96830404e-01] | [8.5224027633667, 0.5764443278312683] |
24cd2f32-713e-4b0f-9115-e8f90f734f55 | spatiotemporal-multi-graph-convolution | null | null | http://202.119.24.249/cache/8/03/www-scf.usc.edu/59dc738bd683a198ef69fb39196799d7/aaai19_multi_graph_convolution.pdf | http://202.119.24.249/cache/8/03/www-scf.usc.edu/59dc738bd683a198ef69fb39196799d7/aaai19_multi_graph_convolution.pdf | Spatiotemporal Multi-Graph Convolution Networkfor Ride-hailing Demand Forecasting | Region-level demand forecasting is an essential task in ridehailing services. Accurate ride-hailing demand forecasting can guide vehicle dispatching, improve vehicle utilization, reduce the wait-time, and mitigate traffic congestion. This task is challenging due to the complicated spatiotemporal dependencies among regions. Existing approaches mainly focus on modeling the Euclidean correlations among spatially adjacent regions while we observe that non-Euclidean pair-wise correlations among possibly distant regions are also critical for accurate forecasting. In this paper, we propose the spatiotemporal multi-graph convolution network (ST-MGCN), a novel deep learning model for ride-hailing demand forecasting. We first encode the non-Euclidean pair-wise correlations among regions into multiple graphs and then explicitly model these correlations using multi-graph convolution. To utilize the global contextual information in modeling the temporal correlation, we further propose contextual gated recurrent neural network which augments recurrent neural network with a contextual-aware gating mechanism to re-weights different historical observations. We evaluate the proposed model on two real-world large scale ride-hailing demand datasets and observe consistent improvement of more than 10% over stateof-the-art baselines. | ['4', '4 Yan Liu 2', '1 Jieping Ye', '4 Qiang Yang', '3 Lingyu Zhang', '1', '∗2 Leye Wang', '∗1 Yaguang Li', 'Xu Geng'] | 2019-01-20 | null | null | null | conference-2019-1 | ['spatio-temporal-forecasting'] | ['time-series'] | [-5.61208487e-01 -5.34883201e-01 -5.41242242e-01 -6.00269854e-01
-4.67515379e-01 -3.49914104e-01 7.16605723e-01 6.18080907e-02
-2.94872038e-02 6.39141917e-01 6.45299017e-01 -6.07261479e-01
-2.16822088e-01 -1.00719523e+00 -7.16520786e-01 -5.39180696e-01
-5.09996474e-01 5.03457367e-01 3.56176525e-01 -6.57260060e-01
1.58419222e-01 8.59408498e-01 -1.46457911e+00 2.15769082e-01
8.79300237e-01 9.25454676e-01 -6.25049397e-02 6.08784080e-01
-4.62361909e-02 1.08756697e+00 -4.18508053e-01 6.63403273e-02
-1.85113139e-02 -1.65704086e-01 -4.04880762e-01 -1.75803676e-01
1.66906893e-01 -3.73545557e-01 -1.00077081e+00 5.20515025e-01
2.07076147e-01 7.74928331e-01 4.83989477e-01 -1.48524559e+00
-7.96147645e-01 3.22082013e-01 -6.35619223e-01 8.54545474e-01
-2.24603981e-01 1.17112972e-01 8.83343577e-01 -5.76412261e-01
3.34631622e-01 1.14982629e+00 4.67588603e-01 1.84327409e-01
-8.49576533e-01 -7.80893505e-01 7.83739448e-01 8.00330937e-01
-1.48670077e+00 -2.38643885e-01 9.81136739e-01 -2.34898359e-01
1.33461618e+00 1.88536912e-01 4.78592306e-01 5.87537467e-01
4.61772889e-01 7.62076378e-01 4.83987302e-01 3.27412903e-01
-1.83144540e-01 -4.28934366e-01 1.75805584e-01 4.11208868e-01
-2.32835427e-01 1.19226336e-01 -4.27774101e-01 2.93985307e-01
3.68881524e-01 4.88186061e-01 -7.00670928e-02 9.99290720e-02
-9.65910256e-01 7.86530435e-01 9.38369036e-01 1.02070086e-01
-6.42590702e-01 3.77304286e-01 5.40037572e-01 1.66565821e-01
9.99944448e-01 -9.87052396e-02 -2.12463245e-01 -1.91500694e-01
-1.05627346e+00 2.70162523e-01 2.62641191e-01 8.50797653e-01
8.61978114e-01 3.77337128e-01 -2.49214500e-01 6.89208448e-01
1.18351832e-01 5.64494729e-01 1.74663931e-01 -3.83185714e-01
7.70816565e-01 5.47934413e-01 6.46120310e-02 -1.42353332e+00
-1.03503239e+00 -5.76510966e-01 -9.60209131e-01 -3.94412339e-01
3.23359556e-02 -3.00331235e-01 -9.28352892e-01 1.71736300e+00
2.08962694e-01 7.54554212e-01 -4.23526794e-01 1.20412612e+00
7.88133442e-01 1.11550605e+00 7.43727684e-02 -5.23097664e-02
7.70840347e-01 -1.40251100e+00 -9.21195567e-01 7.23925186e-03
9.11262691e-01 -3.59809965e-01 6.15559936e-01 -1.18246555e-01
-8.50884497e-01 -5.12493014e-01 -7.51827657e-01 -2.27304220e-01
-8.44894171e-01 4.66633998e-02 5.68212509e-01 6.40901923e-02
-1.14368057e+00 2.02550039e-01 -6.70524538e-01 -4.65361252e-02
2.43126780e-01 3.37000728e-01 -8.72802734e-02 -2.14898482e-01
-1.49575698e+00 9.16888714e-01 -1.16876997e-01 5.06573379e-01
-8.31010282e-01 -7.70330608e-01 -9.62144196e-01 9.54700708e-02
3.78735095e-01 -2.07499951e-01 7.78017104e-01 -3.41431826e-01
-1.02431834e+00 3.13172102e-01 -7.49521613e-01 -5.10567129e-01
1.13130033e-01 8.77159089e-02 -1.37953579e+00 -2.95880258e-01
4.60459813e-02 3.65028888e-01 3.76937598e-01 -9.87379134e-01
-5.79317391e-01 -3.43014359e-01 -8.59360248e-02 1.39156282e-01
9.74315852e-02 -5.10410592e-02 -5.49096167e-01 -7.81795382e-01
-1.82636455e-01 -1.00270796e+00 -2.03040361e-01 -6.71643078e-01
-3.06954116e-01 -6.33792281e-01 1.34886110e+00 -6.92507148e-01
1.63808703e+00 -2.03778696e+00 -2.93292910e-01 4.05218244e-01
4.21427429e-01 2.19099328e-01 -3.65014851e-01 6.15196824e-01
-1.25045506e-02 -1.41553313e-01 1.98346242e-01 -4.16374803e-01
-2.13180599e-03 5.47210693e-01 -5.08872330e-01 7.65040338e-01
3.29963297e-01 1.40304375e+00 -1.00047779e+00 9.24202278e-02
4.16558892e-01 6.40227377e-01 -4.27040368e-01 1.78367808e-01
-8.82994980e-02 4.40562546e-01 -3.29625785e-01 3.60382020e-01
9.18974161e-01 -2.28965372e-01 -9.96405408e-02 -1.86728299e-01
-3.34388584e-01 7.22526670e-01 -6.90601647e-01 1.22332144e+00
-6.62375808e-01 1.09554315e+00 -2.64725983e-01 -1.21546984e+00
1.10316241e+00 -7.38294050e-02 4.93945241e-01 -1.38416803e+00
-7.21823871e-02 -3.54209170e-02 -3.53985205e-02 -2.80084401e-01
1.12017214e+00 1.99592456e-01 -2.08475009e-01 5.50885379e-01
-4.18410808e-01 4.28614438e-01 6.90138340e-02 3.64262164e-01
6.98253870e-01 -2.24426791e-01 -4.29656386e-01 -3.71841371e-01
3.27269316e-01 -2.04909310e-01 7.02179909e-01 1.22098662e-01
-2.49600112e-01 4.86951947e-01 5.82748711e-01 -8.73010397e-01
-7.85313189e-01 -8.73219728e-01 3.61268252e-01 1.20339179e+00
3.86378050e-01 -2.62506723e-01 -1.80050269e-01 -4.05140847e-01
2.09678531e-01 8.54723275e-01 -9.33623254e-01 -2.25951090e-01
-8.92768562e-01 -6.24763787e-01 3.96311402e-01 7.77652025e-01
4.08529371e-01 -7.28078961e-01 -6.40256377e-03 3.82214367e-01
-1.09803796e-01 -1.09755003e+00 -1.09953249e+00 -1.49253458e-01
-3.51908118e-01 -7.20015347e-01 -5.70675731e-01 -6.09779894e-01
5.01055956e-01 9.19163227e-01 1.23860097e+00 2.89397806e-01
1.81713358e-01 -2.32893601e-01 -8.71806145e-02 -7.86175281e-02
1.70230687e-01 4.19305414e-01 -3.10253706e-02 2.66938955e-01
6.14418507e-01 -6.05046749e-01 -8.75425041e-01 7.22935081e-01
-5.91682315e-01 1.24681786e-01 -1.94994397e-02 6.25253737e-01
6.82362139e-01 5.48353828e-02 1.11577046e+00 -6.97972596e-01
8.10851812e-01 -1.14467311e+00 -5.60368776e-01 1.47231594e-01
-5.00122607e-01 -8.99956897e-02 8.74766767e-01 -1.24762781e-01
-8.30669761e-01 -6.16641045e-01 -1.10653639e-01 -6.44140124e-01
3.24735232e-02 6.87194824e-01 9.89181548e-02 1.82550266e-01
-9.43878517e-02 2.72635162e-01 -3.85643423e-01 -1.50829524e-01
6.81976974e-01 3.88312250e-01 4.70795989e-01 -1.67057708e-01
4.02873069e-01 4.93065149e-01 1.17291018e-01 -5.72862864e-01
-4.73787874e-01 -6.88116729e-01 -4.59795296e-01 -4.23485249e-01
6.54615879e-01 -1.06541324e+00 -1.03515065e+00 3.82077336e-01
-9.77952540e-01 -7.23149180e-01 2.24378601e-01 4.20797557e-01
-2.45020837e-01 3.09985057e-02 -6.13794744e-01 -8.43400240e-01
-1.73630863e-02 -1.13757730e+00 9.91596103e-01 1.26935378e-01
2.08911270e-01 -1.38296378e+00 4.37533446e-02 1.31168768e-01
7.93956459e-01 2.62674987e-01 7.78781056e-01 -4.06409472e-01
-4.76626962e-01 -1.76859468e-01 -8.72698247e-01 -2.10564047e-01
1.78924710e-01 -5.41348495e-02 -5.31560063e-01 -8.98795724e-02
-8.33638430e-01 1.37711316e-01 1.14475107e+00 5.53326964e-01
1.44160366e+00 -3.76794994e-01 -5.09676695e-01 6.07607961e-01
9.97445464e-01 2.03900397e-01 5.60770035e-01 -8.61819834e-02
1.17693055e+00 7.52477527e-01 4.97016460e-01 5.18017888e-01
1.42468238e+00 7.88864851e-01 5.36093295e-01 -2.74632841e-01
-2.43007720e-01 -5.21667123e-01 1.00939997e-01 1.13051367e+00
-1.03501072e-02 -6.56118155e-01 -9.65879917e-01 1.16393781e+00
-2.09529400e+00 -1.18225563e+00 -5.30913055e-01 1.80105400e+00
9.69134569e-02 1.24387398e-01 5.15410542e-01 -3.20445508e-01
6.64688706e-01 6.13505661e-01 -7.24005342e-01 -8.82756531e-01
-2.94254005e-01 -3.66280228e-01 1.00201380e+00 7.73082376e-01
-7.90336847e-01 1.02570832e+00 5.86758041e+00 7.51097977e-01
-1.44925570e+00 1.71920761e-01 8.84232700e-01 -4.70001549e-01
-7.05497444e-01 -4.00050849e-01 -6.91668808e-01 7.38495290e-01
1.31169689e+00 -1.27934083e-01 8.26782763e-01 4.52858239e-01
9.21515167e-01 4.84395437e-02 -6.07454062e-01 8.36932540e-01
1.13092877e-01 -1.68692052e+00 2.37786192e-02 1.65791720e-01
1.01954627e+00 7.33556926e-01 2.18784571e-01 3.87227356e-01
3.64470959e-01 -1.22036254e+00 4.75484252e-01 7.25237370e-01
4.59394753e-01 -1.34927654e+00 4.86773133e-01 1.99430242e-01
-1.85508275e+00 -3.07510514e-03 -1.45270810e-01 -1.84218705e-01
7.12523222e-01 5.95960915e-01 -5.64613104e-01 5.47140062e-01
4.47700143e-01 1.29604769e+00 -3.19728315e-01 7.59499013e-01
4.84224297e-02 6.26207530e-01 -2.10126966e-01 4.08655927e-02
7.46422827e-01 -3.78694743e-01 2.26469785e-01 1.34547484e+00
1.70198247e-01 1.95646342e-02 9.27272737e-02 6.67850435e-01
-2.74550021e-01 -1.92820340e-01 -6.87061548e-01 1.45382002e-01
3.76054317e-01 9.45061743e-01 -4.10015225e-01 -3.45219105e-01
-7.08866835e-01 7.38610566e-01 3.78634721e-01 8.44244123e-01
-1.24923289e+00 -4.45073098e-01 1.25477350e+00 2.31315047e-01
4.40346330e-01 -8.20718348e-01 -2.00862259e-01 -9.31111217e-01
-2.91215610e-02 -1.19120106e-01 4.21856761e-01 -7.12720811e-01
-1.17809689e+00 6.18706405e-01 -1.05384126e-01 -1.17893136e+00
-2.81423301e-01 -1.55306429e-01 -9.79927659e-01 1.01485825e+00
-2.45766950e+00 -1.30529368e+00 -2.60253221e-01 8.79526496e-01
6.01164103e-01 7.42077529e-02 1.45120487e-01 7.69794106e-01
-9.13665652e-01 6.41166151e-01 2.06581816e-01 1.70518637e-01
4.91246074e-01 -8.22567403e-01 1.02324045e+00 8.84571791e-01
-2.70182133e-01 3.45988572e-01 3.51563871e-01 -6.40596211e-01
-1.39012396e+00 -1.63549936e+00 1.22725725e+00 -1.39746889e-01
9.01708603e-01 -3.84097576e-01 -1.15629649e+00 7.25347519e-01
1.78681925e-01 4.60535794e-01 4.71164197e-01 1.58097073e-01
-3.83534253e-01 -3.92974973e-01 -6.63356125e-01 5.32828033e-01
1.11648548e+00 -8.34636271e-01 8.93050898e-03 2.81933337e-01
8.55149746e-01 -2.21304640e-01 -6.79005325e-01 1.49583474e-01
3.94360989e-01 -7.86652505e-01 6.87775314e-01 -7.13064253e-01
-8.19945242e-03 -3.40767115e-01 4.13779691e-02 -1.70121706e+00
-3.56044531e-01 -5.35493016e-01 -1.64345920e-01 8.47319603e-01
6.01546526e-01 -9.22603369e-01 6.98603868e-01 7.55838752e-01
-6.00986302e-01 -8.93608212e-01 -9.08332944e-01 -6.36822402e-01
4.58180122e-02 -7.80889511e-01 1.21019495e+00 9.99630749e-01
1.76224839e-02 1.16314277e-01 -8.24465156e-01 2.73413807e-01
1.72268152e-01 3.71805131e-01 6.62743866e-01 -7.17972100e-01
2.84034729e-01 -7.69735277e-01 -3.89217407e-01 -1.62093067e+00
4.40146208e-01 -8.04290771e-01 -5.31538166e-02 -1.68824828e+00
-3.42310458e-01 -6.23969257e-01 -5.73234260e-01 2.53213882e-01
-2.68080309e-02 2.10356921e-01 3.00103091e-02 3.07286121e-02
-8.97855341e-01 8.37576628e-01 1.37087202e+00 -1.56274602e-01
-2.55150259e-01 -1.25793666e-01 -2.59666413e-01 7.15041626e-03
9.16938961e-01 -4.85677645e-02 -6.37344956e-01 -7.92707980e-01
4.68498200e-01 1.22504726e-01 2.01292008e-01 -6.13275468e-01
5.62229395e-01 -4.58433330e-01 -1.03680812e-01 -1.23791850e+00
3.21514904e-01 -6.37299478e-01 5.36957607e-02 -1.33406939e-02
-2.18427166e-01 7.42861211e-01 4.94712025e-01 7.78166473e-01
-4.32540298e-01 8.99893165e-01 1.50350034e-01 4.86314297e-01
-1.17416751e+00 8.50471437e-01 -5.89423478e-01 -8.87890831e-02
9.92490232e-01 -8.02383423e-02 -7.10825741e-01 -7.83256650e-01
-4.80608821e-01 9.68589664e-01 6.73090592e-02 9.20484185e-01
6.60320461e-01 -1.63285398e+00 -7.15486586e-01 4.12391037e-01
1.59450129e-01 -2.34036148e-01 7.97197163e-01 1.11623406e+00
-3.37673128e-01 8.19243729e-01 -1.40594440e-02 -2.95658857e-01
-7.03090847e-01 7.00566709e-01 3.13912630e-01 -3.94312054e-01
-4.33536142e-01 6.46517754e-01 3.06885123e-01 -2.87161201e-01
-1.46405488e-01 -6.87169254e-01 -2.45044321e-01 1.15364529e-01
4.80003148e-01 6.93830729e-01 2.65940815e-01 -1.28070128e+00
-6.35270655e-01 5.57182431e-01 1.42489551e-02 4.19800878e-01
1.30842781e+00 -6.09818995e-01 1.61632359e-01 3.85957509e-01
1.59758210e+00 -2.64981061e-01 -1.32669926e+00 -3.17632914e-01
-1.96045265e-01 -4.22427386e-01 6.25228405e-01 -4.99329686e-01
-1.65850496e+00 1.17750430e+00 1.42125919e-01 4.97606575e-01
1.15181363e+00 -9.87847522e-02 1.54763019e+00 -1.32051045e-02
1.80524260e-01 -1.12437713e+00 -3.00288171e-01 6.99921668e-01
7.84447312e-01 -1.35835421e+00 -3.14935982e-01 -1.22189447e-01
-7.63082147e-01 9.41160560e-01 4.14844960e-01 -4.50027764e-01
1.06278646e+00 -6.13774247e-02 -1.16917744e-01 -3.46438736e-01
-1.05204153e+00 -5.58793247e-01 6.73005581e-01 3.61615270e-01
2.78118432e-01 5.68907142e-01 4.56607528e-02 3.25320870e-01
8.21438208e-02 -1.01600923e-01 1.81862205e-01 2.93899924e-01
-1.43951461e-01 -7.01091051e-01 1.83799565e-01 6.10323012e-01
3.03046722e-02 -2.14161035e-02 1.17643163e-01 6.57790184e-01
-1.40068829e-01 1.18459177e+00 7.19523847e-01 -7.27887154e-01
4.12829578e-01 -5.16204000e-01 -1.61415949e-01 -1.91318229e-01
-4.30679917e-01 -1.42812654e-01 1.90027669e-01 -7.64876068e-01
-2.97221094e-01 -5.53770244e-01 -1.35826576e+00 -9.87225473e-01
4.10312153e-02 9.77599919e-02 5.88700354e-01 1.07135069e+00
9.27349448e-01 7.18033314e-01 9.51740384e-01 -9.27942276e-01
4.08416241e-01 -6.06220722e-01 -6.57944858e-01 3.61643851e-01
9.85887527e-01 -9.30305898e-01 -1.09668620e-01 -5.06378710e-01] | [6.462438106536865, 2.0686116218566895] |
d9d13fa6-0c8d-4fa9-a4e7-d8be597a5243 | exploring-the-robustness-of-large-language | 2306.14583 | null | https://arxiv.org/abs/2306.14583v1 | https://arxiv.org/pdf/2306.14583v1.pdf | Exploring the Robustness of Large Language Models for Solving Programming Problems | Using large language models (LLMs) for source code has recently gained attention. LLMs, such as Transformer-based models like Codex and ChatGPT, have been shown to be highly capable of solving a wide range of programming problems. However, the extent to which LLMs understand problem descriptions and generate programs accordingly or just retrieve source code from the most relevant problem in training data based on superficial cues has not been discovered yet. To explore this research question, we conduct experiments to understand the robustness of several popular LLMs, CodeGen and GPT-3.5 series models, capable of tackling code generation tasks in introductory programming problems. Our experimental results show that CodeGen and Codex are sensitive to the superficial modifications of problem descriptions and significantly impact code generation performance. Furthermore, we observe that Codex relies on variable names, as randomized variables decrease the solved rate significantly. However, the state-of-the-art (SOTA) models, such as InstructGPT and ChatGPT, show higher robustness to superficial modifications and have an outstanding capability for solving programming problems. This highlights the fact that slight modifications to the prompts given to the LLMs can greatly affect code generation performance, and careful formatting of prompts is essential for high-quality code generation, while the SOTA models are becoming more robust to perturbations. | ['Jun Suzuki', 'Yusuke Oda', 'Yuki Nakamura', 'Makoto Morishita', 'Takumi Ito', 'Yutaka Watanobe', 'Atsushi Shirafuji'] | 2023-06-26 | null | null | null | null | ['code-generation'] | ['computer-code'] | [ 9.82575268e-02 -1.87705283e-03 -1.95817664e-01 -2.88268089e-01
-8.01789343e-01 -8.23568523e-01 3.92448485e-01 3.18909705e-01
7.02276751e-02 2.87983567e-01 -5.73392212e-03 -7.11218238e-01
8.45673773e-03 -8.58937204e-01 -8.86003673e-01 -2.47112438e-01
-1.29384965e-01 2.21220359e-01 1.24396704e-01 -5.95296681e-01
6.35944486e-01 -1.06963897e-02 -1.57294035e+00 5.54722190e-01
1.30912280e+00 2.08568081e-01 5.82260013e-01 8.34509373e-01
-6.09955549e-01 9.67653930e-01 -8.51628661e-01 -3.57856423e-01
3.26909237e-02 -2.59717286e-01 -9.91534054e-01 -4.94741768e-01
5.02597809e-01 3.96602275e-03 -5.46549484e-02 1.04002941e+00
4.23198491e-01 -3.94684728e-03 2.16789916e-01 -1.32949197e+00
-8.77582669e-01 1.28140748e+00 -4.98356432e-01 2.15473607e-01
7.89131522e-01 4.39499348e-01 1.05508077e+00 -6.79045141e-01
6.03815258e-01 1.25078607e+00 7.62731671e-01 6.70740783e-01
-1.48563921e+00 -5.21212339e-01 1.75788909e-01 -2.34953582e-01
-1.27159655e+00 -2.65216738e-01 5.54098904e-01 -6.66470945e-01
1.38174105e+00 3.63727480e-01 2.14164093e-01 1.07075799e+00
4.49093968e-01 8.04875493e-01 8.99455369e-01 -5.48952341e-01
-2.79524308e-02 2.29199827e-01 3.24211359e-01 8.75802577e-01
9.47584491e-03 1.10892929e-01 -3.75018954e-01 -4.83416557e-01
5.21170437e-01 -3.05781305e-01 -3.84405226e-01 -2.28391886e-01
-1.16944218e+00 7.93056965e-01 3.31064105e-01 3.94067764e-01
5.49147744e-03 3.38822901e-01 5.69935620e-01 5.92916608e-01
1.91645771e-02 1.11386085e+00 -6.22465372e-01 -6.43312573e-01
-8.43129754e-01 6.95375919e-01 8.91034782e-01 1.39720261e+00
7.58745432e-01 2.11806849e-01 -4.54212904e-01 8.80753815e-01
6.39922023e-02 4.40977722e-01 7.46856868e-01 -7.37987995e-01
1.09491193e+00 8.24724376e-01 -2.68572927e-01 -1.10315990e+00
-2.35188037e-01 -4.42073703e-01 -2.39123538e-01 1.12413660e-01
3.60905200e-01 -8.61327350e-02 -5.15855908e-01 1.90440142e+00
-2.03987852e-01 -3.61769050e-01 2.40676761e-01 4.53106672e-01
8.75703454e-01 8.45348120e-01 1.71859682e-01 2.48222038e-01
1.01151097e+00 -9.26219583e-01 -1.63794473e-01 -6.86123669e-01
1.19023061e+00 -8.55060935e-01 1.62719572e+00 2.45117307e-01
-1.19627666e+00 -7.56662011e-01 -7.63988495e-01 -2.32922453e-02
-1.54489756e-01 7.18996972e-02 9.25974369e-01 7.21071661e-01
-1.23762155e+00 7.58393466e-01 -6.72076404e-01 -3.81669551e-01
8.08488727e-02 1.97229818e-01 -1.08209312e-01 -2.07548112e-01
-9.22619402e-01 7.52782702e-01 4.50385422e-01 -2.34701321e-01
-9.65861022e-01 -1.04872465e+00 -1.11986411e+00 3.82775486e-01
2.84649074e-01 -5.05093217e-01 1.65405202e+00 -8.81396770e-01
-1.44627988e+00 7.42028356e-01 -2.87894189e-01 -2.13693842e-01
2.13995963e-01 -3.06328386e-02 -2.02879548e-01 -2.44700909e-01
1.63613990e-01 6.81550860e-01 6.96818531e-01 -1.21619320e+00
-4.09194171e-01 1.58526734e-01 3.77319276e-01 -1.28516838e-01
-3.15225869e-01 2.60426879e-01 -1.16740227e-01 -7.25774169e-01
-1.19599164e-01 -1.06050932e+00 -2.71717489e-01 -4.44665581e-01
-2.80151218e-01 -4.13667560e-01 3.58553439e-01 -5.39910614e-01
1.60661066e+00 -2.16559005e+00 1.93045482e-01 4.59691398e-02
1.44533098e-01 2.06791803e-01 -6.70884073e-01 6.71436906e-01
-4.13904250e-01 4.97108579e-01 -1.14579685e-01 3.68289724e-02
2.00426668e-01 5.54217771e-02 -7.03539133e-01 -1.08748496e-01
2.19845578e-01 1.09282756e+00 -1.02051640e+00 -1.90677762e-01
-9.83377099e-02 -1.04679480e-01 -1.05568063e+00 3.36815119e-01
-6.96837127e-01 -5.79230711e-02 -4.80708122e-01 4.96606380e-01
4.35193688e-01 -3.66970003e-01 -1.16595805e-01 4.46520329e-01
-2.74486780e-01 7.23600209e-01 -7.97806442e-01 1.70702958e+00
-8.85874569e-01 7.54030585e-01 -2.91463017e-01 -6.38951778e-01
1.01269114e+00 2.07744181e-01 -1.42218709e-01 -7.30346322e-01
-2.40746811e-01 2.02764064e-01 2.60120451e-01 -7.24898577e-01
7.93489814e-01 2.77181685e-01 -3.76336157e-01 7.05426753e-01
-1.30538464e-01 -5.61522543e-01 6.95710897e-01 5.62983036e-01
1.27344131e+00 1.38383299e-01 7.50157908e-02 -3.16210419e-01
3.06420773e-01 2.20923364e-01 3.61166328e-01 1.22593582e+00
2.79523224e-01 5.73059559e-01 8.34385812e-01 -1.48268700e-01
-7.60527790e-01 -6.94193304e-01 1.31370708e-01 1.41326094e+00
-2.49513403e-01 -1.11759996e+00 -8.05836439e-01 -5.40370286e-01
-4.30615917e-02 1.16383827e+00 -3.91544819e-01 -3.99546236e-01
-6.70470238e-01 -6.95745349e-01 6.64973676e-01 6.21410251e-01
1.90589026e-01 -1.15246868e+00 -6.15312815e-01 3.22814226e-01
-2.38928929e-01 -7.37929821e-01 -4.81442511e-01 2.29591727e-01
-9.53853607e-01 -9.61011112e-01 -4.33656037e-01 -8.38608325e-01
9.19292808e-01 1.61411583e-01 1.65056896e+00 5.19710898e-01
-1.45450607e-01 4.44044888e-01 -5.30539691e-01 -2.15637371e-01
-1.14532757e+00 4.88552541e-01 -3.73990715e-01 -8.17602932e-01
1.30407766e-01 -5.69372177e-01 8.03611428e-02 2.26505876e-01
-8.87508512e-01 2.42826521e-01 5.90683758e-01 6.36943519e-01
-8.13056603e-02 1.09406263e-01 2.21325397e-01 -1.18852592e+00
1.06199300e+00 -5.51907063e-01 -7.47985303e-01 2.97408909e-01
-6.24642909e-01 4.30415094e-01 1.00242150e+00 -4.63075697e-01
-1.00560153e+00 -3.70400816e-01 -2.17729405e-01 2.15015803e-02
-7.77481645e-02 9.95828271e-01 1.77263170e-01 -2.93963760e-01
1.20665288e+00 3.90214473e-01 -3.68025601e-01 -3.36823136e-01
6.46978393e-02 3.30729812e-01 2.95462608e-01 -1.35400343e+00
1.07397771e+00 -3.60324919e-01 -4.90550548e-01 -4.55991656e-01
-4.93478447e-01 -1.23782940e-01 -1.28569260e-01 3.03019941e-01
2.80718923e-01 -7.96647370e-01 -4.06175256e-01 5.55408895e-01
-1.38617134e+00 -8.09176922e-01 7.06893280e-02 -3.40612829e-02
-4.17348593e-01 3.63283008e-01 -8.27001154e-01 -3.03337932e-01
-1.92937642e-01 -1.58165181e+00 8.94946516e-01 2.59939462e-01
-7.64293313e-01 -1.12349749e+00 1.95021294e-02 1.71086729e-01
7.97476649e-01 1.61601305e-02 1.69290924e+00 -4.42562997e-01
-7.46542096e-01 -1.74043979e-02 1.17457092e-01 1.36500329e-01
1.53421843e-02 3.91413838e-01 -7.21455634e-01 -3.58276457e-01
-3.71627770e-02 -2.74807751e-01 5.21224976e-01 2.00152993e-02
1.28258920e+00 -4.46592957e-01 -2.96737194e-01 7.12693632e-01
1.39935064e+00 8.33721012e-02 5.51831067e-01 4.35485214e-01
6.52360916e-01 4.16687697e-01 4.81601596e-01 4.03883398e-01
3.96580666e-01 5.19080162e-01 3.28643560e-01 3.12194586e-01
8.72491393e-03 -5.07578611e-01 7.90759027e-01 8.58853757e-01
3.62883300e-01 -1.71711087e-01 -1.38886321e+00 5.69407225e-01
-1.58480787e+00 -7.01830268e-01 -4.35446411e-01 2.04844761e+00
1.31915176e+00 2.95637697e-01 -3.24419975e-01 -6.27191588e-02
3.64435881e-01 1.76313058e-01 -2.62947589e-01 -6.90406263e-01
1.07364014e-01 3.90651137e-01 1.32411078e-01 3.63731533e-01
-5.71134448e-01 9.73328292e-01 7.04902554e+00 8.20672512e-01
-1.20407856e+00 -2.21013248e-01 2.71001935e-01 8.41178596e-02
-6.68832302e-01 3.07826728e-01 -8.45693767e-01 5.32690346e-01
1.06280470e+00 -6.69511318e-01 7.29131699e-01 1.26827216e+00
6.94040656e-02 -1.61937580e-01 -1.60615945e+00 7.91784644e-01
-4.54347357e-02 -1.36387372e+00 1.96826503e-01 -4.24903780e-01
9.82933700e-01 -1.34258652e-02 1.58128574e-01 9.53495264e-01
6.78306341e-01 -1.13182712e+00 9.04569983e-01 1.48398206e-01
4.70665038e-01 -4.73738849e-01 4.85890388e-01 4.86442506e-01
-1.09123898e+00 -1.76366895e-01 -4.72020030e-01 -3.83470982e-01
-2.76240677e-01 4.02280182e-01 -9.40776587e-01 2.03248248e-01
5.80508411e-01 5.37360132e-01 -1.25911283e+00 9.94257092e-01
-5.10883391e-01 7.53226161e-01 3.99629287e-02 -1.31566718e-01
1.52677491e-01 3.79918009e-01 4.13180619e-01 1.13390350e+00
4.63555783e-01 -1.53165266e-01 2.46380925e-01 1.60102010e+00
1.35021210e-01 2.00446285e-02 -4.01998758e-01 -4.44067925e-01
4.42845374e-01 1.06149197e+00 -4.04907256e-01 -2.08420247e-01
-4.49741721e-01 5.23879707e-01 4.17009115e-01 4.82624501e-01
-8.21458459e-01 -6.05789423e-01 6.50258422e-01 7.98256025e-02
6.14318438e-02 -2.94177234e-01 -3.74748439e-01 -1.14583588e+00
1.56015903e-01 -1.57996404e+00 2.08603635e-01 -1.07323992e+00
-8.64442706e-01 5.80581129e-01 1.21758267e-01 -9.05404508e-01
-2.83644438e-01 -5.59777558e-01 -9.60758448e-01 1.00736904e+00
-1.36607075e+00 -6.73028052e-01 -3.83585483e-01 2.42368191e-01
6.52755499e-01 -2.30755538e-01 8.20989251e-01 3.85631621e-02
-5.67412198e-01 7.95589983e-01 -7.72295296e-02 9.45195034e-02
6.04410946e-01 -1.36028397e+00 9.68732476e-01 9.46987629e-01
-1.12155871e-02 1.52177858e+00 9.59760010e-01 -6.17237329e-01
-1.70092082e+00 -1.06155872e+00 9.52587664e-01 -7.99751401e-01
5.95915854e-01 -4.04966354e-01 -1.14584374e+00 7.71992862e-01
4.64609936e-02 -3.27967316e-01 4.67149675e-01 2.84029156e-01
-5.47604144e-01 1.62038743e-01 -7.69724607e-01 7.96012521e-01
8.99728239e-01 -6.86728716e-01 -7.22402096e-01 3.47232044e-01
8.78971696e-01 -8.73745441e-01 -7.70086884e-01 6.99969903e-02
4.89248671e-02 -9.22870159e-01 7.60141611e-01 -5.67848206e-01
1.10084283e+00 -1.40890285e-01 1.09487407e-01 -1.60734332e+00
-4.11828905e-01 -8.04935575e-01 1.58562854e-01 1.50969541e+00
6.58505201e-01 -6.21858716e-01 5.33692837e-01 9.87145901e-01
-3.21213484e-01 -3.82327616e-01 -2.55926639e-01 -7.64524817e-01
3.27348173e-01 -5.51743031e-01 7.70056009e-01 9.90819156e-01
4.01296973e-01 6.97446018e-02 1.25594795e-01 1.31225109e-01
6.57434985e-02 4.40283656e-01 1.03588903e+00 -9.74103808e-01
-7.12754309e-01 -6.44830167e-01 9.81320068e-02 -1.03624785e+00
4.16493863e-01 -1.28189337e+00 6.67869896e-02 -1.27119970e+00
1.70337990e-01 -8.20409775e-01 1.64818197e-01 6.81087017e-01
-6.26266301e-01 -4.96343881e-01 2.93088913e-01 6.06943518e-02
-3.57118815e-01 3.53355050e-01 8.75433445e-01 -2.35765338e-01
-2.87191272e-01 9.36973691e-02 -9.75468218e-01 5.08997917e-01
7.78582692e-01 -7.27408588e-01 -5.30788958e-01 -9.88933444e-01
8.24862480e-01 2.49493480e-01 2.32196972e-01 -1.01957417e+00
2.16871172e-01 -3.58031571e-01 -1.17086694e-01 1.56863369e-02
-3.65414977e-01 -3.94304931e-01 6.10327162e-02 4.60952580e-01
-7.07151175e-01 5.57815969e-01 7.47084320e-01 5.48612177e-02
-2.10001722e-01 -7.98244119e-01 4.58021790e-01 -4.70615268e-01
-8.99226785e-01 8.85403156e-02 -5.95288575e-01 3.88967097e-01
7.22708404e-01 -1.17942795e-01 -5.02503574e-01 -1.52213320e-01
-7.25058913e-02 2.94634312e-01 6.39011502e-01 8.50196719e-01
3.43924582e-01 -1.08386326e+00 -5.73441088e-01 3.52158725e-01
4.32571173e-01 1.83598641e-02 1.04526274e-01 5.63087225e-01
-5.97448826e-01 6.06155217e-01 -4.69989181e-02 -6.66389704e-01
-1.22570789e+00 6.10049307e-01 1.45423770e-01 -3.44475061e-01
-3.80589366e-01 1.10581219e+00 2.67826647e-01 -8.88387442e-01
6.70857430e-02 -9.71680820e-01 1.48081720e-01 -4.36443388e-01
5.01606107e-01 5.59270494e-02 9.04759690e-02 -4.28950340e-02
-1.64243028e-01 3.89261007e-01 -4.03617680e-01 4.15012360e-01
1.32571137e+00 3.21802646e-01 -4.41750616e-01 1.68798983e-01
9.59388375e-01 1.80128515e-01 -9.24689889e-01 -2.23434448e-01
2.29953885e-01 -5.85690916e-01 -2.66574591e-01 -8.50529253e-01
-8.98425281e-01 8.30000818e-01 3.48083749e-02 2.73506612e-01
7.62270153e-01 -1.61872163e-01 5.92501163e-01 5.54526389e-01
6.84375703e-01 -6.72659576e-01 2.83526599e-01 1.01953912e+00
8.98114085e-01 -1.10938370e+00 -4.19582099e-01 -3.67107272e-01
-4.28678274e-01 1.33384800e+00 1.14492917e+00 1.43924788e-01
-3.15744169e-02 5.39593101e-01 -4.91865650e-02 -7.42015243e-02
-1.18850839e+00 3.66837829e-01 -5.26600378e-03 5.99288166e-01
7.60788858e-01 -1.87335655e-01 6.12267526e-03 5.55265129e-01
-7.10528076e-01 -8.37783515e-02 8.89142632e-01 1.12857771e+00
-2.21479610e-01 -1.33482051e+00 -5.53021669e-01 4.18051273e-01
-2.39607021e-01 -6.53246760e-01 -2.24111646e-01 5.88899374e-01
-2.41216525e-01 7.41917193e-01 -3.49656910e-01 -2.50353575e-01
3.74952108e-01 2.12547898e-01 3.97764146e-01 -1.28896832e+00
-1.14608669e+00 -7.00893044e-01 1.32248830e-02 -4.86882389e-01
2.84695566e-01 -4.59734768e-01 -1.21510220e+00 -5.71336389e-01
-2.73446202e-01 3.30629766e-01 3.83861780e-01 7.14523792e-01
4.65203285e-01 6.93671703e-01 2.46599138e-01 -6.40256643e-01
-8.57770503e-01 -7.39939392e-01 -1.25289321e-01 3.48929852e-01
2.62060076e-01 -2.30785072e-01 -3.04966539e-01 1.00826949e-01] | [7.987740516662598, 7.703790664672852] |
7c8beda2-52b5-43c8-b0ad-3983e20d8c06 | pscc-net-progressive-spatio-channel | 2103.10596 | null | https://arxiv.org/abs/2103.10596v2 | https://arxiv.org/pdf/2103.10596v2.pdf | PSCC-Net: Progressive Spatio-Channel Correlation Network for Image Manipulation Detection and Localization | To defend against manipulation of image content, such as splicing, copy-move, and removal, we develop a Progressive Spatio-Channel Correlation Network (PSCC-Net) to detect and localize image manipulations. PSCC-Net processes the image in a two-path procedure: a top-down path that extracts local and global features and a bottom-up path that detects whether the input image is manipulated, and estimates its manipulation masks at multiple scales, where each mask is conditioned on the previous one. Different from the conventional encoder-decoder and no-pooling structures, PSCC-Net leverages features at different scales with dense cross-connections to produce manipulation masks in a coarse-to-fine fashion. Moreover, a Spatio-Channel Correlation Module (SCCM) captures both spatial and channel-wise correlations in the bottom-up path, which endows features with holistic cues, enabling the network to cope with a wide range of manipulation attacks. Thanks to the light-weight backbone and progressive mechanism, PSCC-Net can process 1,080P images at 50+ FPS. Extensive experiments demonstrate the superiority of PSCC-Net over the state-of-the-art methods on both detection and localization. | ['Xiaoming Liu', 'Jun Chen', 'Yaojie Liu', 'Xiaohong Liu'] | 2021-03-19 | null | null | null | null | ['image-manipulation-detection'] | ['computer-vision'] | [ 4.73464072e-01 -7.00365901e-01 -3.41299325e-02 1.13702886e-01
-5.24204910e-01 -7.23547637e-01 3.14191490e-01 2.23601788e-01
-5.11731386e-01 -8.20341930e-02 2.25559995e-02 2.53669135e-02
2.47303173e-01 -7.09365845e-01 -7.96110332e-01 -5.22739947e-01
-4.67837960e-01 -6.24298573e-01 8.73088658e-01 -1.96637779e-01
6.42310739e-01 7.94609725e-01 -1.33879435e+00 8.23858738e-01
1.59547701e-01 1.37994337e+00 2.23501056e-01 1.02698767e+00
3.21513623e-01 9.67578590e-01 -5.47486305e-01 -4.54559326e-01
3.89983952e-01 6.50581494e-02 -3.60051423e-01 1.56468228e-02
4.95910555e-01 -6.40164614e-01 -9.26560700e-01 1.33078074e+00
4.75501597e-01 -2.18864754e-01 2.80213803e-01 -1.12747109e+00
-9.45972621e-01 2.49393508e-01 -1.06938148e+00 6.82396710e-01
3.64170939e-01 7.35977232e-01 7.70247161e-01 -1.04453290e+00
4.77238029e-01 1.46206796e+00 5.55226743e-01 2.64642328e-01
-1.38414693e+00 -8.88298869e-01 2.43334025e-01 2.69033819e-01
-1.31042790e+00 -3.06714058e-01 4.79679078e-01 -2.34211177e-01
8.78140211e-01 -2.03733128e-02 4.72956002e-01 9.88457739e-01
3.72426242e-01 7.96396554e-01 1.11668551e+00 -2.20263693e-02
9.24642906e-02 -5.43491662e-01 1.51447607e-02 8.75238121e-01
2.33808756e-01 1.59021482e-01 -1.09153724e+00 1.66107535e-01
1.08855581e+00 1.32572487e-01 -4.64452684e-01 -2.21551627e-01
-1.31341338e+00 3.18387091e-01 7.57027626e-01 2.41326541e-02
-3.13117683e-01 5.10140896e-01 6.48592651e-01 3.92052591e-01
-1.42271146e-01 4.87946212e-01 -2.99339622e-01 -4.29574028e-02
-9.33032155e-01 -4.02482711e-02 4.24422592e-01 1.08192194e+00
8.03008139e-01 -2.25698933e-01 -5.12140095e-01 2.56620914e-01
-1.16467156e-01 5.47700226e-01 5.62742539e-02 -1.01981986e+00
7.16962397e-01 5.51735640e-01 -1.02125809e-01 -1.44706714e+00
-1.62346900e-01 -3.50227982e-01 -7.89046109e-01 3.16715837e-01
3.99978347e-02 1.37292951e-01 -7.39798665e-01 1.56283510e+00
1.77024990e-01 4.78090882e-01 -3.67426485e-01 9.23411489e-01
4.33742195e-01 4.42600578e-01 3.17254588e-02 1.48093581e-01
1.61512101e+00 -9.91402626e-01 -4.41267937e-01 -2.62073457e-01
1.47790238e-01 -6.59706235e-01 7.69993901e-01 3.25633198e-01
-1.22458053e+00 -6.95944667e-01 -1.15605712e+00 -3.28190237e-01
-5.53011537e-01 1.76929563e-01 2.69770086e-01 2.82236159e-01
-1.09385085e+00 6.99970424e-01 -9.67022955e-01 -3.60353775e-02
7.88424432e-01 3.31237972e-01 -4.97282356e-01 -6.39764816e-02
-9.11759615e-01 5.30117035e-01 2.95700371e-01 2.05106407e-01
-1.11931026e+00 -6.78832531e-01 -8.62043500e-01 4.73342866e-01
4.52583224e-01 -1.99351653e-01 7.07813740e-01 -5.17017603e-01
-1.26500976e+00 6.49922788e-01 -2.35707626e-01 -4.95916516e-01
3.76153320e-01 -2.20900089e-01 -2.53559649e-01 7.86107004e-01
2.48829588e-01 8.71409655e-01 1.27679682e+00 -1.18667066e+00
-7.39187598e-01 -3.67667317e-01 1.86032783e-02 4.80548143e-02
-5.12230694e-01 5.56509078e-01 -1.21104014e+00 -7.66337752e-01
4.08881903e-02 -6.08143985e-01 7.34062046e-02 6.27558231e-01
-5.99276960e-01 5.15536189e-01 1.13622665e+00 -5.14019489e-01
1.30191362e+00 -2.66534925e+00 1.76599752e-02 2.14565694e-01
4.39951628e-01 4.49401915e-01 -6.10599518e-01 5.15321612e-01
-7.14318315e-03 2.07768589e-01 -2.27830615e-02 -5.47230005e-01
-1.88932568e-01 -1.89711586e-01 -7.00343370e-01 7.82624543e-01
6.71362936e-01 1.02309215e+00 -9.53468025e-01 -4.02328610e-01
3.22088271e-01 5.63985646e-01 -5.45868397e-01 5.12522412e-03
1.20087154e-01 2.63326108e-01 -2.80472577e-01 1.05586052e+00
1.00697064e+00 -2.87769496e-01 -1.45293102e-02 -3.38341057e-01
-2.38395408e-01 5.79048879e-02 -9.69310164e-01 1.74786425e+00
-2.48572841e-01 7.86328137e-01 3.92766416e-01 -5.96571028e-01
5.00847280e-01 -1.17650062e-01 1.35273337e-01 -7.69119382e-01
3.57432872e-01 1.37270451e-01 -1.77390039e-01 -2.76718378e-01
6.28461301e-01 7.65845835e-01 -2.33251691e-01 2.32427672e-01
2.28270337e-01 1.39324740e-01 2.84633726e-01 4.73666936e-01
1.60204434e+00 -9.74108502e-02 -1.39320642e-02 2.12974250e-02
6.25168979e-01 -5.70076942e-01 6.08321488e-01 1.06361187e+00
-6.08718276e-01 6.27388954e-01 8.87968123e-01 -2.38883033e-01
-5.13820648e-01 -1.19665372e+00 1.15990579e-01 1.09844923e+00
7.31485784e-01 -5.69502234e-01 -7.25580573e-01 -4.97052312e-01
1.50052756e-01 -5.47087081e-02 -7.03516781e-01 -3.10811132e-01
-5.60719848e-01 -1.00665547e-01 6.72328055e-01 7.01182067e-01
9.56317842e-01 -1.08834815e+00 -9.44805682e-01 1.23422392e-01
9.28905979e-03 -1.47911310e+00 -8.76801491e-01 2.35231340e-01
-4.12369251e-01 -1.23910832e+00 -1.55836955e-01 -8.18356931e-01
6.93987429e-01 7.45767474e-01 7.53702104e-01 3.34218919e-01
-6.79301143e-01 9.87772122e-02 -3.24451745e-01 -5.87370805e-02
7.67026171e-02 -1.29656032e-01 -1.83463052e-01 3.05183947e-01
2.71129698e-01 -6.97441518e-01 -1.03841412e+00 4.90339607e-01
-9.24220622e-01 -2.75942590e-02 8.04216802e-01 7.23733842e-01
5.79553723e-01 3.82310152e-01 -3.19055654e-02 -4.50024724e-01
4.02850270e-01 -1.33037239e-01 -7.73137450e-01 1.59038067e-01
8.51950422e-02 -2.16203108e-01 6.15209520e-01 -5.96402287e-01
-5.51830411e-01 1.49956226e-01 4.00148660e-01 -9.46842909e-01
9.35992673e-02 3.84161994e-02 -3.16899061e-01 -6.07668281e-01
4.84259784e-01 4.67352152e-01 -1.96191534e-01 -9.19297263e-02
5.14822841e-01 3.76180649e-01 1.15872633e+00 -3.80924881e-01
1.22707248e+00 8.28481138e-01 7.48378485e-02 -4.95669246e-01
-5.97521126e-01 -5.76301336e-01 -8.34359884e-01 -2.45987400e-01
1.15325081e+00 -1.10585690e+00 -1.21104598e+00 9.11694586e-01
-1.20077920e+00 -4.05643612e-01 -1.04324436e-02 -4.01094481e-02
-3.31015259e-01 3.06569397e-01 -1.17615652e+00 -5.66122115e-01
-3.28379542e-01 -1.36298943e+00 1.59323847e+00 2.44787320e-01
1.68579027e-01 -4.13934022e-01 -7.00396597e-01 5.01408316e-02
5.90983570e-01 2.05074206e-01 6.30526185e-01 -9.68440473e-02
-1.05571139e+00 -3.68869752e-01 -8.01072598e-01 3.25219691e-01
-1.85023680e-01 -6.05847575e-02 -9.04116273e-01 -6.12534881e-01
-9.11676735e-02 -3.86926442e-01 1.14772642e+00 -1.85310207e-02
1.69351459e+00 -7.54392669e-02 -2.31128216e-01 8.91584694e-01
1.54203582e+00 -1.56757608e-01 9.35297608e-01 2.81969160e-01
6.48494422e-01 7.30501413e-02 3.15094560e-01 2.66914457e-01
1.67352632e-01 5.67021132e-01 7.66918004e-01 -2.74484754e-01
-2.73984075e-01 -3.29290777e-01 5.36861420e-01 2.85727710e-01
4.43756282e-01 -1.70977429e-01 -5.95860124e-01 3.18051577e-01
-1.63801515e+00 -1.12731111e+00 9.46215987e-02 2.07995415e+00
6.50825262e-01 4.03067946e-01 -2.33791366e-01 1.04486719e-01
9.58930254e-01 6.42506361e-01 -6.60832167e-01 -1.21918455e-01
-2.90042996e-01 2.48814374e-01 8.57566357e-01 9.93708745e-02
-1.46441913e+00 1.05510175e+00 6.09545088e+00 9.09556210e-01
-1.35678089e+00 -1.46042779e-01 5.48234642e-01 -2.78340042e-01
4.31866020e-01 -1.13585562e-01 -6.76872015e-01 7.41324663e-01
4.51648772e-01 6.07876241e-01 5.95664620e-01 6.91544533e-01
2.99471375e-02 -3.34436983e-01 -8.81658792e-01 8.58119309e-01
2.05689892e-01 -1.47561610e+00 -9.90758985e-02 8.08037147e-02
6.24314666e-01 2.05183715e-01 4.44537610e-01 7.39569589e-03
2.97603667e-01 -6.88947737e-01 1.01570725e+00 1.88176841e-01
1.09021664e+00 -6.87850535e-01 4.79035884e-01 8.00071880e-02
-1.63252735e+00 -4.92487073e-01 -3.42934757e-01 1.92147225e-01
3.98533158e-02 4.71504897e-01 -1.34631261e-01 2.04390913e-01
9.17410970e-01 8.59630585e-01 -6.68941677e-01 9.09803092e-01
-5.30068874e-01 3.21899086e-01 -7.87246302e-02 3.40946019e-01
1.56642467e-01 1.29376933e-01 4.97777760e-01 1.51695991e+00
-5.60718542e-03 -1.58640370e-02 1.99560106e-01 8.42982829e-01
-5.19313395e-01 -4.35778320e-01 -2.02421978e-01 1.01983704e-01
9.85540032e-01 1.35705018e+00 -8.33234072e-01 -2.21656397e-01
-5.29136121e-01 1.58762705e+00 4.59967792e-01 4.42538619e-01
-9.27989483e-01 -6.79965377e-01 7.84614205e-01 -1.89323910e-02
8.24492395e-01 -3.47848952e-01 -3.71512771e-01 -1.19185996e+00
2.34215975e-01 -9.23448861e-01 6.61300570e-02 -6.94790483e-01
-1.14692259e+00 3.36199403e-01 -4.61015671e-01 -1.25703502e+00
7.95313597e-01 -8.56021285e-01 -8.53500366e-01 8.73871028e-01
-1.89334166e+00 -1.13147521e+00 -5.13545752e-01 7.57539034e-01
2.58618861e-01 4.01626434e-03 4.78153735e-01 1.67647630e-01
-8.86099696e-01 6.72270000e-01 -2.16253683e-01 7.23885715e-01
6.90044880e-01 -1.00918341e+00 7.36405551e-01 1.51063609e+00
-1.09371781e-01 9.83538330e-01 9.80982706e-02 -5.48343539e-01
-1.83932424e+00 -1.42235446e+00 2.95057595e-01 -2.64581472e-01
1.00458276e+00 -7.98482239e-01 -9.08207238e-01 2.92641222e-01
2.39903219e-02 7.68976331e-01 3.39590430e-01 -3.83545697e-01
-1.16213059e+00 -1.50748283e-01 -9.69460487e-01 6.20723069e-01
1.03406799e+00 -1.15700388e+00 -9.53708291e-02 8.54181126e-02
6.14338398e-01 -6.11195743e-01 -6.39477670e-01 1.71847805e-01
5.83867192e-01 -1.19056797e+00 1.12769258e+00 -1.45800203e-01
6.90442562e-01 -6.29805982e-01 -1.63034111e-01 -8.23544025e-01
-5.60831726e-01 -1.03233242e+00 -2.12074906e-01 9.71391618e-01
1.03771888e-01 -4.61971432e-01 5.15942276e-01 7.68661350e-02
-1.49327561e-01 -8.72074246e-01 -7.37413824e-01 -8.00033331e-01
-3.47872853e-01 -4.64417309e-01 5.53861976e-01 4.68992591e-01
-1.28351763e-01 -2.17053995e-01 -2.72752047e-01 8.30305278e-01
4.91687566e-01 -6.45047203e-02 7.37350464e-01 -6.07069194e-01
-4.16722625e-01 -5.76928437e-01 -8.57688189e-01 -1.33576560e+00
-2.17873976e-02 -5.29968679e-01 1.90021247e-01 -9.42450702e-01
2.37785921e-01 -6.19380958e-02 -5.01729310e-01 6.64748251e-01
-3.78004044e-01 6.60930276e-01 5.96939266e-01 4.71981138e-01
-8.74353051e-01 1.89221978e-01 1.02624631e+00 -1.15313753e-01
4.98422235e-03 -6.72768474e-01 -6.85690641e-01 5.93908906e-01
4.31288391e-01 -3.57121885e-01 -4.08529490e-02 -7.55272329e-01
-2.19632406e-02 -8.50186571e-02 6.20060563e-01 -1.32167447e+00
5.87643266e-01 4.03642468e-02 7.28250742e-01 -5.07856727e-01
3.57711166e-01 -6.97251916e-01 -5.41767895e-01 5.39177656e-01
-4.55873489e-01 1.64196029e-01 2.03700200e-01 9.61392820e-01
-3.27230722e-01 2.96625346e-01 9.99778628e-01 8.43869746e-02
-7.97080398e-01 5.60261965e-01 -3.14245701e-01 -2.43753389e-01
1.08103216e+00 -2.28007108e-01 -7.99579322e-01 -6.58892244e-02
-1.74480155e-01 4.23283607e-01 6.82589889e-01 3.61651957e-01
7.49666572e-01 -1.01940703e+00 -3.37387890e-01 5.10097682e-01
2.05706939e-01 1.33283129e-02 4.12979215e-01 1.01122153e+00
-7.04468250e-01 7.19924048e-02 -1.68669671e-01 -5.37836909e-01
-1.18042004e+00 8.40718687e-01 1.73914343e-01 -2.19768584e-01
-6.04710162e-01 1.19036591e+00 3.43436420e-01 2.86938548e-01
3.10094595e-01 -1.92654818e-01 1.38381153e-01 -1.48599386e-01
9.70380008e-01 1.47614092e-01 -1.76944405e-01 -6.14863455e-01
-3.85243446e-01 5.31957150e-01 -2.58852065e-01 4.66798916e-02
9.99570966e-01 -1.35062397e-01 -4.34717476e-01 -8.91966820e-02
1.28986835e+00 -2.62773223e-02 -1.72387671e+00 -4.08928066e-01
-2.79032856e-01 -8.52717161e-01 2.77493477e-01 -7.31541634e-01
-1.35639501e+00 1.00169957e+00 3.97338003e-01 -9.54777747e-02
1.55626535e+00 -2.78204590e-01 8.47784281e-01 1.82929710e-01
4.06328410e-01 -1.00852048e+00 5.27030826e-01 5.44924021e-01
7.97563851e-01 -1.02561069e+00 4.93761944e-03 -6.99021757e-01
-4.62872177e-01 1.15131938e+00 6.06541753e-01 -3.91534030e-01
5.51570952e-01 7.39263058e-01 -2.51902640e-01 -2.29760051e-01
-7.80465364e-01 -1.62120745e-01 -4.63145114e-02 5.36419094e-01
-1.97431341e-01 -1.58343643e-01 3.98542076e-01 1.75570458e-01
2.03894272e-01 -1.33874148e-01 4.50626314e-01 1.10100293e+00
-4.02204990e-01 -6.88892722e-01 -2.48685226e-01 2.54514158e-01
-3.32253844e-01 -2.66319245e-01 -3.17411572e-01 5.40266573e-01
3.19201559e-01 9.58609998e-01 2.41179973e-01 -6.64014041e-01
3.05067956e-01 -5.20633101e-01 1.58072367e-01 -5.38681269e-01
-7.48612523e-01 5.62213212e-02 -3.82702947e-01 -1.38414824e+00
-1.00526765e-01 -6.03857160e-01 -1.16619194e+00 -1.84674382e-01
-3.32581162e-01 -4.95415509e-01 4.80286211e-01 4.52592194e-01
7.52363503e-01 5.51593184e-01 8.26739788e-01 -1.21625483e+00
-4.83505428e-01 -6.92048132e-01 -4.59919900e-01 5.24089992e-01
5.62477291e-01 -5.59296250e-01 -4.37746376e-01 6.02921322e-02] | [12.259861946105957, 0.9214255809783936] |
ae059c45-b0c7-4683-95fb-b7191f0b97ee | catfl-certificateless-authentication-based | 2302.00271 | null | https://arxiv.org/abs/2302.00271v1 | https://arxiv.org/pdf/2302.00271v1.pdf | CATFL: Certificateless Authentication-based Trustworthy Federated Learning for 6G Semantic Communications | Federated learning (FL) provides an emerging approach for collaboratively training semantic encoder/decoder models of semantic communication systems, without private user data leaving the devices. Most existing studies on trustworthy FL aim to eliminate data poisoning threats that are produced by malicious clients, but in many cases, eliminating model poisoning attacks brought by fake servers is also an important objective. In this paper, a certificateless authentication-based trustworthy federated learning (CATFL) framework is proposed, which mutually authenticates the identity of clients and server. In CATFL, each client verifies the server's signature information before accepting the delivered global model to ensure that the global model is not delivered by false servers. On the contrary, the server also verifies the server's signature information before accepting the delivered model updates to ensure that they are submitted by authorized clients. Compared to PKI-based methods, the CATFL can avoid too high certificate management overheads. Meanwhile, the anonymity of clients shields data poisoning attacks, while real-name registration may suffer from user-specific privacy leakage risks. Therefore, a pseudonym generation strategy is also presented in CATFL to achieve a trade-off between identity traceability and user anonymity, which is essential to conditionally prevent from user-specific privacy leakage. Theoretical security analysis and evaluation results validate the superiority of CATFL. | ['Yi Li', 'YuanYuan Zhao', 'Gaolei Li'] | 2023-02-01 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [-2.25363076e-01 1.41834974e-01 -2.15925947e-01 -3.34123224e-01
-6.66039824e-01 -8.62902403e-01 4.65425074e-01 -1.66433409e-01
-4.49085802e-01 7.67977357e-01 -6.29902035e-02 -3.53884131e-01
-7.37888785e-03 -9.77537811e-01 -5.90260267e-01 -8.69427383e-01
3.21528494e-01 1.91751122e-01 3.01893055e-01 -2.19834242e-02
-1.96421295e-02 4.29470956e-01 -1.21607840e+00 2.26630211e-01
7.67055511e-01 1.08451474e+00 -2.09300891e-01 -1.11444771e-01
-4.26317573e-01 1.05486715e+00 -4.96035367e-01 -8.75188529e-01
2.84848988e-01 -2.85971671e-01 -9.59310949e-01 -3.09327424e-01
-1.65682435e-01 -7.02862263e-01 -4.57852066e-01 1.52270329e+00
2.18097165e-01 -6.00450337e-01 2.01274231e-02 -2.26452422e+00
-6.55018270e-01 8.82176816e-01 8.69693160e-02 -6.73934400e-01
1.58765331e-01 1.49438351e-01 7.62532353e-01 -3.59214246e-01
5.54873526e-01 8.59091640e-01 5.43672204e-01 8.94061565e-01
-6.23560965e-01 -1.32145691e+00 3.06928647e-03 3.71592432e-01
-1.58616030e+00 -4.91176367e-01 6.12682223e-01 5.10902330e-02
1.10630572e-01 5.03433526e-01 1.16273366e-01 1.01816750e+00
-3.44960392e-02 5.78948498e-01 1.10138369e+00 -2.25104362e-01
4.51754004e-01 7.38321543e-01 1.59958825e-02 4.57347006e-01
8.59385848e-01 -3.80002744e-02 -5.67625403e-01 -7.25850940e-01
5.09098709e-01 6.92699626e-02 -4.91162300e-01 -5.34625769e-01
-8.62154901e-01 6.56317294e-01 6.56350926e-02 3.52020144e-01
-1.92626297e-01 2.04072490e-01 6.00880623e-01 5.46831965e-01
-3.07105519e-02 -2.70837247e-01 -6.06131554e-01 8.97093341e-02
-3.77913445e-01 1.99604034e-03 1.13670325e+00 1.23667288e+00
5.59082031e-01 -6.87600747e-02 4.09731954e-01 -6.40749037e-02
8.22676122e-01 4.77974921e-01 4.76069033e-01 -9.62080181e-01
1.01598315e-01 7.45523810e-01 4.26826835e-01 -9.36375558e-01
1.93000883e-01 -9.49408039e-02 -5.83614111e-01 9.11689699e-02
5.85364163e-01 -1.15307961e-02 -6.98519275e-02 1.80508614e+00
5.54810703e-01 2.40396753e-01 6.43056750e-01 8.51307452e-01
7.22257435e-01 3.59804958e-01 2.94975698e-01 -3.03136110e-01
1.53756297e+00 -5.69578290e-01 -1.05346382e+00 4.19466794e-01
9.02961075e-01 -5.24584115e-01 6.37452304e-01 3.51077646e-01
-6.55303180e-01 1.01054557e-01 -1.06593263e+00 1.33394286e-01
-3.86986315e-01 -2.85672188e-01 6.36545062e-01 1.13392055e+00
-8.02852273e-01 2.41364419e-01 -7.23543942e-01 -8.91534686e-02
6.41250491e-01 7.13397026e-01 -6.22174144e-01 -3.43201458e-02
-1.64101267e+00 5.20748436e-01 3.94357324e-01 -2.48296082e-01
-9.34393525e-01 -2.58754045e-01 -7.70595908e-01 1.11818619e-01
3.45174015e-01 -3.11528563e-01 1.38104904e+00 -9.18638587e-01
-1.34760964e+00 6.92463219e-01 1.40467346e-01 -6.18245959e-01
7.25663424e-01 4.62329805e-01 -9.60621059e-01 2.29483083e-01
-2.56368984e-02 -1.47955611e-01 6.11174226e-01 -1.45231247e+00
-8.22135806e-01 -6.45911515e-01 1.35621399e-01 5.30890673e-02
-8.66207898e-01 3.25483978e-01 -2.32929215e-01 -6.82716891e-02
2.41219640e-01 -7.31326103e-01 -1.41607404e-01 3.23248208e-01
-2.03544021e-01 -4.33225185e-02 1.42617130e+00 -5.18048048e-01
1.01618433e+00 -2.21768522e+00 -7.29499757e-01 4.71023232e-01
3.49173278e-01 4.78095978e-01 1.89322859e-01 5.17718613e-01
4.22425658e-01 3.34159136e-01 -4.89708744e-02 -2.07315058e-01
2.01767281e-01 3.01998854e-01 -3.21817160e-01 7.24149525e-01
-5.67281485e-01 7.42299676e-01 -8.14764380e-01 -6.32304013e-01
-3.58261429e-02 6.27345681e-01 -1.48132727e-01 -8.01410452e-02
-1.13039754e-01 4.92125005e-01 -8.47517669e-01 6.05533779e-01
1.07440341e+00 -4.12486732e-01 6.07150018e-01 -1.16704553e-02
1.11979179e-01 1.61647588e-01 -1.41539979e+00 1.35720360e+00
-3.82277310e-01 -2.87999988e-01 5.46786010e-01 -3.37408751e-01
9.77132738e-01 1.09167218e+00 5.20285726e-01 -5.93125999e-01
6.08029664e-01 5.24050355e-01 -4.98953849e-01 -3.42106819e-01
2.96948347e-02 -1.40046095e-02 -1.86346158e-01 9.86396909e-01
-2.07620621e-01 8.54470670e-01 -9.16716218e-01 3.08572501e-01
8.13270628e-01 6.68165237e-02 -1.24332225e-02 -4.36042473e-02
1.07106709e+00 -1.91664577e-01 9.70693231e-01 3.77679139e-01
-5.08201599e-01 -1.62350498e-02 2.46329814e-01 -2.98689216e-01
-6.34192050e-01 -8.84455442e-01 2.41923198e-01 6.36871934e-01
7.64620245e-01 -5.69497943e-01 -9.81370211e-01 -1.16097438e+00
2.37496365e-02 5.89201331e-01 1.42976433e-01 -5.54526925e-01
-2.97093809e-01 -2.92802881e-02 1.23074055e+00 2.16175571e-01
9.78514075e-01 -8.44743550e-01 -2.12799460e-01 1.79437980e-01
-5.27902186e-01 -1.26796913e+00 -4.29542452e-01 -3.00423443e-01
-4.42392081e-01 -1.39794433e+00 -6.29791059e-04 -7.25757122e-01
8.07855129e-01 4.72712994e-01 1.27812579e-01 3.57890099e-01
3.81878525e-01 1.05327010e-01 -2.79141486e-01 -4.13416088e-01
-8.21454048e-01 -1.96835011e-01 4.06752855e-01 5.91703951e-01
7.04005003e-01 -3.87973517e-01 -5.34955919e-01 4.89918262e-01
-1.11950254e+00 -3.25005829e-01 4.01790887e-02 4.19122964e-01
2.00481877e-01 2.86992818e-01 8.59711170e-01 -1.04366052e+00
3.88612032e-01 -4.21758741e-01 -8.67671430e-01 5.76955974e-01
-9.91168141e-01 -8.81115422e-02 1.18317556e+00 -4.11496639e-01
-1.24874985e+00 7.66719505e-02 -5.06888665e-02 -4.10731822e-01
-2.19014902e-02 -9.32011846e-03 -1.17179465e+00 -5.65794647e-01
2.12009698e-01 4.88276184e-01 3.54503572e-01 -5.65072477e-01
4.90169972e-01 1.25527453e+00 4.31301624e-01 -3.15709531e-01
1.00473690e+00 7.07831144e-01 -3.02457735e-02 -8.52351859e-02
-7.94556066e-02 -1.41975716e-01 -1.55210152e-01 -8.94407034e-02
6.58717394e-01 -1.06236577e+00 -1.49543786e+00 9.67242241e-01
-1.23028290e+00 4.25583869e-01 7.38769174e-02 5.75700402e-01
-2.98084497e-01 6.38512552e-01 -6.86345041e-01 -1.05914414e+00
-6.86794400e-01 -1.02187109e+00 3.55600744e-01 2.95651019e-01
1.24590009e-01 -7.73924947e-01 -6.04861677e-01 6.78093016e-01
5.22095859e-01 -5.02914116e-02 7.36639619e-01 -1.20068610e+00
-7.98592091e-01 -6.04385018e-01 -1.56367436e-01 4.25890207e-01
3.37780833e-01 -4.00359213e-01 -1.09203780e+00 -3.56391639e-01
4.83054399e-01 -1.27850950e-01 -1.50223315e-01 -5.17865300e-01
8.71304393e-01 -9.95334089e-01 -1.72402665e-01 5.89055836e-01
1.45183659e+00 3.89062017e-01 7.04822183e-01 3.75250667e-01
7.69957602e-01 5.24264514e-01 4.56974208e-01 5.57005167e-01
7.23247588e-01 3.39594156e-01 6.06738746e-01 1.71285048e-01
3.96216810e-01 -6.49682879e-01 3.24433655e-01 4.52460021e-01
4.23940569e-01 4.63852286e-03 -2.13782325e-01 8.67705420e-02
-1.78456116e+00 -8.40478659e-01 -4.90758121e-01 2.47469425e+00
9.03105676e-01 -1.80726618e-01 -1.79663256e-01 4.00143445e-01
7.96446860e-01 -1.70564711e-01 -4.47839767e-01 -2.54404992e-01
-8.91283974e-02 -3.43983769e-02 9.63824511e-01 4.43718076e-01
-6.68799698e-01 1.06007612e+00 4.19655800e+00 7.16500938e-01
-1.10420883e+00 6.97731972e-01 2.33853146e-01 2.76480705e-01
-5.81865370e-01 6.07504547e-01 -7.42606521e-01 7.80626237e-01
8.63583267e-01 -4.80965704e-01 4.99408215e-01 1.08115220e+00
2.00267360e-01 3.96894157e-01 -6.81721687e-01 8.45557690e-01
-1.76833823e-01 -1.27561069e+00 -7.75463432e-02 2.44370237e-01
1.82017788e-01 -4.61218596e-01 -3.01012516e-01 -6.49662316e-02
4.31128353e-01 -5.16982675e-01 7.64961302e-01 4.13073450e-01
8.68615866e-01 -1.11537242e+00 6.57305479e-01 6.03550971e-01
-1.03016400e+00 -1.85252264e-01 -2.69330174e-01 2.60178238e-01
7.50391111e-02 1.76636979e-01 -5.06268799e-01 1.05325782e+00
5.94523609e-01 4.32624035e-02 -2.58412898e-01 6.57084584e-01
-3.53676498e-01 5.13598084e-01 -1.88631877e-01 1.08543962e-01
1.04334436e-01 -3.55199486e-01 1.71938226e-01 5.50388336e-01
1.65527314e-01 1.33242175e-01 -7.65712559e-02 7.79975712e-01
-4.78962064e-01 3.01590741e-01 -7.70149410e-01 7.97534138e-02
1.17665982e+00 1.15386319e+00 -3.73593062e-01 -3.12117845e-01
-4.58724856e-01 1.20212340e+00 -1.61641821e-01 2.02836230e-01
-7.00862885e-01 -4.96016383e-01 7.00224936e-01 6.45539761e-02
4.05470133e-02 1.27765480e-02 -2.05261990e-01 -1.09596455e+00
1.73433587e-01 -1.01334834e+00 5.94217122e-01 -5.03237247e-01
-1.02114522e+00 4.64504898e-01 -5.75397491e-01 -1.40041566e+00
3.59969467e-01 1.48498088e-01 -5.84542274e-01 8.00425529e-01
-1.61941552e+00 -1.63053322e+00 -4.54301499e-02 1.22394991e+00
-5.07109702e-01 -1.31701395e-01 1.22916293e+00 4.51425582e-01
-3.93873274e-01 1.07402241e+00 3.71362381e-02 3.25318009e-01
5.47803819e-01 -4.68995452e-01 1.77755207e-01 8.33912671e-01
-9.89566743e-03 9.17503119e-01 2.69328147e-01 -8.74768555e-01
-1.68563807e+00 -1.17460692e+00 1.41405535e+00 -1.98606491e-01
2.97484457e-01 -4.15081531e-01 -1.12645423e+00 8.43439519e-01
-2.55961746e-01 1.79872259e-01 8.19525003e-01 -6.28132582e-01
-8.48333120e-01 -4.65781271e-01 -1.82679641e+00 4.27398950e-01
7.28608727e-01 -9.02730942e-01 -2.65873168e-02 2.16730461e-01
9.91387963e-01 1.69808224e-01 -7.55122900e-01 1.43972442e-01
6.24541104e-01 -8.80465209e-01 6.28569126e-01 -5.16536832e-01
-5.27171552e-01 -7.08816946e-01 -1.29586771e-01 -3.82418185e-01
2.28684962e-01 -8.51193249e-01 -2.68236607e-01 1.72250307e+00
-3.73550095e-02 -1.11724687e+00 9.42154348e-01 1.04736221e+00
3.99383545e-01 -5.33189401e-02 -1.28907728e+00 -9.38174963e-01
-1.40909880e-01 -4.59368020e-01 1.40565550e+00 1.31932533e+00
1.65558398e-01 -2.32057095e-01 -5.54464400e-01 5.99669516e-01
1.00538826e+00 -3.33753198e-01 7.13972449e-01 -1.32083797e+00
3.25836182e-01 1.06291182e-01 -2.75564820e-01 -4.78529751e-01
2.46548057e-01 -9.83550906e-01 -5.04859149e-01 -1.36222088e+00
-2.46660158e-01 -8.12982976e-01 -4.96468246e-01 8.24534714e-01
3.70340884e-01 -1.63365677e-02 2.01471105e-01 6.11048281e-01
-6.40188992e-01 5.02455413e-01 1.05419695e+00 1.99644849e-01
-1.68452710e-02 3.36463064e-01 -1.08463907e+00 5.12554765e-01
8.91056955e-01 -8.88687313e-01 -7.46872127e-01 -1.08149044e-01
-8.00507292e-02 1.85299471e-01 4.94060606e-01 -8.06983411e-01
6.84102237e-01 -2.70125270e-01 -1.90060809e-01 -1.53431654e-01
-1.12723159e-02 -1.56524563e+00 6.12834096e-01 7.94256866e-01
-3.20903175e-02 -2.60738403e-01 -5.52024543e-01 5.32603562e-01
4.66823652e-02 -7.34514594e-02 6.28468692e-01 -2.11125478e-01
-4.59370345e-01 5.54648399e-01 -3.35672408e-01 -5.26019871e-01
1.37114525e+00 -1.82883829e-01 -3.89594495e-01 -4.36291009e-01
-2.95224190e-01 4.84405071e-01 8.94189298e-01 4.20839757e-01
5.78570902e-01 -1.32812572e+00 -4.02877539e-01 4.35966253e-01
2.04226032e-01 -1.08606249e-01 1.54664800e-01 6.54699206e-01
-3.13407302e-01 2.91228175e-01 1.57772824e-02 8.64444450e-02
-1.46734369e+00 7.68744469e-01 2.90767252e-01 1.33303449e-01
-5.38741887e-01 3.82125288e-01 -3.06329548e-01 -5.24465382e-01
6.00980282e-01 4.43472773e-01 -8.35619215e-03 -2.49621376e-01
7.69318283e-01 4.01570678e-01 1.70834228e-01 -8.27029705e-01
-5.17699540e-01 -6.51101246e-02 -4.01434824e-02 -1.38700962e-01
7.01599300e-01 -5.16710818e-01 -4.54895556e-01 -1.76551536e-01
1.32655418e+00 4.10347916e-02 -7.56117702e-01 -6.40567005e-01
2.35001683e-01 -5.91499269e-01 -6.51729405e-02 -7.24988222e-01
-1.23112226e+00 2.72409797e-01 5.36943376e-01 -1.82953611e-01
8.73988509e-01 -2.75394768e-01 1.36994791e+00 1.18840866e-01
1.09330344e+00 -9.49586093e-01 -4.03857708e-01 -6.58731610e-02
-6.18428923e-02 -8.78507376e-01 -3.07223260e-01 -5.37618160e-01
-8.78462315e-01 6.58205807e-01 7.46015847e-01 5.69187939e-01
6.68682694e-01 1.72162667e-01 4.49340582e-01 -6.69398252e-03
-4.64992881e-01 3.50899160e-01 -5.89823186e-01 8.48643184e-01
-4.10837054e-01 -8.04168433e-02 -6.03025079e-01 1.36079836e+00
-1.76659487e-02 2.64270782e-01 5.93101561e-01 1.32570529e+00
-1.94755554e-01 -1.73163390e+00 -4.15431470e-01 -1.52267650e-01
-7.34860361e-01 1.10427700e-01 -3.92691076e-01 4.07095611e-01
1.49978951e-01 1.25864422e+00 -5.25181890e-01 -5.80801547e-01
1.92858167e-02 3.38503271e-01 -1.22010037e-01 -1.59722567e-01
-6.95747256e-01 -1.46884635e-01 -3.52182575e-02 -5.43602824e-01
-2.26227462e-01 -1.87316954e-01 -1.87650871e+00 -6.40474916e-01
-8.03188741e-01 7.02480078e-01 9.77665305e-01 7.87449658e-01
3.44102502e-01 -2.55023479e-01 1.35849345e+00 4.63293016e-01
-8.51835608e-01 -1.65344864e-01 -8.89578402e-01 4.00484443e-01
1.05825327e-01 -1.50675192e-01 -5.47957063e-01 7.71053182e-03] | [5.8123602867126465, 6.734630584716797] |
ee9e480c-d9b3-4168-b0cd-4bc9f36c9788 | scene-text-recognition-with-full | 2109.01034 | null | https://arxiv.org/abs/2109.01034v1 | https://arxiv.org/pdf/2109.01034v1.pdf | Scene Text recognition with Full Normalization | Scene text recognition has made significant progress in recent years and has become an important part of the work-flow. The widespread use of mobile devices opens up wide possibilities for using OCR technologies in everyday life. However, lack of training data for new research in this area remains relevant. In this article, we present a new dataset consisting of real shots on smartphones and demonstrate the effectiveness of profile normalization in this task. In addition, the influence of various augmentations during the training of models for analyzing document images on smartphones is studied in detail. Our dataset is publicly available. | ['James Amelia', 'Robert Leer', 'Hessi Roma', 'Russell Elijah', 'Gerald Carl', 'Nathan Zachary'] | 2021-07-13 | null | null | null | null | ['scene-text-recognition'] | ['computer-vision'] | [ 6.16525412e-01 -4.80519563e-01 -2.04649940e-01 -4.13919419e-01
-4.06200111e-01 -2.98165023e-01 8.40631366e-01 -1.79153867e-02
-4.49902415e-01 3.10464829e-01 1.39808431e-01 -2.68712699e-01
1.44865826e-01 -3.16068023e-01 -3.26039076e-01 -4.51336950e-01
4.97626215e-01 -8.28273520e-02 3.82357568e-01 -2.09830016e-01
5.53235352e-01 6.58321619e-01 -1.68092704e+00 3.16816270e-01
5.62164724e-01 7.36401200e-01 3.34110469e-01 9.06629264e-01
-1.07410625e-01 6.75928473e-01 -9.51727033e-01 -6.75772369e-01
5.91911301e-02 -3.08837503e-01 -5.21943450e-01 6.23960912e-01
5.39564788e-01 -1.80758134e-01 -5.59514523e-01 8.65520537e-01
6.61945641e-01 3.97321463e-01 5.46178758e-01 -1.05963540e+00
-5.40066779e-01 1.30755231e-01 -6.48935735e-01 4.31223273e-01
4.15983170e-01 -1.45429879e-01 6.22958064e-01 -6.85322285e-01
4.47705507e-01 7.36558676e-01 4.08669353e-01 3.39762419e-01
-6.28353715e-01 -4.54655290e-01 -7.97478035e-02 4.08089459e-01
-1.37711322e+00 -7.11692810e-01 9.08919394e-01 -4.64775890e-01
1.08426154e+00 1.86614364e-01 5.35307944e-01 1.33145010e+00
1.63789988e-01 9.23673213e-01 1.07264948e+00 -8.50125611e-01
4.58431840e-02 3.71209174e-01 4.71229553e-01 5.15476406e-01
2.70751804e-01 -4.90022510e-01 -6.80329204e-01 1.12700835e-01
5.75447857e-01 3.28837544e-01 -1.08403109e-01 -7.97753558e-02
-9.59482372e-01 3.73323351e-01 -1.95032373e-01 6.01134717e-01
4.80482988e-02 -6.73104078e-02 3.14776748e-01 -2.60352399e-02
5.09906828e-01 2.95079052e-01 2.71147545e-02 -8.67578864e-01
-1.07549345e+00 -6.56665638e-02 5.59757173e-01 9.08266604e-01
2.53556818e-01 -1.00022674e-01 -3.63482088e-02 1.21337497e+00
-6.54083341e-02 4.92134750e-01 8.10005665e-01 -7.89745867e-01
5.97957015e-01 8.39253008e-01 2.83729229e-02 -9.39893365e-01
-3.37396622e-01 -2.49010667e-01 -5.75383127e-01 -3.04445535e-01
4.66500551e-01 9.38968956e-02 -8.21280062e-01 9.24184144e-01
2.48635441e-01 4.04669285e-01 -3.24259967e-01 5.93711019e-01
4.00437325e-01 4.36751485e-01 -2.28790835e-01 4.36586440e-02
1.31243634e+00 -9.11338866e-01 -1.09941101e+00 -2.94638366e-01
5.54562569e-01 -1.00458276e+00 1.42119896e+00 6.42961919e-01
-7.59232461e-01 -5.08507788e-01 -1.14121127e+00 -3.51046741e-01
-5.90022743e-01 4.21099842e-01 6.87433898e-01 1.22614467e+00
-6.39510453e-01 4.72876221e-01 -8.52552891e-01 -8.21725965e-01
5.59401810e-01 3.37556124e-01 -3.41560036e-01 -1.32308245e-01
-7.11188078e-01 5.89430571e-01 -7.13309180e-03 -7.46926619e-03
-1.80192694e-01 -1.40446365e-01 -6.73890233e-01 -2.01613188e-01
5.64387202e-01 -5.66933304e-02 1.45754230e+00 -5.69309056e-01
-1.81660819e+00 8.39184046e-01 -2.27601305e-01 -2.47149497e-01
6.36899114e-01 -5.67419410e-01 -6.69320226e-01 7.30019882e-02
-2.70468712e-01 7.01099038e-02 1.06038463e+00 -7.50356197e-01
-5.51752746e-01 -6.07516527e-01 2.87275612e-02 2.79925883e-01
-9.09584284e-01 4.77273345e-01 -1.01186001e+00 -6.33490801e-01
-1.73677459e-01 -9.94906127e-01 1.27847612e-01 -4.56093282e-01
-4.83123332e-01 7.58209499e-04 9.07290757e-01 -7.49576092e-01
1.47208536e+00 -2.35905266e+00 -7.70705566e-02 9.77150574e-02
-2.11868718e-01 3.98820251e-01 3.20965409e-01 4.03819531e-01
1.79794461e-01 1.03447594e-01 -1.00214012e-01 -6.65722370e-01
-1.36184007e-01 -5.00600412e-03 -3.08135927e-01 5.45628667e-01
-2.48083532e-01 6.91077113e-01 -5.57494938e-01 -5.16971111e-01
5.60302913e-01 5.71596682e-01 -4.39880975e-02 -5.70501052e-02
2.01571107e-01 1.16072275e-01 -2.98639357e-01 6.98713005e-01
3.78263116e-01 -1.52068019e-01 -6.04174100e-02 2.22917810e-01
-1.71958700e-01 4.35187891e-02 -1.03665721e+00 1.61443806e+00
-4.30664062e-01 1.12005925e+00 -3.33883703e-01 -7.30197847e-01
6.56294584e-01 1.18268304e-01 1.91122115e-01 -6.73243225e-01
4.59402829e-01 -5.73027954e-02 -9.86659080e-02 -6.27172828e-01
1.19331944e+00 2.51088738e-01 1.36836886e-01 3.61323267e-01
-3.31876338e-01 -1.77634582e-01 6.56610250e-01 3.80519032e-02
8.83342147e-01 1.02286078e-01 2.05603257e-01 3.00655007e-01
4.55544591e-01 -1.18627645e-01 -1.60480142e-01 6.82024121e-01
-3.83436799e-01 8.94684911e-01 2.79403359e-01 -9.58817825e-02
-9.25108016e-01 -5.03504038e-01 -2.57687956e-01 1.07430732e+00
-4.61550476e-03 -5.99405587e-01 -9.62050557e-01 -6.40591860e-01
-2.59707421e-01 6.72869563e-01 -6.75885201e-01 7.68494830e-02
-4.93381709e-01 -8.43305290e-01 6.64574087e-01 5.73329329e-01
5.90931535e-01 -9.77225959e-01 -6.94983542e-01 -1.48212552e-01
-3.06963801e-01 -1.60211217e+00 -4.38096255e-01 -1.36373192e-01
-1.03103006e+00 -1.16094911e+00 -1.02233005e+00 -4.95683968e-01
5.19948304e-01 9.34314430e-01 4.72610861e-01 -1.58472210e-02
-6.42356992e-01 7.64452577e-01 -5.55553734e-01 -7.51001000e-01
-1.66885898e-01 4.38059777e-01 -8.25192556e-02 2.36680388e-01
5.82507968e-01 -1.50798261e-01 -3.15743655e-01 5.89025199e-01
-8.92010629e-01 -1.77811366e-02 2.93757886e-01 3.36848706e-01
2.96559930e-01 1.73951983e-01 1.62166054e-03 -9.76518452e-01
8.91016960e-01 -1.87312692e-01 -5.66649377e-01 2.20115781e-01
-1.63115650e-01 -3.48826677e-01 3.38358670e-01 -3.92397344e-01
-1.23837948e+00 -6.35162890e-02 -1.67936832e-01 -4.17213030e-02
-4.36065614e-01 1.67785749e-01 -6.28830716e-02 7.43101835e-02
6.03897333e-01 1.23932242e-01 -2.90548414e-01 -7.16731429e-01
6.74124658e-02 1.30200040e+00 3.62333685e-01 -2.03791410e-01
5.55866361e-01 6.98469043e-01 -1.50759295e-01 -1.88254118e+00
-7.43409276e-01 -8.61524284e-01 -6.03594005e-01 -3.38067532e-01
6.47589564e-01 -6.18242800e-01 -4.95966107e-01 9.49286103e-01
-8.78014922e-01 -3.26746762e-01 -1.64790392e-01 5.45285583e-01
-2.01791570e-01 6.55531764e-01 -2.20680729e-01 -9.91430163e-01
4.74143773e-02 -1.01981473e+00 1.07866168e+00 3.01824927e-01
-2.91596174e-01 -7.63171971e-01 4.73258942e-02 8.30091357e-01
1.24729708e-01 -1.85897827e-01 2.39774570e-01 -5.60861111e-01
-3.16780031e-01 -7.45707870e-01 -6.55359030e-02 4.38026279e-01
3.27520460e-01 2.20205992e-01 -1.24572515e+00 -5.14026657e-02
1.43277824e-01 1.18755205e-02 6.62449718e-01 2.54930168e-01
1.37071359e+00 4.01215583e-01 -3.48698109e-01 4.53681439e-01
9.22875881e-01 2.14645773e-01 8.58562231e-01 7.45380700e-01
8.45313609e-01 6.30502164e-01 5.72916329e-01 4.44816232e-01
1.96059495e-01 8.64136934e-01 -1.66770797e-02 2.25266322e-01
-1.89522788e-01 -1.17143892e-01 2.25187495e-01 8.19330513e-01
-3.38199496e-01 -4.25787598e-01 -1.10654485e+00 2.18160599e-01
-1.58660090e+00 -8.27155948e-01 -2.61375129e-01 2.28769016e+00
3.27710390e-01 9.82097685e-02 1.17682144e-01 6.40603006e-01
8.29163551e-01 1.05147973e-01 -2.67731905e-01 -2.84609705e-01
-1.69529647e-01 -7.41850659e-02 4.33189392e-01 1.44182220e-01
-1.26256990e+00 8.11715841e-01 6.62093973e+00 8.39173675e-01
-1.06991625e+00 8.27740058e-02 5.88017583e-01 -1.67897552e-01
3.84495497e-01 -3.26969206e-01 -9.25448298e-01 5.96939504e-01
8.74925494e-01 1.24363512e-01 3.82972836e-01 9.31828439e-01
2.35741273e-01 -5.29702246e-01 -8.24392676e-01 1.37571406e+00
5.56230426e-01 -8.63120377e-01 -2.60623366e-01 2.66059101e-01
7.05941260e-01 2.06989259e-03 3.26845646e-01 1.13481797e-01
-3.34005415e-01 -6.77770913e-01 5.45048535e-01 2.78865367e-01
7.27045894e-01 -6.75012469e-01 8.20824444e-01 2.41135806e-01
-7.13425994e-01 -6.99866489e-02 -3.20825040e-01 -2.32493922e-01
6.14614561e-02 5.11246681e-01 -7.57114112e-01 3.96847278e-01
6.72876537e-01 7.33804345e-01 -1.04772317e+00 1.19009614e+00
-2.97193825e-01 7.76647389e-01 -3.23250592e-01 -2.91597158e-01
-2.24293217e-01 -3.06355953e-01 3.19211870e-01 1.46690547e+00
4.18477386e-01 -3.24188858e-01 -3.25261593e-01 -3.28716300e-02
-2.19550073e-01 4.06202227e-01 -6.77626610e-01 -4.89685386e-01
9.55239218e-03 1.24524903e+00 -1.22591412e+00 -2.13711813e-01
-5.85165143e-01 1.26702058e+00 6.55912533e-02 2.85700679e-01
-7.45590508e-01 -5.81386387e-01 4.54380035e-01 2.57685930e-01
5.41818179e-02 -4.48462456e-01 -3.57591748e-01 -1.29907024e+00
2.21919075e-01 -8.41795087e-01 1.53184682e-01 -6.76816881e-01
-8.05165827e-01 1.94416627e-01 -8.83379206e-02 -1.02789807e+00
4.33815345e-02 -8.55189741e-01 -4.95823175e-01 4.60816920e-01
-1.31452930e+00 -7.76403666e-01 -6.27095997e-01 5.26536345e-01
9.03430402e-01 -2.60059893e-01 6.27305806e-01 4.00791228e-01
-1.03847826e+00 5.36423981e-01 4.18113589e-01 5.52768297e-02
8.95271778e-01 -9.68061268e-01 6.63166225e-01 9.97451246e-01
5.37342012e-01 8.38295400e-01 6.41602993e-01 -5.57563007e-01
-1.30886817e+00 -6.54964268e-01 8.72932673e-01 -7.53702581e-01
5.53185046e-01 -6.11780882e-01 -7.90344656e-01 4.57141280e-01
3.03423285e-01 -3.30990046e-01 1.00811303e+00 9.77092981e-02
-9.40585956e-02 3.17412876e-02 -8.14639747e-01 8.93530786e-01
9.26193476e-01 -6.72712386e-01 -5.08883655e-01 2.32662037e-01
2.26650566e-01 -5.24253666e-01 -3.05698991e-01 -7.43481889e-02
7.02425838e-01 -1.06473422e+00 8.09183478e-01 -1.81045577e-01
2.97535658e-01 -5.63232154e-02 -4.36428487e-02 -9.03564692e-01
1.58172414e-01 -7.61119127e-01 -4.27614689e-01 1.33829713e+00
5.52726997e-05 -4.50353563e-01 1.05267894e+00 6.39257669e-01
-2.74573099e-02 -5.01349926e-01 -5.97248793e-01 -7.08606601e-01
-3.71355414e-01 -8.79872441e-01 3.24667722e-01 6.09234154e-01
1.37790576e-01 3.39071125e-01 -5.10726810e-01 -1.85704947e-01
3.28556806e-01 -3.08623254e-01 1.11410022e+00 -1.11368966e+00
-1.51381463e-01 -2.99726427e-01 -5.16670465e-01 -1.08773744e+00
7.58518418e-03 -6.52132452e-01 -3.38834375e-01 -1.60442162e+00
3.32744002e-01 1.41540974e-01 -9.04602632e-02 8.84414315e-02
-1.86473057e-01 5.97786009e-01 4.17363971e-01 1.97895691e-01
-7.37097919e-01 2.63861746e-01 9.86269653e-01 -7.69170225e-02
-2.55112827e-01 2.51821727e-01 -5.80201447e-01 1.01063108e+00
1.28266203e+00 -2.16509938e-01 -4.85191792e-01 -2.76301920e-01
2.59438723e-01 -5.73046386e-01 7.53755309e-03 -1.19487131e+00
2.03188360e-01 1.52126834e-01 4.30008411e-01 -7.60437787e-01
6.70156837e-01 -9.55978811e-01 -2.67076790e-01 -5.08875474e-02
-1.58629596e-01 2.10934654e-02 1.64258704e-01 5.77435255e-01
-6.25585988e-02 -5.83144128e-01 4.44078416e-01 1.71227023e-01
-6.86049879e-01 -1.39450774e-01 -4.82360452e-01 7.42810639e-03
9.40365613e-01 -7.73613632e-01 -1.55921578e-01 -2.53922433e-01
-2.98164159e-01 -3.15980494e-01 6.96491778e-01 7.02331364e-01
4.90670919e-01 -7.48582244e-01 -3.61506976e-02 1.96448863e-01
2.84663409e-01 -4.64276433e-01 3.73358577e-01 7.83364654e-01
-4.80035096e-01 4.96048361e-01 -6.22765124e-02 -4.59053934e-01
-1.65726316e+00 3.49790424e-01 -1.38220444e-01 -5.51264025e-02
-6.83356583e-01 5.34335792e-01 -1.74095422e-01 2.22923607e-02
4.16354448e-01 -1.83417201e-01 -5.56018054e-01 7.51618594e-02
9.56085742e-01 9.46633697e-01 3.53687316e-01 -7.76516378e-01
-1.73908338e-01 6.06978476e-01 -2.08580911e-01 -2.68408567e-01
1.14243472e+00 -2.43040830e-01 1.38194904e-01 7.84936726e-01
1.05234730e+00 1.69158190e-01 -1.08062530e+00 9.01942179e-02
1.11834690e-01 -7.90300071e-01 -9.03385133e-03 -5.38404405e-01
-6.31149054e-01 1.19539833e+00 6.64167523e-01 1.47744790e-01
8.62282574e-01 -3.19417536e-01 6.57508552e-01 6.57366157e-01
3.62263411e-01 -1.60156262e+00 2.55767524e-01 6.44868910e-01
5.14882445e-01 -1.34355426e+00 1.25130266e-01 -5.48547506e-01
-7.18427598e-01 1.08854628e+00 1.72353178e-01 3.08480293e-01
4.35654104e-01 2.85769314e-01 -8.41655582e-03 4.70463634e-02
-2.59137675e-02 -2.29462191e-01 2.47142568e-01 5.85894823e-01
6.69802964e-01 -1.62258759e-01 -2.00171933e-01 4.75302041e-01
-1.98223919e-01 2.01960266e-01 7.56970286e-01 1.24573171e+00
-2.73980677e-01 -1.16188610e+00 -6.26031995e-01 5.41848123e-01
-9.03522313e-01 1.96908060e-02 -6.24561548e-01 7.95339108e-01
-2.08689705e-01 1.18740737e+00 -6.60312548e-02 -3.25246662e-01
4.49947208e-01 3.51865530e-01 6.48945272e-01 -5.99807501e-01
-3.87583703e-01 -3.76675092e-02 1.56036578e-02 -2.82668859e-01
-3.92377496e-01 -8.85862529e-01 -7.91780114e-01 -1.54107451e-01
-3.65498364e-01 -1.20334640e-01 1.15298045e+00 9.09061670e-01
3.01076144e-01 6.14964247e-01 2.10754976e-01 -8.61624002e-01
-3.21258664e-01 -9.82510448e-01 -5.83609462e-01 3.70414495e-01
-7.06880242e-02 -7.00111926e-01 -3.03639352e-01 3.46535891e-01] | [11.82872486114502, 2.5003867149353027] |
176da6f8-466f-4fad-8cef-4fdd1ccc7cc3 | modeling-caricature-expressions-by-3d | 2008.05714 | null | https://arxiv.org/abs/2008.05714v1 | https://arxiv.org/pdf/2008.05714v1.pdf | Modeling Caricature Expressions by 3D Blendshape and Dynamic Texture | The problem of deforming an artist-drawn caricature according to a given normal face expression is of interest in applications such as social media, animation and entertainment. This paper presents a solution to the problem, with an emphasis on enhancing the ability to create desired expressions and meanwhile preserve the identity exaggeration style of the caricature, which imposes challenges due to the complicated nature of caricatures. The key of our solution is a novel method to model caricature expression, which extends traditional 3DMM representation to caricature domain. The method consists of shape modelling and texture generation for caricatures. Geometric optimization is developed to create identity-preserving blendshapes for reconstructing accurate and stable geometric shape, and a conditional generative adversarial network (cGAN) is designed for generating dynamic textures under target expressions. The combination of both shape and texture components makes the non-trivial expressions of a caricature be effectively defined by the extension of the popular 3DMM representation and a caricature can thus be flexibly deformed into arbitrary expressions with good results visually in both shape and color spaces. The experiments demonstrate the effectiveness of the proposed method. | ['Juyong Zhang', 'Jianfei Cai', 'Keyu Chen', 'Jianmin Zheng'] | 2020-08-13 | null | null | null | null | ['caricature'] | ['computer-vision'] | [ 3.92981291e-01 2.55495131e-01 1.88885212e-01 -5.20249046e-02
-1.76503256e-01 -7.00841784e-01 6.28755271e-01 -6.01262510e-01
2.40543768e-01 5.46316087e-01 -1.97751373e-01 8.40600133e-02
9.03837197e-03 -1.05252922e+00 -7.09105730e-01 -7.98084915e-01
2.67239034e-01 2.82915235e-01 -3.35489251e-02 -4.62235630e-01
2.03261636e-02 1.27607489e+00 -1.53356493e+00 4.89220321e-02
9.52309608e-01 9.94354486e-01 -1.00352943e-01 6.39977634e-01
-3.07854444e-01 5.10809302e-01 -7.31317759e-01 -9.45603371e-01
3.09496045e-01 -4.53064084e-01 -4.09709632e-01 4.74388421e-01
4.81492788e-01 -6.12276942e-02 5.81140118e-03 1.13483369e+00
4.08995390e-01 8.66114646e-02 1.04993558e+00 -1.47736597e+00
-1.05246675e+00 -9.98980850e-02 -6.49422050e-01 -7.47526407e-01
1.30881324e-01 -1.06251836e-01 4.08997357e-01 -6.68358326e-01
7.75851727e-01 1.65564609e+00 6.79842949e-01 6.79460764e-01
-1.37202656e+00 -7.41954088e-01 -5.54974750e-02 -2.62670845e-01
-1.62191606e+00 -1.67390063e-01 1.34070122e+00 -4.31161165e-01
-2.92655565e-02 7.76153445e-01 8.24277043e-01 9.19065833e-01
3.23978752e-01 6.56511903e-01 1.13213658e+00 -5.25221646e-01
4.45734113e-02 1.77270800e-01 -5.49465477e-01 7.17314184e-01
-1.20330445e-01 1.59882993e-01 3.38028759e-01 -2.48574913e-01
1.54984593e+00 7.22104236e-02 -1.96308702e-01 -5.21399617e-01
-8.21489930e-01 6.96985662e-01 2.73127526e-01 2.12599471e-01
-2.68529326e-01 1.70776665e-01 -1.12305582e-02 1.14298530e-01
5.01286626e-01 4.82085764e-01 6.71622679e-02 1.06614441e-01
-7.26053894e-01 6.56976044e-01 5.55200040e-01 1.04308617e+00
6.24408484e-01 5.82983315e-01 -1.71614498e-01 1.06783009e+00
2.65367746e-01 8.00157011e-01 1.07221693e-01 -1.02472818e+00
-2.18596101e-01 5.57887793e-01 6.05903715e-02 -1.38443220e+00
1.78380951e-01 -1.58983722e-01 -9.54067945e-01 8.95428360e-01
2.31299251e-01 -6.67239279e-02 -8.48570228e-01 1.67683494e+00
5.58114946e-01 1.89703315e-01 -6.17189929e-02 7.54012465e-01
8.49416316e-01 7.43985415e-01 8.59525949e-02 -6.14088774e-02
1.20651257e+00 -5.78861713e-01 -1.11474812e+00 2.95512110e-01
2.73911767e-02 -1.13149047e+00 1.10126162e+00 2.34611034e-01
-1.36839378e+00 -6.12314343e-01 -1.07333124e+00 -1.13285936e-01
-1.90674499e-01 1.37766525e-01 4.02124077e-01 6.37731671e-01
-9.36676264e-01 3.84536296e-01 -4.55639064e-01 6.89792261e-02
4.88432080e-01 4.25591975e-01 -3.38669211e-01 2.92462736e-01
-9.31241274e-01 8.43655050e-01 6.90692943e-03 2.53168225e-01
-5.34413695e-01 -8.01645994e-01 -9.48415935e-01 -1.80735484e-01
1.83382466e-01 -4.52472240e-01 8.89989436e-01 -1.54699969e+00
-1.87522173e+00 9.99550164e-01 7.77145252e-02 2.13029265e-01
8.67707551e-01 1.96567968e-01 -6.76011503e-01 -1.16631344e-01
-2.19825909e-01 5.42709291e-01 1.23222339e+00 -1.74562907e+00
-2.02310145e-01 -1.87884375e-01 -9.42656363e-04 1.01621650e-01
-1.92380890e-01 -7.51570091e-02 -6.27664506e-01 -1.29980278e+00
-7.23083466e-02 -9.48758245e-01 -1.68821856e-01 5.52469134e-01
-4.14391935e-01 1.49769738e-01 1.56761670e+00 -8.07831526e-01
9.30057645e-01 -2.09943533e+00 4.64326262e-01 4.55953270e-01
1.43259233e-02 2.75272429e-01 -1.57540858e-01 2.57744461e-01
-9.58819911e-02 1.55400798e-01 -4.25260842e-01 -2.47712687e-01
-9.28579941e-02 3.42404187e-01 -3.38495225e-01 3.22598994e-01
5.70004165e-01 1.02721393e+00 -5.75706959e-01 -6.14427388e-01
1.84381396e-01 1.02609205e+00 -5.43178141e-01 3.08123648e-01
-4.30476397e-01 7.01898217e-01 -6.69689536e-01 6.79607630e-01
1.00258958e+00 4.24224049e-01 -2.77386978e-02 -2.49225333e-01
-7.69373728e-03 -8.66430640e-01 -1.28280628e+00 1.24639249e+00
-6.40987098e-01 3.34158063e-01 5.45453012e-01 -6.05376422e-01
1.48070061e+00 1.90255508e-01 3.29280615e-01 -3.58544588e-01
4.17741537e-01 1.64473906e-01 -2.24152848e-01 -3.64177614e-01
4.52418685e-01 -5.12857199e-01 -1.95393205e-01 9.48861837e-02
-4.74999487e-01 -8.38073313e-01 -4.13477570e-01 -2.25912347e-01
1.91763267e-01 3.93284649e-01 -7.33125433e-02 -3.76927793e-01
8.65318537e-01 -2.74993151e-01 4.25086170e-01 -1.19052790e-01
2.03627676e-01 7.20797539e-01 4.24282014e-01 -5.15210509e-01
-1.26412582e+00 -1.09669459e+00 -1.29754856e-01 5.32026470e-01
2.93877333e-01 2.23118246e-01 -1.06111681e+00 -2.59665638e-01
3.78356203e-02 6.96918786e-01 -7.69731879e-01 -3.23902041e-01
-8.70415747e-01 -3.00493658e-01 5.58293641e-01 3.77094597e-01
6.51111722e-01 -1.10415149e+00 -2.37327471e-01 4.18444686e-02
9.71212015e-02 -8.93146396e-01 -7.94704914e-01 -7.73629546e-01
-6.42926037e-01 -9.34027076e-01 -9.85476494e-01 -1.12180257e+00
9.31687236e-01 -2.42884204e-01 7.52337456e-01 1.69858009e-01
-3.97166073e-01 3.74366611e-01 -3.73150036e-02 -4.79702055e-01
-1.05856335e+00 -5.23406088e-01 -1.67699412e-01 7.14850247e-01
-3.17967266e-01 -6.91544473e-01 -5.16507685e-01 3.49398255e-01
-1.32635689e+00 2.41836235e-01 2.97944158e-01 6.40005350e-01
8.39580119e-01 1.77063748e-01 5.52722573e-01 -9.63931203e-01
8.18753660e-01 -6.50823861e-02 -6.25923514e-01 3.11596990e-01
-2.42390454e-01 5.20408563e-02 8.43273818e-01 -8.78359199e-01
-1.37625563e+00 1.72930062e-02 -3.06713015e-01 -8.65330994e-01
1.14231659e-02 -3.77200060e-02 -5.32861948e-01 -3.81705552e-01
3.68213385e-01 1.92878008e-01 6.68530941e-01 -4.13121402e-01
6.12433910e-01 3.27097237e-01 8.57572794e-01 -9.30403769e-01
9.94361103e-01 6.25564873e-01 3.88932317e-01 -9.26726401e-01
-3.58033367e-02 4.08910692e-01 -5.58731675e-01 -4.77895945e-01
9.10355389e-01 -4.68619496e-01 -9.07479525e-01 7.43676007e-01
-1.24540865e+00 -9.76561382e-02 -3.84762049e-01 -8.08151588e-02
-7.90032446e-01 4.95055199e-01 -3.32662046e-01 -8.76470566e-01
-3.95207763e-01 -1.22819471e+00 1.00864124e+00 1.53936654e-01
-2.31689423e-01 -9.89726484e-01 3.87975909e-02 1.31433293e-01
5.11163950e-01 1.35896778e+00 1.17418337e+00 2.14243099e-01
-4.81134802e-01 -3.81036729e-01 6.15393929e-02 4.93992299e-01
3.38031650e-01 4.68273848e-01 -6.98297739e-01 -1.67320877e-01
-6.68348372e-02 3.67401280e-02 1.79750353e-01 2.09453553e-01
1.23374093e+00 -5.29497623e-01 -1.21428706e-01 6.93251431e-01
1.46573770e+00 5.85893452e-01 9.17274237e-01 -3.58214006e-02
8.48652840e-01 7.48770714e-01 4.69353199e-01 3.19342434e-01
1.61357075e-02 1.02853727e+00 5.70172906e-01 -4.80981171e-01
-3.42010558e-01 -4.26073641e-01 1.14076845e-01 3.92308593e-01
-4.46669340e-01 8.48165601e-02 -4.10924673e-01 3.19308847e-01
-1.58127284e+00 -9.38164651e-01 -1.29175410e-01 2.17794299e+00
8.36318135e-01 -3.73917937e-01 -1.25696687e-02 1.22557119e-01
9.89310265e-01 -4.44488004e-02 -4.48319852e-01 -8.34706604e-01
-1.87690243e-01 5.25673091e-01 1.24058887e-01 5.90703309e-01
-8.43768179e-01 9.52086151e-01 6.06740475e+00 1.23664939e+00
-1.42100406e+00 -1.45252526e-01 7.06531227e-01 1.75324038e-01
-6.57402396e-01 -2.99147636e-01 -2.66770363e-01 4.32378501e-01
8.26356634e-02 -1.39447600e-01 4.71241087e-01 8.45788002e-01
3.25561404e-01 5.05178988e-01 -5.19552886e-01 9.35460627e-01
3.29165310e-02 -1.36001086e+00 3.38061392e-01 -1.28226966e-01
9.44348931e-01 -1.14871466e+00 4.85202760e-01 2.26154365e-02
1.79568782e-01 -1.27160418e+00 1.01097667e+00 8.46817493e-01
1.43652546e+00 -1.16945541e+00 2.84089833e-01 1.89458385e-01
-1.11522138e+00 2.43371293e-01 -1.22123361e-01 2.45084852e-01
1.46213293e-01 5.52825220e-02 -4.78827834e-01 3.51496220e-01
3.05576146e-01 2.12099925e-01 -2.40383402e-01 6.93124533e-01
-8.28237366e-03 4.17360924e-02 -1.20991774e-01 -1.24726095e-03
3.80723029e-02 -7.11341619e-01 6.00240052e-01 9.35173631e-01
5.40641963e-01 1.84433088e-01 6.14880882e-02 1.31411111e+00
-1.26737775e-02 3.94582361e-01 -7.69387066e-01 4.46469299e-02
4.95445818e-01 1.15965319e+00 -3.86896789e-01 -1.61652416e-01
-4.52332608e-02 9.79752123e-01 6.05116487e-02 3.23113233e-01
-1.05497551e+00 -3.18012893e-01 7.70083666e-01 5.01954496e-01
7.37939551e-02 -6.35738596e-02 -3.37131441e-01 -8.40119839e-01
-1.19249992e-01 -9.12172496e-01 -1.77975550e-01 -8.67614150e-01
-1.08665121e+00 8.51018488e-01 -9.30207893e-02 -1.33850825e+00
-1.40540749e-01 -5.26441455e-01 -9.81578171e-01 1.06884170e+00
-1.22060287e+00 -1.87431085e+00 -3.65055561e-01 7.63653576e-01
2.76173085e-01 -1.95185646e-01 8.93841505e-01 1.82384059e-01
-3.07590574e-01 7.60640442e-01 4.02945019e-02 2.64049489e-02
4.34548348e-01 -9.19116735e-01 1.02760084e-01 7.44183898e-01
-2.19491154e-01 3.15371245e-01 7.74610281e-01 -6.53852999e-01
-1.46805441e+00 -1.30173540e+00 4.03764278e-01 -4.01666135e-01
1.88520357e-01 -2.69916207e-01 -8.06240916e-01 4.30395067e-01
1.02506578e-01 -1.40380878e-02 4.31149691e-01 -6.41025901e-01
-1.44686878e-01 -5.22716669e-03 -1.60776663e+00 9.91141021e-01
8.58043253e-01 -1.55210271e-01 -2.46644750e-01 -3.21696773e-02
4.90226895e-01 -5.54433763e-01 -1.04622126e+00 5.83142996e-01
7.40469396e-01 -6.76993430e-01 1.08651841e+00 -4.62299854e-01
4.42480326e-01 -5.17691314e-01 7.57735446e-02 -1.13484824e+00
-2.88671553e-01 -1.00527239e+00 -5.74854086e-04 1.36357892e+00
-3.52210999e-02 -3.62265408e-01 7.75615931e-01 7.00430453e-01
-1.46539152e-01 -7.69527078e-01 -7.55370498e-01 -7.60280848e-01
3.42920125e-01 -1.75042346e-01 1.04230344e+00 1.05103946e+00
-5.56595147e-01 -1.58213317e-01 -6.21289730e-01 1.16353892e-01
5.09272218e-01 2.83804774e-01 9.94350731e-01 -1.10696650e+00
-2.39823014e-01 -5.82927406e-01 -3.93590540e-01 -7.00416207e-01
2.76467621e-01 -8.89277935e-01 -2.84246951e-01 -1.00397265e+00
-3.09926122e-01 -8.40879202e-01 3.14103484e-01 1.71396628e-01
6.44051880e-02 4.28203374e-01 3.65304768e-01 8.80923346e-02
4.33407515e-01 8.28567028e-01 2.19523787e+00 -2.57289380e-01
-2.77931690e-01 1.26560584e-01 -7.57114291e-01 7.03154325e-01
4.88290787e-01 2.06300635e-02 -6.17172539e-01 -1.89181581e-01
-3.25683981e-01 2.32988775e-01 5.77291787e-01 -6.56711519e-01
-2.95319974e-01 -2.80732125e-01 4.02200311e-01 -2.82637060e-01
6.27925873e-01 -1.13578045e+00 9.39046144e-01 3.69409472e-01
-1.48994014e-01 6.65440038e-02 3.60833883e-01 3.17660213e-01
-2.76449949e-01 -1.85108501e-02 1.20201337e+00 1.36799350e-01
-5.29424608e-01 5.03874004e-01 7.11979419e-02 -1.73624441e-01
1.39186704e+00 -5.13972998e-01 1.27885312e-01 -4.58006442e-01
-8.45332384e-01 -3.08757573e-01 7.91547298e-01 5.68985164e-01
7.02630222e-01 -1.97496641e+00 -7.00898051e-01 6.92425668e-01
-7.86129087e-02 1.11358017e-01 3.40907723e-01 3.15604687e-01
-1.03540039e+00 -1.99927226e-01 -5.32090783e-01 -2.10445419e-01
-1.35801435e+00 5.65675676e-01 4.91701365e-01 1.10266693e-01
-4.52883780e-01 5.75274646e-01 5.60642123e-01 -4.75228280e-01
-3.93632874e-02 -1.24918722e-01 -3.73805135e-01 -2.99125761e-01
2.80646056e-01 3.44147325e-01 -3.31627637e-01 -1.04232621e+00
1.08998708e-01 1.05094588e+00 3.49079758e-01 1.11717395e-02
1.01525724e+00 1.92478746e-01 -3.37526858e-01 -1.39204308e-01
1.24022877e+00 3.85627061e-01 -1.38276470e+00 1.00132160e-01
-6.71648979e-01 -6.52415097e-01 -9.90324616e-02 -6.55321360e-01
-1.33260858e+00 6.64505303e-01 3.86220157e-01 6.73840195e-02
1.09771526e+00 -2.61105657e-01 8.57922196e-01 -3.53369445e-01
2.49270678e-01 -7.66238391e-01 1.08570568e-01 7.98952132e-02
1.51220262e+00 -6.12222075e-01 -4.32541162e-01 -8.11477959e-01
-7.45509565e-01 1.18427968e+00 4.15454358e-01 -4.01484996e-01
6.99203551e-01 5.12750685e-01 8.60878900e-02 -1.14926539e-01
-1.00595139e-01 4.36396986e-01 5.21768689e-01 7.46292531e-01
1.78586647e-01 1.34236977e-01 -3.77442449e-01 6.95188999e-01
-3.12984377e-01 -1.42577887e-01 3.20141077e-01 7.21305072e-01
-1.27867550e-01 -1.21077788e+00 -7.83668578e-01 -1.10425586e-02
-4.35865104e-01 2.74509698e-01 -3.82527560e-01 1.04972219e+00
5.30442476e-01 5.91188073e-01 1.41169161e-01 -2.31960952e-01
6.30329907e-01 -1.69537827e-01 5.16967893e-01 -1.33696899e-01
-3.19811583e-01 2.42607310e-01 -1.37247548e-01 -1.98375002e-01
-2.87431300e-01 -3.81990761e-01 -1.20333636e+00 -5.49128354e-01
-9.96592045e-02 -6.54603019e-02 4.53545630e-01 3.49131614e-01
4.02228445e-01 3.66479814e-01 8.38463962e-01 -8.27479899e-01
-3.22366685e-01 -5.23154676e-01 -7.46344507e-01 8.55336070e-01
5.74872233e-02 -8.24307263e-01 -8.97761807e-02 3.90509993e-01] | [12.443038940429688, -0.325054407119751] |
daa13aa7-5805-42a4-898f-70b7007f10d3 | lexical-characteristics-analysis-of-chinese | null | null | https://aclanthology.org/W15-3813 | https://aclanthology.org/W15-3813.pdf | Lexical Characteristics Analysis of Chinese Clinical Documents | null | ['Huilong Duan', 'Haomin Li', 'Meizhi Ju'] | 2015-07-01 | null | null | null | ws-2015-7 | ['lexical-analysis'] | ['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] | [-7.137484073638916, 3.750786304473877] |
b14a0632-eb86-4037-8061-9118670b56bb | amrs-assemble-learning-to-ensemble-with | 2306.10786 | null | https://arxiv.org/abs/2306.10786v1 | https://arxiv.org/pdf/2306.10786v1.pdf | AMRs Assemble! Learning to Ensemble with Autoregressive Models for AMR Parsing | In this paper, we examine the current state-of-the-art in AMR parsing, which relies on ensemble strategies by merging multiple graph predictions. Our analysis reveals that the present models often violate AMR structural constraints. To address this issue, we develop a validation method, and show how ensemble models can exploit SMATCH metric weaknesses to obtain higher scores, but sometimes result in corrupted graphs. Additionally, we highlight the demanding need to compute the SMATCH score among all possible predictions. To overcome these challenges, we propose two novel ensemble strategies based on Transformer models, improving robustness to structural constraints, while also reducing the computational time. Our methods provide new insights for enhancing AMR parsers and metrics. Our code is available at \href{https://www.github.com/babelscape/AMRs-Assemble}{github.com/babelscape/AMRs-Assemble}. | ['Roberto Navigli', 'Pere-Lluís Huguet Cabot', 'Abelardo Carlos Martínez Lorenzo'] | 2023-06-19 | null | null | null | null | ['amr-parsing'] | ['natural-language-processing'] | [ 3.31391305e-01 5.23367643e-01 8.95521045e-02 -3.50530505e-01
-1.03841758e+00 -9.95281816e-01 2.48604968e-01 1.71342567e-01
3.46585251e-02 6.47186637e-01 1.86131254e-01 -8.50487113e-01
-6.88363612e-02 -8.96348536e-01 -5.70709348e-01 -3.48753691e-01
2.64938205e-01 4.16911155e-01 3.31862569e-01 -3.85149300e-01
1.40329033e-01 2.08630621e-01 -1.14320219e+00 4.59412098e-01
1.07879400e+00 3.43456924e-01 2.55196542e-01 8.91544342e-01
-3.02176684e-01 8.01404595e-01 -6.35924995e-01 -1.23007810e+00
3.64886671e-02 -5.39611340e-01 -8.33371997e-01 -4.45862621e-01
3.40016901e-01 2.33612016e-01 -2.74997801e-01 1.20292103e+00
5.83002210e-01 5.46715397e-04 2.46391520e-01 -9.48122144e-01
-6.54797971e-01 1.41130710e+00 -4.42761600e-01 3.44948173e-01
5.32568991e-01 5.09247370e-03 1.49069750e+00 -7.96486437e-01
7.38955855e-01 1.03516018e+00 6.93995357e-01 8.20065916e-01
-1.20997214e+00 -6.90673888e-01 4.72598135e-01 1.17926665e-01
-1.05198765e+00 -5.00561714e-01 9.77385581e-01 -2.52694368e-01
1.10623789e+00 5.66084266e-01 2.78908461e-01 1.20825946e+00
2.81003192e-02 7.41773605e-01 1.19420409e+00 -5.52492082e-01
-1.12782016e-01 -4.28458899e-01 4.10709679e-01 9.61846530e-01
3.10735792e-01 -9.39236358e-02 -4.75280374e-01 -8.22066963e-02
4.31554317e-01 -5.41733205e-01 -9.12774801e-02 1.18166335e-01
-9.09225523e-01 8.53612721e-01 5.42551354e-02 4.32038933e-01
-5.55872396e-02 -7.30459094e-02 1.88621998e-01 2.21295968e-01
6.48538053e-01 5.67634165e-01 -7.18958855e-01 -3.25243026e-01
-5.64660132e-01 4.33686711e-02 8.98396254e-01 9.86357272e-01
6.81916773e-01 8.70814268e-03 1.01520836e-01 1.01845837e+00
2.97133118e-01 2.30989799e-01 -5.47856539e-02 -1.00728142e+00
9.00945783e-01 5.70214093e-01 -5.00156999e-01 -8.46571028e-01
-4.52684700e-01 -4.90790188e-01 -6.07160211e-01 -1.16981708e-01
4.55446154e-01 -1.90268725e-01 -9.46641862e-01 1.78945339e+00
3.71915817e-01 -5.99288754e-02 1.50780445e-02 5.33783972e-01
9.95828450e-01 5.01247764e-01 2.49465659e-01 -2.12844491e-01
1.20527351e+00 -1.00547647e+00 -6.71318710e-01 -4.92776901e-01
1.14211094e+00 -8.53313446e-01 9.36806262e-01 3.76975209e-01
-1.21322763e+00 -4.45445716e-01 -7.79588819e-01 -3.43463719e-02
-1.82039693e-01 2.21972153e-01 7.48576403e-01 7.68875778e-01
-1.11701930e+00 9.66038585e-01 -1.03745651e+00 -3.87513265e-02
1.49201840e-01 3.52226049e-01 -4.84258771e-01 3.51223461e-02
-1.08391666e+00 9.71795261e-01 5.16886115e-01 1.14432737e-01
-2.90343672e-01 -3.55295807e-01 -1.01963747e+00 -1.01560421e-01
5.53729475e-01 -4.62327152e-01 1.35603416e+00 -8.18328202e-01
-1.53931725e+00 6.50600612e-01 -2.20950216e-01 -2.13338420e-01
4.96091545e-01 -2.95642942e-01 -5.02791107e-01 -2.08670527e-01
-1.34276941e-01 1.43116057e-01 2.75140822e-01 -1.10736620e+00
-1.80970550e-01 -4.04213071e-01 8.86493549e-02 -1.52196571e-01
6.44923225e-02 2.69020796e-01 -4.19238985e-01 -7.55233049e-01
3.77106398e-01 -9.86066580e-01 -3.46626252e-01 -1.01700211e+00
-7.29977012e-01 -2.70745069e-01 3.21127713e-01 -1.03099084e+00
1.67215645e+00 -1.85414875e+00 4.17250037e-01 7.84503967e-02
3.37739378e-01 3.16258103e-01 -3.21698725e-01 7.04270899e-01
-2.26696417e-01 6.04958236e-01 -5.26538134e-01 -4.68631536e-01
-1.09660625e-01 3.27166170e-01 -3.25972810e-02 1.45608804e-03
4.61479157e-01 9.02990460e-01 -9.05586839e-01 -5.21153271e-01
1.55174062e-01 3.12034041e-01 -5.23284674e-01 2.00827286e-01
-3.52986038e-01 6.24692082e-01 -4.29597437e-01 6.60734236e-01
7.05893576e-01 -1.81858838e-01 8.53185654e-01 6.69701956e-03
-1.15355439e-01 8.64795744e-01 -1.06436598e+00 1.71290267e+00
-2.89095581e-01 2.21272528e-01 -4.79667559e-02 -9.88322377e-01
1.09669650e+00 6.63091987e-03 8.44960362e-02 -3.65697920e-01
1.50879100e-01 4.08304930e-01 3.00822705e-01 -7.27379248e-02
5.14039040e-01 3.80148292e-01 -1.94825128e-01 3.13539654e-01
1.16959035e-01 -1.10100664e-01 5.66859305e-01 3.58510077e-01
1.48024392e+00 7.26141393e-01 4.71953541e-01 -3.25754918e-02
4.43974674e-01 -1.27529912e-02 8.54070783e-01 7.33989954e-01
3.57862958e-03 5.90173304e-01 6.87997639e-01 -3.27599138e-01
-1.09672344e+00 -9.49527025e-01 1.45453811e-01 1.11914444e+00
-3.45169216e-01 -9.34629679e-01 -7.93546140e-01 -1.05441666e+00
-3.84191722e-01 1.04710639e+00 -5.02372324e-01 4.91042174e-02
-1.15948403e+00 -8.23182881e-01 6.67466581e-01 6.44643545e-01
9.45632905e-02 -1.04106033e+00 -1.06841862e-01 3.04364681e-01
-3.67057323e-01 -1.21320522e+00 -1.07367836e-01 2.65501648e-01
-8.18672478e-01 -1.04031038e+00 -1.70894817e-01 -4.47919786e-01
6.05771184e-01 -1.57072112e-01 1.49854434e+00 5.27043819e-01
5.24428561e-02 1.10287189e-01 -9.00935829e-01 -2.23487809e-01
-9.55881476e-01 4.41569269e-01 -3.32377940e-01 -5.65322101e-01
2.52183080e-01 -7.92150795e-01 -2.64976114e-01 1.60155714e-01
-5.07954597e-01 3.94759119e-01 3.65703076e-01 6.53745174e-01
5.92435181e-01 -3.90520960e-01 3.78246576e-01 -1.33793950e+00
4.01333749e-01 -4.31562424e-01 -7.38983452e-01 4.16655362e-01
-5.47808886e-01 1.36928737e-01 7.12359786e-01 -9.89407375e-02
-8.75654519e-01 5.69568798e-02 -6.99437141e-01 -3.63553502e-02
-5.22053987e-02 5.76775372e-01 -2.28597865e-01 1.49464607e-01
4.28553462e-01 6.13903776e-02 -2.89744258e-01 -7.13172853e-01
5.00745893e-01 5.04267454e-01 2.26141378e-01 -8.38362932e-01
7.42798865e-01 -1.47905618e-01 -1.15697078e-01 -3.65408927e-01
-8.62656176e-01 -1.42425582e-01 -7.87972093e-01 -1.50365666e-01
7.93324232e-01 -6.66776180e-01 -2.60321826e-01 7.09531233e-02
-1.35197067e+00 -4.58076596e-01 9.95602086e-02 1.64947584e-01
-4.46313649e-01 7.15561628e-01 -7.87677050e-01 -9.24177051e-01
-5.82583547e-01 -1.19251323e+00 9.18140292e-01 7.38109499e-02
-3.28879893e-01 -7.90820599e-01 2.51110613e-01 6.10665500e-01
4.00943458e-02 2.70181984e-01 1.03520596e+00 -1.08129966e+00
-4.39947963e-01 8.44834372e-02 1.48761980e-02 2.20861882e-01
-3.54546495e-02 3.32838744e-01 -7.83077359e-01 -1.78379223e-01
-6.36796296e-01 1.76524267e-01 9.11192715e-01 8.66361409e-02
1.25029969e+00 -1.36754349e-01 -2.86495477e-01 4.19098824e-01
1.15641308e+00 1.08626403e-01 7.42929399e-01 2.99595416e-01
8.50917816e-01 6.55852616e-01 5.24588346e-01 1.90195113e-01
5.25303125e-01 7.24896669e-01 4.31675106e-01 2.37839177e-01
-3.49916637e-01 -4.45090503e-01 4.65202302e-01 1.47182071e+00
-5.30064940e-01 -6.57998502e-01 -1.12983811e+00 4.10130799e-01
-1.91609550e+00 -8.66646707e-01 -5.18999338e-01 1.95394671e+00
6.03926003e-01 2.32008263e-01 7.76654482e-02 3.06164712e-01
7.31943727e-01 3.66320372e-01 -5.53743690e-02 -8.17622840e-01
-1.56556904e-01 5.61351120e-01 3.54162455e-01 7.38017023e-01
-9.93503928e-01 1.49394739e+00 5.56814957e+00 6.69343948e-01
-5.25411785e-01 3.57948273e-01 4.73878860e-01 2.62148261e-01
-6.99701428e-01 3.46081436e-01 -9.62356269e-01 3.61572713e-01
1.17182136e+00 5.88890202e-02 3.36419582e-01 6.86457038e-01
-1.37473702e-01 3.17321807e-01 -8.16309810e-01 3.51870626e-01
-1.39564082e-01 -1.28474009e+00 -1.36029541e-01 1.56902033e-03
3.95541996e-01 2.97791600e-01 -2.32013449e-01 4.14621532e-01
8.42564702e-01 -8.48602772e-01 5.84388912e-01 1.29137129e-01
6.42673373e-01 -6.28651440e-01 6.89485013e-01 1.04327545e-01
-1.41990113e+00 1.28546327e-01 -2.56758630e-01 7.22790509e-03
2.40393519e-01 8.06077719e-01 -7.21784830e-01 1.12360966e+00
5.77902913e-01 5.56650996e-01 -7.69101977e-01 6.08098030e-01
-7.33167291e-01 8.86812687e-01 -2.23811254e-01 -9.59433094e-02
-2.31649190e-01 -2.67619342e-01 8.30497324e-01 1.44639516e+00
4.62401330e-01 2.27259427e-01 1.99063420e-01 7.49371171e-01
-4.17844243e-02 3.83597851e-01 -6.89172149e-01 -2.54769474e-01
5.65140963e-01 1.34730172e+00 -8.13154459e-01 -9.49250460e-02
-4.35609162e-01 8.13476026e-01 8.99499893e-01 1.54573292e-01
-9.62992311e-01 -1.34654135e-01 4.95249540e-01 7.94212148e-03
2.16763631e-01 -5.13227463e-01 -2.93961704e-01 -1.31715202e+00
1.38131633e-01 -1.10970211e+00 6.58844233e-01 -7.00145185e-01
-1.13978815e+00 9.47128952e-01 -5.92738912e-02 -8.51970077e-01
-3.58854949e-01 -7.91776359e-01 -6.43670857e-01 6.07906342e-01
-1.15107715e+00 -1.27075493e+00 -1.65773984e-02 1.43029705e-01
6.54977977e-01 -1.78311616e-02 1.00672746e+00 2.08374679e-01
-8.15256596e-01 6.89154267e-01 -4.33952838e-01 2.94557810e-01
3.52337301e-01 -1.43851674e+00 1.16069686e+00 1.22039986e+00
5.62947512e-01 4.25273478e-01 7.77121842e-01 -9.39679146e-01
-1.30598724e+00 -1.01107836e+00 9.26769674e-01 -8.10990334e-01
9.45800006e-01 -4.60048854e-01 -1.01107836e+00 1.02667046e+00
8.70832652e-02 -9.38933790e-02 8.02904487e-01 5.42323351e-01
-6.39418125e-01 2.66412765e-01 -7.58152544e-01 6.79578245e-01
1.54122329e+00 -2.66924024e-01 -5.96875191e-01 1.54382601e-01
9.60471809e-01 -6.45839810e-01 -1.00496590e+00 7.06283271e-01
3.50899339e-01 -1.03549349e+00 5.77190757e-01 -5.74214220e-01
5.59623361e-01 -1.81470901e-01 -4.00341541e-01 -1.43907106e+00
-4.19518441e-01 -8.54426324e-01 -2.68422782e-01 1.46995866e+00
1.17303944e+00 -8.76151443e-01 7.16841936e-01 3.41677815e-01
-4.71777022e-01 -6.75983310e-01 -1.05660927e+00 -7.16722071e-01
2.16341779e-01 -8.16337049e-01 6.57822251e-01 9.37660038e-01
8.65156725e-02 2.64789551e-01 -2.85328299e-01 3.90914321e-01
5.20970941e-01 2.06634286e-03 6.43441498e-01 -1.15307403e+00
-5.93491435e-01 -2.80533522e-01 -2.86679775e-01 -4.23910201e-01
3.43946695e-01 -1.19580889e+00 -2.69780397e-01 -1.69413781e+00
1.26248270e-01 -4.13117647e-01 -3.15858275e-01 8.11917305e-01
-3.29060882e-01 1.74829081e-01 5.49810231e-01 -4.52140272e-02
-5.81020057e-01 -2.08163355e-02 9.15247440e-01 1.65080294e-01
-1.54847965e-01 -9.23049971e-02 -7.70459890e-01 7.26988494e-01
1.35475922e+00 -7.34136939e-01 -7.83412345e-03 -6.73107266e-01
6.07551277e-01 -6.96548913e-03 6.68612793e-02 -7.21893489e-01
-1.63235858e-01 4.56747301e-02 8.11829269e-02 -5.32508552e-01
1.75725162e-01 -3.13458025e-01 5.05815506e-01 3.40615869e-01
-1.69696212e-01 6.15335226e-01 3.62815529e-01 2.61823595e-01
-1.87337678e-03 -4.09561962e-01 4.84470576e-01 -2.56201386e-01
-4.64416653e-01 1.36208907e-01 -7.81344101e-02 4.06855643e-02
7.73629010e-01 1.61410436e-01 -7.21541584e-01 -1.11255407e-01
-9.31745291e-01 3.27510051e-02 3.80493075e-01 4.38302010e-01
4.20736760e-01 -8.92696857e-01 -9.39443350e-01 5.10787368e-02
-8.24477803e-03 -1.39988065e-01 2.28084058e-01 8.23399842e-01
-5.73210001e-01 9.15787742e-03 1.67322561e-01 -2.52279341e-01
-1.54575849e+00 3.55541080e-01 1.40107676e-01 -7.64323890e-01
-5.59456944e-01 8.99039447e-01 -2.57940233e-01 -8.19650054e-01
-2.79684633e-01 3.55393402e-02 -2.78421402e-01 -1.48232132e-01
1.71676040e-01 3.41278583e-01 2.42595956e-01 -6.16953194e-01
-5.33329964e-01 6.17408395e-01 -1.66507706e-01 9.02718753e-02
1.38899219e+00 3.47282887e-02 -2.86266029e-01 9.69626307e-02
6.69558287e-01 5.90417862e-01 -6.27640843e-01 -3.16891856e-02
3.42890620e-01 -1.70222998e-01 -2.46727288e-01 -1.00884807e+00
-9.71529126e-01 7.64937937e-01 1.46507546e-01 4.33073401e-01
1.11549819e+00 2.11141512e-01 5.99142611e-01 9.60101411e-02
4.14173692e-01 -9.22786593e-01 -2.79449970e-01 6.63583994e-01
7.39549756e-01 -1.09841418e+00 1.21212760e-02 -1.04722166e+00
-7.79359400e-01 1.11512649e+00 6.29051208e-01 -5.84426560e-02
3.34971130e-01 4.17627484e-01 2.75454838e-02 -1.10558860e-01
-8.74584436e-01 -3.90939653e-01 1.76780134e-01 6.39166415e-01
8.25668037e-01 4.13278848e-01 -6.70961976e-01 7.37458169e-01
-7.15183735e-01 -5.49387157e-01 6.38905346e-01 6.84626281e-01
-1.91168055e-01 -1.78591681e+00 -1.88846439e-02 3.01864773e-01
-8.83378446e-01 -4.85024869e-01 -7.61799872e-01 7.74839699e-01
-2.02363864e-01 1.17786574e+00 -2.36791700e-01 -7.41245985e-01
4.84122634e-01 3.49547982e-01 5.92824459e-01 -7.59420455e-01
-7.50757873e-01 1.04673900e-01 9.13383663e-01 -5.49312711e-01
-3.11705858e-01 -5.77783644e-01 -1.09477067e+00 -3.29851687e-01
-4.54497546e-01 2.96519160e-01 5.57472885e-01 6.84688270e-01
5.12281954e-01 7.32876301e-01 4.30674195e-01 -4.26776379e-01
-4.21076596e-01 -1.13300145e+00 -1.24459520e-01 9.65192914e-02
-2.15849131e-01 -3.74522835e-01 -4.93476152e-01 -3.30238253e-01] | [10.452759742736816, 9.558238983154297] |
72cdf4c2-fac3-4b50-97a2-fb9a1cc50bec | synthetic-target-domain-supervision-for-open | 2204.09248 | null | https://arxiv.org/abs/2204.09248v1 | https://arxiv.org/pdf/2204.09248v1.pdf | Synthetic Target Domain Supervision for Open Retrieval QA | Neural passage retrieval is a new and promising approach in open retrieval question answering. In this work, we stress-test the Dense Passage Retriever (DPR) -- a state-of-the-art (SOTA) open domain neural retrieval model -- on closed and specialized target domains such as COVID-19, and find that it lags behind standard BM25 in this important real-world setting. To make DPR more robust under domain shift, we explore its fine-tuning with synthetic training examples, which we generate from unlabeled target domain text using a text-to-text generator. In our experiments, this noisy but fully automated target domain supervision gives DPR a sizable advantage over BM25 in out-of-domain settings, making it a more viable model in practice. Finally, an ensemble of BM25 and our improved DPR model yields the best results, further pushing the SOTA for open retrieval QA on multiple out-of-domain test sets. | ['Salim Roukos', 'Radu Florian', 'Vittorio Castelli', 'Avirup Sil', 'Rong Zhang', 'Md Arafat Sultan', 'Bhavani Iyer', 'Revanth Gangi Reddy'] | 2022-04-20 | null | null | null | null | ['passage-retrieval'] | ['natural-language-processing'] | [-5.16857244e-02 4.55922037e-02 -1.14528090e-01 -1.63257960e-03
-1.86655056e+00 -8.28331828e-01 9.34694767e-01 3.32670199e-04
-5.85461020e-01 1.26232541e+00 5.57588935e-01 -4.49391395e-01
-3.58496100e-01 -6.65420353e-01 -8.33082259e-01 -1.60049096e-01
1.25147432e-01 1.29432523e+00 5.46213627e-01 -1.06662869e+00
6.62470609e-02 3.05378810e-02 -1.24746466e+00 5.64428687e-01
1.08530831e+00 5.52269876e-01 1.33475184e-01 6.38035595e-01
-1.02822974e-01 7.15128005e-01 -8.33320439e-01 -4.63023603e-01
2.72554278e-01 -3.62062693e-01 -1.34904754e+00 -7.25555778e-01
6.68833554e-01 -4.48398829e-01 -6.44860923e-01 6.13624156e-01
9.77084935e-01 5.22484303e-01 1.16080630e+00 -4.78646547e-01
-1.40627384e+00 6.00605190e-01 -1.47354394e-01 4.33018327e-01
8.16493988e-01 2.38636397e-02 1.39169431e+00 -7.99918294e-01
1.08188510e+00 1.28337336e+00 2.98480392e-01 7.48546124e-01
-1.09942293e+00 -3.39738727e-01 -2.79500008e-01 1.32973313e-01
-1.36561036e+00 -3.02317530e-01 3.17087889e-01 5.35103306e-02
1.38631654e+00 2.20279232e-01 -1.30497158e-01 1.49950695e+00
-6.66192546e-02 1.10617936e+00 7.99641550e-01 -7.62426436e-01
2.20071718e-01 -4.09972630e-02 4.05895382e-01 6.45561591e-02
3.00704926e-01 -4.24763635e-02 -1.40568167e-01 -5.05543172e-01
5.27093887e-01 -2.67865300e-01 -6.86360121e-01 -5.83930016e-02
-1.17796648e+00 9.56649482e-01 4.79055762e-01 3.83833051e-01
-2.71091759e-01 -1.72361225e-01 4.67711836e-01 9.20765102e-01
7.29043245e-01 1.09617460e+00 -7.81288683e-01 -1.10394601e-02
-1.10270202e+00 8.60623121e-01 1.13633168e+00 8.34295571e-01
4.87258613e-01 -4.56678927e-01 -9.16869640e-01 1.46235859e+00
8.52559358e-02 7.54001617e-01 9.24445808e-01 -7.61811972e-01
7.07448661e-01 2.11944103e-01 5.40370703e-01 -2.35571787e-01
2.82824561e-02 -3.21819812e-01 -4.61435050e-01 -4.36974376e-01
6.03638172e-01 -1.28623128e-01 -1.10281026e+00 1.51729655e+00
-1.22449301e-01 -1.75077483e-01 6.00742340e-01 9.06693399e-01
1.05394375e+00 9.60641503e-01 -6.86799139e-02 1.06314704e-01
1.33276689e+00 -1.29373407e+00 -4.79355365e-01 -2.38009796e-01
7.18132496e-01 -6.52038813e-01 1.38537073e+00 1.35564879e-01
-9.68824685e-01 -3.46648484e-01 -8.56989503e-01 -5.01025975e-01
-5.88010192e-01 9.51668769e-02 1.47445723e-01 2.56892800e-01
-1.25079799e+00 4.49717134e-01 -2.58392930e-01 -7.42709577e-01
-7.95530807e-03 5.71608543e-03 -2.57034868e-01 -8.80893707e-01
-1.84819579e+00 1.12988091e+00 3.85603905e-01 -5.59955239e-01
-1.01392019e+00 -7.03164339e-01 -6.31201923e-01 2.72475153e-01
4.11457658e-01 -1.14553463e+00 1.76456141e+00 -3.13822359e-01
-1.46610332e+00 9.13373232e-01 8.62902123e-03 -8.25295806e-01
2.36559317e-01 -4.49445963e-01 -6.44088745e-01 4.92178798e-01
1.95838094e-01 8.99755418e-01 9.06717360e-01 -1.04895556e+00
-1.25485331e-01 -2.01413691e-01 3.03153217e-01 3.97636950e-01
-1.20643593e-01 3.74358892e-02 -3.66833389e-01 -4.74408954e-01
-3.99184734e-01 -7.03494251e-01 -1.43581986e-01 -5.79663217e-01
-1.44369274e-01 -7.57458091e-01 3.54101032e-01 -7.43365467e-01
1.20141995e+00 -1.78562427e+00 -9.97226909e-02 -1.79173604e-01
-1.03172541e-01 6.59604490e-01 -6.89584196e-01 8.79184246e-01
7.37943873e-02 1.16043627e-01 -6.25206083e-02 -2.53252953e-01
1.97582930e-01 1.34807706e-01 -8.21250379e-01 -1.00507975e-01
3.56231838e-01 1.19243228e+00 -1.02523458e+00 -5.10012031e-01
-3.84823382e-01 1.44960657e-01 -4.89735305e-01 2.98034936e-01
-9.51912940e-01 1.29791528e-01 -7.84060121e-01 5.31993568e-01
4.81797636e-01 -4.59067225e-01 -3.84231567e-01 6.31677330e-01
6.03411257e-01 6.34830654e-01 -6.86013639e-01 2.25494742e+00
-4.39800620e-01 6.40035570e-01 -2.09821582e-01 -6.15083158e-01
9.01157200e-01 4.11244571e-01 2.27554124e-02 -1.19773388e+00
-3.98641527e-02 4.98828322e-01 -2.65804857e-01 -4.87771958e-01
1.22977221e+00 -7.26191550e-02 -1.82265803e-01 6.45755708e-01
5.10957479e-01 -3.36169422e-01 5.08341730e-01 6.08864546e-01
1.54959750e+00 1.04067279e-02 7.84593076e-03 -2.68164277e-01
4.34384525e-01 3.03982615e-01 -1.47211358e-01 1.25393260e+00
-1.22547314e-01 1.09395015e+00 -7.32985660e-02 1.66397229e-01
-9.20664966e-01 -1.14972699e+00 -2.41271436e-01 1.25635970e+00
-8.94284248e-02 -1.08906463e-01 -6.89947069e-01 -8.61914814e-01
-9.22760373e-05 8.98252726e-01 -3.54090869e-01 -2.00904235e-01
-4.49164093e-01 -3.21117997e-01 8.77808988e-01 1.89554393e-01
2.09143490e-01 -1.34942448e+00 1.08064763e-01 3.50903213e-01
-6.51004791e-01 -9.26210999e-01 -3.20166469e-01 1.11396573e-02
-7.28686392e-01 -7.40738153e-01 -1.74252081e+00 -9.10726786e-01
1.32755339e-01 5.00436544e-01 1.88472533e+00 -1.99916959e-01
3.04537471e-02 6.67107105e-01 -8.52533877e-01 -3.02259207e-01
-6.06460512e-01 6.33764863e-01 -1.71608716e-01 -6.39606714e-01
6.42859221e-01 -3.13289732e-01 -6.80006742e-01 3.60714257e-01
-1.26564860e+00 -7.09629655e-01 5.19967020e-01 1.11009014e+00
3.66102248e-01 -4.75778073e-01 1.09573293e+00 -6.99122071e-01
1.58174253e+00 -7.78596342e-01 -3.76801372e-01 7.44126141e-01
-4.06391412e-01 2.51297027e-01 5.13891697e-01 -5.07972300e-01
-1.21390009e+00 -9.86650705e-01 -2.14169592e-01 -4.33902621e-01
3.95852700e-02 6.61319733e-01 2.17121348e-01 3.92178684e-01
1.61113191e+00 1.76672950e-01 -1.92684591e-01 -7.00949728e-01
8.46760273e-01 1.02465451e+00 3.54617298e-01 -9.74470913e-01
5.18397212e-01 7.30942562e-02 -6.72268689e-01 -6.81165099e-01
-9.95781839e-01 -1.02701175e+00 -1.58663556e-01 3.73448789e-01
5.49898565e-01 -1.26562226e+00 7.56119266e-02 9.02473480e-02
-1.17883718e+00 -4.86299247e-01 -6.16515756e-01 1.80136412e-01
-4.56384540e-01 3.93289477e-01 -8.23085487e-01 -4.22079682e-01
-8.50640118e-01 -8.19385469e-01 1.39676344e+00 2.37093180e-01
-2.00473085e-01 -1.10125470e+00 7.29536772e-01 5.79974532e-01
6.23455226e-01 -4.68240082e-01 6.98481083e-01 -1.40275478e+00
-3.57285857e-01 -3.61281961e-01 -3.44516754e-01 5.97775102e-01
-3.64352137e-01 -4.82041538e-01 -9.30076659e-01 -5.40723741e-01
-2.95881927e-01 -1.11407995e+00 1.20993710e+00 5.42528331e-02
6.91825390e-01 4.39754687e-02 -1.81369826e-01 -6.17244020e-02
1.34382832e+00 -2.90348232e-01 8.78249288e-01 4.47168499e-01
-1.83424354e-03 3.08912963e-01 8.56870413e-01 1.26269087e-01
3.89336318e-01 4.14737403e-01 -1.78808436e-01 1.05992906e-01
-4.08443749e-01 -4.27592784e-01 3.77086133e-01 6.98787093e-01
3.77777189e-01 -7.94454098e-01 -9.17900980e-01 8.84476781e-01
-1.67010081e+00 -8.77789319e-01 2.07972512e-01 2.13524413e+00
1.26306081e+00 2.80318744e-02 -1.37191594e-01 -4.47575003e-01
4.00334060e-01 1.41228110e-01 -3.13141257e-01 -3.42704326e-01
-2.04461619e-01 8.28955829e-01 1.65994570e-01 3.23254079e-01
-7.46183395e-01 1.17166889e+00 6.45519638e+00 1.35766637e+00
-6.45263374e-01 1.83762252e-01 3.38758171e-01 8.26325174e-03
-4.10862416e-01 -1.19917259e-01 -1.03755140e+00 2.34054178e-01
1.31569564e+00 -2.59000599e-01 4.66227293e-01 8.74950051e-01
-3.26491743e-01 -1.52073398e-01 -1.21374428e+00 6.20277762e-01
4.19045746e-01 -1.26739740e+00 5.03706813e-01 -1.74884185e-01
9.08048630e-01 7.43575037e-01 -9.11417603e-02 1.25593305e+00
8.18436265e-01 -9.85347450e-01 1.85349882e-01 2.48085633e-01
8.69114816e-01 -2.40716994e-01 7.11271048e-01 5.88456929e-01
-2.68141001e-01 1.12603093e-02 -8.99280071e-01 2.35697493e-01
1.61786273e-01 3.95172715e-01 -1.09849930e+00 8.52564752e-01
5.64537942e-01 3.07015479e-01 -7.75641203e-01 1.05415881e+00
-2.45379329e-01 6.37966931e-01 -4.27033037e-01 -2.12582394e-01
4.14799303e-01 7.29675516e-02 6.53226018e-01 1.27410114e+00
1.83997080e-01 -3.17550786e-02 -7.90775493e-02 7.54622042e-01
-5.40794969e-01 1.56464651e-01 -8.23190451e-01 -7.14140385e-02
3.55433643e-01 9.28492904e-01 -1.21047543e-02 -5.79395235e-01
-1.68700635e-01 1.02601588e+00 5.55941105e-01 7.03963459e-01
-4.22575176e-01 -5.72971165e-01 3.73646915e-01 -6.35555014e-02
3.94940406e-01 6.59381822e-02 3.64141703e-01 -1.66953218e+00
2.02684626e-01 -1.20796049e+00 7.86961734e-01 -1.06554878e+00
-1.84546661e+00 7.30064690e-01 1.71499297e-01 -1.30942357e+00
-7.96736121e-01 -5.06315291e-01 -2.38291085e-01 1.10767186e+00
-2.11282372e+00 -1.03044200e+00 1.88475341e-01 8.39020371e-01
7.98875868e-01 -1.45283774e-01 1.09267664e+00 2.88265228e-01
-5.46017848e-02 8.13403606e-01 1.06542361e+00 1.98089957e-01
1.38538480e+00 -1.19936788e+00 5.81086874e-01 5.16688943e-01
3.20447922e-01 9.05942857e-01 6.21681929e-01 -4.30017382e-01
-1.14321172e+00 -9.60954726e-01 8.90943050e-01 -1.01697302e+00
8.68439794e-01 -2.54989684e-01 -1.19258213e+00 5.28228819e-01
6.19204104e-01 -1.21162109e-01 6.37161016e-01 3.78773123e-01
-5.70656717e-01 1.20505035e-01 -1.13300538e+00 6.87795877e-01
7.56538570e-01 -8.07907403e-01 -1.51513469e+00 6.63267672e-01
1.25643039e+00 -4.44230795e-01 -9.16003585e-01 3.31844449e-01
8.71616378e-02 -4.29417640e-01 1.24471581e+00 -8.81023586e-01
6.18706644e-01 3.14933695e-02 -1.58543944e-01 -1.65655971e+00
-1.05424650e-01 -5.95207214e-01 -1.20352343e-01 1.31947660e+00
7.40177155e-01 -6.39320254e-01 3.96329224e-01 4.26301807e-01
-7.96611086e-02 -3.24284375e-01 -1.07278299e+00 -9.79255080e-01
8.93638909e-01 -2.12434039e-01 4.04098749e-01 8.44557047e-01
2.41602778e-01 9.93430018e-01 1.25950679e-01 -2.59696275e-01
1.19137973e-01 -6.85644969e-02 8.13763797e-01 -1.07869124e+00
-5.39751947e-01 -1.90254137e-01 2.62167901e-01 -1.66411746e+00
3.07963759e-01 -8.91447663e-01 5.72876222e-02 -1.82962894e+00
1.29429594e-01 -2.92578399e-01 -3.66286576e-01 1.39947847e-01
-3.49280268e-01 1.46866143e-01 -8.32991395e-03 3.13217461e-01
-1.21697676e+00 7.51617908e-01 1.57092953e+00 -4.70603675e-01
-9.82932076e-02 -9.31240469e-02 -8.67386043e-01 -4.09764387e-02
4.43711996e-01 -3.92133266e-01 -6.29667580e-01 -7.68438399e-01
1.18341282e-01 5.44296026e-01 7.11222962e-02 -7.54589081e-01
8.20546150e-02 3.31587672e-01 7.54093677e-02 -5.16610861e-01
4.16195452e-01 -4.81789410e-01 -7.36726582e-01 -1.53957710e-01
-6.61882341e-01 -1.19691774e-01 2.46818841e-01 6.64974570e-01
-5.69296300e-01 -5.65529048e-01 3.18291545e-01 -4.04602170e-01
-5.44616759e-01 1.07387923e-01 -2.71129489e-01 8.84186804e-01
2.60144085e-01 1.60261899e-01 -9.14932728e-01 -6.90328240e-01
-3.83492082e-01 6.49331152e-01 3.02363783e-01 6.95890069e-01
3.97288740e-01 -9.69321132e-01 -1.18459356e+00 -2.49568507e-01
7.11983025e-01 2.48606279e-01 2.37858221e-01 1.51519433e-01
-3.67450625e-01 1.08885205e+00 2.33040586e-01 -4.34593171e-01
-7.27978945e-01 4.85858321e-01 2.10948110e-01 -1.19386220e+00
-4.45005953e-01 7.40044832e-01 -1.25718059e-03 -8.01166534e-01
1.08683489e-01 -2.94394135e-01 -4.44884181e-01 -4.60755788e-02
7.62197196e-01 1.45246699e-01 4.57029521e-01 -9.92633253e-02
1.84852183e-01 1.11995963e-02 -6.19024277e-01 -5.44596076e-01
1.06264758e+00 -8.88200179e-02 4.23482619e-02 5.32253273e-02
1.25531876e+00 -2.14861318e-01 -6.72271550e-01 -7.18985319e-01
-3.38676609e-02 -1.89018667e-01 -1.27940640e-01 -1.38095355e+00
-1.21558048e-01 8.71718764e-01 3.18933189e-01 2.57932216e-01
8.67242098e-01 1.79139212e-01 1.11720347e+00 1.09189153e+00
5.88051319e-01 -1.02093279e+00 2.03511521e-01 1.15788269e+00
1.07426751e+00 -1.35148859e+00 -2.28257298e-01 2.93621212e-01
-6.09082818e-01 6.92732990e-01 4.65891749e-01 -2.17981815e-01
3.45843285e-01 -5.24254978e-01 1.90140679e-01 -2.73660421e-01
-9.27399158e-01 -3.62560093e-01 4.57462728e-01 6.70345902e-01
2.45153964e-01 -1.12246305e-01 -3.68423820e-01 5.18282354e-01
-1.88978478e-01 2.39479452e-01 3.83473128e-01 9.00976896e-01
-3.29743743e-01 -1.18455553e+00 -3.30806196e-01 4.85369831e-01
-5.74065804e-01 -4.96036530e-01 -5.30929804e-01 8.78936589e-01
-7.95025051e-01 1.11199737e+00 -1.94642752e-01 7.15292096e-02
3.76758724e-01 4.00349587e-01 4.66235191e-01 -8.30592334e-01
-7.39247799e-01 -2.34168619e-01 4.64750141e-01 -1.48440987e-01
-4.76224981e-02 -3.97371203e-01 -7.81888723e-01 1.02860816e-01
-7.87259996e-01 6.48107708e-01 4.07204062e-01 8.86257708e-01
6.36632502e-01 1.94601104e-01 3.07309240e-01 -4.16772753e-01
-1.28342652e+00 -1.67472744e+00 -4.52329874e-01 6.45330906e-01
3.23961109e-01 -4.47968364e-01 -3.68811190e-01 -3.54981869e-01] | [11.437810897827148, 7.845363140106201] |
2b05bb6a-68a5-4923-9868-c13955438e6c | context-aware-neural-model-for-temporal | null | null | https://aclanthology.org/P18-1049 | https://aclanthology.org/P18-1049.pdf | Context-Aware Neural Model for Temporal Information Extraction | We propose a context-aware neural network model for temporal information extraction. This model has a uniform architecture for event-event, event-timex and timex-timex pairs. A Global Context Layer (GCL), inspired by Neural Turing Machine (NTM), stores processed temporal relations in narrative order, and retrieves them for use when relevant entities come in. Relations are then classified in context. The GCL model has long-term memory and attention mechanisms to resolve irregular long-distance dependencies that regular RNNs such as LSTM cannot recognize. It does not require any new input features, while outperforming the existing models in literature. To our knowledge it is also the first model to use NTM-like architecture to process the information from global context in discourse-scale natural text processing. We are going to release the source code in the future. | ['Anna Rumshisky', 'Yuanliang Meng'] | 2018-07-01 | null | null | null | acl-2018-7 | ['temporal-information-extraction'] | ['natural-language-processing'] | [ 1.32047758e-01 4.98389512e-01 -4.28762168e-01 -4.02408957e-01
-5.52423477e-01 -5.89252055e-01 1.26601601e+00 4.81674790e-01
-8.42984796e-01 8.38758826e-01 7.60126173e-01 -5.63526630e-01
-3.50668520e-01 -1.08392715e+00 -6.33259594e-01 -7.81008527e-02
-6.65376246e-01 4.64750022e-01 3.52714747e-01 -2.47124344e-01
5.16335070e-01 2.66673654e-01 -1.09145927e+00 8.72723043e-01
3.75184983e-01 9.84367728e-01 3.23350638e-01 7.38306582e-01
-4.86609459e-01 1.84137905e+00 -7.50032425e-01 -2.87224889e-01
-3.17414194e-01 -5.15744328e-01 -1.71774256e+00 -8.61073077e-01
-1.79681718e-01 -2.45729014e-01 -7.38985598e-01 4.01781678e-01
1.72471598e-01 7.80241370e-01 5.35506487e-01 -5.46121836e-01
-1.00300038e+00 1.25908709e+00 1.45951271e-01 8.51793647e-01
6.97503150e-01 -1.17711894e-01 9.47136939e-01 -8.92903328e-01
1.27005374e+00 1.21817315e+00 4.40625876e-01 5.24321735e-01
-6.42373800e-01 -8.03461373e-02 4.43569779e-01 9.47143376e-01
-1.03865373e+00 -2.62540609e-01 5.50501585e-01 -1.15129404e-01
2.24923778e+00 4.06646997e-01 5.76378167e-01 1.60169780e+00
6.30414546e-01 8.87707531e-01 8.51543665e-01 -5.09008110e-01
2.10353553e-01 -7.91594923e-01 4.51868445e-01 2.85110652e-01
-4.25365448e-01 2.81669021e-01 -8.01106453e-01 2.82701671e-01
6.45176649e-01 -4.18088168e-01 7.38197640e-02 9.01419580e-01
-1.49770975e+00 5.25034964e-01 5.89370787e-01 9.80828881e-01
-5.51047862e-01 3.61695856e-01 7.68448770e-01 6.28241479e-01
3.83000284e-01 7.50495851e-01 -5.11846006e-01 -3.77067924e-01
-9.48635280e-01 3.34145650e-02 8.18894804e-01 9.18191552e-01
1.31576300e-01 -2.32290685e-01 -5.23708880e-01 4.98493850e-01
-1.24345906e-01 -2.72930831e-01 8.11579764e-01 -1.06752670e+00
7.32802808e-01 5.88952601e-01 -1.64409176e-01 -1.15096557e+00
-8.80349517e-01 -3.79664414e-02 -8.20401967e-01 -3.77848536e-01
1.19327001e-01 -1.74878895e-01 -7.10686266e-01 1.76866734e+00
-6.09745122e-02 5.15574038e-01 2.82418251e-01 8.31051886e-01
1.25549901e+00 1.16045904e+00 3.97571921e-01 -6.00656450e-01
1.46482003e+00 -8.67531002e-01 -1.31292212e+00 -6.15613222e-01
6.25520766e-01 -4.98587221e-01 5.50401688e-01 9.74088833e-02
-1.52185857e+00 -2.93220878e-01 -1.24020624e+00 -7.89532661e-01
-9.00841653e-01 -2.05113500e-01 8.12841952e-01 -4.40823376e-01
-1.09388316e+00 1.05367005e+00 -9.06974077e-01 -6.34370446e-01
1.82727445e-02 2.67810822e-01 -3.11717868e-01 5.62666476e-01
-2.05281806e+00 1.61239517e+00 1.08596671e+00 6.76037073e-01
-7.92352319e-01 -1.09009050e-01 -9.94668782e-01 3.69228125e-02
7.30772018e-01 -6.03534937e-01 1.67546761e+00 -4.68522370e-01
-1.53591573e+00 8.82182479e-01 -4.80656713e-01 -7.80660391e-01
-6.21251501e-02 -3.69313598e-01 -7.36006796e-01 1.39800906e-01
-3.47497724e-02 3.92751604e-01 2.81801999e-01 -6.02991879e-01
-5.06190717e-01 -1.03308395e-01 1.03187390e-01 8.23892504e-02
9.24488530e-02 6.45180285e-01 -1.80913523e-01 -8.03079545e-01
1.90369651e-01 -4.72381592e-01 -1.74740687e-01 -9.63311136e-01
-3.46552640e-01 -8.80582988e-01 6.47018492e-01 -8.74717355e-01
1.82381165e+00 -1.63102829e+00 3.09424281e-01 -2.74417758e-01
-1.34311303e-01 2.05121905e-01 -3.01358163e-01 1.01863360e+00
-2.52936006e-01 4.42661792e-01 -2.86201209e-01 -1.67001709e-01
-1.57754961e-02 4.28723216e-01 -6.42537177e-01 -1.99128091e-02
6.28356218e-01 1.41542244e+00 -1.16586399e+00 -6.86086059e-01
-1.01628274e-01 6.23917691e-02 2.45829478e-01 1.82355225e-01
-6.97067857e-01 1.34552240e-01 -2.53705680e-01 2.76521415e-01
-3.06307618e-02 -2.91738600e-01 5.65106153e-01 3.24478835e-01
-4.30976778e-01 1.33284533e+00 -8.22436273e-01 1.88679171e+00
-6.14467204e-01 1.18014526e+00 -3.40947181e-01 -9.65612054e-01
6.89029634e-01 1.04130757e+00 -3.55180390e-02 -9.28706169e-01
3.20826143e-01 3.26571614e-02 -5.46224117e-02 -7.63669014e-01
8.73405516e-01 4.30568196e-02 -4.33154702e-01 6.04499876e-01
3.70039284e-01 3.64276111e-01 4.91689503e-01 2.24373668e-01
1.20104373e+00 3.43109936e-01 6.18477404e-01 8.12490806e-02
4.81469125e-01 -5.84392324e-02 5.90631545e-01 6.52623177e-01
7.34650269e-02 2.46885300e-01 8.30776572e-01 -8.75843108e-01
-8.73028040e-01 -7.77617097e-01 2.00449899e-01 1.28232098e+00
-1.97639897e-01 -7.88878024e-01 -3.35733235e-01 -4.59943801e-01
-8.29443336e-01 1.13373983e+00 -8.56591463e-01 -3.38309444e-02
-1.34132159e+00 -3.77617925e-01 6.03116810e-01 8.53510916e-01
5.26088119e-01 -2.06462455e+00 -7.76627362e-01 8.10840428e-01
-6.36721849e-01 -1.10943055e+00 -1.77615359e-01 5.67148983e-01
-9.29841280e-01 -9.47886229e-01 -2.41210777e-03 -1.06641579e+00
-1.56187162e-01 -6.01749480e-01 1.30744553e+00 -1.88466370e-01
2.47346267e-01 -1.60498261e-01 -5.03717899e-01 -4.03310388e-01
-2.35670045e-01 3.66614342e-01 -2.44577974e-01 -5.22919595e-01
6.07876897e-01 -8.34347129e-01 -3.77511770e-01 -7.51087442e-02
-7.24226415e-01 1.07480235e-01 2.83197492e-01 7.06070840e-01
5.57096541e-01 6.55633956e-02 6.70079231e-01 -7.21322954e-01
8.32857370e-01 -7.58197784e-01 -2.20581010e-01 3.34082752e-01
-3.41433048e-01 1.70155555e-01 5.72717309e-01 -5.04070818e-01
-1.31600428e+00 -4.79742259e-01 8.41040313e-02 2.04794630e-01
-2.52668947e-01 1.22326934e+00 2.93146640e-01 9.90517855e-01
6.53352976e-01 4.22934033e-02 -8.68602812e-01 -2.16535375e-01
5.48956037e-01 9.89108607e-02 9.69591498e-01 -7.50871301e-01
8.18028077e-02 1.89526170e-01 4.71881889e-02 -4.79298949e-01
-9.63605165e-01 -1.47599399e-01 -7.84142554e-01 -2.88692862e-01
1.30616677e+00 -4.67524976e-01 -7.56915867e-01 2.59550929e-01
-1.78895140e+00 -7.61246443e-01 -3.61176372e-01 6.27193868e-01
-7.35167861e-01 1.07868552e-01 -1.40131867e+00 -9.27968085e-01
-4.21431363e-01 -2.38107249e-01 6.26911879e-01 2.98215985e-01
-7.26106226e-01 -1.14596975e+00 -1.41049707e-02 -2.52195984e-01
1.68008268e-01 4.57243353e-01 8.02716017e-01 -7.55194485e-01
-6.58865511e-01 1.02998398e-01 1.29892364e-01 -4.29125965e-01
-2.50431806e-01 -1.41542349e-02 -7.80842006e-01 3.86652350e-01
2.04450846e-01 -2.47463986e-01 1.07714128e+00 4.51300323e-01
8.78234863e-01 -6.84634328e-01 -5.09352148e-01 2.09041461e-01
1.01924062e+00 6.51716650e-01 8.91425908e-01 7.46547520e-01
3.14263731e-01 7.26929903e-01 6.62931502e-01 7.64071196e-02
6.91180289e-01 2.40587234e-01 2.50524223e-01 3.31675023e-01
5.42806834e-02 -1.77280366e-01 4.91077244e-01 1.09621346e+00
-4.06238556e-01 -4.45780158e-01 -1.24081647e+00 9.27343130e-01
-2.24561143e+00 -1.57845891e+00 -3.33244115e-01 1.53180623e+00
1.13182628e+00 4.10010606e-01 -3.23615581e-01 -2.21062694e-02
5.46803772e-01 4.96308506e-01 -2.42258698e-01 -1.11703980e+00
-3.35128695e-01 4.16930377e-01 -1.13920562e-01 6.22863650e-01
-1.26976824e+00 1.31319737e+00 5.98520947e+00 5.64355135e-01
-7.56105244e-01 2.78949529e-01 4.57080632e-01 -2.95166969e-01
1.40944757e-02 3.01121056e-01 -2.97380090e-01 1.08977407e-01
1.70372355e+00 -1.49595261e-01 4.08032686e-01 3.12484473e-01
3.05661172e-01 -2.90944457e-01 -1.48403585e+00 6.19946837e-01
-2.04185814e-01 -1.61254776e+00 -2.87448049e-01 -2.33630806e-01
3.77185732e-01 1.32124379e-01 -3.66128743e-01 6.12421751e-01
3.90544802e-01 -1.20288515e+00 7.92140067e-01 1.01974547e+00
5.38857639e-01 -6.18364871e-01 6.97510362e-01 4.47133750e-01
-1.39159405e+00 -2.58312494e-01 -1.02630347e-01 -5.03939986e-01
5.51297903e-01 2.93956965e-01 -7.49307096e-01 8.70036364e-01
5.95683575e-01 6.92309439e-01 -4.60823238e-01 5.31914175e-01
-7.16356218e-01 6.32005155e-01 -2.52490669e-01 -2.70832032e-01
4.78156596e-01 1.04642771e-02 7.15656042e-01 1.63736916e+00
1.71449795e-01 6.65781021e-01 -1.84980348e-01 6.98507011e-01
-1.33252442e-01 -5.90868741e-02 -6.29980505e-01 -9.33205411e-02
5.62005043e-01 9.30020869e-01 -8.43554258e-01 -6.20880008e-01
-1.18747830e-01 1.09595633e+00 6.73308790e-01 4.50453490e-01
-7.05233335e-01 -3.84368718e-01 1.40285464e-02 -2.62558341e-01
1.76892817e-01 -6.91058278e-01 -3.23208034e-01 -9.34618235e-01
1.60733476e-01 -3.20753992e-01 1.05250919e+00 -9.81845260e-01
-1.24506116e+00 9.74044979e-01 3.01086932e-01 -6.93021834e-01
-8.57429147e-01 -4.65489089e-01 -9.24810946e-01 7.38087654e-01
-1.08661103e+00 -1.06150615e+00 4.33760822e-01 4.53470826e-01
5.92226207e-01 1.73096478e-01 1.01279211e+00 3.12218070e-02
-4.20210093e-01 6.94968775e-02 -4.27158356e-01 5.49838781e-01
4.74202245e-01 -1.41233301e+00 8.78234446e-01 1.04397309e+00
9.76205766e-02 9.88622248e-01 3.80994976e-01 -8.43111217e-01
-9.11501288e-01 -1.12301600e+00 2.04501128e+00 -6.98591650e-01
8.78313601e-01 -3.51871669e-01 -1.01408517e+00 1.16558313e+00
9.79178846e-01 -4.65303898e-01 5.54550231e-01 4.29320633e-01
-3.47652793e-01 2.68861413e-01 -4.04616237e-01 8.32621157e-01
1.34214270e+00 -9.63162363e-01 -1.46430552e+00 3.49039882e-01
1.22916591e+00 -5.93144417e-01 -1.20985126e+00 2.99373627e-01
1.42784014e-01 -5.26312590e-01 7.18831778e-01 -8.90271366e-01
7.52238214e-01 -2.36265227e-01 -4.87493761e-02 -8.33768249e-01
-5.04200697e-01 -9.17702734e-01 -9.19877946e-01 1.43481648e+00
7.13073730e-01 -3.96722972e-01 1.14899538e-01 5.37190139e-01
-4.15232301e-01 -7.29236841e-01 -1.23847210e+00 -9.03395295e-01
2.50530750e-01 -8.57222140e-01 6.05401337e-01 1.24403870e+00
6.37435913e-01 8.42058718e-01 -2.61155576e-01 7.48175457e-02
-1.19243450e-01 2.47769877e-01 -2.38278344e-01 -9.11660731e-01
-2.89185364e-02 -7.55533934e-01 9.88020003e-02 -6.24216676e-01
4.85345840e-01 -1.11060035e+00 -1.34184524e-01 -2.14781356e+00
-1.28687859e-01 2.60741532e-01 -3.30249578e-01 7.38897860e-01
-1.21945471e-01 -2.95877308e-01 1.41312554e-01 -3.06567717e-02
-8.68749380e-01 4.55095172e-01 1.09570360e+00 -1.30321994e-01
-4.99129862e-01 -4.07092094e-01 -2.65532792e-01 5.61704814e-01
1.13605297e+00 -4.82997507e-01 -2.85246849e-01 -5.91960490e-01
7.58754790e-01 4.15262789e-01 2.52845079e-01 -8.13909292e-01
7.54997849e-01 -3.96435082e-01 4.37396914e-01 -1.01309597e+00
2.90369332e-01 -3.92643929e-01 8.41166750e-02 1.86217073e-02
-6.93163812e-01 4.16523427e-01 2.34755650e-01 2.10838929e-01
-5.07173598e-01 -2.47906730e-01 -5.69580831e-02 -4.68034178e-01
-9.38786566e-01 3.15259546e-02 -1.12102556e+00 -2.09339359e-03
6.24609768e-01 1.22493125e-01 -7.14575410e-01 -4.70427692e-01
-1.03538740e+00 3.18467170e-01 -3.38686019e-01 6.36716902e-01
6.61327779e-01 -1.40015101e+00 -6.50336683e-01 -6.16096497e-01
-2.61335462e-01 3.32406998e-01 8.38894323e-02 6.46863103e-01
-2.12022960e-01 7.33841300e-01 1.66849330e-01 -4.35011350e-02
-6.50990725e-01 1.25699699e+00 2.40378037e-01 -7.53185689e-01
-6.17406905e-01 6.45804763e-01 -3.45888197e-01 -1.39922142e-01
1.21428734e-02 -9.67139959e-01 -6.53422654e-01 3.87304038e-01
7.07971632e-01 2.14314371e-01 1.63030192e-01 -3.95273894e-01
-3.77666920e-01 2.14157999e-01 6.25191033e-02 -6.40692830e-01
1.45359433e+00 -1.63218126e-01 -6.26747847e-01 1.00902128e+00
9.56365943e-01 -5.04014611e-01 -6.30341768e-01 -2.45987937e-01
1.00496984e+00 4.67676938e-01 -2.69622445e-01 -9.71067607e-01
-3.62037271e-01 6.98692262e-01 -1.69796407e-01 6.48929238e-01
1.31571567e+00 2.46123090e-01 9.60885823e-01 7.52169609e-01
5.45961112e-02 -1.51000285e+00 -1.17340526e-02 1.51961589e+00
1.27935040e+00 -6.61021173e-01 -1.59954444e-01 1.35836184e-01
-4.05346692e-01 1.30533171e+00 7.26493597e-01 -1.18074633e-01
4.41271216e-01 3.58267605e-01 -1.71379030e-01 -5.14168143e-01
-1.56130493e+00 -3.47662151e-01 3.55596602e-01 3.90820950e-01
6.82168901e-01 -1.50221204e-02 -7.77259290e-01 9.84135211e-01
-2.70266742e-01 -3.75565700e-02 4.84123886e-01 1.08057976e+00
-7.03731477e-02 -1.28362453e+00 -1.83715150e-01 1.37517452e-01
-5.16476393e-01 -3.70938778e-01 -6.95852101e-01 8.36759508e-01
3.70439813e-02 1.22363615e+00 2.64218450e-01 -2.91313440e-01
3.71498495e-01 6.99970007e-01 4.01393116e-01 -4.47268844e-01
-1.04885364e+00 -1.04250766e-01 7.08248138e-01 -8.11132431e-01
-6.93954349e-01 -7.02331603e-01 -1.86132812e+00 -2.10987866e-01
5.82403205e-02 -1.33961355e-02 1.87761217e-01 9.62347567e-01
2.93824464e-01 7.88010836e-01 1.52916729e-01 -6.31603003e-01
2.84006387e-01 -1.21113384e+00 5.49927354e-02 1.15851119e-01
3.72406751e-01 -3.10653239e-01 1.65330186e-01 1.16800897e-01] | [9.094268798828125, 9.172510147094727] |
84cca294-aab2-4e5a-a835-e1678a354862 | aesthetic-text-logo-synthesis-via-content | 2204.02701 | null | https://arxiv.org/abs/2204.02701v1 | https://arxiv.org/pdf/2204.02701v1.pdf | Aesthetic Text Logo Synthesis via Content-aware Layout Inferring | Text logo design heavily relies on the creativity and expertise of professional designers, in which arranging element layouts is one of the most important procedures. However, few attention has been paid to this task which needs to take many factors (e.g., fonts, linguistics, topics, etc.) into consideration. In this paper, we propose a content-aware layout generation network which takes glyph images and their corresponding text as input and synthesizes aesthetic layouts for them automatically. Specifically, we develop a dual-discriminator module, including a sequence discriminator and an image discriminator, to evaluate both the character placing trajectories and rendered shapes of synthesized text logos, respectively. Furthermore, we fuse the information of linguistics from texts and visual semantics from glyphs to guide layout prediction, which both play important roles in professional layout design. To train and evaluate our approach, we construct a dataset named as TextLogo3K, consisting of about 3,500 text logo images and their pixel-level annotations. Experimental studies on this dataset demonstrate the effectiveness of our approach for synthesizing visually-pleasing text logos and verify its superiority against the state of the art. | ['Zhouhui Lian', 'Hongwen Kang', 'Pengfei Xiong', 'Yexin Wang', 'Wenhan Luo', 'Guo Pu', 'Yizhi Wang'] | 2022-04-06 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Wang_Aesthetic_Text_Logo_Synthesis_via_Content-Aware_Layout_Inferring_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Wang_Aesthetic_Text_Logo_Synthesis_via_Content-Aware_Layout_Inferring_CVPR_2022_paper.pdf | cvpr-2022-1 | ['layout-design'] | ['computer-vision'] | [ 2.83056289e-01 -4.34952438e-01 2.65130401e-01 -3.11229140e-01
-4.28221896e-02 -6.07278645e-01 3.55677336e-01 -5.26182167e-02
1.52842514e-02 2.36150146e-01 3.62121135e-01 -4.26744014e-01
-2.36385651e-02 -6.69842362e-01 -6.76793039e-01 -4.81518000e-01
7.22506106e-01 2.60020971e-01 1.96192876e-01 -2.10067242e-01
5.70183575e-01 4.02107179e-01 -1.35274971e+00 5.76588988e-01
1.31777489e+00 9.37761128e-01 5.64302564e-01 5.04963636e-01
-6.41675413e-01 6.30452633e-01 -8.04065406e-01 -5.04360020e-01
3.24666537e-02 -5.88730633e-01 -2.53560841e-01 6.62729323e-01
4.42043692e-01 3.26436199e-03 -3.90616059e-01 9.95094836e-01
6.39415801e-01 -1.03244942e-03 9.43680465e-01 -9.58667040e-01
-1.07647228e+00 7.65406191e-01 -7.87643015e-01 -2.65061677e-01
8.13659504e-02 5.67989349e-01 9.76770282e-01 -8.61664236e-01
5.66869557e-01 1.36202812e+00 2.40608588e-01 2.43215069e-01
-1.17245233e+00 -5.83088279e-01 2.37775847e-01 2.09771439e-01
-1.21803188e+00 -3.10094863e-01 1.16996312e+00 -6.71053350e-01
3.19944620e-01 2.66966671e-01 7.93598831e-01 1.01989961e+00
1.69590041e-01 1.15651131e+00 1.07014155e+00 -5.62485397e-01
1.05374314e-01 4.79454100e-01 -3.47026348e-01 6.33358359e-01
3.06890830e-02 -4.24234062e-01 -4.33790594e-01 5.01547337e-01
7.73172677e-01 -9.81868580e-02 -2.59218067e-01 -3.99744123e-01
-1.22576749e+00 3.55158001e-01 3.68454635e-01 -2.03087945e-02
1.91694684e-02 -1.13175295e-01 6.11864269e-01 -9.60694402e-02
1.99073002e-01 4.34365898e-01 -2.36433316e-02 7.52248168e-02
-6.23741031e-01 1.43420175e-01 4.13465649e-01 1.29926443e+00
4.01365787e-01 2.48414893e-02 -3.67064536e-01 1.13742507e+00
4.00622219e-01 5.68857133e-01 5.09609103e-01 -3.75022918e-01
9.99664426e-01 9.45255339e-01 -4.74175252e-02 -1.31351733e+00
-2.12129131e-01 -1.56513289e-01 -9.29202497e-01 1.42154038e-01
1.47370800e-01 -1.18072882e-01 -8.24374616e-01 1.10649586e+00
4.89349142e-02 -4.01059657e-01 -1.86076254e-01 8.81723225e-01
7.94114351e-01 8.57866347e-01 -1.16026662e-01 -3.40144266e-03
1.54728556e+00 -1.32942677e+00 -9.42829311e-01 -4.10611123e-01
4.02940810e-01 -1.23368692e+00 1.81496143e+00 5.44218063e-01
-9.02998030e-01 -7.73062944e-01 -1.14670467e+00 -1.94926620e-01
-2.17559755e-01 9.24302638e-01 3.89757127e-01 5.02132356e-01
-4.71650988e-01 2.87816584e-01 -3.20443064e-01 -2.37929881e-01
5.10924995e-01 -1.37427926e-01 1.36419430e-01 8.52802768e-02
-8.91296864e-01 2.72827029e-01 5.92513919e-01 4.49783564e-01
-4.71222818e-01 -3.55690986e-01 -5.87851167e-01 1.87703371e-01
5.53753495e-01 -4.87176329e-01 9.34400022e-01 -8.71390581e-01
-1.48338032e+00 5.72122872e-01 2.19535857e-01 1.58033222e-01
7.44896710e-01 -2.20743157e-02 -5.60970604e-01 -3.87410596e-02
1.29659763e-02 7.15917587e-01 8.47310960e-01 -1.42975259e+00
-7.96987832e-01 -1.16670758e-01 -1.31957635e-01 3.69921535e-01
-6.95064068e-01 -2.59269059e-01 -8.58644903e-01 -1.00937653e+00
-7.82596916e-02 -6.70627058e-01 -5.61795458e-02 1.54788136e-01
-9.25109863e-01 -1.04435362e-01 6.10740185e-01 -8.26552391e-01
1.73825407e+00 -2.25580764e+00 4.25771736e-02 3.67861301e-01
5.76479435e-02 1.64332226e-01 -1.56688076e-02 5.77614307e-01
1.21757783e-01 1.65199831e-01 -8.05370286e-02 -2.93015420e-01
3.24858814e-01 -1.65595055e-01 -2.78964400e-01 -1.10176891e-01
8.77922401e-02 1.10306191e+00 -6.03307962e-01 -9.21460032e-01
2.79674441e-01 2.10187271e-01 -4.68617558e-01 1.62738696e-01
-6.51069462e-01 4.25632209e-01 -6.20626211e-01 5.57156146e-01
5.92225134e-01 -3.40221167e-01 3.02149415e-01 -4.99784797e-01
-2.47841939e-01 -2.85230651e-02 -1.07131767e+00 1.72400737e+00
-5.87350667e-01 8.00552964e-01 -4.77625161e-01 -5.85031450e-01
1.29028094e+00 -7.55841136e-02 -1.15385633e-02 -1.00856829e+00
2.79358089e-01 3.23899150e-01 -9.02442634e-02 -9.10304904e-01
5.67651749e-01 3.30163985e-01 -3.79766412e-02 5.16431391e-01
-7.19825447e-01 -3.88116552e-03 4.67440188e-01 1.32315993e-01
6.77476406e-01 3.81556749e-01 -4.55969200e-02 -2.16779351e-01
7.00762331e-01 -2.42017936e-02 3.36918980e-01 2.85789788e-01
3.24682266e-01 8.52317572e-01 8.89021099e-01 -4.16781694e-01
-1.48400223e+00 -6.93557143e-01 9.90628898e-02 8.56003225e-01
4.05417323e-01 -4.62898999e-01 -1.02989459e+00 -6.10241294e-01
-2.23885641e-01 5.51655233e-01 -5.96793532e-01 -1.38728186e-01
-8.19862247e-01 -4.78652209e-01 2.95993567e-01 6.96716785e-01
5.97350717e-01 -1.47775280e+00 -5.57403803e-01 9.48121697e-02
-1.59113079e-01 -9.39968228e-01 -1.12358010e+00 -1.40263349e-01
-5.97328663e-01 -8.49886894e-01 -9.16621268e-01 -1.16371512e+00
1.04976940e+00 2.69345284e-01 9.29057777e-01 8.52450952e-02
-4.29746062e-01 -1.79579079e-01 -4.20599550e-01 -2.45571688e-01
-3.37472051e-01 2.70653933e-01 -4.91907775e-01 2.60528445e-01
-1.86035354e-02 -5.04738688e-01 -8.65745723e-01 5.51031649e-01
-1.19174898e+00 8.08436453e-01 1.02115464e+00 6.59921110e-01
5.56305289e-01 2.76957721e-01 1.27200156e-01 -1.20844865e+00
9.31199074e-01 1.66386422e-02 -5.75672984e-01 7.03914344e-01
-4.13554370e-01 1.78427875e-01 1.06019926e+00 -5.83333254e-01
-1.21087193e+00 1.53526470e-01 7.19033405e-02 -3.79918098e-01
-1.21413141e-01 4.09253716e-01 -8.93606603e-01 3.47154558e-01
5.03029644e-01 4.22960490e-01 -2.20463589e-01 -7.26017118e-01
5.41396379e-01 8.87828410e-01 4.66350228e-01 -6.46643937e-01
8.73927057e-01 1.23202421e-01 -2.44325921e-01 -7.42258906e-01
-6.79194987e-01 5.22369556e-02 -6.44224644e-01 -2.61010796e-01
8.21071982e-01 -4.17391956e-01 -7.43319571e-01 5.47119260e-01
-1.20958006e+00 -3.95658880e-01 -2.04071581e-01 -4.95503284e-02
-3.93945724e-01 4.00068939e-01 -2.17307240e-01 -5.92226207e-01
-1.82180792e-01 -1.35269690e+00 1.07345712e+00 4.80426371e-01
8.10413212e-02 -8.83817852e-01 -3.49288881e-01 2.56633937e-01
-9.08268243e-03 2.32803047e-01 1.39829290e+00 -8.86795968e-02
-8.22305381e-01 1.27634883e-01 -5.97265959e-01 3.17923129e-01
2.83064336e-01 3.27858627e-01 -8.23227763e-01 1.65886506e-01
-5.74000120e-01 -1.25250205e-01 6.11856401e-01 2.36804277e-01
1.59005475e+00 -5.32438099e-01 -3.12676966e-01 7.14707017e-01
1.25713861e+00 5.32796085e-01 8.04017842e-01 4.52706873e-01
1.31197226e+00 9.54838514e-01 7.41625905e-01 6.58949494e-01
3.69081616e-01 6.65395498e-01 1.34779602e-01 -1.30835682e-01
-3.20798367e-01 -7.48869836e-01 1.28719702e-01 1.05994403e+00
2.94984668e-01 -7.80159175e-01 -8.75690877e-01 2.70291924e-01
-1.81951118e+00 -6.27267182e-01 -3.05863619e-01 1.84903133e+00
8.82758856e-01 3.40671331e-01 7.45897517e-02 1.94384664e-01
8.95542502e-01 2.43782759e-01 -6.19317651e-01 -3.19711715e-01
-2.35839143e-01 -2.95626730e-01 2.87025273e-01 5.55505119e-02
-8.21378946e-01 1.00611401e+00 4.66684723e+00 1.39816654e+00
-1.08779752e+00 -4.05396521e-01 9.30819929e-01 1.29060715e-01
-8.24654281e-01 -1.38162985e-01 -7.42646754e-01 9.11427200e-01
6.94948584e-02 -2.33298019e-02 6.43878102e-01 6.72874212e-01
4.86076057e-01 1.46195097e-02 -8.46332371e-01 1.18760407e+00
2.01799855e-01 -1.27223885e+00 2.76043206e-01 5.64627945e-02
9.22371626e-01 -1.09736741e+00 3.71230423e-01 3.41037870e-04
-8.64898562e-02 -1.05801475e+00 1.25195432e+00 6.11771405e-01
9.73432243e-01 -7.70561934e-01 2.61946678e-01 1.84831738e-01
-1.46048474e+00 -9.88606270e-03 -4.30397898e-01 4.22292084e-01
1.62678435e-01 5.14194191e-01 -7.68579841e-01 4.07161087e-01
5.61265886e-01 8.86311531e-01 -9.63616073e-01 1.24449015e+00
-5.72245955e-01 3.87091339e-01 5.57698756e-02 -5.06447196e-01
1.84742346e-01 -3.59163493e-01 7.03500807e-02 1.17158043e+00
3.54054153e-01 -1.85481280e-01 9.93308201e-02 1.11054671e+00
-1.65055305e-01 5.62511027e-01 -2.46084735e-01 -3.73779774e-01
5.90719163e-01 1.40965247e+00 -1.02542138e+00 -2.77598590e-01
-1.93785578e-01 1.10277843e+00 7.34976903e-02 5.24576187e-01
-8.97693515e-01 -8.83144796e-01 2.20873013e-01 3.73499840e-01
3.15863580e-01 -6.96133971e-02 -8.31581771e-01 -8.85863304e-01
4.13944304e-01 -1.05000710e+00 -6.15338385e-02 -1.13124585e+00
-1.23358083e+00 5.36322176e-01 -3.25841635e-01 -1.65629816e+00
4.04370070e-01 -8.23506236e-01 -8.33609819e-01 8.49007130e-01
-1.13034058e+00 -1.42294097e+00 -6.79774404e-01 6.31280169e-02
8.66360605e-01 -1.71261325e-01 -3.96520197e-02 5.42312324e-01
-8.17768753e-01 6.06286824e-01 2.01163575e-01 2.05295891e-01
7.69125164e-01 -1.25259972e+00 8.08278799e-01 6.83901310e-01
9.97320488e-02 5.65085113e-01 5.71026027e-01 -7.62660861e-01
-1.21327198e+00 -1.03072238e+00 6.38734281e-01 -1.07657291e-01
5.80262244e-01 -7.86972761e-01 -7.73104310e-01 3.43240559e-01
2.02023014e-01 -4.86544669e-01 2.57984430e-01 -2.20087647e-01
-8.81380588e-02 -2.31371462e-01 -3.80517721e-01 1.00771058e+00
1.23484898e+00 -2.27493167e-01 -1.64405346e-01 4.49207574e-01
6.63332462e-01 -6.20898724e-01 -3.85531783e-01 7.57893100e-02
7.36677706e-01 -1.07107508e+00 7.98567116e-01 -1.86260328e-01
9.33982253e-01 -5.08139372e-01 4.39529926e-01 -1.21849418e+00
-1.35331079e-01 -4.99029011e-01 3.16054910e-01 1.66108930e+00
3.75814229e-01 -1.78826839e-01 8.62216592e-01 1.85624689e-01
-3.59819591e-01 -8.43494058e-01 -8.01500678e-02 -4.62112397e-01
-3.57208133e-01 -2.50773579e-01 9.12892938e-01 5.85707009e-01
-2.66143054e-01 5.25966227e-01 -6.21583223e-01 -1.49462670e-01
2.28048474e-01 3.85265470e-01 9.32265162e-01 -9.38080549e-01
-2.05161005e-01 -7.74328351e-01 -2.94599328e-02 -1.47827327e+00
-1.28522336e-01 -7.18765557e-01 2.06752077e-01 -1.85581672e+00
3.97112295e-02 -6.07628047e-01 2.18601525e-01 3.27155203e-01
-3.35610867e-01 1.52207509e-01 1.68007597e-01 1.85036808e-01
-6.13740921e-01 7.78528094e-01 1.91196954e+00 -3.46700847e-01
-3.55658144e-01 -2.00863063e-01 -8.97214115e-01 6.36975825e-01
5.97849131e-01 -2.26697735e-02 -7.48485029e-01 -7.27741420e-01
5.54717898e-01 -9.58273336e-02 1.55331746e-01 -9.54807401e-01
1.72373608e-01 -2.89158344e-01 6.74966991e-01 -9.22689974e-01
-6.30095601e-02 -7.51141071e-01 1.96700711e-02 2.28882328e-01
-5.85729599e-01 3.58601540e-01 3.85115370e-02 4.68074560e-01
1.12817092e-02 -3.56169522e-01 5.48658073e-01 -1.00533277e-01
-7.50553668e-01 2.55591333e-01 -7.80593976e-02 1.68673530e-01
8.56710374e-01 -4.41719592e-01 -3.92918020e-01 -9.19872746e-02
-3.28655183e-01 3.54693264e-01 6.72441483e-01 5.41936159e-01
7.69699574e-01 -1.55538476e+00 -4.28046227e-01 3.02347749e-01
3.46937120e-01 2.22384602e-01 5.50097287e-01 5.88782489e-01
-1.09399629e+00 3.59747291e-01 -2.35275820e-01 -3.56663555e-01
-1.07195902e+00 6.37202799e-01 -3.59790772e-02 -1.37812242e-01
-7.51639783e-01 6.94244385e-01 6.74892962e-01 -1.45831764e-01
3.01079452e-01 -4.02353317e-01 -4.21279252e-01 1.04832388e-01
4.61053282e-01 3.15059543e-01 -2.06014499e-01 -2.22520858e-01
7.83909336e-02 8.27752054e-01 -1.45955354e-01 -7.38840625e-02
9.19477046e-01 -3.43058974e-01 -2.20910534e-02 4.08702075e-01
9.45790172e-01 2.35219747e-01 -1.46448195e+00 2.90318741e-03
7.77753070e-02 -6.85089052e-01 -4.48855639e-01 -8.53027582e-01
-1.08260977e+00 1.41145015e+00 3.05889428e-01 1.30533949e-01
1.15243578e+00 -3.48648936e-01 1.06045830e+00 3.92054319e-02
-4.70867753e-02 -1.23910177e+00 6.41479135e-01 1.07395656e-01
9.58592951e-01 -9.96365011e-01 -1.03531688e-01 -6.10850751e-01
-9.85629201e-01 1.17557216e+00 9.14443135e-01 5.37065081e-02
-1.83385436e-03 5.03878556e-02 1.97346792e-01 1.54693890e-03
-3.93712521e-01 4.72521074e-02 7.78071225e-01 3.90296727e-01
4.32918221e-01 -6.69400990e-02 -2.56939888e-01 5.91487527e-01
-3.77769589e-01 -2.62312561e-01 3.01482290e-01 6.51884496e-01
-4.40498710e-01 -1.25191414e+00 -3.34258974e-01 3.75127256e-01
6.11456260e-02 -3.35633874e-01 -4.25677031e-01 4.54003870e-01
3.41961622e-01 5.82035065e-01 -5.58673113e-04 -6.29085660e-01
5.77688396e-01 -3.40064138e-01 1.71295866e-01 -5.34069896e-01
-4.76532876e-01 4.19979960e-01 -1.72272995e-01 -8.67639929e-02
9.01368633e-02 -4.22037721e-01 -1.14178991e+00 -2.85725027e-01
-1.16318621e-01 3.22138891e-02 6.44569099e-01 6.33446455e-01
1.96196809e-01 1.14935279e+00 7.47335851e-01 -6.24311149e-01
-2.70651102e-01 -7.55247355e-01 -7.14608490e-01 4.54351604e-01
-2.02149764e-01 -2.94411421e-01 5.43647967e-02 3.73536676e-01] | [11.545196533203125, -0.30243536829948425] |
b418673b-b967-46f1-80b2-de76580d48ad | object-discovery-from-motion-guided-tokens | 2303.15555 | null | https://arxiv.org/abs/2303.15555v1 | https://arxiv.org/pdf/2303.15555v1.pdf | Object Discovery from Motion-Guided Tokens | Object discovery -- separating objects from the background without manual labels -- is a fundamental open challenge in computer vision. Previous methods struggle to go beyond clustering of low-level cues, whether handcrafted (e.g., color, texture) or learned (e.g., from auto-encoders). In this work, we augment the auto-encoder representation learning framework with two key components: motion-guidance and mid-level feature tokenization. Although both have been separately investigated, we introduce a new transformer decoder showing that their benefits can compound thanks to motion-guided vector quantization. We show that our architecture effectively leverages the synergy between motion and tokenization, improving upon the state of the art on both synthetic and real datasets. Our approach enables the emergence of interpretable object-specific mid-level features, demonstrating the benefits of motion-guidance (no labeling) and quantization (interpretability, memory efficiency). | ['Martial Hebert', 'Adrien Gaidon', 'Yu-Xiong Wang', 'Pavel Tokmakov', 'Zhipeng Bao'] | 2023-03-27 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Bao_Object_Discovery_From_Motion-Guided_Tokens_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Bao_Object_Discovery_From_Motion-Guided_Tokens_CVPR_2023_paper.pdf | cvpr-2023-1 | ['object-discovery'] | ['computer-vision'] | [ 4.34173971e-01 1.33531585e-01 -2.34613940e-01 -3.10364723e-01
-7.57779598e-01 -8.17415595e-01 8.80072176e-01 1.86190233e-01
-4.08269614e-01 4.51216966e-01 2.46804893e-01 -2.52385825e-01
-1.65088788e-01 -5.83186507e-01 -8.67603123e-01 -6.52273536e-01
-2.07761183e-01 2.42824674e-01 4.14433360e-01 1.04258478e-01
2.62958795e-01 4.78952259e-01 -1.89473093e+00 4.06466424e-01
5.08531988e-01 1.10095084e+00 2.34911069e-01 7.59280264e-01
-1.47326946e-01 8.52400780e-01 -1.75745994e-01 -1.27090693e-01
3.09727401e-01 -2.73402840e-01 -9.08074617e-01 4.38964248e-01
7.14429140e-01 -5.20651996e-01 -1.47545576e-01 9.46646214e-01
1.35143459e-01 -5.47240339e-02 5.32832563e-01 -1.28153372e+00
-7.94805825e-01 2.87117273e-01 -4.41884547e-01 7.38363015e-03
-1.31481692e-01 3.84204715e-01 1.32240081e+00 -8.13229501e-01
7.72072911e-01 1.33401096e+00 5.78406096e-01 4.32551980e-01
-1.44125235e+00 -1.36417851e-01 1.94982514e-01 3.94295692e-01
-1.20915711e+00 -5.94462812e-01 5.78259170e-01 -8.76887083e-01
7.98378229e-01 7.41151944e-02 5.19198418e-01 9.99717414e-01
-1.52526811e-01 9.24461842e-01 1.06573057e+00 -4.93780345e-01
4.18814361e-01 7.77983442e-02 5.44202588e-02 9.33637381e-01
4.21887934e-01 2.18620822e-01 -6.57458425e-01 3.24567139e-01
9.26911056e-01 -1.03160515e-01 -1.46508366e-01 -7.96590388e-01
-1.51355827e+00 6.10056520e-01 3.29780251e-01 1.10225335e-01
-1.93327457e-01 7.14064181e-01 3.28602850e-01 -5.29160611e-02
1.42559782e-01 4.09807831e-01 -6.15748227e-01 -3.38848323e-01
-1.08080077e+00 -7.05153272e-02 4.90155995e-01 1.09381139e+00
1.06316876e+00 9.45230573e-02 -4.26149994e-01 3.51214767e-01
4.88439858e-01 5.72132051e-01 2.42892459e-01 -1.17544007e+00
1.55188128e-01 5.76925874e-01 2.35979319e-01 -8.37045908e-01
-3.40877503e-01 -3.02632958e-01 -6.17037714e-01 4.57883805e-01
6.22313499e-01 6.47401363e-02 -1.10412359e+00 1.94351292e+00
2.68172860e-01 2.69180745e-01 -1.07250072e-01 9.37556207e-01
5.43798387e-01 2.99948454e-01 1.53731927e-02 1.37121737e-01
1.56370878e+00 -1.07520390e+00 -4.77032542e-01 -1.46142095e-01
6.86578274e-01 -3.21465552e-01 1.13246012e+00 4.92316067e-01
-9.82251763e-01 -5.48921764e-01 -1.02338541e+00 -3.53897810e-01
-4.20521319e-01 3.22213620e-01 7.75306463e-01 7.17501819e-01
-1.22627711e+00 6.81271136e-01 -1.15739954e+00 -4.54007685e-01
6.58843338e-01 5.09229422e-01 -6.15310311e-01 -7.95679986e-02
-6.20954096e-01 6.35133684e-01 3.05502862e-01 -1.69002727e-01
-1.00138402e+00 -5.55542946e-01 -8.88884604e-01 6.00485764e-02
5.29229999e-01 -9.73861516e-01 1.20158434e+00 -9.85429287e-01
-1.63620055e+00 8.10360134e-01 -2.04332486e-01 -5.58462441e-01
5.04521966e-01 -3.99923712e-01 1.78288013e-01 3.95502836e-01
7.97259435e-02 1.19592214e+00 1.13974655e+00 -1.25380421e+00
-9.00343180e-01 -2.22648948e-01 2.93374389e-01 -1.68921262e-01
-5.72975874e-01 -2.94508010e-01 -6.40084565e-01 -5.70437074e-01
-1.54886231e-01 -9.68013585e-01 -2.58666962e-01 4.43063587e-01
-3.93994868e-01 8.80736858e-03 8.23874652e-01 -5.16634881e-01
8.76939654e-01 -2.20626521e+00 2.13628381e-01 -5.88024557e-02
3.24121416e-01 1.94656089e-01 -1.92744121e-01 8.65323395e-02
6.33540377e-02 2.60127693e-01 -2.67725885e-01 -5.07587075e-01
2.67130852e-01 3.94297659e-01 -3.80483419e-02 5.05271375e-01
7.29878366e-01 1.19066012e+00 -1.06808114e+00 -4.80385751e-01
5.28234601e-01 6.14600480e-01 -7.98648775e-01 -7.63592944e-02
-4.05633032e-01 5.04407406e-01 -3.66526067e-01 6.81635082e-01
4.58437145e-01 -5.22852659e-01 1.00662053e-01 -4.41507518e-01
-1.15857594e-01 3.40156198e-01 -1.28223658e+00 1.96560252e+00
-4.01216120e-01 8.06536376e-01 7.21460506e-02 -9.01848257e-01
4.79193389e-01 8.37506726e-02 5.78759074e-01 -5.40316522e-01
6.91705495e-02 4.52203751e-02 -4.47224686e-03 -5.04347920e-01
5.63223124e-01 2.26816356e-01 4.33503389e-02 3.82712036e-01
3.07712913e-01 8.30530822e-02 2.62336195e-01 1.60800382e-01
1.15881336e+00 5.67512810e-01 3.40909630e-01 -3.27930808e-01
2.81744421e-01 1.95256010e-01 4.41821009e-01 8.05963457e-01
-1.41120613e-01 7.69615352e-01 4.78437334e-01 -1.90669999e-01
-1.16953993e+00 -1.00419438e+00 -9.27493814e-03 1.23425674e+00
1.52311549e-01 -6.78265035e-01 -7.85378039e-01 -8.21159899e-01
1.61831558e-01 3.66423547e-01 -7.90308714e-01 -1.26769200e-01
-4.49703068e-01 -5.10414302e-01 3.63379866e-01 7.61278629e-01
2.14521348e-01 -7.08446622e-01 -1.33514905e+00 2.19573528e-01
1.85327947e-01 -1.44535422e+00 -1.09511808e-01 5.03673613e-01
-8.22335839e-01 -8.13845932e-01 -3.95264804e-01 -3.29773515e-01
5.40304303e-01 4.46522593e-01 1.21288359e+00 -2.21745484e-02
-5.49353540e-01 6.98335886e-01 -4.09727812e-01 -8.59702229e-02
-4.16346155e-02 1.01574063e-01 -9.42790210e-02 2.57474303e-01
6.14966154e-02 -6.53458834e-01 -7.16640115e-01 6.79672360e-02
-1.09193623e+00 2.21925274e-01 9.40064549e-01 8.18769574e-01
5.63998699e-01 -2.30102301e-01 2.05788255e-01 -7.71382868e-01
5.55878179e-03 -2.71504730e-01 -7.27009594e-01 1.77041128e-01
-4.54853028e-01 5.19357204e-01 3.32769841e-01 -2.97496349e-01
-6.24192834e-01 4.98414010e-01 1.67808905e-01 -4.96461898e-01
-4.18917030e-01 1.95662394e-01 -2.94339180e-01 -1.43273368e-01
3.65451247e-01 1.26526073e-01 -9.38274786e-02 -5.28709471e-01
1.12970805e+00 4.00613934e-01 8.38575602e-01 -7.40137815e-01
8.47419500e-01 8.83197010e-01 2.37820432e-01 -7.40606070e-01
-6.94879174e-01 -4.93175656e-01 -1.06295824e+00 -1.64555723e-03
1.22035539e+00 -8.69032025e-01 -8.41014504e-01 3.86167131e-02
-1.18246126e+00 -3.34177226e-01 -6.61902070e-01 3.68062377e-01
-7.41169989e-01 3.61497372e-01 -4.39111322e-01 -8.55570734e-01
1.07667439e-01 -1.23061872e+00 1.52931070e+00 3.55267040e-02
-1.82373807e-01 -7.81278133e-01 -2.96941161e-01 1.08000286e-01
3.73020440e-01 4.52967435e-01 8.72793734e-01 -3.12229872e-01
-1.24091542e+00 2.20580697e-01 -5.96515119e-01 2.20773429e-01
1.54637486e-01 4.88493107e-02 -1.28940225e+00 -8.44912603e-02
-3.68497252e-01 -3.75756025e-01 1.16266787e+00 2.85509437e-01
1.09045553e+00 -3.58797580e-01 -2.99100935e-01 7.94773221e-01
1.42751658e+00 -1.41732812e-01 3.95748496e-01 5.25452435e-01
9.75385904e-01 5.76131880e-01 4.11825955e-01 6.57879949e-01
4.94310558e-01 7.97396541e-01 7.09425092e-01 -9.32040140e-02
-4.81255949e-01 -5.83568811e-02 3.10371667e-01 5.13278008e-01
-1.34576678e-01 -5.91100529e-02 -9.69853699e-01 7.66601264e-01
-2.02446556e+00 -7.26503730e-01 -7.89021477e-02 1.97602057e+00
6.78136110e-01 1.64198413e-01 1.25037849e-01 2.60059685e-01
2.65466094e-01 1.28268125e-02 -5.25228798e-01 -5.80773503e-02
-1.41037583e-01 7.57776424e-02 6.28252447e-01 6.14001215e-01
-1.42066586e+00 1.07922840e+00 5.97694016e+00 4.64862704e-01
-1.10742700e+00 1.90072358e-01 3.71778429e-01 -5.82646206e-02
-3.14833760e-01 9.71476287e-02 -8.55653763e-01 2.97889084e-01
7.45753586e-01 2.59405404e-01 3.54754388e-01 8.60623837e-01
1.64922420e-02 8.27338081e-03 -1.63065553e+00 1.02848554e+00
3.10325641e-02 -1.52829492e+00 1.16367131e-01 2.21195459e-01
6.07152700e-01 -2.87436079e-02 2.59029508e-01 1.73192695e-01
3.56295943e-01 -1.08332527e+00 1.16421878e+00 4.31147277e-01
8.67237985e-01 -3.64052624e-01 4.65841115e-01 3.59563008e-02
-1.30138040e+00 -1.15313777e-03 -7.52185658e-02 8.58114138e-02
1.25002027e-01 6.15695834e-01 -9.25604105e-01 7.93155789e-01
5.49317479e-01 9.50761974e-01 -9.13916230e-01 7.47797132e-01
-1.77957088e-01 6.00179911e-01 -3.52550626e-01 2.96924412e-01
4.41478133e-01 -5.42020090e-02 3.66433233e-01 1.58864212e+00
8.69963989e-02 -2.10957617e-01 2.05395430e-01 9.31800604e-01
1.74521491e-01 -3.18806380e-01 -4.67772394e-01 -1.71927646e-01
2.25542888e-01 1.33848715e+00 -8.95625710e-01 -4.41045702e-01
-4.96996760e-01 1.09813821e+00 2.59710819e-01 2.62256742e-01
-7.80102670e-01 -1.54056251e-01 7.88572431e-01 3.76566648e-02
9.34097946e-01 -5.65489948e-01 -4.33316112e-01 -1.06153429e+00
-4.23178524e-02 -7.65718818e-01 9.34885629e-03 -5.79100907e-01
-1.07340038e+00 3.53572398e-01 -1.96229190e-01 -1.23571324e+00
-4.24608171e-01 -1.02191412e+00 -1.01051189e-01 3.90344948e-01
-1.70856595e+00 -1.45597553e+00 -4.06400502e-01 4.47259516e-01
3.44702661e-01 1.96226820e-01 6.79238737e-01 2.76795924e-01
-3.82260680e-01 4.79146302e-01 8.99429247e-02 1.46006688e-01
4.95424122e-01 -1.49023914e+00 3.43941867e-01 8.29198956e-01
5.47786951e-01 5.70973635e-01 4.53519791e-01 -1.78393036e-01
-1.79147983e+00 -1.28977954e+00 3.79659712e-01 -7.83271015e-01
6.67358875e-01 -7.20200896e-01 -6.74813986e-01 6.05569839e-01
2.50039808e-02 4.64971602e-01 4.45374131e-01 1.44010698e-02
-6.21690929e-01 1.41784397e-03 -7.30631113e-01 6.28490865e-01
1.18580759e+00 -6.89237416e-01 -4.06429291e-01 1.21403441e-01
8.50314200e-01 -5.81088923e-02 -7.21318483e-01 3.37495655e-01
6.31451249e-01 -1.00929487e+00 1.21883750e+00 -7.71283984e-01
4.68650907e-01 -6.95872903e-01 -4.83812600e-01 -8.29942167e-01
-4.03336734e-01 -6.92898154e-01 -4.55709130e-01 1.41289282e+00
1.55189142e-01 -1.07598796e-01 7.67489612e-01 3.08422148e-01
-1.71982616e-01 -7.25878894e-01 -7.34240234e-01 -9.10511911e-01
-2.77046084e-01 -6.42965198e-01 3.19935501e-01 9.29633796e-01
-1.87414467e-01 5.63695967e-01 -4.31080550e-01 2.58071631e-01
5.60589135e-01 2.70618439e-01 9.05214548e-01 -1.19212723e+00
-5.02890229e-01 -5.79857051e-01 -7.98703492e-01 -1.34614921e+00
2.89740767e-02 -9.71140981e-01 2.19664469e-01 -1.57060754e+00
1.58047378e-01 -2.64428169e-01 -2.82659471e-01 7.65348077e-01
-1.27521470e-01 3.64154011e-01 3.90998453e-01 7.07336934e-03
-9.99394953e-01 5.59175909e-01 1.06754005e+00 -1.59673527e-01
1.10961008e-03 -6.11776173e-01 -6.34342968e-01 7.74566948e-01
5.33917904e-01 -4.81460661e-01 -2.95795232e-01 -7.00519443e-01
5.37465811e-02 -1.53807014e-01 7.34570801e-01 -1.17601418e+00
1.93309143e-01 -2.29905788e-02 2.83904791e-01 -4.68142778e-01
4.05051917e-01 -8.37793052e-01 -2.95063734e-01 3.94233465e-01
-2.82572687e-01 -7.17939213e-02 2.87590265e-01 8.38742733e-01
-1.14072137e-01 1.20382942e-01 6.07409000e-01 2.18397900e-02
-1.11046171e+00 2.05834523e-01 -3.28206539e-01 3.88846360e-02
9.50593650e-01 -3.02639961e-01 -2.74682999e-01 -1.62530079e-01
-6.44631565e-01 -9.57503077e-03 4.79122043e-01 3.73411119e-01
4.67140853e-01 -1.20130336e+00 -5.25454164e-01 2.16973677e-01
2.62317121e-01 8.28712136e-02 -1.76833719e-01 9.70212281e-01
-3.54088157e-01 5.98501086e-01 -3.09799522e-01 -1.20712674e+00
-9.27788019e-01 6.76789403e-01 3.68536301e-02 -1.27792805e-01
-6.87724948e-01 7.84654856e-01 4.62667137e-01 -7.95959309e-03
2.39464834e-01 -8.76563609e-01 2.47236207e-01 2.86222361e-02
4.15902138e-01 2.03159750e-01 2.39373203e-02 -4.07750189e-01
-5.28094649e-01 8.15224648e-01 7.54903406e-02 -3.20334435e-01
1.32416511e+00 -3.27649921e-01 4.10412848e-02 5.63979208e-01
1.08074188e+00 -3.19483839e-02 -1.75059891e+00 -1.58117011e-01
5.27021408e-01 -4.76264030e-01 7.72554725e-02 -6.20302916e-01
-8.75575960e-01 1.24102628e+00 6.12108827e-01 1.01817884e-01
9.62254763e-01 3.42510715e-02 6.14608407e-01 6.08410954e-01
4.26446855e-01 -7.99688816e-01 4.34330791e-01 5.07182062e-01
4.12142128e-01 -1.32428849e+00 -2.35339999e-01 -3.98917258e-01
-4.97783303e-01 1.09928119e+00 3.27518225e-01 -7.60866255e-02
4.67280179e-01 4.67232406e-01 -9.27489847e-02 -3.27540338e-02
-9.05322254e-01 -8.44898403e-01 4.42129731e-01 7.16696084e-01
4.59965020e-01 -1.36484250e-01 2.83765942e-01 4.54287320e-01
7.49811623e-03 -4.46277633e-02 3.58300090e-01 9.84093428e-01
-5.02728045e-01 -1.06570613e+00 -1.96828827e-01 2.22131804e-01
-3.21458489e-01 -2.20103219e-01 -3.04144412e-01 7.88159311e-01
5.65416217e-01 8.45841110e-01 8.84767398e-02 -4.31927413e-01
-7.18495017e-03 3.29181477e-02 6.18068874e-01 -7.78455973e-01
-2.96840996e-01 3.81132066e-02 -2.18775775e-02 -9.13110197e-01
-6.42260969e-01 -5.94242752e-01 -1.01791704e+00 2.19517663e-01
-2.13063926e-01 -2.54086494e-01 7.60997474e-01 9.72960889e-01
5.61205566e-01 7.49009967e-01 1.40686527e-01 -1.00966167e+00
-4.47609186e-01 -5.93662202e-01 -3.69026899e-01 5.17452598e-01
7.82847762e-01 -7.60141969e-01 -1.96484298e-01 6.94008946e-01] | [9.489168167114258, 0.5140320658683777] |
3543ce1e-07ee-4015-b911-4eac8baf4d2e | face-identification-by-means-of-a-neural-net | 2204.00305 | null | https://arxiv.org/abs/2204.00305v1 | https://arxiv.org/pdf/2204.00305v1.pdf | Face identification by means of a neural net classifier | This paper describes a novel face identification method that combines the eigenfaces theory with the Neural Nets. We use the eigenfaces methodology in order to reduce the dimensionality of the input image, and a neural net classifier that performs the identification process. The method presented recognizes faces in the presence of variations in facial expression, facial details and lighting conditions. A recognition rate of more than 87% has been achieved, while the classical method of Turk and Pentland achieves a 75.5%. | ['Marcos Faundez-Zanuy', 'Virginia Espinosa-Duro'] | 2022-04-01 | null | null | null | null | ['face-identification'] | ['computer-vision'] | [-3.20095211e-01 -2.74602883e-02 1.21257193e-01 -5.44783533e-01
5.61296701e-01 -2.94835746e-01 5.60496986e-01 -7.79313445e-01
-4.50706631e-01 4.08027917e-01 -1.75645411e-01 -1.50085837e-01
3.12060546e-02 -5.45031250e-01 -1.13252020e-02 -6.48424625e-01
-2.97289509e-02 1.43418118e-01 -4.25850868e-01 -3.09941351e-01
2.46222302e-01 1.20235920e+00 -2.12651467e+00 2.52916008e-01
7.29157552e-02 1.17713535e+00 -4.98888969e-01 4.43284035e-01
-1.79224640e-01 3.18807602e-01 -4.28066909e-01 -3.19367826e-01
3.41298640e-01 -2.27971859e-02 -5.78376770e-01 2.10542783e-01
3.46876174e-01 -2.75594234e-01 -8.76488537e-02 1.05937839e+00
4.79429185e-01 5.53191826e-02 9.89992619e-01 -9.95889723e-01
-6.95457518e-01 2.19292790e-01 -5.22967219e-01 -3.44034955e-02
6.17493689e-01 -2.11851299e-01 1.61226824e-01 -1.03309548e+00
4.73800808e-01 1.54542530e+00 6.99918628e-01 8.70669484e-01
-1.10635829e+00 -7.88938880e-01 -5.85610867e-01 4.01002496e-01
-1.75167108e+00 -6.29158795e-01 8.10122788e-01 -5.87472975e-01
8.39248121e-01 1.51769862e-01 5.10089219e-01 6.84760630e-01
-9.87222791e-02 1.55908465e-01 1.36449397e+00 -9.88420010e-01
6.96260408e-02 4.23672974e-01 2.32441306e-01 8.46853018e-01
1.24610588e-01 2.79008925e-01 -1.92778453e-01 -1.07308261e-01
7.42325127e-01 -7.58583322e-02 2.90289193e-01 1.00033857e-01
-4.11684304e-01 7.35650361e-01 4.32728976e-02 8.30352485e-01
-6.40146494e-01 -3.43995243e-01 5.46649992e-01 3.32989812e-01
1.45459756e-01 1.78619936e-01 -2.20286831e-01 8.51895437e-02
-8.22397470e-01 -3.35345715e-01 9.95554805e-01 3.92936617e-01
6.78706706e-01 4.05185908e-01 3.68846387e-01 7.98655808e-01
4.94391203e-01 5.65290570e-01 7.72479355e-01 -7.78082550e-01
-2.94842273e-01 8.24757755e-01 -1.85536277e-02 -1.38562047e+00
-2.96934307e-01 3.36632013e-01 -8.24216068e-01 7.71846473e-01
1.19118407e-01 -2.59297788e-01 -9.47715223e-01 1.22252440e+00
4.39291149e-01 -1.46586552e-01 1.89516142e-01 7.33879745e-01
1.01011109e+00 3.56313020e-01 -6.43202215e-02 -3.69464397e-01
1.30197322e+00 -2.85951018e-01 -1.05482078e+00 2.94318497e-01
-1.52375236e-01 -9.11256433e-01 5.72544158e-01 5.38880408e-01
-8.09054911e-01 -8.83544028e-01 -1.04488039e+00 5.25992751e-01
-7.25359440e-01 6.85536385e-01 4.32279915e-01 1.37606180e+00
-1.18705833e+00 4.86803442e-01 -3.46034169e-01 -5.31576991e-01
5.82450889e-02 9.40288901e-01 -1.03982389e+00 4.73688513e-01
-8.27113271e-01 1.06184840e+00 4.53228980e-01 5.19800305e-01
-2.06897661e-01 1.53080091e-01 -7.62293875e-01 -4.81775030e-02
-3.46372157e-01 -1.53218694e-02 8.15737307e-01 -1.47116721e+00
-1.84748542e+00 1.13955474e+00 -4.76305604e-01 1.08451001e-01
4.07236069e-01 5.05577922e-02 -7.99441695e-01 2.16999441e-01
-4.06445146e-01 4.37459171e-01 1.12513340e+00 -8.96097898e-01
-1.11746050e-01 -6.99712396e-01 -5.27320921e-01 -2.50944674e-01
-3.71281415e-01 4.69657660e-01 -1.60632879e-01 -5.27266338e-02
2.95308769e-01 -7.93153286e-01 2.67338276e-01 -4.93238062e-01
-1.57565147e-01 -4.00198877e-01 1.10496759e+00 -6.95052803e-01
8.53454113e-01 -2.40462995e+00 -1.91413134e-01 8.58490944e-01
-4.51575853e-02 5.97682953e-01 6.45599514e-02 2.00870320e-01
-4.51987624e-01 2.29889564e-02 3.09946328e-01 -2.06716090e-01
-4.05879803e-02 2.94135928e-01 3.09643567e-01 4.17901218e-01
8.04927647e-02 2.94606596e-01 6.31731302e-02 -5.19192278e-01
4.10630763e-01 8.74753833e-01 -1.37397408e-01 1.77778468e-01
6.79134727e-01 -7.55715296e-02 -7.97848105e-02 9.03367817e-01
9.79191720e-01 5.63298345e-01 3.87326360e-01 -3.70353609e-01
-2.56327480e-01 -6.14052832e-01 -1.56878626e+00 5.25165498e-01
-1.74198031e-01 6.78738236e-01 3.81331533e-01 -8.80052686e-01
1.46332002e+00 6.34192050e-01 3.45411330e-01 -3.63896787e-01
7.49233127e-01 2.85086334e-01 7.32700992e-03 -9.69341815e-01
6.66236952e-02 -2.13536486e-01 6.05340242e-01 1.79245159e-01
2.91065186e-01 5.86435556e-01 1.43430412e-01 -5.79545021e-01
3.72537643e-01 -4.71596681e-02 4.27459627e-01 -1.96106777e-01
1.13632500e+00 -5.30511558e-01 2.67958015e-01 1.34538546e-01
-3.76204461e-01 1.06406100e-01 3.73567015e-01 -9.68377471e-01
-8.10573757e-01 -4.99922931e-01 -3.26240212e-01 7.43964732e-01
-4.34587210e-01 2.91700512e-01 -1.06331360e+00 -4.04435039e-01
1.69406623e-01 2.17839450e-01 -8.53327036e-01 1.57239083e-02
-1.89021736e-01 -5.92648923e-01 6.23016834e-01 2.25338578e-01
6.92836344e-01 -1.20967031e+00 -4.92455870e-01 -1.71972439e-01
2.77231544e-01 -9.88162041e-01 1.29203200e-01 -4.00606170e-02
-7.37652659e-01 -1.10595298e+00 -3.34957361e-01 -8.92747283e-01
9.77328420e-01 -3.33831072e-01 7.53679812e-01 1.37315303e-01
-6.47276342e-01 2.14071274e-01 -2.70911843e-01 -3.91879141e-01
-5.41653216e-01 -3.38162541e-01 5.88815093e-01 5.61925292e-01
1.13358438e+00 -4.46469486e-01 -1.82844058e-01 2.98501253e-01
-4.86875713e-01 -7.15843856e-01 5.17946601e-01 7.01664567e-01
2.20841676e-01 2.70166904e-01 3.28417152e-01 -6.92536175e-01
7.08205640e-01 -1.71881169e-01 -5.85752487e-01 1.29101649e-01
-7.45318711e-01 -1.57860070e-01 3.85154188e-01 -3.49970937e-01
-1.02652800e+00 7.09333777e-01 -3.44011694e-01 -3.60663116e-01
-7.70879388e-01 1.08262196e-01 -1.42909735e-01 -9.02384937e-01
7.93887019e-01 1.10072471e-01 6.70358241e-01 -6.13194168e-01
2.22070590e-02 1.09780705e+00 5.08324862e-01 -1.41025469e-01
5.77227473e-01 3.66676658e-01 1.80628359e-01 -1.05127954e+00
1.34274423e-01 -5.72454214e-01 -1.13830626e+00 -6.76219046e-01
8.61046851e-01 -4.81215328e-01 -1.27781415e+00 6.75294995e-01
-1.11869502e+00 4.61681038e-01 1.88145161e-01 5.85517943e-01
-3.91016603e-01 3.03918779e-01 -5.70543051e-01 -1.38039625e+00
-5.15063107e-01 -1.02030361e+00 4.73063976e-01 3.97718608e-01
-2.21540228e-01 -7.73022056e-01 -1.00711845e-01 1.84420049e-01
2.35548928e-01 2.67047495e-01 5.85545123e-01 -6.97652459e-01
2.87673980e-01 -7.07028031e-01 -2.15354368e-01 6.29658103e-01
3.83838177e-01 5.30152440e-01 -1.32001925e+00 3.80519032e-02
7.22560585e-02 -1.49393737e-01 3.95435363e-01 1.13554925e-01
7.20515013e-01 -2.85432130e-01 1.13903312e-02 4.72066075e-01
1.51518142e+00 8.63199472e-01 7.82898247e-01 2.57820129e-01
2.61767983e-01 7.76410341e-01 2.14356487e-03 3.41762424e-01
-2.39004463e-01 3.62071395e-01 1.67248368e-01 -1.41633242e-01
2.09677041e-01 2.25102812e-01 1.07568741e-01 4.36656117e-01
-6.18311465e-01 4.38701659e-01 -1.06828797e+00 2.95328889e-02
-1.27387691e+00 -1.33957314e+00 1.00896552e-01 1.77378702e+00
4.26994711e-01 -4.01448369e-01 4.09166902e-01 6.30193293e-01
8.68072510e-01 -4.56292093e-01 -1.12154350e-01 -1.08389390e+00
-9.10981372e-02 2.88992375e-01 2.48057589e-01 4.88713682e-01
-1.23430204e+00 7.33976066e-01 8.17973804e+00 3.22368354e-01
-1.46488845e+00 -2.72826225e-01 5.52706718e-01 3.44906330e-01
5.97673476e-01 -6.45712674e-01 -7.33721375e-01 2.04422608e-01
8.52356195e-01 2.03993559e-01 5.35072267e-01 1.10260642e+00
1.84645981e-01 -4.78989892e-02 -6.27701819e-01 1.48353779e+00
5.57936370e-01 -7.85521507e-01 4.86477390e-02 1.28796175e-01
4.68848705e-01 -4.47679937e-01 1.89875394e-01 -7.99021032e-03
-2.59741783e-01 -1.20101500e+00 2.30412781e-01 7.27253258e-01
7.85096467e-01 -9.27049637e-01 1.06505227e+00 2.56699584e-02
-1.00652623e+00 -3.00078839e-01 -4.43998516e-01 -2.70919025e-01
-5.60915947e-01 5.88481650e-02 -1.03382957e+00 2.26449594e-01
6.72476768e-01 1.76292866e-01 -4.86628115e-01 7.19261110e-01
8.74134526e-02 2.39191502e-01 -5.74780226e-01 -3.27054322e-01
2.28403434e-02 -4.41554695e-01 1.70867607e-01 1.32459438e+00
2.88686901e-01 2.07265601e-01 -2.35532895e-01 7.44828820e-01
2.12784901e-01 6.74603820e-01 -7.30476320e-01 -7.49346465e-02
3.81301612e-01 1.44158387e+00 -6.49572492e-01 -1.61266834e-01
-2.42875949e-01 8.76495898e-01 -1.52022943e-01 2.85696268e-01
-1.41313553e-01 -5.01897097e-01 6.06670439e-01 -1.07220858e-01
2.84399629e-01 3.11912713e-03 -3.53612214e-01 -7.12877393e-01
-1.80617664e-02 -8.63639891e-01 2.29184061e-01 -5.09674132e-01
-1.09046185e+00 8.50892782e-01 -1.96433160e-02 -8.16022038e-01
-5.57728469e-01 -1.23983669e+00 -7.38756001e-01 1.15172267e+00
-8.17183793e-01 -9.77150857e-01 -1.70108601e-01 5.94505966e-01
3.82600985e-02 -1.00508153e+00 1.11475658e+00 2.55796343e-01
-8.13659966e-01 7.12579966e-01 2.56806642e-01 5.01984537e-01
1.95177734e-01 -9.20696914e-01 -2.25713670e-01 4.71240014e-01
-2.47564018e-01 8.44017446e-01 5.76759279e-01 -3.07461530e-01
-1.35232854e+00 -2.77964652e-01 1.16648328e+00 -6.52980730e-02
1.64501116e-01 -5.42930923e-02 -5.75293422e-01 1.90455213e-01
2.20489755e-01 -1.20297000e-01 9.56984222e-01 1.23920284e-01
-6.06465816e-01 -2.78910011e-01 -1.79756343e+00 3.60614747e-01
3.62185776e-01 -7.73322642e-01 -6.38663828e-01 -1.19304009e-01
-4.52208519e-01 1.72546133e-01 -9.44394588e-01 2.48610467e-01
1.18420374e+00 -1.15235770e+00 9.70262170e-01 -6.01762891e-01
-2.17828035e-01 -2.02195823e-01 -1.06409259e-01 -7.46401906e-01
-4.63498414e-01 -3.89889598e-01 2.46499747e-01 1.27983069e+00
3.40521067e-01 -7.47386813e-01 9.41793084e-01 8.90957057e-01
6.74563944e-01 -3.62112373e-01 -1.04199946e+00 -7.18240738e-01
-4.01133776e-01 -4.37646843e-02 5.88447809e-01 8.73511791e-01
2.28182316e-01 1.42127037e-01 -2.10539356e-01 -2.11671323e-01
6.06017172e-01 -4.37032014e-01 4.04858232e-01 -1.47726655e+00
3.47478718e-01 -5.80240607e-01 -1.12453628e+00 3.28195721e-01
6.84260905e-01 -3.40125889e-01 -3.51998627e-01 -8.46716821e-01
-1.14621013e-01 1.58937439e-01 -2.11745232e-01 5.63369632e-01
4.15581822e-01 6.23793364e-01 1.94737300e-01 1.36517435e-01
2.22214535e-01 8.79426748e-02 6.45350218e-01 7.46894702e-02
-2.47813120e-01 -6.26324713e-02 -2.83562928e-01 9.48075354e-01
1.03529668e+00 -2.76050493e-02 -2.19687670e-01 1.26180574e-01
-2.33766302e-01 -3.83322567e-01 1.29919574e-01 -1.24296331e+00
7.62127712e-03 -2.15152260e-02 1.08383369e+00 -3.41858447e-01
3.53908688e-01 -1.27270257e+00 5.63087881e-01 5.87505341e-01
7.51689598e-02 1.96851507e-01 2.79221416e-01 -4.06388827e-02
-3.88139784e-01 -5.38228512e-01 9.98227954e-01 -1.10700175e-01
-8.28369141e-01 -7.36835599e-02 -6.31661654e-01 -9.65909719e-01
1.04172003e+00 -8.36490810e-01 1.86877221e-01 -1.26219586e-01
-1.10908663e+00 -5.78721642e-01 2.44016141e-01 2.21839026e-01
7.02167153e-01 -1.43735182e+00 -4.82037336e-01 1.03780127e+00
-2.56427109e-01 -1.14845526e+00 2.10210204e-01 5.40012479e-01
-8.51045132e-01 3.47019076e-01 -9.86252367e-01 -2.06986561e-01
-2.02734351e+00 3.84121925e-01 7.13339448e-01 6.22589290e-01
-9.02339518e-02 6.66745186e-01 -4.82067466e-01 -1.98802873e-01
2.34113395e-01 5.31902373e-01 -1.11033452e+00 2.80332267e-01
8.58651161e-01 6.32727921e-01 5.72370365e-02 -1.36255682e+00
-5.82712054e-01 9.42463756e-01 1.44300207e-01 -1.41605705e-01
1.01334929e+00 1.60497352e-01 -6.20003223e-01 2.89060563e-01
1.36569500e+00 -1.48335189e-01 -4.57617462e-01 1.77072093e-01
-4.60195281e-02 -4.82262194e-01 -5.55230901e-02 -7.47740448e-01
-9.98119891e-01 7.30228245e-01 1.36724198e+00 2.37951145e-01
1.42802882e+00 -4.38987821e-01 8.50639716e-02 8.08804691e-01
-2.55284775e-02 -1.45265353e+00 -4.31704462e-01 6.55331612e-01
9.35422838e-01 -1.16062665e+00 -1.70005322e-01 -3.59150290e-01
-2.87780732e-01 1.74730873e+00 5.08812606e-01 -1.92958936e-01
1.10002780e+00 3.35477740e-01 5.07694364e-01 -1.80564687e-01
-3.42746228e-01 -2.35630780e-01 4.68281925e-01 9.04253066e-01
6.69702351e-01 2.09586382e-01 -5.27076244e-01 5.28908193e-01
-2.13560641e-01 2.22892374e-01 1.33856796e-02 6.86994731e-01
-5.05473018e-01 -9.63086247e-01 -1.14824092e+00 1.56224549e-01
-6.05349422e-01 3.86242568e-01 -9.14556444e-01 7.32100666e-01
5.95846474e-01 8.08054686e-01 1.33618146e-01 -8.88358831e-01
4.82042968e-01 7.11874425e-01 3.70241493e-01 -2.10946687e-02
-6.96575701e-01 6.78685158e-02 -4.03862447e-02 -3.66435856e-01
-6.59820795e-01 -7.04429567e-01 -1.06013680e+00 -6.66005969e-01
-1.08007334e-01 2.72502571e-01 1.21020842e+00 7.43234634e-01
1.13830566e-01 -1.45069584e-01 8.60248983e-01 -8.99854600e-01
-3.70512336e-01 -1.11025918e+00 -7.44032085e-01 3.86336088e-01
1.99235529e-01 -7.89140821e-01 -4.33683574e-01 3.24845850e-01] | [13.306941032409668, 0.9074081778526306] |
00200902-0bcb-4a1d-b8d7-d64c1e02bba9 | difformer-scalable-graph-transformers-induced | 2301.09474 | null | https://arxiv.org/abs/2301.09474v4 | https://arxiv.org/pdf/2301.09474v4.pdf | DIFFormer: Scalable (Graph) Transformers Induced by Energy Constrained Diffusion | Real-world data generation often involves complex inter-dependencies among instances, violating the IID-data hypothesis of standard learning paradigms and posing a challenge for uncovering the geometric structures for learning desired instance representations. To this end, we introduce an energy constrained diffusion model which encodes a batch of instances from a dataset into evolutionary states that progressively incorporate other instances' information by their interactions. The diffusion process is constrained by descent criteria w.r.t.~a principled energy function that characterizes the global consistency of instance representations over latent structures. We provide rigorous theory that implies closed-form optimal estimates for the pairwise diffusion strength among arbitrary instance pairs, which gives rise to a new class of neural encoders, dubbed as DIFFormer (diffusion-based Transformers), with two instantiations: a simple version with linear complexity for prohibitive instance numbers, and an advanced version for learning complex structures. Experiments highlight the wide applicability of our model as a general-purpose encoder backbone with superior performance in various tasks, such as node classification on large graphs, semi-supervised image/text classification, and spatial-temporal dynamics prediction. | ['Junchi Yan', 'David Wipf', 'Yixuan He', 'Wentao Zhao', 'Chenxiao Yang', 'Qitian Wu'] | 2023-01-23 | null | null | null | null | ['image-text-classification'] | ['miscellaneous'] | [ 4.18852955e-01 4.64447200e-01 -4.63212401e-01 -4.56524521e-01
-4.26445633e-01 -6.67657435e-01 8.54188859e-01 2.09300086e-01
6.08078875e-02 7.34986663e-01 1.39276087e-01 -1.77617386e-01
-7.25854039e-01 -9.90030766e-01 -9.50388074e-01 -1.08845258e+00
-5.35015047e-01 8.15723777e-01 -1.36615830e-02 -2.10509494e-01
2.71237940e-01 5.13102293e-01 -1.22287571e+00 1.26814350e-01
9.60724473e-01 9.66643512e-01 5.92688248e-02 5.91863930e-01
-2.22616300e-01 7.87137091e-01 -3.40152472e-01 -6.47610426e-01
1.52513534e-01 -5.69451153e-01 -9.66200590e-01 3.51702601e-01
3.77349734e-01 1.21674754e-01 -7.13265777e-01 1.10820699e+00
1.88982815e-01 6.72343895e-02 1.24060810e+00 -1.23003137e+00
-1.26621270e+00 7.14429200e-01 -6.78062975e-01 3.21635544e-01
-2.68030469e-03 -1.16779916e-01 1.31811821e+00 -5.33910513e-01
1.03418779e+00 1.04452002e+00 6.06497109e-01 7.61160016e-01
-1.48048985e+00 -3.97831351e-01 3.42929780e-01 2.56895691e-01
-1.24907482e+00 -3.54250550e-01 1.15691900e+00 -5.98313570e-01
9.09847379e-01 8.34884942e-02 7.51810431e-01 1.32468617e+00
3.93208295e-01 7.77002811e-01 8.65171552e-01 -2.37509564e-01
4.03621376e-01 1.26590088e-01 1.86229527e-01 1.16858745e+00
6.03981555e-01 -1.27819225e-01 -7.11440384e-01 -2.27940992e-01
8.16304564e-01 -2.34383568e-01 -3.63107055e-01 -7.58979797e-01
-9.88112807e-01 8.57341826e-01 8.24158788e-01 3.11443865e-01
-3.67198318e-01 2.31390804e-01 2.77442392e-02 2.32500985e-01
6.57089293e-01 2.75569439e-01 -2.84137994e-01 2.88500190e-01
-7.27011681e-01 2.60617435e-01 7.33119726e-01 8.92464340e-01
6.77821755e-01 -7.96242431e-02 1.75964415e-01 5.88587701e-01
3.33399594e-01 1.36192858e-01 3.92744154e-01 -7.06434011e-01
4.87873673e-01 7.05069542e-01 -2.58117735e-01 -1.15789461e+00
-2.60916859e-01 -7.63218939e-01 -1.24741220e+00 -2.22485155e-01
2.13387579e-01 -1.51991248e-02 -6.53664708e-01 2.19802213e+00
3.33037257e-01 1.66995674e-01 -9.35872197e-02 5.73580980e-01
3.02465528e-01 8.19194376e-01 -1.85995162e-01 -5.45765102e-01
8.63365173e-01 -7.77396441e-01 -6.94987655e-01 -1.90361455e-01
5.06892681e-01 -5.28417453e-02 6.62724018e-01 2.32615098e-01
-1.29671109e+00 -2.68775493e-01 -9.94984090e-01 -6.73813298e-02
-4.51492846e-01 -2.91977406e-01 9.84406471e-01 3.00055534e-01
-1.29613519e+00 1.05172420e+00 -1.03019440e+00 -2.06988767e-01
5.23321271e-01 6.37211323e-01 -3.09260935e-01 8.74757171e-02
-9.53482091e-01 6.51055098e-01 2.92031795e-01 1.53221637e-01
-7.57589757e-01 -5.11139691e-01 -7.53137290e-01 -2.63452120e-02
-3.91151197e-02 -9.60880220e-01 8.77223909e-01 -1.07497394e+00
-1.27866888e+00 7.44704843e-01 -2.26373300e-01 -4.13523972e-01
3.93024057e-01 1.77614167e-01 -2.21501574e-01 1.62203595e-01
-6.34494573e-02 7.58441389e-01 1.02092612e+00 -1.32106245e+00
-2.93123741e-02 -5.55556715e-01 -8.69917795e-02 3.04897845e-01
-6.85488701e-01 -6.26295924e-01 -1.60195872e-01 -6.38733566e-01
2.34276146e-01 -9.51873899e-01 -3.72850925e-01 2.53521502e-01
-5.80848575e-01 -4.70968992e-01 7.25589812e-01 -2.95047760e-01
1.22954273e+00 -1.77638483e+00 8.25343728e-01 3.93683642e-01
4.70612228e-01 -2.44750246e-01 -2.38748282e-01 6.40517056e-01
-4.23742794e-02 1.44441068e-01 -5.22566736e-01 -2.23010957e-01
-1.04421362e-01 4.21943069e-01 -2.68741041e-01 5.28879881e-01
3.61646563e-01 1.18169618e+00 -1.07723010e+00 -5.01925051e-01
-1.68004051e-01 5.53563774e-01 -5.94057739e-01 1.24949425e-01
-4.32687998e-01 3.14835995e-01 -5.52088559e-01 3.58665854e-01
2.99685657e-01 -9.13010061e-01 3.74734312e-01 -1.54941157e-01
3.40311766e-01 2.40711153e-01 -9.23393130e-01 1.81765819e+00
-1.48865893e-01 6.83085144e-01 4.98813167e-02 -1.28258705e+00
7.99321890e-01 1.17882257e-02 5.56948543e-01 -2.04442307e-01
-1.89153373e-01 3.48478138e-01 -4.89187539e-02 -3.65282774e-01
3.32776755e-01 -1.92320764e-01 4.85722627e-03 4.20284301e-01
3.07559609e-01 1.01474196e-01 1.64873123e-01 6.07313454e-01
1.21419227e+00 -5.17763421e-02 8.55462030e-02 -5.91893137e-01
6.43256158e-02 -2.71848470e-01 2.80582368e-01 7.35543370e-01
8.05280432e-02 2.45777756e-01 5.06502807e-01 -5.16768217e-01
-9.42547262e-01 -1.15731966e+00 -2.28923932e-01 5.38510084e-01
2.65487731e-01 -3.29933077e-01 -6.59388900e-01 -8.38888884e-01
5.11085577e-02 4.85138535e-01 -9.14794803e-01 -5.27354419e-01
-4.17227685e-01 -1.06328106e+00 1.44249797e-01 4.64589268e-01
2.87625343e-01 -7.44696856e-01 -1.33631557e-01 2.53331542e-01
4.08722647e-03 -7.94336379e-01 -5.70957124e-01 4.28431064e-01
-1.08851326e+00 -9.46416676e-01 -6.29756033e-01 -8.32055926e-01
1.07251954e+00 -6.69459626e-02 1.11527276e+00 -1.03682749e-01
-4.66172129e-01 4.65961993e-01 7.14564547e-02 3.67756635e-02
-4.26703483e-01 3.97140294e-01 1.44217402e-01 1.01526059e-01
-2.26067286e-02 -7.63715267e-01 -8.63147914e-01 -6.25312477e-02
-9.19733167e-01 1.85934231e-01 7.54939437e-01 9.60627615e-01
6.70506001e-01 2.28535876e-01 5.09064615e-01 -9.56967652e-01
7.68870115e-01 -7.79276490e-01 -3.48322541e-01 5.18188238e-01
-7.91436493e-01 7.07003534e-01 6.30172253e-01 -5.36305547e-01
-9.77058411e-01 7.47764930e-02 1.35170683e-01 -2.59922564e-01
1.90471753e-01 4.91893560e-01 4.79899384e-02 -5.30200265e-02
6.45872295e-01 5.03118038e-01 -5.80764711e-02 -2.12790087e-01
5.45043051e-01 3.77970189e-01 4.72219378e-01 -8.64554822e-01
8.65777016e-01 4.73420501e-01 3.40406805e-01 -9.72691298e-01
-1.04402256e+00 -2.73892172e-02 -7.89996505e-01 -1.25464350e-01
8.18495214e-01 -4.83425111e-01 -3.70645016e-01 3.18173587e-01
-1.13478005e+00 -4.21163589e-01 -4.70219135e-01 8.22211727e-02
-5.65884233e-01 2.92678237e-01 -8.82286489e-01 -7.66193151e-01
-3.51880461e-01 -8.49475384e-01 1.02012873e+00 4.14079018e-02
2.60469224e-02 -1.50489116e+00 3.92858535e-01 5.75585589e-02
1.87193617e-01 3.10024709e-01 1.39468706e+00 -3.59924287e-01
-8.23812008e-01 -5.36984578e-02 -8.05100705e-03 4.00550328e-02
5.45080490e-02 2.29014620e-01 -5.58014393e-01 -3.99585992e-01
-2.23607913e-01 -3.29990298e-01 1.11358058e+00 5.02078235e-01
1.35752285e+00 -6.58415079e-01 -7.17202842e-01 6.60222530e-01
1.39292324e+00 -2.42558196e-02 1.92646667e-01 -2.62329936e-01
6.40780389e-01 6.17393613e-01 -2.68665999e-01 3.30039710e-01
3.76572818e-01 5.83145320e-01 4.34069574e-01 1.61523640e-01
-6.91975001e-03 -4.95401144e-01 2.25814149e-01 1.42064273e+00
-1.50378481e-01 -5.78265667e-01 -6.50708735e-01 4.52723593e-01
-1.85030019e+00 -1.05605710e+00 -6.34685233e-02 1.99129677e+00
9.03023779e-01 2.43423179e-01 -8.50345716e-02 -1.04789466e-01
6.62710488e-01 4.57912803e-01 -9.94986355e-01 -1.14941604e-01
1.72326248e-02 2.04070643e-01 3.36219221e-01 4.40709710e-01
-7.90765464e-01 5.64879954e-01 6.63748026e+00 6.16545856e-01
-7.90103436e-01 -8.29644725e-02 9.47279572e-01 4.79997648e-03
-6.76342487e-01 -1.27125829e-01 -7.99761117e-01 3.77638251e-01
1.02556431e+00 -2.53428102e-01 5.49211621e-01 6.33863866e-01
-1.61257163e-01 4.60816085e-01 -1.26721227e+00 7.47751534e-01
-5.83492452e-03 -1.76884496e+00 3.75256211e-01 4.26812947e-01
1.07877851e+00 -1.28243463e-02 2.51053452e-01 -2.03578882e-02
4.80106562e-01 -8.63578379e-01 5.68138659e-01 5.50293922e-01
7.72148609e-01 -5.83474934e-01 8.61894190e-02 4.69406933e-01
-1.21884739e+00 -7.39531815e-02 -4.76144522e-01 1.65771678e-01
1.08340152e-01 6.02526724e-01 -4.23494041e-01 3.86526823e-01
1.71045363e-01 1.12807989e+00 -4.33283448e-01 5.95000982e-01
2.43578870e-02 4.77654606e-01 -4.02964950e-01 -2.33093098e-01
2.83974737e-01 -3.19453299e-01 4.57019210e-01 1.00306261e+00
3.90514374e-01 1.61268204e-01 -6.48420081e-02 1.04215908e+00
-4.65382636e-01 -8.73976424e-02 -8.06181371e-01 -2.29913518e-01
5.03434658e-01 1.23224211e+00 -9.35287535e-01 -1.52933627e-01
-9.64434743e-02 1.16322887e+00 9.32194769e-01 3.39074194e-01
-6.65977240e-01 -9.36331302e-02 4.81100529e-01 1.91768482e-01
1.52006656e-01 -3.63467187e-01 -4.64326032e-02 -1.42899585e+00
4.20645401e-02 -4.33360547e-01 4.50834185e-01 -4.37656879e-01
-1.74536014e+00 7.54128814e-01 6.44695461e-02 -7.51637459e-01
-2.31304005e-01 -8.23508501e-01 -7.38807440e-01 4.13706750e-01
-1.29459071e+00 -8.92006040e-01 -1.38892606e-01 6.81284606e-01
5.16001463e-01 -9.28302947e-03 7.33395219e-01 -7.76540413e-02
-7.92395949e-01 5.15963078e-01 3.05357605e-01 -1.37387961e-01
1.71130642e-01 -1.57741511e+00 3.49026799e-01 5.63483953e-01
6.92742884e-01 6.46083653e-01 5.88785887e-01 -6.47632241e-01
-1.62428045e+00 -1.09624624e+00 8.20996284e-01 -5.20414114e-01
9.39394891e-01 -5.21094024e-01 -1.05558741e+00 7.32208610e-01
2.68994272e-01 1.42860740e-01 5.44623077e-01 4.22967523e-02
-3.10187042e-01 -8.24191421e-02 -7.93157995e-01 5.05600512e-01
1.55196583e+00 -5.84058523e-01 -2.75112599e-01 7.79118121e-01
4.87778574e-01 -2.17499748e-01 -1.01065361e+00 3.23810846e-01
3.81273806e-01 -7.19238222e-01 8.62604976e-01 -8.44515443e-01
7.90124595e-01 1.54036745e-01 4.52628136e-02 -1.34541690e+00
-4.44125831e-01 -7.43433893e-01 -8.10126603e-01 8.65911782e-01
6.10599577e-01 -5.47247589e-01 8.81529152e-01 6.48262441e-01
1.04906566e-01 -1.23171854e+00 -9.08367634e-01 -6.81607306e-01
2.77520269e-01 6.22591423e-03 1.97077796e-01 1.12577188e+00
7.25055486e-02 5.78812003e-01 -1.76338837e-01 1.79740176e-01
1.10323644e+00 1.69608101e-01 1.24534808e-01 -1.36521053e+00
-6.84115887e-01 -7.08111286e-01 -4.85323578e-01 -1.44104433e+00
4.48544323e-01 -1.24639976e+00 -2.72946149e-01 -1.57021940e+00
4.54298943e-01 -5.53440630e-01 -2.99424022e-01 8.59449208e-02
-8.09750333e-02 -5.07250167e-02 -3.27370167e-01 6.31425142e-01
-5.31829476e-01 7.59134412e-01 1.18414974e+00 -2.02156648e-01
-1.31643787e-01 -4.78890315e-02 -5.94198167e-01 5.28954089e-01
5.65617800e-01 -4.48087156e-01 -8.97670865e-01 -5.77390313e-01
5.61416090e-01 1.00388058e-01 2.87364721e-01 -7.60095835e-01
2.98421621e-01 -1.14514165e-01 4.15697336e-01 -2.65633255e-01
2.59101748e-01 -6.11097455e-01 2.64258832e-01 4.78380829e-01
-9.93783236e-01 1.69261694e-01 -4.08126652e-01 1.05496204e+00
7.10474849e-02 -2.89353937e-01 7.56066799e-01 -1.52489513e-01
-3.89879793e-01 7.74743855e-01 -1.19721349e-02 2.19421655e-01
1.14094150e+00 -4.63281199e-03 -4.33394104e-01 -4.92391139e-01
-7.69150555e-01 1.37105146e-02 3.02348018e-01 2.96677381e-01
5.73026717e-01 -1.38983572e+00 -6.16377294e-01 2.47400180e-01
-1.18537501e-01 5.85839003e-02 1.11887716e-01 5.85486591e-01
-2.51251012e-01 1.80842072e-01 1.40271425e-01 -6.54770255e-01
-7.88977563e-01 5.42949617e-01 5.23101270e-01 -5.75270414e-01
-6.41957521e-01 1.07929075e+00 3.11080068e-01 -1.63328901e-01
1.44305620e-02 -1.05888449e-01 1.72183454e-01 -1.42211765e-01
1.51465923e-01 1.57034650e-01 -1.47223696e-01 -5.89966238e-01
-1.59024626e-01 4.88557339e-01 -3.21551144e-01 2.44474728e-02
1.45213974e+00 -3.42867747e-02 -2.14274570e-01 5.64190924e-01
1.35173047e+00 -3.74376923e-01 -1.55900931e+00 -3.80154997e-01
3.22984792e-02 -5.41915409e-02 4.56714295e-02 -5.14796317e-01
-1.20046747e+00 8.68154943e-01 2.69203097e-01 6.14944220e-01
8.73988748e-01 4.98363107e-01 5.55170000e-01 5.66895723e-01
4.37243223e-01 -7.76860178e-01 4.37680632e-01 1.87791273e-01
7.82759726e-01 -8.70871842e-01 -9.31518972e-02 -2.95162231e-01
-4.17001218e-01 1.18921196e+00 6.10960722e-01 -2.30330750e-01
8.48944187e-01 2.46957585e-01 -6.64596498e-01 -5.00911951e-01
-1.27064300e+00 2.28952244e-01 4.76159245e-01 4.21774626e-01
4.01141256e-01 -4.75664549e-02 -3.74551900e-02 1.78318828e-01
-7.96051845e-02 -4.01527047e-01 1.53798953e-01 6.88448131e-01
-3.85714144e-01 -9.79934454e-01 3.05823207e-01 7.37720668e-01
-8.89372677e-02 9.72680077e-02 -5.41669369e-01 5.95747709e-01
-1.67260561e-02 3.92045766e-01 1.65768057e-01 -1.59273878e-01
-6.13912269e-02 2.64980882e-01 7.90435553e-01 -4.18613106e-01
-1.93197548e-01 -2.63791144e-01 -1.72015905e-01 -2.14827359e-01
-5.10335326e-01 -6.70732200e-01 -9.20081437e-01 -2.33018309e-01
-5.38770735e-01 1.86198339e-01 5.07075012e-01 8.68218780e-01
5.68290830e-01 4.15727824e-01 7.82445550e-01 -8.41142356e-01
-8.15712929e-01 -6.55905187e-01 -4.42895681e-01 4.96093720e-01
2.59873748e-01 -6.12175763e-01 -4.07060981e-01 1.38948813e-01] | [6.961780071258545, 6.128242492675781] |
8af996a9-9c16-430f-9bd8-5bc1dc9d8990 | temporal-decoupling-graph-convolutional | null | null | https://ieeexplore.ieee.org/abstract/document/10113233/ | https://www.researchgate.net/profile/Mengyuan-Liu-2/publication/370452642_Temporal_Decoupling_Graph_Convolutional_Network_for_Skeleton-based_Gesture_Recognition/links/645bb28739c408339b3ace97/Temporal-Decoupling-Graph-Convolutional-Network-for-Skeleton-based-Gesture-Recognition.pdf | Temporal Decoupling Graph Convolutional Network for Skeleton-based Gesture Recognition | Skeleton-based gesture recognition methods have achieved high success using Graph Convolutional Network (GCN), which commonly uses an adjacency matrix to model the spatial topology of skeletons. However, previous methods use the same adjacency matrix for skeletons from different frames, which limits the flexibility of GCN to model temporal information. To solve this problem, we propose a Temporal Decoupling Graph Convolutional Network (TD-GCN), which applies different adjacency matrices for skeletons from different frames. The main steps of each convolution layer in our proposed TD-GCN are as follows. To extract deep spatiotemporal information from skeleton joints, we first extract high-level spatiotemporal features from skeleton data. Then, channel-dependent and temporal-dependent adjacency matrices corresponding to different channels and frames are calculated to capture the spatiotemporal dependencies between skeleton joints. Finally, to fuse topology information from neighbor skeleton joints, spatiotemporal features of skeleton joints are fused based on channel-dependent and temporal-dependent adjacency matrices. To the best of our knowledge, we are the first to use temporal-dependent adjacency matrices for temporal-sensitive topology learning from skeleton joints. The proposed TD-GCN effectively improves the modeling ability of GCN and achieves state-of-the-art results on gesture datasets including SHREC'17 Track and DHG-14/28. Our code is available at: https://github.com/liujf69/TD-GCN-Gesture . | ['Mengyuan Liu', 'Yuan Gao', 'Can Wang', 'Xinshun Wang', 'Jinfu Liu'] | 2023-05-01 | null | null | null | ieee-transactions-on-multimedia-2023-5 | ['hand-gesture-recognition', 'skeleton-based-action-recognition', 'gesture-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-1.84941709e-01 -6.32752180e-01 -2.03565091e-01 -2.78332591e-01
-3.11455250e-01 -2.90499538e-01 2.05012545e-01 -3.50542754e-01
-4.83051687e-01 2.39168316e-01 4.76898760e-01 6.93695247e-02
-5.37087396e-02 -8.76654804e-01 -4.92664665e-01 -5.81096768e-01
-3.87415111e-01 -3.71782528e-03 6.66088045e-01 -1.89205363e-01
-5.66781790e-04 7.72468030e-01 -1.03311801e+00 2.94289261e-01
3.04718018e-01 8.74675274e-01 -7.58483112e-02 6.67541206e-01
-2.68517375e-01 3.69509429e-01 -2.90923297e-01 -2.17635818e-02
2.84310699e-01 -4.75719690e-01 -5.76501071e-01 -2.84719020e-01
4.20405775e-01 -5.97151220e-01 -1.27243543e+00 7.47963965e-01
7.22225189e-01 3.66750121e-01 4.23729509e-01 -1.08294511e+00
-3.11343491e-01 6.12229764e-01 -7.48730898e-01 4.22679245e-01
1.72525749e-01 4.78467852e-01 8.43878567e-01 -6.28852248e-01
7.59108007e-01 1.33788478e+00 5.71198583e-01 7.08504379e-01
-8.84501040e-01 -9.92144346e-01 4.22993958e-01 2.92812228e-01
-1.31997871e+00 -1.67874798e-01 8.56077313e-01 -1.68269023e-01
9.87400472e-01 5.66137396e-03 1.16237223e+00 1.29834068e+00
2.14729831e-02 1.03705275e+00 5.53832471e-01 3.52035537e-02
-9.69714597e-02 -1.56774485e+00 7.87153989e-02 1.00535023e+00
-2.10015535e-01 1.02289230e-01 -8.28472257e-01 2.38732249e-01
1.43598104e+00 2.57418424e-01 -2.30526939e-01 -1.58684820e-01
-1.43596900e+00 5.53648829e-01 7.78169274e-01 3.38885546e-01
-2.90962607e-01 8.11498404e-01 5.56114614e-01 2.15496376e-01
2.20631268e-02 -3.66600811e-01 -2.56679684e-01 -3.71343404e-01
-6.24204993e-01 4.57090400e-02 5.32562256e-01 8.35804105e-01
4.71797973e-01 -2.30410341e-02 -2.50343472e-01 6.38601661e-01
5.13307333e-01 4.95549232e-01 4.97185498e-01 -9.60774064e-01
6.41297460e-01 7.40006685e-01 -5.31805396e-01 -1.04134965e+00
-6.95715010e-01 -6.61566993e-03 -8.65844369e-01 -2.44422294e-02
7.44500816e-01 -1.76233843e-01 -1.32198405e+00 1.72761416e+00
2.29682088e-01 5.97892165e-01 -3.83310527e-01 1.29132843e+00
1.08328676e+00 3.97825688e-01 1.06177099e-01 1.94266140e-01
1.13500476e+00 -9.01102722e-01 -4.55181241e-01 2.43851319e-01
5.08443773e-01 -5.50828755e-01 1.06632185e+00 4.77187932e-02
-7.68174708e-01 -4.35756564e-01 -8.05505157e-01 -2.07833573e-01
-8.92554149e-02 2.81019449e-01 7.02785313e-01 1.85642615e-01
-8.19013953e-01 4.87625539e-01 -1.63380802e+00 -5.41955233e-01
5.20346701e-01 5.27982593e-01 -3.83396298e-01 -1.08771518e-01
-1.09960830e+00 2.16219291e-01 1.74654230e-01 6.02496207e-01
-7.54172623e-01 -3.00025195e-01 -8.67945731e-01 -2.11763233e-01
2.75508493e-01 -5.84595084e-01 9.58208501e-01 -4.49810892e-01
-1.61515570e+00 2.72091359e-01 -1.95819512e-01 -1.75229624e-01
6.40833259e-01 -1.70917660e-01 -3.77688468e-01 4.37236667e-01
-1.01930257e-02 8.07806671e-01 6.08982801e-01 -6.61342561e-01
-4.77777302e-01 -5.38057089e-01 1.20650092e-02 1.71401158e-01
-2.17169881e-01 -1.34905472e-01 -1.24846745e+00 -9.49209988e-01
5.32224119e-01 -1.02603137e+00 -2.62077868e-01 4.18734103e-01
-5.09402335e-01 -1.49549201e-01 1.07494259e+00 -7.19915450e-01
1.42085719e+00 -2.11612034e+00 2.91317314e-01 3.35702419e-01
2.99124986e-01 3.41686815e-01 -5.72459221e-01 3.31229985e-01
1.79013982e-01 -1.96658239e-01 -1.55304238e-01 -1.75436929e-01
-1.56246156e-01 5.57978988e-01 1.55929834e-01 3.49164665e-01
9.61255953e-02 1.25048685e+00 -9.21494663e-01 -6.67365849e-01
5.45499682e-01 7.43687689e-01 -2.23939940e-01 -2.05255255e-01
-2.36717120e-01 6.96364224e-01 -8.44360888e-01 7.43699908e-01
3.99084419e-01 -1.33150592e-01 3.14348787e-01 -6.13543391e-01
-8.36843327e-02 9.47563201e-02 -1.12902498e+00 2.39183259e+00
-5.78357317e-02 6.56478167e-01 -7.65331462e-02 -8.78243923e-01
7.20286429e-01 1.52157545e-01 1.01171792e+00 -7.08449006e-01
2.99359828e-01 9.33589786e-02 1.49050683e-01 -5.66931069e-01
-4.07208614e-02 3.39398623e-01 4.06176634e-02 5.55258393e-01
1.77849047e-02 4.96171355e-01 3.43748242e-01 2.40673199e-01
1.41100848e+00 4.17523354e-01 -1.65961593e-01 1.62226111e-01
2.48098910e-01 -2.74117291e-01 7.70728171e-01 4.06344086e-01
-3.98282945e-01 8.46393526e-01 4.61681128e-01 -7.48968780e-01
-5.44390023e-01 -1.28323054e+00 3.43184680e-01 8.01498950e-01
3.27278078e-01 -5.01163244e-01 -5.17264366e-01 -7.80811727e-01
2.53207963e-02 -2.27612093e-01 -6.51168108e-01 8.45794845e-03
-1.25566614e+00 -4.98957992e-01 9.37396049e-01 9.77339447e-01
9.51460958e-01 -9.81517076e-01 -7.75296450e-01 4.43002015e-01
-4.09068257e-01 -1.03094673e+00 -9.63899791e-01 -2.66462117e-01
-1.10466874e+00 -1.26915455e+00 -7.95022905e-01 -5.45341849e-01
3.83985400e-01 2.40367517e-01 5.77821016e-01 3.40050459e-01
-5.81646681e-01 3.75946462e-01 -4.80439007e-01 1.73560709e-01
3.86607528e-01 1.96465373e-01 -2.98800409e-01 1.28675476e-01
3.51519883e-01 -1.03682172e+00 -9.78506148e-01 4.15220916e-01
-9.40063596e-01 8.84414557e-03 6.46378279e-01 6.23152971e-01
6.20640516e-01 -3.45501661e-01 -1.72181241e-02 -2.17626512e-01
4.19490755e-01 -4.30593863e-02 -3.10594440e-01 2.75904059e-01
-1.12641761e-02 2.23692104e-01 3.66071105e-01 -6.15882814e-01
-7.17852652e-01 4.06430364e-01 -7.66984746e-02 -8.07774007e-01
-1.44476742e-01 5.20474017e-01 -2.44365826e-01 -1.48421168e-01
2.29424998e-01 2.31722862e-01 9.87043381e-02 -6.56065345e-01
3.54669124e-01 1.58255190e-01 6.81613326e-01 -7.13087320e-01
4.45076883e-01 7.17136443e-01 2.87472159e-01 -7.83856332e-01
-3.55378777e-01 -4.99614060e-01 -9.55166101e-01 -4.80367094e-01
1.03843784e+00 -6.62649691e-01 -7.68655360e-01 9.96309698e-01
-1.04725420e+00 -8.02205086e-01 -1.05890863e-01 8.03013444e-01
-3.18374306e-01 7.20508635e-01 -9.65973556e-01 -3.44020844e-01
-2.88061619e-01 -1.18547511e+00 1.19803023e+00 2.84154356e-01
-1.11917406e-02 -9.71664727e-01 2.53598616e-02 1.40099674e-01
1.74374759e-01 5.50284386e-01 4.83061105e-01 -6.76413812e-03
-7.35732317e-01 2.54599769e-02 -3.57797623e-01 -4.65224013e-02
3.87433708e-01 2.49609113e-01 -3.38475615e-01 -1.78585827e-01
-7.87374556e-01 -8.04912597e-02 1.26386166e+00 7.02593446e-01
1.07005656e+00 7.93745667e-02 -4.40538198e-01 8.34982216e-01
1.19788015e+00 9.59788859e-02 7.85962343e-01 1.44633159e-01
1.24386930e+00 2.05609575e-01 3.37545931e-01 4.73019421e-01
4.18208271e-01 8.77028883e-01 4.00008887e-01 -4.14014608e-02
-6.81514680e-01 -2.68170804e-01 4.59384352e-01 7.85004199e-01
-6.95087910e-01 -1.02169245e-01 -8.45314443e-01 5.22229850e-01
-2.18139482e+00 -7.77519822e-01 -3.52964401e-01 1.78105593e+00
5.86243212e-01 -9.19652730e-02 3.33270282e-01 -2.38280907e-01
7.27456212e-01 4.48626667e-01 -7.91645885e-01 2.87817031e-01
-1.01066589e-01 4.35090631e-01 4.10721451e-01 3.74763757e-01
-1.25945652e+00 1.11542535e+00 5.05067444e+00 7.67658114e-01
-1.27047479e+00 8.90686177e-03 -2.62925141e-02 -6.04655862e-01
3.33352126e-02 -2.67313011e-02 -3.45632941e-01 5.46919107e-01
3.61410737e-01 3.72593522e-01 3.95595521e-01 2.67793328e-01
3.81727874e-01 6.73384368e-02 -7.97743142e-01 1.06564617e+00
-2.21953794e-01 -1.22607195e+00 2.56517440e-01 1.63576499e-01
4.53264534e-01 3.53427380e-01 -1.34268790e-01 -1.87674105e-01
3.33436579e-01 -8.84424686e-01 4.51220483e-01 7.15057135e-01
7.48086691e-01 -5.70014179e-01 4.60171074e-01 -1.14686430e-01
-1.92251563e+00 9.91408676e-02 -1.05736688e-01 -3.49950567e-02
3.32183778e-01 3.98741007e-01 -1.55189827e-01 6.91910923e-01
8.42203081e-01 1.20338392e+00 -4.14885789e-01 1.05315220e+00
-3.58751982e-01 5.90416729e-01 -7.08192050e-01 8.90854374e-02
4.54344124e-01 -1.73347607e-01 4.34253305e-01 1.22730434e+00
3.08367670e-01 4.28055137e-01 1.56582460e-01 5.07668674e-01
-1.62921492e-02 -1.59781843e-01 -3.45281392e-01 -4.83136736e-02
3.44281048e-01 1.06105900e+00 -1.01323366e+00 -1.35329932e-01
-5.83649218e-01 1.17305636e+00 2.35203788e-01 5.87698996e-01
-7.79269159e-01 -3.15218538e-01 9.01313543e-01 -1.00088306e-01
2.20469847e-01 -8.40448916e-01 -1.37833744e-01 -1.43964946e+00
2.25427300e-01 -4.90653306e-01 6.85296535e-01 -3.76472741e-01
-1.17467821e+00 1.73008949e-01 -9.80782807e-02 -1.31776774e+00
1.55483067e-01 -5.70476472e-01 -7.29019642e-01 5.83724976e-01
-1.15888667e+00 -1.40179706e+00 -5.46150506e-01 1.13536727e+00
5.07772684e-01 4.33519147e-02 5.30108511e-01 4.03186917e-01
-6.39292061e-01 6.51777506e-01 -2.47401536e-01 1.10553741e+00
5.75779676e-01 -9.50846076e-01 7.42819667e-01 9.49098766e-01
3.08160692e-01 7.92140484e-01 -3.57512608e-02 -9.28765237e-01
-1.52775145e+00 -1.16913974e+00 4.78336036e-01 -1.75674826e-01
5.90644896e-01 -5.21333963e-02 -7.25305080e-01 6.64053023e-01
-3.22994351e-01 4.90323305e-01 4.28814083e-01 -2.15083912e-01
-5.90422928e-01 -4.07555401e-02 -6.07087195e-01 7.26056457e-01
1.76867676e+00 -4.45745230e-01 -1.96683824e-01 9.22202840e-02
4.01021183e-01 -6.22010469e-01 -1.03566647e+00 3.95908028e-01
1.17426443e+00 -8.44065726e-01 9.84373331e-01 -5.39898753e-01
2.97404230e-01 -5.60950935e-01 -1.13279276e-01 -9.77136433e-01
-1.29025012e-01 -4.25416857e-01 -1.90386206e-01 7.24422693e-01
1.15196176e-01 -3.91558290e-01 1.08654940e+00 4.41030502e-01
-2.21646279e-01 -8.48368943e-01 -1.15726924e+00 -8.58954966e-01
-1.70634702e-01 -6.99999154e-01 5.58283687e-01 8.20761681e-01
-1.62497699e-01 1.96111202e-03 -2.08625540e-01 6.66970834e-02
6.42678201e-01 2.36601442e-01 7.42048264e-01 -9.97304142e-01
-1.10029906e-01 -6.17916822e-01 -7.16741502e-01 -1.43410540e+00
3.68495882e-02 -7.86000192e-01 -1.25846416e-01 -1.74350941e+00
-1.78207207e-04 -3.15145403e-01 -3.34386259e-01 8.60197365e-01
8.75047222e-02 4.15536880e-01 4.43250239e-01 3.09044898e-01
-5.79343081e-01 5.33295274e-01 1.71221948e+00 -2.99275666e-01
-4.60708827e-01 -2.27527753e-01 1.19642034e-01 5.60321569e-01
9.23696160e-01 -1.79507390e-01 -3.84498507e-01 -7.79526889e-01
-2.48547778e-01 1.41649693e-01 6.91237748e-01 -1.06574881e+00
5.09031296e-01 -2.51218110e-01 4.86658454e-01 -7.10995793e-01
4.37399924e-01 -6.53644800e-01 3.53572607e-01 6.91011667e-01
-1.87649041e-01 2.28049196e-02 1.34266717e-02 7.34057486e-01
-2.18899250e-01 5.18025339e-01 4.39119905e-01 -2.98393756e-01
-1.07803154e+00 9.36849892e-01 -3.54725212e-01 -1.57700881e-01
6.12448633e-01 -3.20719212e-01 -1.75791383e-01 -3.00158143e-01
-1.01214325e+00 2.82182634e-01 2.53507525e-01 5.28431118e-01
1.01040781e+00 -1.60684586e+00 -5.14303267e-01 1.49992883e-01
-7.22772032e-02 2.04917014e-01 5.63481688e-01 1.18236089e+00
-7.29248762e-01 2.59143770e-01 -5.66977799e-01 -7.20347106e-01
-1.32021868e+00 -3.33261564e-02 4.11110640e-01 -1.67967364e-01
-1.01127589e+00 9.51244056e-01 4.03465331e-02 -2.85584807e-01
3.02599996e-01 -6.98115289e-01 6.56098276e-02 -9.52804089e-02
2.38298878e-01 5.27408242e-01 -1.65934876e-01 -7.15754092e-01
-5.59777617e-01 1.05193019e+00 1.30189508e-01 -3.72651786e-01
1.15855718e+00 4.94585447e-02 -8.10782611e-02 1.66279361e-01
1.32797062e+00 -5.03998578e-01 -1.66297162e+00 -3.21411401e-01
-2.47993499e-01 -4.90143448e-01 -1.23916365e-01 -5.57646394e-01
-1.86878836e+00 1.19967866e+00 6.02657020e-01 -5.66150427e-01
1.20671010e+00 -6.40205517e-02 1.36346185e+00 3.36852551e-01
4.60969597e-01 -1.07337725e+00 3.72663736e-01 6.33501768e-01
7.31526315e-01 -7.65677810e-01 -1.63984284e-01 -3.18490833e-01
-3.20794702e-01 1.49245393e+00 6.68248355e-01 -1.21783830e-01
9.10386622e-01 2.18463898e-01 2.20949054e-01 -4.53783751e-01
-2.35479787e-01 -4.56559598e-01 3.12209904e-01 5.91563940e-01
3.86033446e-01 2.48883590e-01 -4.94849265e-01 3.22953343e-01
1.44580290e-01 1.95478395e-01 -3.04238964e-03 1.05316257e+00
-8.59908089e-02 -1.40311837e+00 -4.97148596e-02 4.49648470e-01
-1.13102369e-01 2.17299927e-02 -6.00000560e-01 9.61130202e-01
5.73078617e-02 7.23793089e-01 -1.43673485e-02 -8.19225550e-01
3.88522983e-01 -9.34385136e-02 7.41901398e-01 -3.63626927e-01
-3.84448916e-01 4.75787878e-01 -4.03214432e-02 -1.17436182e+00
-7.98650205e-01 -9.05167997e-01 -1.90807974e+00 -3.69738758e-01
6.06373101e-02 -2.70966560e-01 3.55349928e-01 7.88944960e-01
4.77497280e-01 4.15148288e-01 1.10133693e-01 -9.72100198e-01
8.16557324e-04 -8.34567547e-01 -6.26430631e-01 3.94499332e-01
1.44056275e-01 -8.51551533e-01 1.62391633e-01 6.22180407e-04] | [7.6502203941345215, 0.18366886675357819] |
98a97ae7-82b2-4bc8-ae21-c39a904f8359 | understanding-the-tradeoff-between-cost-and | null | null | https://aclanthology.org/2020.law-1.7 | https://aclanthology.org/2020.law-1.7.pdf | Understanding the Tradeoff between Cost and Quality of Expert Annotations for Keyphrase Extraction | Generating expert ground truth annotations of documents can be a very expensive process. However, such annotations are essential for training domain-specific keyphrase extraction models, especially when utilizing data-intensive deep learning models in unique domains such as real-estate. Therefore, it is critical to optimize the manual annotation process to maximize the quality of the annotations while minimizing the cost of manual labor. To address this need, we explore multiple annotation strategies including self-review and peer-review as well as various methods of resolving annotator disagreements. We evaluate these annotation strategies with respect to their cost and on the task of learning keyphrase extraction models applied with an experimental dataset in the real-estate domain. The results demonstrate that different annotation strategies should be considered depending on specific metrics such as precision and recall. | ['Ondrej Linda', 'Kai Liu', 'Saeid Balaneshin', 'Hung Chau'] | null | null | null | null | coling-law-2020-12 | ['keyphrase-extraction'] | ['natural-language-processing'] | [ 1.15151040e-01 3.65123093e-01 -2.69275337e-01 -5.34645855e-01
-9.49372470e-01 -8.53497267e-01 6.95563078e-01 6.82654560e-01
-7.02493727e-01 9.51341987e-01 1.30327553e-01 -1.76152766e-01
-1.67421907e-01 -7.45024323e-01 -3.46744210e-01 -3.51104677e-01
5.00784934e-01 5.96038043e-01 -5.42682558e-02 -3.11981216e-02
3.70442152e-01 3.57797980e-01 -1.20900166e+00 3.55153918e-01
8.85799646e-01 9.18685794e-01 -9.88889858e-02 4.04491842e-01
-3.50031614e-01 7.18959153e-01 -1.05006385e+00 -7.79470682e-01
3.00868332e-01 3.07273567e-02 -8.96495759e-01 2.69837845e-02
3.02805483e-01 -4.45866823e-01 1.96533855e-02 1.05103016e+00
4.21560019e-01 -2.03196555e-01 7.79483259e-01 -1.38795447e+00
-3.55516464e-01 8.91302228e-01 -3.77533674e-01 5.39559834e-02
9.29737017e-02 -2.23189935e-01 1.34720814e+00 -7.53864408e-01
4.06051725e-01 4.70087498e-01 5.27051270e-01 2.24087685e-01
-1.02163565e+00 -7.46254325e-01 8.88966545e-02 1.60422996e-01
-1.42680490e+00 -2.94406176e-01 5.63094854e-01 -5.89200974e-01
7.03982711e-01 4.28146385e-02 3.62121195e-01 9.78125691e-01
1.63855925e-02 8.57003868e-01 8.73781919e-01 -6.26846611e-01
3.08597207e-01 4.52639818e-01 3.29600811e-01 5.59078991e-01
7.88947999e-01 -6.75493300e-01 -7.30459809e-01 -3.07921290e-01
4.38884407e-01 -4.17718023e-01 -2.51408130e-01 -1.18248813e-01
-9.88685489e-01 7.16598809e-01 -1.94754943e-01 3.90980512e-01
-5.73853850e-01 5.99571913e-02 5.08713305e-01 -2.87654670e-03
6.44380510e-01 1.09113121e+00 -8.07439446e-01 -5.23991995e-02
-1.18535233e+00 4.16826427e-01 1.05419934e+00 9.21362162e-01
8.11922908e-01 -3.49969923e-01 -3.45893800e-01 6.44829571e-01
2.22753108e-01 -7.89964199e-02 1.25692159e-01 -8.25762510e-01
7.12089121e-01 9.87957597e-01 5.88067472e-01 -1.04178929e+00
-3.14109176e-01 -4.65111345e-01 -6.98727846e-01 -1.72853336e-01
4.04749364e-01 -5.38168132e-01 -7.96292901e-01 1.28349888e+00
2.48541847e-01 -2.16752887e-01 -1.53939640e-02 7.25005746e-01
8.75083804e-01 3.77877265e-01 4.13934261e-01 -2.40776464e-01
1.46146286e+00 -6.80756152e-01 -1.12662625e+00 -1.15489133e-01
8.70798588e-01 -8.52657557e-01 9.58555460e-01 4.34685588e-01
-7.08653450e-01 -3.13406825e-01 -1.14014518e+00 -8.62549767e-02
-6.11606836e-01 8.72622669e-01 5.27651727e-01 5.62511086e-01
-4.58693236e-01 4.27442610e-01 -7.13487148e-01 -1.42863065e-01
3.32377404e-01 3.22852433e-01 -2.87421316e-01 3.08029115e-01
-1.51975441e+00 9.16507840e-01 6.80611193e-01 8.17594752e-02
-6.79707408e-01 -6.87214434e-01 -4.94922906e-01 5.00451565e-01
6.00601554e-01 -4.92069751e-01 1.61152518e+00 -7.28648901e-01
-1.14630091e+00 7.56524742e-01 1.36851206e-01 -4.82303768e-01
6.03912890e-01 -5.98166108e-01 -1.96828440e-01 4.91418764e-02
2.03937486e-01 3.48704278e-01 5.55671573e-01 -1.15662062e+00
-9.86313820e-01 -6.20654523e-02 4.04417068e-01 2.18781739e-01
-9.06855166e-01 6.00379668e-02 -3.32700431e-01 -5.89656055e-01
-3.34210962e-01 -9.22075033e-01 -1.43696398e-01 -2.41086513e-01
-5.59091389e-01 -4.76513416e-01 8.15921903e-01 -9.35584366e-01
1.39372587e+00 -1.81134820e+00 -3.07857782e-01 3.13809514e-01
2.99261361e-01 4.03003633e-01 1.48445457e-01 4.44714218e-01
3.23724300e-01 4.09556538e-01 2.41687265e-03 -2.35699669e-01
-4.85946015e-02 8.99127685e-03 -2.42341399e-01 -4.34344560e-02
3.43720406e-01 5.21723807e-01 -8.82695496e-01 -6.41009569e-01
-7.23899528e-02 3.10595572e-01 -1.12696916e-01 4.06782836e-01
-5.36975443e-01 8.22073519e-02 -8.28638732e-01 5.78505635e-01
2.11442709e-01 -5.73359430e-01 4.47255135e-01 -5.62388659e-01
8.64331052e-02 4.18634534e-01 -1.32039976e+00 1.32723033e+00
-5.86617291e-01 8.49861264e-01 2.10262425e-02 -9.28725898e-01
8.89416933e-01 8.17584217e-01 6.95208848e-01 -2.62637109e-01
3.53092968e-01 3.27691078e-01 -3.81130546e-01 -4.04882163e-01
1.01851487e+00 3.70736778e-01 -9.15252268e-02 6.96979165e-01
3.23735438e-02 -4.35440466e-02 4.66215193e-01 3.19842309e-01
1.06491971e+00 -3.72382514e-02 5.46037436e-01 -3.04324597e-01
2.77714491e-01 2.72422045e-01 3.89454246e-01 5.58984101e-01
9.58753303e-02 2.90044278e-01 6.22346222e-01 -4.54290032e-01
-1.06145632e+00 -1.57136604e-01 2.17944011e-01 9.17241871e-01
-1.62893862e-01 -7.07440317e-01 -9.48224187e-01 -1.08968329e+00
-2.66000301e-01 7.06494391e-01 -4.33291197e-01 7.20936134e-02
-2.96832770e-01 -8.40703726e-01 5.09940624e-01 5.64505219e-01
6.20145738e-01 -7.67384708e-01 -9.16604102e-01 3.46872628e-01
-4.10674810e-01 -1.48636830e+00 -1.05389915e-01 4.40541685e-01
-3.33123893e-01 -1.28795004e+00 -5.05678475e-01 -2.41756797e-01
7.14838743e-01 6.20320402e-02 1.16691983e+00 1.14389397e-01
4.40585539e-02 4.94168043e-01 -4.41078931e-01 -6.63631141e-01
-3.93802792e-01 7.70905375e-01 -1.71264485e-01 -2.30182111e-02
7.93170512e-01 -2.11516574e-01 -5.00672519e-01 2.34275863e-01
-9.33918536e-01 5.84653802e-02 6.83282375e-01 6.83391035e-01
4.53755409e-01 4.38061267e-01 6.15109682e-01 -9.45791364e-01
1.25601411e+00 -2.94859797e-01 -8.68262887e-01 7.27084398e-01
-9.37856197e-01 1.26979589e-01 4.52205658e-01 -5.26755750e-01
-1.21283472e+00 2.18728572e-01 3.61987114e-01 -3.02311819e-04
1.26663328e-03 8.44272733e-01 -2.53185213e-01 1.55172139e-01
7.89620876e-01 -2.87255734e-01 -5.21413922e-01 -3.77648860e-01
2.85474569e-01 9.77275252e-01 2.75963128e-01 -7.43190765e-01
6.75643146e-01 1.21825486e-01 -2.15058863e-01 -5.68500578e-01
-1.29840744e+00 -4.04436082e-01 -8.02177131e-01 -1.43718958e-01
8.50012481e-01 -9.71915185e-01 -4.06290770e-01 5.80236539e-02
-1.30697739e+00 -8.08714852e-02 -2.14278966e-01 3.60738963e-01
-1.31600633e-01 1.32093474e-01 -1.77441627e-01 -6.58474743e-01
-7.38067687e-01 -1.00711429e+00 1.00578666e+00 4.85412031e-01
-7.77921915e-01 -7.14285493e-01 -1.92945197e-01 5.01871109e-01
1.37516543e-01 1.16097070e-01 9.23991621e-01 -1.17051804e+00
-6.52901232e-01 -5.30003488e-01 -5.20088315e-01 4.96089518e-01
3.30324233e-01 1.90729216e-01 -8.49313974e-01 5.51951081e-02
-4.48317200e-01 -5.04862726e-01 4.10456806e-01 9.14592519e-02
1.15975225e+00 -4.08453196e-01 -3.12887460e-01 -1.15797020e-01
1.23084140e+00 2.09927976e-01 1.52839571e-01 5.52573204e-01
7.13959157e-01 7.33586371e-01 1.02706254e+00 5.83047628e-01
3.54692489e-01 5.62709749e-01 1.44142911e-01 1.07509591e-01
2.29034305e-01 -1.61453247e-01 -1.49829224e-01 3.38067323e-01
-8.55882391e-02 -5.88482618e-01 -1.21244335e+00 7.82867551e-01
-1.95815933e+00 -5.11978924e-01 7.92609751e-02 1.92612839e+00
1.19991398e+00 2.41535217e-01 -1.39856935e-01 4.53972608e-01
5.52422345e-01 6.47409782e-02 -1.79886162e-01 -3.32492664e-02
2.53867269e-01 1.78889155e-01 5.11714041e-01 2.33646154e-01
-1.21790433e+00 8.62493455e-01 5.92197752e+00 6.02714181e-01
-1.01230371e+00 -1.29702473e-02 6.66893363e-01 8.73208195e-02
-2.05258027e-01 1.05912960e-03 -1.21183586e+00 4.47019786e-01
8.27964902e-01 -1.92648128e-01 -1.02207042e-01 9.98921633e-01
-6.43176073e-03 -1.52778462e-01 -1.15563834e+00 6.51119292e-01
-1.57307833e-01 -1.40276802e+00 -4.51487191e-02 4.56735212e-03
6.91329479e-01 -3.92186612e-01 -3.73736560e-01 2.45018616e-01
5.51802278e-01 -8.01638186e-01 5.31317234e-01 2.18251824e-01
4.84981030e-01 -7.04510391e-01 1.00048530e+00 2.81780064e-01
-8.96362126e-01 1.12085015e-01 1.43518960e-02 2.03162298e-01
1.43339977e-01 8.75506759e-01 -1.23208654e+00 3.70434135e-01
6.37610316e-01 2.30279803e-01 -4.75448608e-01 7.13762820e-01
-6.10903561e-01 5.00972867e-01 -3.33433449e-01 -2.32184589e-01
2.01552823e-01 2.64732540e-01 7.34747276e-02 1.31133091e+00
3.06020200e-01 2.24991128e-01 2.50081807e-01 7.23540485e-01
-4.37742442e-01 1.94182619e-01 -2.86187381e-01 -5.77982187e-01
8.71444166e-01 1.58550572e+00 -8.90598834e-01 -4.05070603e-01
-3.26327085e-01 6.03762865e-01 6.18680604e-02 3.41949016e-01
-6.57738090e-01 -4.52890992e-01 4.33477998e-01 4.27528948e-01
5.29847927e-02 -2.91669339e-01 -4.55464125e-01 -8.98428619e-01
1.97642997e-01 -1.00819254e+00 3.53122622e-01 -7.11825788e-01
-9.27228272e-01 5.45856833e-01 2.62865931e-01 -1.10665011e+00
-2.65646756e-01 -4.90035355e-01 -1.87860027e-01 6.39334023e-01
-1.46539140e+00 -1.27255893e+00 -6.55718744e-01 1.46323681e-01
4.27524328e-01 -1.90047711e-01 7.62949526e-01 4.03562278e-01
-4.29673284e-01 5.22671580e-01 -3.25379133e-01 3.47306937e-01
7.80193269e-01 -1.23700643e+00 3.04727942e-01 7.29971170e-01
3.20503175e-01 5.96261799e-01 6.92712843e-01 -6.71791136e-01
-7.51835167e-01 -8.33788455e-01 1.11300814e+00 -3.69617641e-01
7.04100311e-01 -1.43405065e-01 -9.63185370e-01 5.44314623e-01
3.36672425e-01 -2.42164046e-01 1.01370335e+00 3.08918327e-01
-1.78755462e-01 2.18557529e-02 -9.79931653e-01 5.90237558e-01
4.29585695e-01 -6.08313739e-01 -4.31266546e-01 4.18344378e-01
5.92169762e-01 -3.55431139e-01 -9.05323148e-01 4.92758662e-01
5.21685183e-01 -3.21433067e-01 6.23441637e-01 -4.29729521e-01
3.87097538e-01 -3.76855254e-01 6.01118095e-02 -1.22737098e+00
2.73581203e-02 -3.81353050e-01 -3.77209455e-01 1.53608143e+00
6.86942697e-01 -6.43214583e-02 9.45558786e-01 1.21529019e+00
3.29294592e-01 -7.95297861e-01 -5.04504621e-01 -3.38367403e-01
-5.80631077e-01 -2.21083820e-01 6.40529037e-01 1.23971856e+00
-2.07770079e-01 5.10487139e-01 -3.23375970e-01 2.60494024e-01
1.41950369e-01 -2.00293064e-01 9.49522316e-01 -1.61869335e+00
5.19423299e-02 -1.52908012e-01 6.14571497e-02 -4.96345758e-01
2.19520777e-01 -3.91281754e-01 -6.56639934e-02 -1.72940516e+00
8.06777626e-02 -4.57904339e-01 -3.63218904e-01 9.14475262e-01
-2.63038039e-01 -8.24130476e-02 -1.40868470e-01 2.42015883e-01
-6.14601374e-01 1.96971193e-01 7.85767138e-01 -3.13290060e-01
-2.14433789e-01 9.31427479e-02 -9.31365609e-01 7.75562644e-01
8.48856449e-01 -6.35370910e-01 -5.82388937e-01 -4.89790350e-01
7.36855924e-01 -1.87382773e-01 1.40817553e-01 -7.34545112e-01
3.43341678e-01 -3.07387829e-01 2.66117245e-01 -5.34395158e-01
1.06209710e-01 -1.02115023e+00 -1.92495361e-01 6.19104831e-03
-5.49772680e-01 2.39169155e-03 9.21763331e-02 2.96037078e-01
-3.97662610e-01 -5.65893352e-01 4.96152848e-01 -2.06469595e-01
-5.89446068e-01 1.89736895e-02 -3.71491075e-01 -1.08430065e-01
1.02352178e+00 2.02487577e-02 -3.28652054e-01 -4.28321272e-01
-4.67981398e-01 2.33200148e-01 8.17077905e-02 4.37236726e-01
2.17738345e-01 -1.00171399e+00 -6.27409637e-01 -2.08549052e-01
2.59269029e-01 3.07704687e-01 -2.08457664e-01 3.12970996e-01
-4.33570802e-01 4.84336942e-01 -9.54640731e-02 -1.68040752e-01
-1.44496560e+00 7.99923986e-02 1.35799959e-01 -9.76625025e-01
-7.34646022e-02 6.95663273e-01 -3.12409103e-01 -1.10840864e-01
5.34979284e-01 -2.22206205e-01 -5.51538527e-01 5.67427397e-01
3.90028715e-01 2.70293653e-01 3.95005107e-01 -9.77790877e-02
-1.09840713e-01 3.32464129e-02 -4.02134925e-01 -6.12531602e-03
1.17753398e+00 3.91211770e-02 2.50077005e-02 1.71052497e-02
6.07738793e-01 -7.81976879e-02 -9.44035947e-01 -3.11990410e-01
6.33078992e-01 -2.17092767e-01 3.40537041e-01 -9.53076601e-01
-8.68375063e-01 6.95560396e-01 3.91944200e-01 3.51992756e-01
1.05677080e+00 -3.81267518e-01 6.36373639e-01 7.99947739e-01
2.88936049e-01 -1.70841920e+00 2.60191500e-01 2.49424398e-01
6.85086966e-01 -1.34389973e+00 4.51062888e-01 -2.96041876e-01
-5.71429074e-01 9.40648198e-01 8.77107024e-01 3.13321352e-01
5.98558009e-01 3.37864161e-01 1.75567552e-01 -3.63332629e-01
-5.49764574e-01 4.18616943e-02 5.27062118e-01 1.92359224e-01
8.47743452e-01 -2.04133475e-03 -6.61273301e-01 7.28874207e-01
-4.91954237e-02 1.80338830e-01 5.80055237e-01 1.15798950e+00
-1.87643543e-01 -1.20814002e+00 -2.76616570e-02 5.24775922e-01
-8.43306065e-01 3.05594318e-02 -8.83705795e-01 7.35494733e-01
7.11164847e-02 7.90375888e-01 -3.33970517e-01 -2.74434775e-01
1.75038278e-01 2.34908924e-01 1.02312438e-01 -8.81087422e-01
-7.58751690e-01 -1.82127863e-01 5.27146161e-01 -2.89741792e-02
-7.09333181e-01 -2.43159771e-01 -1.01207161e+00 2.14546472e-02
-6.54922962e-01 4.50281769e-01 9.07784283e-01 1.19367003e+00
4.37821984e-01 4.49999481e-01 1.64699331e-01 -4.27038133e-01
-5.45451045e-01 -1.12191021e+00 -1.75432697e-01 2.33031988e-01
-2.63140462e-02 -4.91923958e-01 -3.28005850e-02 2.40168959e-01] | [9.70946979522705, 4.828636646270752] |
ed3061f9-36da-4aaa-abb9-437848b99b8d | assumption-questioning-latent-copying-and | null | null | https://openreview.net/forum?id=r1lM_sA5Fm | https://openreview.net/pdf?id=r1lM_sA5Fm | Assumption Questioning: Latent Copying and Reward Exploitation in Question Generation | Question generation is an important task for improving our ability to process natural language data, with additional challenges over other sequence transformation tasks. Recent approaches use modifications to a Seq2Seq architecture inspired by advances in machine translation, but unlike translation the input and output vocabularies overlap significantly, and there are many different valid questions for each input. Approaches using copy mechanisms and reinforcement learning have shown promising results, but there are ambiguities in the exact implementation that have not yet been investigated. We show that by removing inductive bias from the model and allowing the choice of generation path to become latent, we achieve substantial improvements over implementations biased with both naive and smart heuristics. We perform a human evaluation to confirm these findings. We show that although policy gradient methods may be used to decouple training from the ground truth and optimise directly for quality metrics that have previously been assumed to be good choices, these objectives are poorly aligned with human judgement and the model simply learns to exploit the weaknesses of the reward source. Finally, we show that an adversarial objective learned directly from the ground truth data is not able to generate a useful training signal. | ['Sebastian Riedel', 'Tom Hosking'] | 2018-09-27 | null | null | null | null | ['policy-gradient-methods'] | ['methodology'] | [ 6.13674045e-01 6.43306315e-01 -9.88387540e-02 -3.68393093e-01
-1.18256629e+00 -9.84311461e-01 9.67802584e-01 -2.31080614e-02
-6.40241206e-01 1.25143909e+00 6.16016805e-01 -3.86106461e-01
9.35653225e-02 -6.84409678e-01 -8.45727623e-01 -4.74738330e-01
3.72947991e-01 8.57703567e-01 -9.96587519e-03 -6.20240092e-01
3.82355690e-01 2.69110966e-02 -1.18867719e+00 4.58651304e-01
9.26088929e-01 4.47787732e-01 5.27433567e-02 9.07593191e-01
-2.18780905e-01 8.78872573e-01 -8.46689522e-01 -4.77224678e-01
4.55258995e-01 -8.83257568e-01 -1.22516716e+00 -3.34631443e-01
3.57874066e-01 -2.93057948e-01 1.03349909e-01 9.75772738e-01
6.00937545e-01 1.03665136e-01 5.74661374e-01 -1.13015592e+00
-7.65728295e-01 6.90653980e-01 9.62091833e-02 1.17056683e-01
6.29846513e-01 6.02063298e-01 1.20950568e+00 -2.87573129e-01
9.03026044e-01 1.31073868e+00 3.85040224e-01 8.49546313e-01
-1.44926322e+00 -4.67535496e-01 -1.37772989e-02 -5.38333412e-03
-7.39017069e-01 -5.51739395e-01 3.31845999e-01 -3.32355976e-01
9.16388154e-01 4.09802109e-01 5.17585218e-01 1.44675148e+00
2.95340598e-01 6.69008911e-01 1.54071128e+00 -6.08448684e-01
2.60999352e-01 3.82265419e-01 -4.16250229e-01 5.40124118e-01
-1.63164362e-02 4.53299940e-01 -4.23595041e-01 -2.32408881e-01
6.42901123e-01 -5.51101744e-01 -4.48174208e-01 -4.93254244e-01
-1.50698197e+00 1.01606202e+00 3.88169140e-01 3.57314318e-01
-3.64644229e-01 3.40490758e-01 1.95398673e-01 8.25626791e-01
3.45331244e-02 1.45308661e+00 -5.80280960e-01 -4.26197261e-01
-9.14422452e-01 5.84815085e-01 1.11054432e+00 8.71312499e-01
5.92648625e-01 -7.89703950e-02 -6.62630498e-01 5.10155082e-01
-5.82134798e-02 5.44149816e-01 5.97230256e-01 -1.34892595e+00
5.12036383e-01 3.21809918e-01 5.43191791e-01 -5.67674279e-01
-1.80759043e-01 -3.15241963e-01 -3.25906575e-01 3.28570426e-01
9.12136972e-01 -4.57698971e-01 -7.81526864e-01 1.86321199e+00
-1.58229657e-02 -3.66218179e-01 2.06307054e-01 9.48138535e-01
8.10588375e-02 5.84247589e-01 4.02472652e-02 -1.05634078e-01
1.13663471e+00 -8.98594558e-01 -5.34554422e-01 -5.05136430e-01
6.24151766e-01 -8.16112220e-01 1.29978263e+00 4.09551769e-01
-1.41446149e+00 -3.27146590e-01 -9.81555045e-01 -1.74120396e-01
-2.40415484e-01 -3.24543476e-01 4.54307944e-01 5.62432945e-01
-1.20171082e+00 7.72037208e-01 -4.92942780e-01 -3.55426133e-01
1.03438959e-01 3.55169117e-01 -1.15929350e-01 -9.10066161e-03
-1.53692710e+00 1.33289230e+00 2.75873989e-01 -3.77193093e-02
-7.49811232e-01 -4.04055804e-01 -5.07099509e-01 1.47924037e-03
5.81517100e-01 -9.74717557e-01 1.65193939e+00 -1.43375409e+00
-1.92746961e+00 5.09634078e-01 1.47069646e-02 -5.47420442e-01
7.76994288e-01 -4.82535595e-03 7.64973983e-02 -8.33439678e-02
6.35174811e-02 9.97547686e-01 8.21987152e-01 -1.02074492e+00
-5.92909515e-01 7.36919567e-02 4.17591423e-01 2.68533558e-01
1.08539879e-01 -6.84127063e-02 2.17021540e-01 -5.28156221e-01
-4.32033986e-01 -1.03288519e+00 -3.87609482e-01 -1.37912720e-01
-6.97565675e-02 -6.99709058e-02 -1.37163829e-02 -7.10912704e-01
8.42465580e-01 -1.42533886e+00 2.01711521e-01 1.30474418e-01
-2.43692741e-01 1.98590890e-01 -4.53194171e-01 5.30941546e-01
2.32773706e-01 4.40084875e-01 -3.45088154e-01 1.18952908e-01
2.21253321e-01 2.52166718e-01 -5.25092423e-01 7.49321654e-02
5.14483154e-01 1.30447543e+00 -1.34072602e+00 -2.67371416e-01
-3.78087610e-01 2.44675279e-02 -7.21324623e-01 4.03363228e-01
-8.00562382e-01 4.68055099e-01 -3.57540995e-01 9.48515385e-02
3.70635316e-02 -1.11091219e-01 2.21761227e-01 3.89288038e-01
8.56960490e-02 7.15443194e-01 -9.43672717e-01 1.64624512e+00
-6.05925024e-01 6.04459345e-01 -8.25025886e-02 -7.03970373e-01
7.08940685e-01 4.47497636e-01 -2.06847489e-02 -9.42532659e-01
4.86756638e-02 3.36295605e-01 4.21954006e-01 -5.65965593e-01
6.09978795e-01 -5.23535848e-01 9.11657698e-03 7.20830917e-01
1.02926426e-01 -6.88026249e-01 1.68996498e-01 3.91233824e-02
1.18661785e+00 5.64410329e-01 1.16990089e-01 -2.54729211e-01
2.77609318e-01 4.22268718e-01 4.86809999e-01 1.03953481e+00
-1.00614682e-01 7.63848305e-01 8.65281105e-01 -1.02502279e-01
-1.41031623e+00 -8.65968704e-01 2.75132239e-01 1.19022787e+00
-1.95380613e-01 -3.85406017e-02 -8.85119617e-01 -8.37167740e-01
-2.04509869e-01 1.24374402e+00 -6.37091279e-01 -2.30311543e-01
-7.08067358e-01 -5.52198708e-01 6.06696665e-01 3.97033274e-01
-2.68781614e-02 -1.41130030e+00 -7.70433545e-01 5.17025054e-01
-3.49786133e-01 -9.77702141e-01 -6.51496232e-01 2.07871541e-01
-7.68309474e-01 -7.87962437e-01 -7.27855384e-01 -5.12511611e-01
5.63947022e-01 -3.01543206e-01 1.45879197e+00 5.21692447e-02
6.78718239e-02 2.53953725e-01 -2.19424382e-01 -5.45521855e-01
-1.01265109e+00 3.49332392e-01 -2.86993235e-01 -3.75546932e-01
2.01927677e-01 -3.08167756e-01 -5.44732988e-01 3.55218500e-01
-8.72168243e-01 9.21775699e-02 7.26991951e-01 1.12687206e+00
1.04898453e-01 -5.87823868e-01 8.86973500e-01 -1.06108868e+00
1.13106668e+00 -3.72279704e-01 -5.40657222e-01 1.74573794e-01
-8.43751013e-01 7.85946488e-01 7.66399026e-01 -4.32107776e-01
-9.53371346e-01 -2.19768211e-01 -1.09457575e-01 -2.75206449e-03
-1.55647770e-01 1.57334670e-01 6.09933920e-02 2.28407010e-01
9.52659726e-01 1.41869634e-01 3.22168678e-01 -8.73685926e-02
6.21438146e-01 4.59596813e-01 1.52515307e-01 -8.54074359e-01
7.06639588e-01 1.09819517e-01 -2.91383564e-01 -2.78856486e-01
-7.48588204e-01 -5.82649894e-02 -3.77967983e-01 1.04785167e-01
7.98423171e-01 -6.34052217e-01 -5.12299418e-01 -1.23061046e-01
-1.26063132e+00 -7.19291151e-01 -6.43088162e-01 3.03748667e-01
-9.68063354e-01 3.36731039e-02 -3.28081280e-01 -6.39386177e-01
-1.84102640e-01 -1.17869592e+00 9.29761469e-01 1.20034747e-01
-6.94222033e-01 -9.49047565e-01 2.25577831e-01 4.09294844e-01
7.71916091e-01 3.43337320e-02 1.05779183e+00 -9.28367317e-01
-7.87914276e-01 2.54316274e-02 4.85078879e-02 3.94605219e-01
1.02668032e-01 -1.79557443e-01 -8.00702572e-01 -2.14931756e-01
7.39054307e-02 -5.90670764e-01 5.60164750e-01 -1.11881778e-01
5.01278222e-01 -6.44684911e-01 2.32273024e-02 1.02743573e-01
1.10542011e+00 5.47745004e-02 7.39558220e-01 5.27519166e-01
2.65879810e-01 1.00536942e+00 4.36818600e-01 -1.71532467e-01
3.23758930e-01 6.56714976e-01 1.48628607e-01 9.70139056e-02
-9.75526646e-02 -4.20078516e-01 6.37791693e-01 5.03576279e-01
1.02417350e-01 -3.50894719e-01 -7.74426758e-01 7.46477664e-01
-1.69196427e+00 -1.15502226e+00 7.97181129e-02 2.25924897e+00
1.35349500e+00 4.64735836e-01 1.37381643e-01 -1.94987327e-01
4.09777254e-01 4.35742695e-04 -5.10099053e-01 -8.60874832e-01
8.84966552e-03 5.01691699e-01 6.66061044e-01 8.01192760e-01
-4.34229434e-01 1.00048387e+00 7.12210846e+00 4.33892399e-01
-1.00651371e+00 2.23022792e-02 5.46122551e-01 -3.10476303e-01
-8.10508847e-01 3.83010298e-01 -3.91972780e-01 3.78673345e-01
1.18295562e+00 -2.00831220e-01 8.11977386e-01 3.50810021e-01
3.37752968e-01 -4.37105894e-02 -1.38499367e+00 2.93175220e-01
-2.67570447e-02 -1.06194198e+00 5.03962161e-03 6.86477199e-02
9.01325047e-01 -4.75387499e-02 -1.46099910e-01 5.00334382e-01
8.79511416e-01 -1.32995152e+00 7.62754142e-01 7.93410480e-01
3.66759270e-01 -6.29566491e-01 7.24249363e-01 5.57915807e-01
-2.11574212e-01 -8.67040306e-02 -4.43498433e-01 -1.62427664e-01
7.75751472e-02 1.58548698e-01 -1.27054286e+00 3.98443490e-01
9.46019068e-02 -7.78604895e-02 -4.61909622e-01 7.35113025e-01
-6.52036309e-01 7.70060480e-01 -2.03737810e-01 -4.39978063e-01
4.61942732e-01 -1.14530101e-01 5.70681632e-01 1.02541578e+00
1.87683225e-01 -1.51822478e-01 -9.46112052e-02 1.14343941e+00
-1.55304909e-01 1.04965858e-01 -5.92829704e-01 -2.21198425e-01
9.45558026e-02 9.27018046e-01 -2.55095005e-01 -2.26285294e-01
-1.45924777e-01 9.93465424e-01 3.19245905e-01 4.06947404e-01
-6.72614515e-01 -3.25251997e-01 5.55620790e-01 1.12373076e-01
2.18120143e-01 -1.53609246e-01 -3.41882914e-01 -1.09552050e+00
-5.79064758e-03 -1.45365012e+00 1.09558344e-01 -8.87471855e-01
-1.10576630e+00 5.16222477e-01 -1.51439577e-01 -8.82643402e-01
-9.22116816e-01 -4.51687038e-01 -4.93231982e-01 1.22464967e+00
-1.61461806e+00 -7.33662307e-01 3.02428693e-01 1.12159982e-01
5.09812057e-01 1.57501977e-02 8.00330520e-01 -2.75167793e-01
-4.78358865e-02 5.86597025e-01 1.91001564e-01 4.88845520e-02
9.42231715e-01 -1.54633629e+00 6.96200490e-01 6.82764888e-01
2.56053388e-01 6.14281774e-01 1.04571807e+00 -4.79261547e-01
-1.34622085e+00 -8.45051110e-01 1.10919392e+00 -9.31182563e-01
6.28904402e-01 -4.50518757e-01 -9.03425395e-01 5.70503235e-01
6.45884395e-01 -4.80296761e-01 3.58445227e-01 -5.26699759e-02
-2.12032452e-01 2.44787857e-01 -1.10229230e+00 7.83448696e-01
8.97865415e-01 -4.38577384e-01 -9.24991608e-01 2.40546748e-01
8.73097360e-01 -4.17769492e-01 -6.21469915e-01 9.50374175e-03
4.97479051e-01 -7.22750962e-01 5.76608479e-01 -1.08379710e+00
6.27108872e-01 -1.61775097e-01 1.58571347e-03 -1.59038627e+00
-2.59597838e-01 -8.84339035e-01 2.38032088e-01 1.17274857e+00
8.02704513e-01 -7.14347959e-01 8.00306618e-01 7.71031439e-01
1.83628678e-01 -7.30520248e-01 -8.01182806e-01 -7.74244845e-01
6.10818386e-01 -1.14256348e-02 7.66509712e-01 7.09524632e-01
3.59401777e-02 7.11824298e-01 -3.80068153e-01 -1.75799608e-01
2.78162569e-01 8.12239498e-02 7.18567371e-01 -8.23482513e-01
-5.89158237e-01 -6.83255613e-01 4.78112511e-02 -1.06922734e+00
6.97848853e-03 -9.81520653e-01 4.10370976e-01 -1.64375305e+00
-1.65842846e-01 -4.14554268e-01 -1.14366934e-01 4.04180676e-01
-3.55527341e-01 6.85173925e-03 3.70715559e-01 -4.78044152e-03
-2.79661328e-01 5.18211722e-01 1.45088768e+00 -1.51044711e-01
-1.24416813e-01 -4.63719927e-02 -1.11442268e+00 2.95785338e-01
8.76080096e-01 -5.87743700e-01 -5.12578607e-01 -6.74142838e-01
6.38088763e-01 1.69378713e-01 2.67275929e-01 -5.98142207e-01
8.65806416e-02 -5.01458764e-01 1.80327415e-01 2.57155240e-01
8.10697153e-02 -5.78568876e-01 -1.12031199e-01 4.56363022e-01
-1.04897237e+00 1.64211646e-01 5.61663248e-02 4.95698869e-01
7.58193657e-02 -5.68780065e-01 6.49759471e-01 -4.23534870e-01
-1.61017865e-01 -1.20762289e-01 -4.84955639e-01 5.36921203e-01
6.76775336e-01 3.18999141e-02 -2.09176466e-01 -7.19833076e-01
-4.80269760e-01 4.59521502e-01 5.76550782e-01 6.13416135e-01
1.65658012e-01 -1.20207143e+00 -1.04411983e+00 -1.55833364e-01
-5.27380444e-02 -1.26835257e-01 -4.37797219e-01 5.17113864e-01
-3.65410388e-01 5.82512558e-01 -2.21766964e-01 -2.55273849e-01
-6.74933732e-01 5.99637151e-01 5.29018044e-01 -5.83984077e-01
-1.29244208e-01 5.95167518e-01 -1.13656364e-01 -6.64934039e-01
8.09859391e-03 -5.03369689e-01 -1.83984023e-02 5.53620197e-02
2.98016220e-01 8.99085328e-02 1.25590920e-01 -1.22225128e-01
1.75448526e-02 1.80472478e-01 -9.24640149e-02 -5.51846504e-01
1.11114860e+00 -1.52873844e-02 1.74213260e-01 2.71640986e-01
8.71241689e-01 7.67815784e-02 -1.28485715e+00 -1.75995454e-01
2.80667126e-01 -4.09156531e-01 -3.47406477e-01 -1.27457488e+00
-4.31180418e-01 9.45036590e-01 3.51975530e-01 2.79869348e-01
7.25335777e-01 -3.00321370e-01 8.23759675e-01 5.24508059e-01
3.50292891e-01 -1.19726014e+00 1.47245988e-01 5.27986944e-01
1.09691584e+00 -1.19510901e+00 -3.08258533e-01 1.46169856e-01
-7.81779587e-01 1.03164005e+00 5.49753487e-01 -2.03682445e-02
-1.64020211e-01 6.27758056e-02 3.29373360e-01 1.61895275e-01
-1.17589509e+00 -2.09032014e-01 4.37152013e-02 6.05745137e-01
5.10102093e-01 -8.00557733e-02 -4.58898753e-01 1.86302021e-01
-6.21363103e-01 1.22307636e-01 7.48972476e-01 9.11564827e-01
-3.60036641e-01 -1.66560841e+00 -4.83177245e-01 5.45697331e-01
-5.47243655e-01 -2.02232555e-01 -4.64647442e-01 6.16739392e-01
-9.02577043e-02 9.07024503e-01 -1.35085568e-01 -1.14004351e-01
4.31192398e-01 4.68471855e-01 7.11063027e-01 -8.24712276e-01
-8.75686288e-01 -3.45432490e-01 3.29395741e-01 -4.70375478e-01
-3.17033939e-02 -7.94246674e-01 -1.05175769e+00 -2.40673106e-02
-2.96548367e-01 4.26898181e-01 7.09419131e-01 1.04384029e+00
3.68999690e-01 4.45578247e-01 5.58779180e-01 -4.24102455e-01
-1.28036177e+00 -7.88343906e-01 1.79168329e-01 4.39886630e-01
4.42265183e-01 -1.65268183e-01 -1.90524504e-01 8.46526548e-02] | [11.763516426086426, 9.047136306762695] |
73e25017-799e-4358-9618-d32ab73516ca | sparse-multi-modal-graph-transformer-with | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Nakhli_Sparse_Multi-Modal_Graph_Transformer_With_Shared-Context_Processing_for_Representation_Learning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Nakhli_Sparse_Multi-Modal_Graph_Transformer_With_Shared-Context_Processing_for_Representation_Learning_CVPR_2023_paper.pdf | Sparse Multi-Modal Graph Transformer With Shared-Context Processing for Representation Learning of Giga-Pixel Images | Processing giga-pixel whole slide histopathology images (WSI) is a computationally expensive task. Multiple instance learning (MIL) has become the conventional approach to process WSIs, in which these images are split into smaller patches for further processing. However, MIL-based techniques ignore explicit information about the individual cells within a patch. In this paper, by defining the novel concept of shared-context processing, we designed a multi-modal Graph Transformer that uses the cellular graph within the tissue to provide a single representation for a patient while taking advantage of the hierarchical structure of the tissue, enabling a dynamic focus between cell-level and tissue-level information. We benchmarked the performance of our model against multiple state-of-the-art methods in survival prediction and showed that ours can significantly outperform all of them including hierarchical vision Transformer (ViT). More importantly, we show that our model is strongly robust to missing information to an extent that it can achieve the same performance with as low as 20% of the data. Finally, in two different cancer datasets, we demonstrated that our model was able to stratify the patients into low-risk and high-risk groups while other state-of-the-art methods failed to achieve this goal. We also publish a large dataset of immunohistochemistry (IHC) images containing 1,600 tissue microarray (TMA) cores from 188 patients along with their survival information, making it one of the largest publicly available datasets in this context. | ['Ali Bashashati', 'Blake Gilks', 'Alexander Baras', 'Hossein Farahani', 'Haoyang Mi', 'Puria Azadi Moghadam', 'Ramin Nakhli'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['multiple-instance-learning'] | ['methodology'] | [ 3.85800838e-01 1.55616805e-01 1.39808301e-02 -5.40313050e-02
-1.32768512e+00 -4.86203462e-01 4.31109101e-01 6.35191083e-01
-3.51598024e-01 4.97241616e-01 2.23431081e-01 -3.26918066e-01
-1.97624415e-01 -7.57054806e-01 -4.83937472e-01 -1.33689523e+00
-1.43208563e-01 5.93051076e-01 3.52423012e-01 -1.69908732e-01
-2.17716163e-03 6.60096169e-01 -1.22079086e+00 7.56276369e-01
5.13419509e-01 9.11357820e-01 2.32491776e-01 8.53344798e-01
-1.79230645e-01 6.08189702e-01 -2.03800321e-01 -1.35961562e-01
-1.02505609e-01 -5.39631024e-02 -8.43823731e-01 6.61871359e-02
6.29534721e-01 3.07976365e-01 -3.01221162e-01 7.33024061e-01
4.92519766e-01 -5.86976767e-01 6.85441375e-01 -9.41047072e-01
-1.33373737e-02 4.16819215e-01 -7.17004299e-01 1.10609569e-01
-7.27462247e-02 2.80421883e-01 1.10824394e+00 -7.36416876e-01
9.85855758e-01 6.97993696e-01 8.71393025e-01 3.74380171e-01
-1.59425402e+00 -4.26230490e-01 -2.09576320e-02 1.60866171e-01
-1.45599842e+00 -2.13479415e-01 5.71315110e-01 -4.85637337e-01
9.63558197e-01 5.31851232e-01 8.74355495e-01 7.36483991e-01
7.72229016e-01 5.56828320e-01 1.53626895e+00 -4.38220114e-01
1.11234039e-01 -1.61753193e-01 4.36473787e-01 9.85568166e-01
7.86300600e-02 -2.03958422e-01 -4.24205869e-01 -3.37678194e-01
4.38345402e-01 3.06589484e-01 -2.53127545e-01 -2.74604350e-01
-1.46045911e+00 6.47430241e-01 6.57655954e-01 5.82345545e-01
-1.85856506e-01 1.02651365e-01 3.19687992e-01 1.33891851e-01
5.50404668e-01 1.51193812e-01 -3.91886145e-01 4.02593434e-01
-9.84651506e-01 -1.85666144e-01 6.64593279e-01 2.22792774e-01
7.68040359e-01 -6.47723496e-01 -3.93955499e-01 4.95137662e-01
1.72955394e-01 4.57436964e-02 5.46847403e-01 -5.37917018e-01
-7.45510459e-02 9.74299967e-01 -5.13913512e-01 -8.66700053e-01
-8.96131396e-01 -7.17680335e-01 -1.19519496e+00 1.70169547e-01
6.41798556e-01 2.04620957e-01 -1.17213249e+00 1.51994574e+00
3.57383341e-01 4.18664396e-01 -9.38549638e-02 6.23182058e-01
9.44047809e-01 3.30460370e-01 2.29081884e-01 -3.42768103e-01
1.64936709e+00 -7.17571616e-01 -3.62812191e-01 -3.41416076e-02
1.08194649e+00 -5.73250592e-01 7.60602951e-01 3.33446681e-01
-7.17362940e-01 -1.22384451e-01 -8.26144576e-01 4.19157334e-02
-4.49082911e-01 3.05616725e-02 9.20191348e-01 3.68158221e-01
-1.45956063e+00 5.34031332e-01 -9.06704128e-01 -6.45228744e-01
6.28947556e-01 5.53001821e-01 -8.52107048e-01 -1.73484579e-01
-5.89097679e-01 6.83658183e-01 -1.32696956e-01 -1.20911643e-01
-9.91608620e-01 -1.31354213e+00 -5.46541095e-01 1.90320902e-03
2.33277842e-01 -1.00637555e+00 6.87828183e-01 -5.59499383e-01
-1.09016836e+00 1.34157073e+00 -3.66111994e-01 -3.01127672e-01
3.53479207e-01 5.19230843e-01 -1.16891585e-01 4.12038654e-01
-1.46235600e-01 5.59744298e-01 4.67801273e-01 -1.18762851e+00
-6.74272835e-01 -7.92107940e-01 -2.54740298e-01 -6.10820241e-02
-4.08005267e-01 -3.55430990e-01 -5.57769239e-01 -4.46082264e-01
1.80707633e-01 -1.05252421e+00 -5.00641584e-01 1.53454721e-01
-5.87001681e-01 -6.14770539e-02 6.80918515e-01 -5.93154848e-01
9.99989033e-01 -2.14624405e+00 3.82899255e-01 3.75272334e-01
6.12022996e-01 -1.41612902e-01 -2.06989199e-01 4.33369219e-01
-9.65564921e-02 1.98563039e-01 -1.50034457e-01 -6.27190351e-01
-3.28517616e-01 1.52769133e-01 1.83048978e-01 7.60278761e-01
-7.24762902e-02 1.03299749e+00 -7.98695266e-01 -9.02488291e-01
7.83731788e-02 5.51474273e-01 -3.93764794e-01 7.79731348e-02
8.71347412e-02 4.36845124e-01 -1.04164407e-01 7.37949789e-01
4.01262373e-01 -6.36905909e-01 5.05933106e-01 -4.98155177e-01
1.39509872e-01 -2.50154495e-01 -7.66849518e-01 1.59377253e+00
-3.21297705e-01 4.36271340e-01 3.57108980e-01 -9.10416722e-01
3.67608011e-01 2.54713297e-01 9.88165617e-01 -3.88321668e-01
-1.30033074e-02 1.50185063e-01 6.37220517e-02 -2.69081473e-01
-1.51680604e-01 -3.84853363e-01 -2.46201362e-02 2.96227574e-01
1.48562677e-02 -7.60918260e-02 1.20352328e-01 3.14841479e-01
1.63239157e+00 -4.71488267e-01 3.28333527e-01 -4.50555176e-01
5.62576890e-01 2.48816282e-01 5.10559082e-01 5.91969132e-01
-1.69409916e-01 7.78349996e-01 7.05910146e-01 -4.08396900e-01
-7.27132380e-01 -9.91982877e-01 -1.99485138e-01 8.63560438e-01
-1.46572053e-01 -4.41324681e-01 -5.43878078e-01 -7.95390785e-01
1.13135144e-01 2.03898065e-02 -1.15002334e+00 8.25687572e-02
-3.77780557e-01 -1.44717586e+00 3.83451194e-01 2.68207639e-01
6.91207312e-03 -4.82210726e-01 -2.45349661e-01 1.67560562e-01
-1.42990559e-01 -1.01363921e+00 -3.73547286e-01 2.92501301e-01
-9.51357543e-01 -1.40247977e+00 -5.96086979e-01 -9.76018667e-01
1.09702003e+00 3.87874573e-01 1.14343286e+00 4.87574756e-01
-1.03770542e+00 2.41975203e-01 -9.18377489e-02 -2.71262288e-01
-4.79936808e-01 1.04393229e-01 -5.06017148e-01 2.42591709e-01
6.81604445e-02 -4.97974873e-01 -7.92161226e-01 1.16209038e-01
-8.45535457e-01 3.84335011e-01 9.58481193e-01 1.21211243e+00
1.20836675e+00 1.08376756e-01 2.47021362e-01 -1.21743369e+00
9.91179422e-02 -4.04782206e-01 -2.49338746e-01 4.48543638e-01
-5.30695975e-01 -1.81075215e-01 5.16662300e-01 -1.13532387e-01
-7.48525560e-01 3.21369588e-01 -8.00121799e-02 -3.04971952e-02
-4.36824299e-02 7.27065384e-01 1.28090933e-01 -4.97540534e-01
3.90254438e-01 1.82163045e-01 3.73421609e-01 -2.60534197e-01
-3.13655138e-02 2.66426772e-01 3.00151676e-01 -8.03983286e-02
5.59960425e-01 9.31283534e-01 7.94960320e-01 -7.41127968e-01
-8.79845560e-01 -8.05667400e-01 -7.16248989e-01 -1.38134941e-01
7.95598984e-01 -6.94197714e-01 -7.43985951e-01 6.74919903e-01
-5.27826846e-01 -5.14724910e-01 -8.68633837e-02 7.15460405e-02
-3.75194997e-01 4.03229594e-01 -1.14789212e+00 -2.82659858e-01
-4.82615113e-01 -1.15723610e+00 1.44510674e+00 -6.67658597e-02
5.41828349e-02 -1.25239563e+00 2.96593457e-01 5.02033889e-01
4.58844721e-01 6.84606493e-01 1.32640004e+00 -4.64121878e-01
-4.80057597e-01 -3.59878063e-01 -1.88183725e-01 -2.60353476e-01
2.81073689e-01 2.54443467e-01 -1.04573059e+00 -5.69773376e-01
-4.05737072e-01 -5.14024533e-02 1.32317066e+00 4.68601763e-01
1.35572159e+00 5.20185847e-03 -1.05405092e+00 7.56741166e-01
1.86218262e+00 -4.62080657e-01 6.53777301e-01 1.66745618e-01
7.38176763e-01 6.93895280e-01 3.26043904e-01 8.31843391e-02
4.74203378e-01 6.07908726e-01 5.54508865e-01 -6.53151691e-01
-3.72019529e-01 1.08351290e-01 -4.99343546e-03 5.18951535e-01
-2.85232160e-02 -6.98490161e-03 -1.15057242e+00 6.48689687e-01
-1.74634969e+00 -6.93980634e-01 -2.79314786e-01 2.00440192e+00
8.14092278e-01 -1.89082101e-02 -1.35158692e-02 1.77090347e-01
3.07242066e-01 -4.72685546e-02 -3.05161297e-01 1.79869965e-01
-2.28134185e-01 2.41418764e-01 5.15224934e-01 4.33018416e-01
-1.11152947e+00 6.21824443e-01 6.75100613e+00 9.43806708e-01
-1.24257493e+00 6.83382824e-02 1.17564344e+00 -4.43281159e-02
-3.04310560e-01 -1.54463962e-01 -6.97709084e-01 1.46524042e-01
8.78776848e-01 -5.72950915e-02 -3.89763638e-02 4.11287397e-01
8.04106221e-02 -2.05217823e-01 -1.09365797e+00 7.57781565e-01
-2.44829115e-02 -1.74971533e+00 -3.82344122e-03 5.36909699e-01
5.98497808e-01 2.18513772e-01 -6.67261705e-02 1.36863058e-02
2.63759911e-01 -1.23541856e+00 2.18542628e-02 7.10977793e-01
8.57552230e-01 -4.25685257e-01 1.13918436e+00 2.47486621e-01
-1.06607914e+00 -1.62797309e-02 -1.55789822e-01 2.81035721e-01
-2.05594048e-01 9.33998644e-01 -1.12712848e+00 7.31884480e-01
5.69132030e-01 7.77612686e-01 -1.09892333e+00 9.83759999e-01
4.30904120e-01 5.92271686e-01 -3.24385285e-01 2.57483751e-01
8.73211846e-02 1.70102954e-01 2.10368037e-01 1.41289639e+00
4.16955799e-01 8.86431783e-02 2.39377111e-01 2.08027050e-01
1.08502299e-01 1.76013038e-01 -2.72055835e-01 2.23157004e-01
-2.48445235e-02 1.91547441e+00 -9.61130857e-01 -1.72549173e-01
-4.33860600e-01 6.11557066e-01 5.88547111e-01 1.32774696e-01
-4.62842196e-01 1.32940516e-01 4.92131859e-01 3.82412046e-01
1.18607759e-01 6.38298839e-02 -3.80832613e-01 -9.14763033e-01
-3.51611018e-01 -5.20015717e-01 8.84436965e-01 -3.69227469e-01
-1.61640835e+00 4.38068300e-01 -4.04349804e-01 -1.16622007e+00
2.53877968e-01 -8.67401302e-01 -6.26621962e-01 7.05610931e-01
-1.90991247e+00 -1.68996704e+00 -6.07607007e-01 6.10666573e-01
3.53488326e-02 7.26329088e-02 1.09760749e+00 -6.58479799e-03
-5.90516806e-01 5.66766381e-01 1.06500089e-01 6.15886562e-02
8.02551866e-01 -1.38848662e+00 -3.49171072e-01 3.76111209e-01
-8.48608930e-03 4.06045735e-01 5.48060656e-01 -4.30567831e-01
-1.77806067e+00 -1.14472902e+00 7.51246333e-01 -4.36336905e-01
9.28073287e-01 -3.18686545e-01 -8.66571963e-01 6.08028114e-01
8.14528614e-02 4.34003979e-01 1.34571207e+00 7.96893388e-02
-1.33845836e-01 -4.03857738e-01 -1.20176852e+00 5.70251346e-01
8.74904394e-01 -3.01804334e-01 -7.71487057e-02 5.19763291e-01
3.99218947e-01 -4.10248220e-01 -1.24017239e+00 4.65829402e-01
3.84402454e-01 -9.68546212e-01 9.07794476e-01 -3.00590485e-01
2.55522013e-01 -3.54032397e-01 -6.53256997e-02 -1.29534066e+00
-6.50894105e-01 -1.06531344e-01 4.49239537e-02 1.02619863e+00
5.69444478e-01 -7.82567263e-01 9.67356980e-01 2.47246563e-01
-3.64051908e-01 -1.20096099e+00 -1.14539707e+00 -2.96786606e-01
2.15822592e-01 -1.23448372e-01 3.19996685e-01 6.60362840e-01
3.08843344e-01 4.15769033e-02 1.46800935e-01 2.09896281e-01
8.41323256e-01 3.44495028e-01 5.48438787e-01 -1.35533059e+00
-3.92085314e-01 -5.85557640e-01 -6.34845257e-01 -1.38733864e-01
1.19957201e-01 -1.27873373e+00 -3.14585537e-01 -1.71955216e+00
8.35012853e-01 -4.95141178e-01 -7.52792656e-01 6.99815869e-01
-3.38852644e-01 7.76240110e-01 2.15144712e-03 4.21744853e-01
-5.66934049e-01 -7.31256306e-02 1.40549529e+00 -5.26587129e-01
2.28402346e-01 -3.60139012e-01 -8.99012029e-01 4.50311482e-01
5.73632538e-01 -2.29861498e-01 -2.45269015e-02 -9.92950052e-03
1.72390305e-02 1.94395453e-01 5.45059264e-01 -9.88169551e-01
5.46777368e-01 -1.62863165e-01 6.12404525e-01 -5.43851614e-01
2.87295312e-01 -7.52408445e-01 4.93076921e-01 7.21550763e-01
-1.88611850e-01 -4.09523100e-01 2.11977243e-01 5.83711505e-01
-2.48621434e-01 3.62983018e-01 7.72443950e-01 -1.98788032e-01
-5.05324602e-01 6.21234834e-01 -3.37095201e-01 -4.95153695e-01
1.34415674e+00 -1.80523470e-01 -6.94478273e-01 1.02615178e-01
-9.88017678e-01 3.52459252e-01 6.63679600e-01 -1.54359430e-01
4.67384368e-01 -9.81220841e-01 -9.70012724e-01 9.55673978e-02
4.28387225e-01 4.39206697e-02 7.39658654e-01 1.35301030e+00
-5.58415830e-01 5.09784877e-01 -3.65946963e-02 -1.02703738e+00
-1.64318645e+00 5.27732432e-01 5.69420457e-01 -1.09430909e+00
-6.41237140e-01 7.60303974e-01 6.19104028e-01 -3.11242908e-01
-1.14097856e-01 -2.20111579e-01 -3.47207636e-01 8.81802216e-02
4.89395827e-01 -3.91591340e-02 3.80331278e-01 -5.16640306e-01
-4.98848498e-01 8.55740190e-01 -3.52475971e-01 2.78833032e-01
1.42298079e+00 1.73039008e-02 -5.27718127e-01 4.33937192e-01
1.22335839e+00 -8.42402428e-02 -1.07954991e+00 -1.40418991e-01
-5.98381832e-02 -1.62117779e-01 3.07605177e-01 -8.70695591e-01
-1.23721814e+00 6.51598215e-01 5.27411282e-01 1.45889848e-01
1.30829751e+00 2.36505806e-01 5.54120421e-01 -8.41395855e-02
5.63615799e-01 -8.07991624e-01 3.41523625e-02 1.27864689e-01
5.81653476e-01 -1.24590790e+00 5.04091643e-02 -8.39675307e-01
-3.02391112e-01 1.10851097e+00 3.19401026e-01 1.63055921e-03
8.14746797e-01 5.05875230e-01 2.26744413e-01 -5.64791858e-01
-1.13690376e+00 -1.03822775e-01 2.31306374e-01 4.37820584e-01
4.88172203e-01 3.04822028e-01 -1.01798579e-01 5.22518158e-01
-6.88723400e-02 5.72335646e-02 4.56909090e-01 7.82023549e-01
-2.30349377e-01 -9.93622541e-01 -2.41936043e-01 7.19406784e-01
-6.34067416e-01 -7.39813149e-02 -3.75282258e-01 8.43614340e-01
-6.59960657e-02 6.42779350e-01 -5.10980412e-02 -2.41573274e-01
1.85764641e-01 -6.69049546e-02 4.88856316e-01 -6.30412936e-01
-7.03364491e-01 1.67818308e-01 -1.27717018e-01 -5.80340326e-01
-3.56402397e-01 -7.15745807e-01 -1.31657982e+00 -2.12495387e-01
-5.24236597e-02 -1.01567686e-01 4.61764574e-01 8.30154061e-01
3.87213439e-01 7.40575552e-01 6.17750943e-01 -7.22349226e-01
-1.13917500e-01 -6.85416222e-01 -7.78979421e-01 4.32011306e-01
4.93704230e-01 -2.79501110e-01 -6.31469369e-01 9.40674990e-02] | [15.104144096374512, -3.0024819374084473] |
8220c514-9ba9-4d55-9f65-bdf62962a503 | learning-non-autoregressive-models-from-2 | 2205.14521 | null | https://arxiv.org/abs/2205.14521v1 | https://arxiv.org/pdf/2205.14521v1.pdf | Learning Non-Autoregressive Models from Search for Unsupervised Sentence Summarization | Text summarization aims to generate a short summary for an input text. In this work, we propose a Non-Autoregressive Unsupervised Summarization (NAUS) approach, which does not require parallel data for training. Our NAUS first performs edit-based search towards a heuristically defined score, and generates a summary as pseudo-groundtruth. Then, we train an encoder-only non-autoregressive Transformer based on the search result. We also propose a dynamic programming approach for length-control decoding, which is important for the summarization task. Experiments on two datasets show that NAUS achieves state-of-the-art performance for unsupervised summarization, yet largely improving inference efficiency. Further, our algorithm is able to perform explicit length-transfer summary generation. | ['Lili Mou', 'Chenyang Huang', 'Puyuan Liu'] | 2022-05-28 | learning-non-autoregressive-models-from-1 | https://aclanthology.org/2022.acl-long.545 | https://aclanthology.org/2022.acl-long.545.pdf | acl-2022-5 | ['abstractive-sentence-summarization', 'unsupervised-sentence-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 7.03151584e-01 5.28100908e-01 -2.11784661e-01 -3.83664817e-01
-1.52928817e+00 -4.76695627e-01 5.56929469e-01 4.92457658e-01
-3.57110202e-01 8.06011617e-01 7.65323997e-01 -1.55585125e-01
2.99988419e-01 -7.95187235e-01 -9.58321810e-01 -4.34862375e-01
3.24050725e-01 8.00008178e-01 5.45618944e-02 -2.29464397e-02
5.53120255e-01 3.52787669e-03 -1.14062357e+00 3.93802524e-01
1.42506444e+00 4.62790757e-01 4.52720344e-01 1.24590755e+00
-8.94671828e-02 7.58260906e-01 -8.87698472e-01 -5.30224741e-01
-2.37402752e-01 -1.03992248e+00 -8.86091173e-01 1.00544505e-01
4.42964315e-01 -3.62285763e-01 -2.05598146e-01 8.11872363e-01
5.44663906e-01 3.22048157e-01 8.82022858e-01 -3.67580086e-01
-3.99787426e-01 1.16138518e+00 -2.76066035e-01 -2.45904620e-03
5.08327544e-01 -3.18354964e-02 1.40364110e+00 -6.39237225e-01
5.97758055e-01 1.05781996e+00 4.01528299e-01 5.81433594e-01
-1.28903329e+00 -1.88284427e-01 9.00605991e-02 -1.08620644e-01
-9.41879392e-01 -6.05455697e-01 6.48623407e-01 -8.88370723e-03
1.22147346e+00 5.85584283e-01 6.36424780e-01 1.00906920e+00
3.20077360e-01 1.14559412e+00 3.59161437e-01 -5.99611402e-01
5.16160607e-01 -3.35394233e-01 3.32164317e-01 5.86660087e-01
3.39559078e-01 -5.29747963e-01 -5.81718266e-01 -2.03022391e-01
2.53255218e-01 -4.39288795e-01 -1.03472658e-01 1.55618012e-01
-1.13192856e+00 7.75872767e-01 -1.11324556e-01 9.80430990e-02
-8.21591377e-01 1.81115568e-01 6.79340065e-01 1.00729309e-01
7.46151865e-01 6.74708068e-01 -2.49040410e-01 -4.24338132e-01
-1.53259826e+00 2.56947279e-01 1.12882233e+00 9.60079312e-01
6.11328900e-01 2.78400570e-01 -7.57713199e-01 8.22055697e-01
-9.63983387e-02 4.79471058e-01 6.71316385e-01 -1.01038492e+00
6.08520627e-01 4.09281611e-01 2.58741342e-02 -6.96355462e-01
-1.36118144e-01 -4.39069897e-01 -9.96522784e-01 -4.99471128e-01
-1.12846605e-01 -4.20978069e-01 -7.91758537e-01 1.51995265e+00
-1.02412038e-01 1.80019215e-01 4.01723027e-01 3.90132189e-01
8.64972353e-01 9.83792245e-01 -8.76049474e-02 -7.05759466e-01
1.02818346e+00 -1.33082855e+00 -7.47926772e-01 -3.55556548e-01
7.50022411e-01 -3.44134569e-01 9.05650020e-01 3.89351636e-01
-1.59184217e+00 -2.76691496e-01 -1.02397811e+00 -3.60232979e-01
3.34429473e-01 5.25127590e-01 2.42890090e-01 3.53776008e-01
-1.06528664e+00 7.19523430e-01 -1.08355415e+00 -2.73306429e-01
1.41753033e-01 2.12949648e-01 -7.05421492e-02 3.08676362e-01
-8.19810987e-01 6.83366656e-01 8.61235201e-01 -1.83320269e-01
-5.80060899e-01 -4.94409114e-01 -1.09125113e+00 4.61845160e-01
4.49361056e-01 -1.14404953e+00 1.51350808e+00 -7.83982813e-01
-2.04177141e+00 4.98756170e-01 -7.90601194e-01 -1.04745317e+00
3.68706226e-01 -4.53101128e-01 1.29157335e-01 2.92529821e-01
5.94857298e-02 4.10636306e-01 6.82797968e-01 -1.10694647e+00
-4.65118855e-01 -1.19649813e-01 -2.71419376e-01 4.10509646e-01
-2.50210077e-01 3.11627146e-02 -7.25894272e-01 -7.84190714e-01
-1.85638592e-01 -8.16949129e-01 -5.10965765e-01 -1.00447333e+00
-8.39957237e-01 -4.09716994e-01 1.89910695e-01 -1.16303194e+00
1.87316811e+00 -1.69808805e+00 5.87911010e-01 5.49509749e-03
-6.78960085e-02 4.69340980e-01 -3.09789807e-01 7.51367927e-01
3.11940789e-01 8.59271437e-02 -7.84024000e-01 -8.80927444e-01
-7.16272965e-02 1.00516200e-01 -4.48866338e-01 -1.74535617e-01
3.53015423e-01 1.13661480e+00 -1.07413614e+00 -6.87569857e-01
-1.26895279e-01 -1.02919908e-02 -8.61094654e-01 4.38683152e-01
-6.98091745e-01 2.98251927e-01 -4.58581716e-01 1.04172312e-01
2.77671129e-01 -4.82642651e-02 4.23824608e-01 3.83292288e-02
-2.13267520e-01 6.94968760e-01 -6.70804501e-01 1.98482645e+00
-6.17062092e-01 5.74255526e-01 -4.02456224e-01 -1.05698264e+00
9.08736527e-01 8.08217973e-02 1.65239230e-01 -3.00119311e-01
1.94664270e-01 1.98327258e-01 -3.41766685e-01 -2.34045833e-01
1.27363086e+00 2.68498361e-01 -2.31286496e-01 7.34112024e-01
1.21487089e-01 -4.12373215e-01 6.53142989e-01 5.07225990e-01
1.24411201e+00 1.18464112e-01 5.68174779e-01 -6.60702074e-03
6.17069066e-01 -1.81171217e-03 5.67334116e-01 1.05236053e+00
5.00211298e-01 9.61559296e-01 8.11000824e-01 1.46828175e-01
-1.20168328e+00 -8.03594649e-01 3.93904567e-01 1.02572823e+00
-2.53776699e-01 -8.76233339e-01 -1.22433722e+00 -7.01382995e-01
-4.10231084e-01 1.44354856e+00 -3.01618814e-01 -2.38845557e-01
-8.63403738e-01 -5.81712127e-01 6.25292659e-01 4.98156339e-01
2.96132356e-01 -1.12859833e+00 -6.20673537e-01 5.21245718e-01
-4.79317933e-01 -8.77472520e-01 -7.57483304e-01 5.21537568e-03
-1.06384194e+00 -5.81797302e-01 -7.34624505e-01 -7.26850331e-01
6.44582987e-01 -8.02686960e-02 1.08778155e+00 -1.86048508e-01
3.18789542e-01 2.07958117e-01 -3.49158913e-01 -4.53033745e-01
-9.62791264e-01 7.35680401e-01 -2.69769490e-01 -1.09430179e-01
-2.28252308e-03 -5.54561615e-01 -3.34668189e-01 -3.58885974e-01
-9.01225269e-01 3.09831738e-01 8.60664666e-01 8.34814787e-01
8.17329288e-01 -2.72420347e-01 7.99807906e-01 -1.15716887e+00
1.03317010e+00 -2.87409455e-01 -3.24457377e-01 3.91673654e-01
-4.59250420e-01 4.88782942e-01 9.05340850e-01 -1.74580172e-01
-1.16679955e+00 2.50599254e-02 -3.05678725e-01 -1.11326866e-01
1.11589685e-01 8.55192602e-01 -4.10458408e-02 7.27871120e-01
6.12466097e-01 7.37775326e-01 -1.02809541e-01 -3.52406949e-01
4.71165687e-01 8.81522834e-01 8.04686129e-01 -3.58965933e-01
5.37556350e-01 1.18884966e-01 -2.15807110e-01 -1.08563960e+00
-9.82402682e-01 -4.37067568e-01 -5.48760772e-01 2.24948838e-01
7.24967659e-01 -8.51969421e-01 -2.30986819e-01 2.97195047e-01
-1.47369421e+00 -5.88153124e-01 -4.03870106e-01 3.71203661e-01
-9.48536575e-01 7.91527331e-01 -5.72504580e-01 -8.30577195e-01
-1.23587883e+00 -7.02776492e-01 1.22524261e+00 2.01039389e-01
-5.90885580e-01 -7.77220905e-01 3.19559425e-01 3.13119978e-01
1.48452997e-01 1.66419231e-05 7.34333158e-01 -1.14903104e+00
-4.86529619e-01 -2.32983291e-01 -1.85505394e-03 5.59435308e-01
-1.73292652e-01 7.46237412e-02 -5.41002214e-01 -1.09845281e-01
-1.69582978e-01 -1.62087172e-01 1.25845897e+00 5.64459503e-01
1.19738376e+00 -7.57299542e-01 -1.40647992e-01 5.00515938e-01
1.03960657e+00 -6.78267628e-02 7.29691863e-01 4.92340587e-02
6.38582110e-01 3.47167343e-01 6.67120218e-01 5.83808422e-01
5.20754397e-01 4.46490347e-01 -1.92580651e-02 3.64934474e-01
5.31098945e-03 -5.13930023e-01 5.94660163e-01 1.31895554e+00
-4.16554436e-02 -6.78206623e-01 -6.05057716e-01 6.04730487e-01
-2.06887007e+00 -1.07557487e+00 -2.23098263e-01 2.26353621e+00
1.18089104e+00 1.02549732e-01 1.62173640e-02 -9.32603329e-02
5.77589154e-01 3.86925876e-01 -5.94682992e-01 -8.74862075e-01
3.04925572e-02 4.43939209e-01 4.16292638e-01 7.03617692e-01
-9.23616052e-01 1.28023291e+00 6.03663778e+00 8.83239567e-01
-8.37465644e-01 -1.52642474e-01 4.62546706e-01 -1.58354640e-01
-5.43186128e-01 9.12385155e-03 -6.86013520e-01 4.68019009e-01
1.37888825e+00 -7.07233846e-01 2.67035216e-01 8.00261259e-01
5.00636280e-01 -2.65458077e-01 -9.84032154e-01 6.52962148e-01
3.81833792e-01 -1.43380892e+00 4.40688491e-01 -2.58273929e-01
8.65178049e-01 -1.26805529e-01 -4.74883020e-01 4.56363350e-01
3.24972510e-01 -6.25477493e-01 5.47736704e-01 6.00617528e-01
7.41476715e-01 -1.00004768e+00 6.37668788e-01 6.63356483e-01
-7.67358065e-01 9.87831429e-02 -3.97376299e-01 1.10594042e-01
4.82217878e-01 6.78419054e-01 -8.99713933e-01 8.63485575e-01
-1.19729407e-01 7.53262103e-01 -4.57926720e-01 8.58871639e-01
-6.05424762e-01 1.04567194e+00 -1.79221690e-01 -2.18122706e-01
1.20821394e-01 -3.93000066e-01 9.13209975e-01 1.70516109e+00
3.88863683e-01 2.14264706e-01 1.05823360e-01 6.19789302e-01
-4.23451930e-01 2.91975707e-01 -2.70956725e-01 -2.69424498e-01
3.81271034e-01 1.00518692e+00 -4.37630385e-01 -8.03546488e-01
1.52207404e-01 1.53573489e+00 4.00644869e-01 2.62736738e-01
-6.73028529e-01 -6.41104162e-01 5.41787632e-02 -2.76989967e-01
3.64493430e-01 -2.57802755e-01 -4.04220700e-01 -1.43296409e+00
5.85418120e-02 -8.81577730e-01 2.84023762e-01 -6.16311133e-01
-6.19945109e-01 5.28654158e-01 -7.32189193e-02 -8.21746647e-01
-9.17949677e-01 2.35722512e-01 -1.13784325e+00 6.72483146e-01
-1.24003994e+00 -9.32551742e-01 -4.83747385e-02 -2.27606058e-01
1.12550449e+00 -3.36021706e-02 8.71393442e-01 -2.34482348e-01
-7.23041654e-01 6.39300883e-01 3.15186232e-01 -9.36568528e-02
6.81935549e-01 -1.48797023e+00 9.13476408e-01 1.12270355e+00
9.41180214e-02 6.39621437e-01 1.03101921e+00 -9.14585710e-01
-1.45017767e+00 -1.26491547e+00 1.35452724e+00 -1.85794726e-01
4.66899782e-01 -1.15418091e-01 -9.41757619e-01 7.61468232e-01
5.08513749e-01 -9.70654190e-01 6.88765526e-01 4.98707183e-02
5.85325807e-02 2.10723534e-01 -6.01869106e-01 6.51495636e-01
9.36572075e-01 -2.67279297e-01 -1.03139877e+00 2.98318475e-01
9.78984356e-01 -6.50288045e-01 -5.89938402e-01 1.33460507e-01
4.71418381e-01 -7.77967334e-01 6.09713614e-01 -5.99278748e-01
1.04465961e+00 1.00103721e-01 1.91051260e-01 -1.67453742e+00
-2.88262367e-01 -1.02721345e+00 -5.16829252e-01 1.40673316e+00
5.19364595e-01 -3.60501140e-01 8.41954768e-01 3.12135398e-01
-5.29632866e-01 -7.33855903e-01 -4.41010892e-01 -6.43588066e-01
1.12973724e-03 -2.28297234e-01 3.87775332e-01 2.49281123e-01
7.24548921e-02 7.99480259e-01 -3.95461887e-01 5.18549718e-02
6.03499472e-01 1.61318168e-01 1.03969085e+00 -8.95485103e-01
-4.17444527e-01 -4.92610604e-01 6.98869079e-02 -1.45535803e+00
4.94548589e-01 -8.98750007e-01 2.68939763e-01 -2.11760974e+00
4.33724433e-01 3.14812154e-01 1.49387628e-01 1.78220853e-01
-5.91007233e-01 -1.73077390e-01 1.35511369e-01 6.72750548e-02
-9.23659682e-01 9.66284931e-01 8.74248028e-01 -9.73104686e-02
-5.90852082e-01 2.33288199e-01 -8.19224656e-01 4.86092985e-01
9.84928250e-01 -4.39764500e-01 -2.91220367e-01 -4.06626046e-01
1.70924425e-01 4.09757078e-01 -1.69365570e-01 -8.77407551e-01
3.80380303e-01 9.26316716e-03 -1.05508670e-01 -1.13956201e+00
4.60785776e-02 1.04356781e-02 -1.76739663e-01 4.21039641e-01
-7.60763705e-01 -8.99328366e-02 -9.57217719e-03 6.43183529e-01
-2.18543962e-01 -6.80846989e-01 4.44192559e-01 1.44505352e-02
-6.92439899e-02 5.29283434e-02 -5.35810292e-01 2.45157614e-01
5.18478811e-01 -3.15128677e-02 -7.61213899e-02 -5.36725760e-01
-4.26251084e-01 4.62175220e-01 4.87837166e-01 -3.78066972e-02
4.24398184e-01 -8.08394969e-01 -1.26721692e+00 -1.34393588e-01
-1.09331623e-01 4.00662392e-01 1.66431665e-01 5.12075841e-01
-5.53689718e-01 5.76827347e-01 3.49865645e-01 -3.06278020e-01
-1.20104921e+00 2.36634359e-01 -2.24595904e-01 -8.21007371e-01
-6.60260737e-01 4.71538633e-01 -7.60918260e-02 -2.43097425e-01
-3.55655663e-02 -3.52296084e-01 -1.39237344e-01 2.64329970e-01
6.91170156e-01 5.13899744e-01 1.27931967e-01 -2.82703727e-01
1.48297548e-01 1.68847412e-01 -2.85067856e-01 -3.74278516e-01
1.40569329e+00 -1.67647257e-01 -2.34202310e-01 4.44830865e-01
1.01660180e+00 2.93642372e-01 -1.00257969e+00 -1.49636850e-01
1.48914576e-01 9.38005000e-02 -9.77852270e-02 -6.09241664e-01
-4.45422798e-01 6.04743958e-01 -4.84770358e-01 1.21505514e-01
1.10368419e+00 -1.18644737e-01 1.13138390e+00 8.28774571e-01
-1.22697562e-01 -1.13254821e+00 -6.48939284e-03 9.38461006e-01
9.93520498e-01 -7.48170733e-01 1.59302369e-01 -2.33811453e-01
-8.52480948e-01 1.11784041e+00 1.93626985e-01 -2.13355169e-01
-2.29061976e-01 -3.92594039e-02 -4.46802676e-01 1.25674397e-01
-1.10123074e+00 -1.37934580e-01 3.13678890e-01 2.74956733e-01
4.26914304e-01 4.60572317e-02 -7.55274773e-01 6.88877702e-01
-6.53950512e-01 -4.47508693e-02 8.39902461e-01 6.93953156e-01
-7.30147898e-01 -1.11917269e+00 -1.36378668e-02 7.71347463e-01
-4.95336860e-01 -3.75875711e-01 -5.05672514e-01 1.50832400e-01
-8.19404781e-01 9.69448388e-01 1.63312271e-01 -1.25042126e-01
3.29395145e-01 2.30564386e-01 3.89172822e-01 -9.98100698e-01
-7.45248139e-01 1.56842813e-01 4.38316554e-01 -3.23581547e-01
5.99080930e-03 -7.71237373e-01 -1.38595974e+00 -9.24253464e-02
-2.98781186e-01 4.73871976e-01 7.13826299e-01 8.45874012e-01
6.21900022e-01 7.13080704e-01 7.03315139e-01 -8.82100046e-01
-7.91792333e-01 -1.26860511e+00 -1.03555851e-01 1.22021697e-01
1.51976854e-01 2.91229993e-01 -1.88202411e-01 1.88262746e-01] | [12.465516090393066, 9.431222915649414] |
f433ecb2-c4d5-4cf3-8f75-ae10da610b38 | comma-icon-multilingual-gender-biased-and | null | null | https://aclanthology.org/2021.icon-multigen.1 | https://aclanthology.org/2021.icon-multigen.1.pdf | ComMA@ICON: Multilingual Gender Biased and Communal Language Identification Task at ICON-2021 | This paper presents the findings of the ICON-2021 shared task on Multilingual Gender Biased and Communal Language Identification, which aims to identify aggression, gender bias, and communal bias in data presented in four languages: Meitei, Bangla, Hindi and English. The participants were presented the option of approaching the task as three separate classification tasks or a multi-label classification task or a structured classification task. If approached as three separate classification tasks, the task includes three sub-tasks: aggression identification (sub-task A), gender bias identification (sub-task B), and communal bias identification (sub-task C). For this task, the participating teams were provided with a total dataset of approximately 12,000, with 3,000 comments across each of the four languages, sourced from popular social media sites such as YouTube, Twitter, Facebook and Telegram and the the three labels presented as a single tuple. For the test systems, approximately 1,000 comments were provided in each language for every sub-task. We attracted a total of 54 registrations in the task, out of which 11 teams submitted their test runs. The best system obtained an overall instance-F1 of 0.371 in the multilingual test set (it was simply a combined test set of the instances in each individual language). In the individual sub-tasks, the best micro f1 scores are 0.539, 0.767 and 0.834 respectively for each of the sub-task A, B and C. The best overall, averaged micro f1 is 0.713. The results show that while systems have managed to perform reasonably well in individual sub-tasks, especially gender bias and communal bias tasks, it is substantially more difficult to do a 3-class classification of aggression level and even more difficult to build a system that correctly classifies everything right. It is only in slightly over 1/3 of the instances that most of the systems predicted the correct class across the board, despite the fact that there was a significant overlap across the three sub-tasks. | ['Akanksha Bansal', 'Bornini Lahiri', 'Yogesh Dawer', 'Akash Bhagat', 'Laishram Niranjana Devi', 'Enakshi Nandi', 'Siddharth Singh', 'Shyam Ratan', 'Ritesh Kumar'] | null | null | null | null | icon-2021-12 | ['aggression-identification'] | ['natural-language-processing'] | [-7.03364611e-01 1.27247885e-01 -8.23785141e-02 -6.22272730e-01
-7.92198777e-01 -7.24235475e-01 8.57613981e-01 3.17354977e-01
-8.27203453e-01 9.84004796e-01 2.23047987e-01 -4.67238396e-01
-2.06857055e-01 -2.25059152e-01 6.39172364e-03 -5.58310449e-01
2.36361414e-01 1.07890010e+00 -3.82740125e-02 -5.54195762e-01
6.73304081e-01 2.54775882e-01 -1.59281433e+00 1.91484243e-01
8.05734456e-01 5.14200091e-01 -4.06127721e-01 5.64319789e-01
-1.02010131e-01 7.54871666e-01 -7.87480295e-01 -1.05197990e+00
-7.74388313e-02 -1.34540245e-01 -1.28311002e+00 -2.89949566e-01
6.63920641e-01 -2.33790979e-01 1.09818295e-01 9.10988152e-01
7.65115201e-01 -8.16811770e-02 8.49269211e-01 -1.61104989e+00
-1.67771339e-01 6.31376624e-01 -7.79017508e-01 1.99526712e-01
6.42548025e-01 6.08486384e-02 1.19726360e+00 -8.70357871e-01
5.61002374e-01 1.66712928e+00 7.34577119e-01 8.00402522e-01
-1.30457115e+00 -1.27722895e+00 -3.38534750e-02 -1.16665117e-01
-1.27193117e+00 -5.43213487e-01 1.03062227e-01 -9.79813695e-01
7.73010254e-01 4.29862857e-01 4.04794395e-01 1.39423931e+00
-5.67611158e-02 3.22939634e-01 1.65982270e+00 1.00059502e-01
-3.52574624e-02 9.31669414e-01 5.92668056e-01 2.23338634e-01
3.16840947e-01 -5.03819399e-02 -6.21141016e-01 -5.37037969e-01
1.19199663e-01 -5.53585947e-01 4.59281892e-01 4.18123156e-01
-1.00512314e+00 8.79177272e-01 -9.67343450e-02 4.89165723e-01
1.10845849e-01 -3.36435199e-01 1.04687405e+00 4.97712106e-01
6.51253641e-01 4.50725526e-01 -4.12494957e-01 -5.32715201e-01
-9.47745025e-01 6.58192337e-01 8.91317248e-01 5.24803996e-01
6.08340442e-01 -2.52247807e-02 -1.31161496e-01 1.48411298e+00
1.46610662e-01 6.40156448e-01 6.82822406e-01 -7.96655118e-01
7.27683842e-01 6.27037942e-01 1.74191281e-01 -1.08533549e+00
-9.67472970e-01 -2.38471225e-01 -5.21873713e-01 3.37019451e-02
9.38474953e-01 -3.48797679e-01 -3.39081317e-01 1.91770303e+00
2.42563650e-01 -7.74917364e-01 -2.63898134e-01 7.64108419e-01
8.99720371e-01 3.23503703e-01 4.09880817e-01 -1.39031738e-01
1.43819737e+00 -5.42365789e-01 -6.27675056e-01 -2.38079339e-01
1.08439326e+00 -8.47634554e-01 1.03233695e+00 5.74333549e-01
-1.15812325e+00 -5.45896471e-01 -7.52194762e-01 1.26230836e-01
-7.51426876e-01 -1.21836416e-01 3.89954150e-01 1.12476265e+00
-1.10186684e+00 3.00451871e-02 -6.44553676e-02 -7.00239718e-01
-2.71646939e-02 6.14304066e-01 -6.65880084e-01 2.10048527e-01
-1.09296489e+00 1.10992157e+00 -1.00306630e-01 -2.80510753e-01
-6.27846420e-01 -4.20829237e-01 -6.09475136e-01 -4.59241778e-01
-1.91520959e-01 -1.01750165e-01 1.07328558e+00 -9.97301817e-01
-8.77422273e-01 1.68574476e+00 -1.59354389e-01 2.57331908e-01
6.74317122e-01 1.73925146e-01 -5.65564334e-01 -5.75982273e-01
7.65906334e-01 7.06334829e-01 4.17704105e-01 -1.24583018e+00
-1.19250834e+00 -7.80750096e-01 1.00209258e-01 1.68743253e-01
-3.70755434e-01 9.37007189e-01 2.05146149e-01 -4.41662252e-01
-1.17068611e-01 -9.94363785e-01 2.67809540e-01 -8.32340479e-01
-5.15437424e-01 -6.85300469e-01 6.26833797e-01 -9.65157390e-01
1.45240748e+00 -2.10994482e+00 2.38236248e-01 2.15216562e-01
4.11321670e-01 -8.47182274e-02 3.25250179e-01 4.86891180e-01
-4.17936236e-01 5.78816593e-01 2.57196054e-02 -5.45568764e-01
2.64509946e-01 4.47008610e-02 1.15500011e-01 6.36901140e-01
-3.52928609e-01 1.96104184e-01 -7.64397740e-01 -3.02190900e-01
-3.42157304e-01 -1.16242573e-01 -5.04971087e-01 3.15622725e-02
6.45309806e-01 6.72499895e-01 2.13408887e-01 5.88250637e-01
6.56265557e-01 5.75319946e-01 4.33647260e-02 4.68762666e-01
-4.29798633e-01 3.87084752e-01 -1.13098025e+00 8.27392161e-01
-5.26172340e-01 8.02053332e-01 5.48937440e-01 -7.67675579e-01
1.14644647e+00 3.30878466e-01 4.10771012e-01 -8.37711394e-01
1.15279995e-01 1.01533782e+00 4.57576215e-01 -4.17425156e-01
6.77767277e-01 -3.48936468e-01 -7.54703403e-01 7.58014739e-01
5.06672189e-02 -2.32184101e-02 6.23317540e-01 2.44756013e-01
7.26481855e-01 -6.04319990e-01 4.09493819e-02 -6.28854454e-01
7.36555219e-01 -2.00666070e-01 3.72945756e-01 6.94381773e-01
-5.98500371e-01 3.59182835e-01 9.69227731e-01 -5.50588131e-01
-9.36681926e-01 -7.11771905e-01 -4.35763121e-01 1.75502253e+00
-2.67973781e-01 -5.12693465e-01 -6.53540611e-01 -8.96533072e-01
1.33125901e-01 7.05875397e-01 -5.96126974e-01 -3.96740809e-02
-3.20257485e-01 -9.82988060e-01 1.03375566e+00 1.89178120e-02
4.00893956e-01 -8.74869347e-01 4.53275219e-02 -9.40041319e-02
-4.72641408e-01 -8.69524181e-01 -2.50509679e-01 2.02802807e-01
-3.55846733e-01 -9.72874582e-01 -5.81514597e-01 -6.31158531e-01
3.86198670e-01 -2.59073138e-01 1.01899207e+00 1.99815482e-01
-2.18703076e-02 1.97437093e-01 -1.01476073e-01 -3.96404892e-01
-5.34530044e-01 3.34117383e-01 4.87576485e-01 3.00758926e-04
6.51292443e-01 -3.42737079e-01 2.16691215e-02 7.07948864e-01
-2.41896272e-01 -2.95954794e-01 -5.62011302e-02 7.89447725e-01
-4.58035111e-01 -2.85986602e-01 6.84073031e-01 -9.62165892e-01
9.61418569e-01 -8.89672935e-01 -1.03213705e-01 -4.10699278e-01
-5.38909674e-01 -7.16061354e-01 3.62399280e-01 -2.44572416e-01
-7.62233555e-01 -5.03894269e-01 -3.76318365e-01 4.62035507e-01
-4.83823627e-01 3.57425719e-01 2.72029847e-01 1.66498125e-01
7.88345575e-01 -2.99158275e-01 6.09831214e-02 -4.61682141e-01
-3.53457987e-01 1.40053403e+00 1.88970000e-01 -8.43148887e-01
7.00241745e-01 -1.43913537e-01 -3.82958382e-01 -7.46326506e-01
-6.72142088e-01 -5.94370902e-01 -7.10575700e-01 -5.52467525e-01
9.62313771e-01 -1.02463603e+00 -1.02291858e+00 7.76395023e-01
-8.53107452e-01 -3.57430190e-01 2.59625286e-01 4.43757951e-01
-3.13816577e-01 -5.55285551e-02 -5.45431316e-01 -1.10280263e+00
-5.41080646e-02 -1.28238893e+00 1.06898570e+00 -3.47744264e-02
-1.04990864e+00 -1.08663559e+00 1.14290267e-01 7.52922177e-01
2.97325820e-01 -6.80711269e-02 1.05076444e+00 -1.41863120e+00
7.80060410e-01 -2.82136202e-01 -3.46826822e-01 2.16764376e-01
6.88577220e-02 1.12932995e-01 -1.03267336e+00 -5.02694905e-01
-2.86487639e-01 -9.04756784e-01 4.03764486e-01 -1.27461538e-01
8.72940898e-01 -2.20031068e-01 4.72711818e-03 1.51938960e-01
7.97762632e-01 2.71408856e-01 2.81239688e-01 5.24715602e-01
7.41405427e-01 1.12822092e+00 6.52344763e-01 4.24940199e-01
8.31347108e-01 1.00669849e+00 9.95134637e-02 3.73063190e-03
1.79483607e-01 9.93820205e-02 5.05685091e-01 7.35195994e-01
-2.57387906e-01 3.19528580e-01 -1.55095315e+00 5.80283344e-01
-1.66167819e+00 -9.59799469e-01 -5.76791584e-01 2.34244394e+00
7.55486190e-01 1.61688805e-01 1.02668130e+00 4.64656591e-01
8.24288249e-01 -7.64188617e-02 -1.84911534e-01 -1.29872108e+00
-5.51399328e-02 -1.46931142e-01 3.35724741e-01 8.43145609e-01
-9.76417303e-01 7.16726482e-01 6.62676811e+00 7.89049208e-01
-1.09232628e+00 5.78454323e-02 1.04561365e+00 -1.88890159e-01
1.48605332e-01 -3.43117952e-01 -1.17928350e+00 8.23750675e-01
1.20530915e+00 -1.72506958e-01 3.72833669e-01 5.55269599e-01
1.27066448e-01 -4.68102396e-01 -9.94032919e-01 1.16829371e+00
9.72537091e-05 -2.64030248e-01 -8.49947259e-02 4.94063228e-01
5.27583361e-01 -7.95820877e-02 1.13720506e-01 7.58565366e-01
3.10190976e-01 -1.26518571e+00 1.05113721e+00 1.35966642e-02
9.91812825e-01 -8.59122634e-01 1.06872106e+00 6.87706232e-01
-4.77057457e-01 -6.54021561e-01 -9.78911296e-02 -6.15207314e-01
-3.06096345e-01 1.60723418e-01 -4.54640985e-01 3.24540228e-01
1.08271027e+00 4.94605809e-01 -8.36158991e-01 6.51045859e-01
1.63943321e-01 7.16421604e-01 -1.41545311e-01 -6.75381795e-02
3.02675247e-01 -1.39958248e-01 5.54278612e-01 1.45962882e+00
5.10767996e-02 -3.95464510e-01 8.25514346e-02 1.84406802e-01
1.49624348e-01 3.47693712e-01 -5.73818445e-01 1.91237167e-01
3.98275673e-01 1.53114450e+00 -5.40140331e-01 -1.86220646e-01
-2.76736379e-01 4.13023114e-01 4.89820600e-01 2.00781003e-01
-6.72650456e-01 -3.02972347e-01 7.06783056e-01 3.47906590e-01
-5.67296028e-01 9.93501171e-02 -6.38977885e-01 -7.19156504e-01
-1.97259858e-01 -1.35951579e+00 6.58615530e-01 -3.26547533e-01
-1.48836362e+00 2.16512457e-01 2.15385377e-01 -8.37545693e-01
-5.09779871e-01 -7.67259598e-01 -2.94966310e-01 1.19057214e+00
-6.45247340e-01 -1.06757724e+00 -2.28333980e-01 5.30605733e-01
3.15909177e-01 -6.25473797e-01 8.46775413e-01 8.31838369e-01
-9.27705407e-01 8.40937197e-01 -2.28946105e-01 1.31760135e-01
1.38656628e+00 -1.44950318e+00 -2.44402647e-01 7.67565519e-02
-3.71151954e-01 5.69705009e-01 7.42426097e-01 -4.92053300e-01
-4.90151197e-01 -6.01041019e-01 1.51449680e+00 -6.78177595e-01
7.80826032e-01 -7.08229363e-01 -4.02047873e-01 6.94369376e-01
1.13608286e-01 -5.03986895e-01 1.04035950e+00 6.13802791e-01
-4.02588397e-01 -1.16274454e-01 -1.31829321e+00 4.15550470e-01
1.05351520e+00 -6.59833848e-01 -3.38933259e-01 4.48235095e-01
-1.93118021e-01 -4.45733309e-01 -8.28102291e-01 -8.41782335e-03
7.21400499e-01 -1.17585599e+00 6.65622711e-01 -8.15901577e-01
8.38407695e-01 2.00171471e-01 -2.38163799e-01 -1.43118060e+00
-3.38451535e-01 -3.28187734e-01 6.60320044e-01 1.74710250e+00
6.85621917e-01 -8.82994831e-01 3.64808112e-01 8.11777949e-01
5.67558631e-02 -5.28987706e-01 -1.27616560e+00 -5.38088262e-01
7.77861714e-01 -3.37811589e-01 3.36587816e-01 1.33073306e+00
2.82797813e-01 5.53854227e-01 -2.54719913e-01 -3.68334502e-01
2.47241989e-01 -6.56244338e-01 1.14030111e+00 -1.50198984e+00
1.50642172e-01 -8.31684351e-01 -4.76967007e-01 -2.09399775e-01
4.64554995e-01 -1.13343942e+00 -4.98536587e-01 -1.23034537e+00
4.37961668e-01 -6.94791615e-01 6.82320297e-02 5.02192140e-01
-1.32465214e-01 5.60056448e-01 5.87217286e-02 3.02161247e-01
-3.02662849e-01 -5.27097024e-02 7.40555525e-01 -6.84906915e-02
-4.38764915e-02 2.83444207e-02 -1.18786609e+00 6.93278193e-01
7.93531954e-01 -5.84017932e-01 -1.42221525e-01 -2.34144926e-01
4.44686830e-01 -1.57821104e-01 2.71365881e-01 -9.38244045e-01
5.37306331e-02 -5.10333031e-02 3.56408536e-01 -3.56695980e-01
3.06638896e-01 -4.78378356e-01 -7.05088153e-02 2.92003959e-01
-2.35511154e-01 2.35993236e-01 2.66052216e-01 -2.65692055e-01
-2.63396174e-01 -2.86778808e-01 6.99249327e-01 1.58016826e-03
-1.50299609e-01 -1.21273711e-01 -7.78331399e-01 4.43619341e-01
1.08249974e+00 -2.07658425e-01 -3.97816032e-01 -5.30983508e-01
-8.55132163e-01 3.62271339e-01 2.56471962e-01 5.75365007e-01
-1.08767435e-01 -1.22279596e+00 -1.04018462e+00 1.30414099e-01
1.45388216e-01 -6.57770813e-01 -1.53327649e-02 1.14337635e+00
-1.12342507e-01 4.06803071e-01 -5.55658817e-01 -5.61649621e-01
-1.59639251e+00 -6.98430985e-02 5.18417597e-01 -3.50274861e-01
3.12718034e-01 7.49416471e-01 1.19258411e-01 -1.19459283e+00
1.16279475e-01 5.49544215e-01 -6.67665541e-01 9.27649200e-01
4.08008873e-01 7.78567314e-01 1.41857639e-01 -1.48818028e+00
-4.38582659e-01 5.30211210e-01 -2.24167883e-01 -5.30027926e-01
1.01808035e+00 -1.71427950e-01 -6.89954460e-01 8.79591346e-01
1.14820623e+00 3.78069699e-01 -1.21924892e-01 4.62141871e-01
3.11188459e-01 -5.02645612e-01 -3.48007202e-01 -1.01264906e+00
-6.74725354e-01 7.61408210e-01 2.72424430e-01 6.59581184e-01
5.13180912e-01 -1.04102075e-01 9.75028276e-02 -1.39623578e-03
3.60913932e-01 -1.48664069e+00 -8.70125070e-02 9.57777739e-01
1.00948632e+00 -1.12626636e+00 -5.23408316e-02 -1.91710889e-01
-6.04941010e-01 9.38781142e-01 1.06695175e+00 2.71058112e-01
4.01513189e-01 -1.52493091e-02 3.25426668e-01 -3.43673497e-01
-5.93435168e-01 1.79545134e-01 1.08096145e-01 5.38457334e-01
9.96618569e-01 2.11831897e-01 -8.81722391e-01 9.10140514e-01
-8.89367163e-01 -6.23069584e-01 4.87043440e-01 4.40785885e-01
-1.36703849e-01 -1.08393109e+00 -6.98881924e-01 7.25880027e-01
-7.69350708e-01 2.89821029e-01 -9.75127459e-01 1.09511805e+00
2.89623022e-01 1.18742716e+00 3.73732150e-01 -7.90451705e-01
4.52000737e-01 3.72459859e-01 -3.09528708e-02 -4.74731594e-01
-1.38173008e+00 -4.00865138e-01 8.27574670e-01 -3.70092243e-01
-1.70220837e-01 -9.96350765e-01 -8.33381176e-01 -9.35086548e-01
1.27457261e-01 2.61406362e-01 7.34038472e-01 8.03265512e-01
-3.19950759e-01 -6.14722148e-02 7.41185129e-01 -8.41694415e-01
-4.25032437e-01 -1.35851300e+00 -8.97196949e-01 7.38510251e-01
1.77801087e-01 -9.66183245e-01 -7.77561903e-01 -5.98220825e-01] | [9.093523025512695, 10.490554809570312] |
1665f9db-2867-4f05-8956-1756f03aaebb | performance-of-humans-in-iris-recognition-the | 1807.05245 | null | http://arxiv.org/abs/1807.05245v2 | http://arxiv.org/pdf/1807.05245v2.pdf | Performance of Humans in Iris Recognition: The Impact of Iris Condition and Annotation-driven Verification | This paper advances the state of the art in human examination of iris images
by (1) assessing the impact of different iris conditions in identity
verification, and (2) introducing an annotation step that improves the accuracy
of people's decisions. In a first experimental session, 114 subjects were asked
to decide if pairs of iris images depict the same eye (genuine pairs) or two
distinct eyes (impostor pairs). The image pairs sampled six conditions: (1)
easy for algorithms to classify, (2) difficult for algorithms to classify, (3)
large difference in pupil dilation, (4) disease-affected eyes, (5) identical
twins, and (6) post-mortem samples. In a second session, 85 of the 114 subjects
were asked to annotate matching and non-matching regions that supported their
decisions. Subjects were allowed to change their initial classification as a
result of the annotation process. Results suggest that: (a) people improve
their identity verification accuracy when asked to annotate matching and
non-matching regions between the pair of images, (b) images depicting the same
eye with large difference in pupil dilation were the most challenging to
subjects, but benefited well from the annotation-driven classification, (c)
humans performed better than iris recognition algorithms when verifying genuine
pairs of post-mortem and disease-affected eyes (i.e., samples showing
deformations that go beyond the distortions of a healthy iris due to pupil
dilation), and (d) annotation does not improve accuracy of analyzing images
from identical twins, which remain confusing for people. | ['Kevin W. Bowyer', 'Adam Czajka', 'Mateusz Trokielewicz', 'Daniel Moreira', 'Patrick J. Flynn'] | 2018-07-13 | null | null | null | null | ['pupil-dilation'] | ['computer-vision'] | [ 3.90636384e-01 4.21821140e-02 1.59269303e-01 -4.37083036e-01
-4.58767772e-01 -8.16448092e-01 4.39694494e-01 1.71818241e-01
-5.08773923e-01 7.14374185e-01 8.45642935e-04 -4.84945536e-01
-2.01639950e-01 -2.41823226e-01 -3.74562413e-01 -7.49214172e-01
5.92670739e-02 4.99349594e-01 -3.97257134e-02 2.56447762e-01
4.94040161e-01 6.32359922e-01 -1.90543270e+00 1.56370759e-01
9.33955908e-01 5.89022756e-01 -4.36919779e-01 7.01890886e-01
2.19604731e-01 2.97296643e-01 -6.10238135e-01 -5.42207897e-01
7.23309338e-01 -9.00733888e-01 -9.18268561e-01 3.43075573e-01
1.22122455e+00 -4.24251199e-01 2.42845640e-01 1.20539820e+00
6.03617668e-01 -1.22901894e-01 4.35644627e-01 -9.40426707e-01
-5.08292794e-01 -1.99543089e-02 -7.12259769e-01 1.19049482e-01
6.33961141e-01 7.55851328e-01 4.00608242e-01 -4.14272010e-01
7.14021921e-01 8.67594004e-01 6.56867445e-01 7.96021938e-01
-1.43308675e+00 -7.09919214e-01 -3.66943538e-01 -1.74193159e-01
-1.61717749e+00 -7.80682147e-01 2.30418950e-01 -7.79899240e-01
6.43427372e-01 5.66534936e-01 8.57994199e-01 4.21106935e-01
-7.36506060e-02 -5.23451082e-02 1.61247301e+00 -4.08754945e-01
-1.28045931e-01 4.83973861e-01 5.44061475e-02 4.40008879e-01
3.12767833e-01 6.27307773e-01 -1.30887032e-01 -3.49745125e-01
6.18980885e-01 -4.26305532e-01 -4.48818386e-01 -2.00588658e-01
-1.34199476e+00 3.90135795e-01 1.84410870e-01 4.06680167e-01
-2.88019300e-01 -7.09943295e-01 2.29759991e-01 4.78182822e-01
-3.24873589e-02 9.81076896e-01 -2.20843628e-02 -3.42955217e-02
-9.59327340e-01 1.08824484e-01 3.93659681e-01 4.15863723e-01
5.22412300e-01 -4.26860392e-01 -2.15013951e-01 5.69476247e-01
2.08031740e-02 2.94100463e-01 7.19430923e-01 -6.88435495e-01
7.87280127e-02 8.70465338e-01 5.08519948e-01 -6.73352063e-01
-3.38855326e-01 -2.33885616e-01 -6.81892514e-01 7.22814679e-01
1.13984418e+00 -1.25691086e-01 -1.00252426e+00 1.48245215e+00
3.22348684e-01 -1.24681652e-01 2.64539476e-02 1.27765131e+00
7.24048555e-01 -1.13683291e-01 -2.23961901e-02 -3.46291840e-01
1.71519446e+00 -4.59924757e-01 -5.74907720e-01 6.05221577e-02
6.72762394e-01 -1.32645404e+00 8.99205744e-01 2.40157560e-01
-1.25841367e+00 -7.92367876e-01 -8.84572625e-01 1.19153388e-01
-2.21285358e-01 2.42888778e-01 3.97703588e-01 9.98326898e-01
-1.24655342e+00 4.90751982e-01 -3.53381068e-01 -6.36936486e-01
2.70374715e-01 6.28131151e-01 -7.34978378e-01 2.15380549e-01
-8.41743529e-01 1.02007008e+00 2.01727062e-01 3.19465250e-02
-2.25928910e-02 -5.48863113e-01 -7.88492858e-01 -2.77170271e-01
-1.13892876e-01 -6.30114913e-01 7.76779354e-01 -1.45193756e+00
-1.13756752e+00 1.81394506e+00 -5.50742269e-01 -2.49489501e-01
6.95953548e-01 2.47214437e-01 -6.68739796e-01 3.34276468e-01
1.12078130e-01 6.33334398e-01 8.02914858e-01 -1.09457135e+00
-7.44934678e-01 -6.97088063e-01 -9.94250849e-02 1.08082034e-01
3.26786458e-01 5.69859028e-01 -5.99791855e-02 -4.06647414e-01
1.88094199e-01 -9.44993675e-01 2.44274065e-01 2.33517393e-01
-4.86780941e-01 -1.60067677e-01 4.93083388e-01 -8.21829140e-01
1.01489878e+00 -2.33959007e+00 -5.28886020e-01 3.33074212e-01
4.38981831e-01 8.24021637e-01 -9.95561406e-02 2.80355420e-02
-5.08053958e-01 4.95848358e-01 8.95899311e-02 -7.28698671e-02
-2.16747880e-01 -8.86327326e-02 7.70007148e-02 7.77408123e-01
1.20518185e-01 6.18374765e-01 -6.94125056e-01 -5.11275113e-01
1.40018925e-01 2.32815385e-01 6.19736426e-02 7.67107159e-02
4.96184677e-01 7.44596064e-01 1.47865325e-01 6.61720693e-01
8.02157640e-01 -1.94554210e-01 -8.84813294e-02 -2.89157063e-01
-1.47785828e-01 8.80946741e-02 -1.27821231e+00 7.99700499e-01
1.42577603e-01 8.69774401e-01 1.74544919e-02 -5.66535473e-01
7.80317008e-01 6.51202738e-01 3.26034613e-02 -5.37100792e-01
7.99338371e-02 3.66590500e-01 6.82583511e-01 -6.65939152e-01
3.50115895e-01 -3.18854511e-01 7.01054633e-01 9.41063464e-01
-4.54837352e-01 -3.86140868e-02 2.63823539e-01 -2.54937023e-01
5.00155210e-01 -1.85956702e-01 4.58229870e-01 -9.72349569e-02
6.68678701e-01 -2.18494341e-01 4.84123915e-01 6.73168838e-01
-7.13392317e-01 8.94944668e-01 5.83782911e-01 -8.85681927e-01
-1.08079815e+00 -8.75466228e-01 -6.34134114e-01 1.96875721e-01
1.54345676e-01 8.07239711e-02 -7.92993605e-01 -5.53173006e-01
8.93683136e-02 4.34915513e-01 -7.16471851e-01 2.02192396e-01
-7.64328316e-02 -6.75339699e-01 6.03395462e-01 1.19417295e-01
4.36522752e-01 -4.83613640e-01 -7.35953689e-01 -4.18519229e-01
-2.47374192e-01 -7.12241709e-01 -8.43497336e-01 -5.39053559e-01
-6.71882451e-01 -1.63094842e+00 -8.23581040e-01 -6.25493109e-01
1.45226777e+00 1.58263966e-01 7.58868992e-01 6.96987867e-01
-6.11771226e-01 2.10981190e-01 -8.76854956e-02 -2.36303896e-01
-5.50594211e-01 -6.95505083e-01 4.22519714e-01 4.33107764e-01
6.53614402e-01 -6.31938875e-02 -5.81659019e-01 9.56963599e-01
-6.21102750e-01 -2.41875842e-01 4.24354702e-01 8.95466864e-01
4.82769996e-01 1.30238205e-01 5.85356615e-02 -6.88843846e-01
4.73993540e-01 2.00844660e-01 -5.88947058e-01 4.85174805e-01
-9.25845742e-01 -3.44721198e-01 1.35832518e-01 -7.19640672e-01
-6.69192433e-01 2.47655567e-02 3.90316099e-01 -2.24201366e-01
-6.59690917e-01 -7.53471702e-02 4.02892113e-01 -6.46094143e-01
9.63301778e-01 2.36841477e-02 5.04856408e-01 -1.83803663e-01
-3.47089618e-01 1.05417776e+00 7.18527853e-01 -3.68278712e-01
8.17946553e-01 4.53647256e-01 -6.08919859e-02 -6.90391123e-01
-2.69276589e-01 -5.83506048e-01 -9.73712742e-01 -1.36729613e-01
8.79769564e-01 -5.88440299e-01 -8.55558276e-01 7.76852906e-01
-8.39955449e-01 -1.24828510e-01 -2.58357882e-01 9.52543914e-01
-8.44779611e-02 6.17713153e-01 -2.06954047e-01 -9.00793672e-01
-1.93449467e-01 -1.36412036e+00 8.26917231e-01 7.71754086e-01
-6.92903161e-01 -8.47413838e-01 -1.38146818e-01 8.79269361e-01
2.49033868e-01 3.52565855e-01 1.00200963e+00 -7.72992611e-01
-2.40961134e-01 -6.26075566e-01 -3.19544613e-01 2.77824312e-01
5.46646118e-01 4.43640649e-01 -1.08627272e+00 -4.31723356e-01
-2.45239735e-01 -2.71172345e-01 1.97274730e-01 2.74127424e-01
4.99919832e-01 -1.78215623e-01 -3.54980320e-01 6.02676570e-01
9.83993530e-01 4.34901208e-01 1.06098199e+00 5.71315587e-02
1.28592074e-01 1.18510556e+00 3.02091181e-01 -3.52305025e-01
2.38190964e-02 8.03216934e-01 2.29377002e-02 -6.38536870e-01
-2.98172772e-01 -4.21034284e-02 -3.12250219e-02 -5.37725352e-02
-5.02703786e-01 1.89687669e-01 -9.49012220e-01 6.93884909e-01
-1.36854482e+00 -9.98173952e-01 -6.30212605e-01 2.85529208e+00
8.84748757e-01 -1.47486702e-01 3.70093852e-01 -3.01121548e-02
1.20528984e+00 -4.51820612e-01 -5.57177842e-01 -4.48965400e-01
-2.15388790e-01 2.88249940e-01 1.86435029e-01 5.99720061e-01
-8.91400218e-01 4.98600811e-01 6.65970945e+00 3.60323966e-01
-1.17061901e+00 -3.32423538e-01 8.19369733e-01 -4.16715443e-02
1.10661127e-01 2.61326134e-01 -6.35028958e-01 5.24502993e-01
5.57792306e-01 -1.35000601e-01 3.19913656e-01 2.13123366e-01
1.14232622e-01 -4.70310807e-01 -1.10406744e+00 1.04883504e+00
2.51966000e-01 -1.00209451e+00 -1.35057598e-01 3.18938822e-01
5.92697799e-01 -4.23461825e-01 -1.24568306e-03 -3.88233244e-01
-1.60210997e-01 -1.36408472e+00 2.21596479e-01 7.81187296e-01
1.15974283e+00 -2.19104886e-01 1.10832334e+00 2.03501545e-02
-6.76715970e-01 2.98107654e-01 -7.54020289e-02 -3.03755313e-01
-1.53288424e-01 8.15940499e-02 -9.52527344e-01 3.18419367e-01
6.20114446e-01 1.34385467e-01 -8.46280634e-01 1.27709091e+00
-2.41280943e-01 2.83256918e-01 -2.66912639e-01 2.44691864e-01
-2.28707224e-01 -4.06967849e-01 7.32662857e-01 5.38429081e-01
6.57709911e-02 1.57722041e-01 -4.08711284e-01 1.05157030e+00
2.62847990e-01 8.11446607e-02 -2.96553403e-01 -6.52977824e-02
2.40635082e-01 9.81208682e-01 -8.68721604e-01 -1.83015540e-01
-3.39566946e-01 7.81056106e-01 -2.24206269e-01 4.42627996e-01
-1.07210390e-01 -5.87814152e-01 8.14072907e-01 4.47755933e-01
-9.54398364e-02 5.33996880e-01 -3.16900045e-01 -8.66778672e-01
1.71380356e-01 -9.63791192e-01 2.74370432e-01 -1.17759907e+00
-1.11026382e+00 5.03108144e-01 -4.08028185e-01 -1.45220923e+00
-3.27069730e-01 -3.59180003e-01 -6.37676716e-01 1.59506309e+00
-1.07573605e+00 -9.50290442e-01 -4.14100736e-01 3.96925449e-01
-2.39115402e-01 -2.46483669e-01 9.20056045e-01 1.88636974e-01
-3.84607553e-01 8.07328105e-01 -3.55308264e-01 5.40925205e-01
1.18337297e+00 -1.18454134e+00 1.61505342e-01 9.13060784e-01
3.46370898e-02 1.06282163e+00 4.07269984e-01 -6.12531960e-01
-6.31937563e-01 -5.13599455e-01 1.45407319e+00 -5.88813722e-01
9.28374156e-02 1.46515548e-01 -8.46354365e-01 4.09654260e-01
-1.28733357e-02 -1.64366104e-02 1.06906223e+00 1.80111647e-01
-2.31081694e-01 9.42724571e-02 -1.59208965e+00 5.45699000e-01
5.17691374e-01 -7.31197357e-01 -7.40974367e-01 3.49168897e-01
-2.14513287e-01 -6.31885350e-01 -9.03319716e-01 2.16063976e-01
1.03351796e+00 -1.22951186e+00 6.53971553e-01 -8.90144229e-01
2.16355324e-01 -8.25968206e-01 4.98299062e-01 -8.16827357e-01
-1.66305572e-01 -8.22400093e-01 7.49549925e-01 1.10468626e+00
4.94097739e-01 -1.02024102e+00 4.39810127e-01 1.25350177e+00
3.39027345e-01 -6.27422094e-01 -9.06767666e-01 -6.57820523e-01
-1.29521444e-01 3.74233991e-01 7.05934286e-01 1.22844541e+00
3.27940546e-02 -2.28327915e-01 -5.37338145e-02 3.07202071e-01
3.39894742e-01 3.15298378e-01 1.02465606e+00 -1.31650865e+00
-5.61765023e-02 -6.33662283e-01 -9.76191700e-01 -5.70007443e-01
-3.48473638e-01 -5.08222878e-01 -3.57288629e-01 -1.14219713e+00
1.71890646e-01 -3.65171283e-01 1.76496595e-01 6.70368493e-01
-4.57232594e-01 6.96835697e-01 1.05562611e-02 5.67664206e-01
2.10937187e-01 -5.93949974e-01 1.24105275e+00 8.69032815e-02
-2.37866521e-01 3.58967245e-01 -8.22183311e-01 4.83335584e-01
5.27515233e-01 -1.13842875e-01 -1.66656733e-01 -3.50924619e-02
2.37341061e-01 9.66959149e-02 7.39063501e-01 -6.20684683e-01
3.08291972e-01 1.49753734e-01 7.18661487e-01 -1.06264561e-01
-1.63499549e-01 -5.31117737e-01 3.34997118e-01 6.23321712e-01
-1.47495672e-01 -1.30802169e-01 3.53224307e-01 -9.01223943e-02
-2.80271292e-01 -3.67262155e-01 1.20160711e+00 -5.11278883e-02
-3.43309641e-01 -8.65181759e-02 -1.86403558e-01 6.53473213e-02
1.03402984e+00 -9.57536995e-01 -5.21019101e-01 -2.87424207e-01
-9.88515198e-01 4.49773967e-02 1.19576693e+00 8.52701291e-02
3.13903987e-01 -8.13089311e-01 -7.86248624e-01 6.57628000e-01
2.59224504e-01 -2.70189524e-01 2.02902541e-01 1.44764698e+00
-6.50214016e-01 2.29672000e-01 -2.04702407e-01 -7.05145657e-01
-1.87265480e+00 2.87733138e-01 8.53974164e-01 1.65003657e-01
-4.75204229e-01 1.00295067e+00 7.00678527e-02 -2.64228016e-01
2.33315289e-01 -1.37753025e-01 -1.15820996e-01 1.67005509e-01
8.50532591e-01 1.81595445e-01 1.76357687e-01 -1.03363168e+00
-1.96422502e-01 1.01246309e+00 -4.01476383e-01 2.57030845e-01
4.97882962e-01 -1.26036689e-01 -2.75229603e-01 2.18265932e-02
6.56202972e-01 1.80757180e-01 -8.07040453e-01 -1.98131770e-01
-4.88545537e-01 -1.06637704e+00 -4.74042185e-02 -1.25645578e+00
-8.40175807e-01 6.95007980e-01 1.18883491e+00 2.31673568e-01
1.12990034e+00 -2.71880567e-01 4.47042823e-01 -2.94436157e-01
2.58121043e-01 -8.59988809e-01 -7.09510207e-01 -3.07440907e-01
5.31806588e-01 -1.38358951e+00 2.85994798e-01 -1.80646360e-01
-6.24533772e-01 1.22966921e+00 3.35629404e-01 4.56625342e-01
3.06472450e-01 -2.61612236e-01 5.36462545e-01 -3.96578342e-01
-1.77309543e-01 -3.22336078e-01 8.27712715e-01 8.17552030e-01
4.73786801e-01 -3.38036232e-02 -5.56908250e-01 1.15623429e-01
-4.28755075e-01 2.31578816e-02 4.90026653e-01 4.83434916e-01
1.78163245e-01 -1.08146870e+00 -6.55004859e-01 7.68613100e-01
-2.30199128e-01 -2.98165791e-02 -7.91494012e-01 9.11418915e-01
5.27563870e-01 1.02174985e+00 4.63695228e-01 -1.03677854e-01
2.50612557e-01 2.11836666e-01 5.23969829e-01 -5.05421340e-01
-8.19118857e-01 -1.12130828e-01 4.29892763e-02 -3.65646422e-01
-6.19187534e-01 -1.08197987e+00 -8.57031643e-01 -3.51359695e-01
-5.02203763e-01 1.22042827e-01 4.34680730e-01 1.02452862e+00
4.18643773e-01 -1.07270531e-01 3.76420587e-01 -3.63505155e-01
-2.46433049e-01 -9.42212939e-01 -8.43429267e-01 6.48404300e-01
6.82956755e-01 -4.72514689e-01 -8.80226731e-01 5.02101183e-01] | [3.741464376449585, -3.6319007873535156] |
ff5600ff-f51e-4d74-9999-0761dd167dbe | strategic-resource-selection-with-homophilic | 2305.00843 | null | https://arxiv.org/abs/2305.00843v1 | https://arxiv.org/pdf/2305.00843v1.pdf | Strategic Resource Selection with Homophilic Agents | The strategic selection of resources by selfish agents is a classic research direction, with Resource Selection Games and Congestion Games as prominent examples. In these games, agents select available resources and their utility then depends on the number of agents using the same resources. This implies that there is no distinction between the agents, i.e., they are anonymous. We depart from this very general setting by proposing Resource Selection Games with heterogeneous agents that strive for joint resource usage with similar agents. So, instead of the number of other users of a given resource, our model considers agents with different types and the decisive feature is the fraction of same-type agents among the users. More precisely, similarly to Schelling Games, there is a tolerance threshold $\tau \in [0,1]$ which specifies the agents' desired minimum fraction of same-type agents on a resource. Agents strive to select resources where at least a $\tau$-fraction of those resources' users have the same type as themselves. For $\tau=1$, our model generalizes Hedonic Diversity Games with a peak at $1$. For our general model, we consider the existence and quality of equilibria and the complexity of maximizing social welfare. Additionally, we consider a bounded rationality model, where agents can only estimate the utility of a resource, since they only know the fraction of same-type agents on a given resource, but not the exact numbers. Thus, they cannot know the impact a strategy change would have on a target resource. Interestingly, we show that this type of bounded rationality yields favorable game-theoretic properties and specific equilibria closely approximate equilibria of the full knowledge setting. | ['Alexander Skopalik', 'Pascal Lenzner', 'Simon Krogmann', 'Jonathan Gadea Harder'] | 2023-05-01 | null | null | null | null | ['type'] | ['speech'] | [-5.20226717e-01 2.86492944e-01 -5.21052063e-01 3.35220575e-01
4.62381728e-02 -8.89266670e-01 -1.02198660e-01 -1.06691934e-01
-1.02039862e+00 1.02640522e+00 -3.53741199e-02 -1.39210775e-01
-4.33330864e-01 -1.30596471e+00 -1.68836325e-01 -8.44781458e-01
-3.90967846e-01 6.97710454e-01 2.13984460e-01 -4.46816832e-01
-3.14209075e-03 1.75412357e-01 -1.10903764e+00 -5.54129004e-01
6.37847483e-01 8.11949253e-01 2.03120157e-01 4.73080367e-01
5.12065776e-02 6.35027587e-01 -8.52468729e-01 -5.10265231e-01
9.11074877e-01 -4.35167402e-01 -4.77454901e-01 -4.86544473e-03
-6.58450186e-01 -5.31696618e-01 -4.82528359e-01 1.25994813e+00
4.13831413e-01 2.50678092e-01 5.66108227e-01 -1.47439098e+00
-4.73679096e-01 1.09394836e+00 -6.45859361e-01 1.52157500e-01
1.46495834e-01 5.87593094e-02 1.44795036e+00 9.48863924e-02
6.14234030e-01 7.12546647e-01 -1.09794503e-02 8.46360683e-01
-1.08508790e+00 -6.47996962e-01 1.82997242e-01 -2.98083007e-01
-1.38059998e+00 -2.66284466e-01 2.83451349e-01 -3.64519417e-01
4.57832724e-01 3.58140439e-01 6.74374998e-01 2.89179146e-01
-2.68514901e-01 2.58344978e-01 8.62080097e-01 -3.42263877e-01
5.52560389e-01 2.00835958e-01 -1.44218922e-01 2.36035287e-01
7.62280226e-01 9.65495706e-02 -4.64717746e-01 -3.25723231e-01
9.48024333e-01 1.37309864e-01 -4.22129452e-01 -6.42707527e-01
-6.80050313e-01 9.15828884e-01 1.86874226e-01 3.71199310e-01
-6.28413558e-01 6.01057410e-02 -7.59153394e-03 7.21827269e-01
4.85007018e-02 7.45508194e-01 -2.71455854e-01 -1.98872373e-01
-3.97759289e-01 7.72702396e-02 1.08616841e+00 1.06389797e+00
9.74864364e-01 -6.26649633e-02 -2.13795081e-01 5.62029958e-01
-1.32285669e-01 4.50110734e-01 1.00494837e-02 -1.27016258e+00
4.49270368e-01 6.03352070e-01 9.75588381e-01 -7.07398415e-01
-3.83180261e-01 -5.39005160e-01 -7.32004166e-01 3.84886205e-01
7.47039080e-01 -7.98296571e-01 -9.02172625e-02 2.50998330e+00
-1.31917551e-01 -3.37553263e-01 1.75876487e-02 1.08178151e+00
5.24548702e-02 4.70335156e-01 -2.72162825e-01 -8.00308526e-01
1.15933204e+00 -4.80582476e-01 -2.35431746e-01 -2.57368460e-02
3.53075296e-01 -2.73903519e-01 6.40410304e-01 -3.89619470e-02
-1.40365803e+00 3.23655903e-01 -6.81247652e-01 8.34271729e-01
-4.46676947e-02 -5.79888225e-01 4.79022592e-01 1.03841054e+00
-1.04021168e+00 4.02714431e-01 -2.06117392e-01 -1.66241616e-01
2.47304693e-01 5.17989695e-01 6.12129420e-02 2.23087981e-01
-1.22342670e+00 6.12833083e-01 -6.87530637e-03 -4.25781876e-01
-7.90873468e-01 -6.57083035e-01 -2.18219399e-01 6.00665152e-01
1.19528079e+00 -8.38946760e-01 1.15828931e+00 -1.11694968e+00
-1.50943935e+00 6.29120946e-01 2.53870338e-01 -3.92086774e-01
9.48137343e-01 4.70387578e-01 1.08227842e-01 -7.05358833e-02
4.45587099e-01 1.40899383e-02 2.36895353e-01 -1.06964660e+00
-1.01339102e+00 -8.87978449e-02 9.39108491e-01 7.69535422e-01
-3.63218635e-01 4.18951828e-03 -2.71422952e-01 3.45382020e-02
-4.66646105e-01 -8.71257484e-01 -4.47280467e-01 -2.05779299e-01
-5.10451794e-01 -3.88342119e-03 -1.61681339e-01 3.79285991e-01
1.12187397e+00 -2.10880566e+00 1.52601616e-03 5.88009894e-01
6.10875487e-01 -7.46524110e-02 -1.76940531e-01 5.27844906e-01
5.85338116e-01 5.87073326e-01 1.69813499e-01 -1.38702005e-01
5.41119337e-01 1.53848693e-01 1.59191087e-01 3.66314977e-01
-7.67063737e-01 4.20666337e-01 -8.48149955e-01 2.76536882e-01
-2.06395954e-01 -3.59714240e-01 -6.87220335e-01 1.65883407e-01
-1.62919573e-02 1.73264351e-02 -8.42662573e-01 -6.68313876e-02
5.76363206e-01 -2.18166202e-01 5.70975125e-01 6.76389039e-01
-3.47157508e-01 1.33793041e-01 -1.55704939e+00 6.93433881e-01
-3.78221095e-01 1.16252810e-01 6.53114498e-01 -7.31272995e-01
2.45930031e-01 1.79618880e-01 6.23951674e-01 -6.47708535e-01
3.53709728e-01 2.99791157e-01 5.22412121e-01 6.96384301e-03
5.39067566e-01 -3.07304889e-01 -2.95466632e-01 1.03139448e+00
-4.52572048e-01 3.44254196e-01 3.80718231e-01 5.04372597e-01
1.03087258e+00 -7.21043825e-01 6.84207380e-01 -4.28146273e-01
1.07057750e-01 -1.33839026e-01 1.10624921e+00 1.40076458e+00
-4.15374488e-01 6.12754487e-02 1.11490119e+00 2.10700203e-02
-1.02909398e+00 -9.81067300e-01 4.56229925e-01 1.22881794e+00
8.59942853e-01 -1.42870024e-02 -6.07899666e-01 -1.17621519e-01
1.89062446e-01 3.18497896e-01 -8.13432217e-01 1.58432499e-01
1.17614403e-01 -5.86641550e-01 9.05728564e-02 -4.75974493e-02
7.52160549e-01 -1.05875671e+00 -9.53777194e-01 1.30974665e-01
-3.74786719e-03 -7.33123541e-01 -1.02324891e+00 -8.41981918e-02
-7.23571926e-02 -1.24236548e+00 -7.44827688e-01 -3.07586849e-01
5.54515004e-01 7.06485808e-01 9.17258620e-01 2.00105414e-01
1.12270117e-01 5.07026136e-01 -3.05230916e-01 -2.81894982e-01
-7.62199238e-02 1.35301352e-01 3.68483633e-01 2.98101967e-03
1.14973910e-01 -4.99438405e-01 -9.51220572e-01 5.53778291e-01
-6.24970675e-01 -1.66876301e-01 2.28761673e-01 4.81287926e-01
1.40719518e-01 5.83492219e-01 8.80282700e-01 -9.35884833e-01
8.69139552e-01 -7.42558002e-01 -7.82950282e-01 2.23598540e-01
-2.44732499e-01 -2.96144754e-01 1.04021609e+00 -3.41763914e-01
-7.59168029e-01 -5.14294982e-01 7.78505981e-01 1.19970545e-01
3.03510934e-01 3.41901094e-01 -5.51237464e-01 6.87749162e-02
3.22161615e-01 4.92007881e-02 7.36163780e-02 -1.22098655e-01
2.50500292e-01 5.41038334e-01 2.53690891e-02 -6.63353562e-01
7.36039042e-01 3.90164256e-01 7.29746222e-02 -6.44369960e-01
-1.36558220e-01 -7.49915466e-02 5.56373671e-02 -2.35495456e-02
3.67336661e-01 -7.36679912e-01 -1.68473172e+00 5.75152934e-01
-6.94889128e-01 -4.42230821e-01 -5.66347003e-01 4.08336878e-01
-5.88764071e-01 1.95534497e-01 -3.83930326e-01 -1.37236404e+00
1.46475166e-01 -9.69791114e-01 -1.48980916e-01 6.75528705e-01
2.24526420e-01 -6.36736631e-01 -1.55342013e-01 4.79816012e-02
8.80729973e-01 -3.78017336e-01 7.04598248e-01 -7.40667582e-01
-1.08610344e+00 -1.51531212e-02 -6.50262311e-02 -3.61217171e-01
2.87925303e-01 -3.81183803e-01 -1.14066832e-01 -4.73891020e-01
-3.54506731e-01 4.67688292e-02 3.15712482e-01 6.70679629e-01
5.42464375e-01 -8.27748537e-01 -3.86470318e-01 1.77856714e-01
1.40383482e+00 6.52263880e-01 2.65894800e-01 3.06656212e-01
8.60656053e-02 5.62830865e-01 1.34955823e-01 9.53969896e-01
5.16073048e-01 7.82156289e-01 5.75469792e-01 2.13982984e-01
6.12925470e-01 -2.72599701e-02 9.97339785e-02 -1.27260864e-01
-4.55642372e-01 -4.98080552e-01 -3.56206834e-01 5.40358543e-01
-1.86920369e+00 -1.18784976e+00 6.39023960e-01 2.97902822e+00
7.93530047e-01 2.72285610e-01 9.11227167e-01 -2.27475673e-01
8.85634422e-01 -1.05838843e-01 -8.89803410e-01 -2.43243366e-01
-3.42198461e-01 -1.24468297e-01 1.06807804e+00 8.77147079e-01
-4.08759683e-01 7.69883335e-01 5.79181337e+00 7.59631574e-01
-6.06911242e-01 9.62810814e-02 8.56881976e-01 -8.70567620e-01
-4.23368007e-01 -6.06057495e-02 -5.57481766e-01 8.36966038e-01
4.81728494e-01 -1.10009384e+00 1.11774802e+00 6.61717057e-01
4.88231719e-01 -4.98540044e-01 -8.03708553e-01 6.85208023e-01
-4.85865653e-01 -1.19514692e+00 -4.03678417e-01 5.48575878e-01
6.47783041e-01 -2.28499368e-01 2.33784467e-02 -1.70073491e-02
1.17249036e+00 -8.29575717e-01 7.07508385e-01 2.04394743e-01
6.07779205e-01 -1.12607276e+00 6.97429538e-01 5.48453748e-01
-1.13953066e+00 -5.01181483e-01 -3.72312427e-01 -5.75542867e-01
2.32209697e-01 3.95052493e-01 -2.53742695e-01 4.33180392e-01
4.38277423e-01 -1.77044168e-01 2.30512902e-01 1.24027097e+00
7.04038367e-02 2.82321721e-02 -4.18361723e-01 -2.35917225e-01
5.58988331e-03 -6.03433132e-01 5.12760758e-01 3.46593350e-01
4.03987974e-01 7.64207721e-01 5.65951586e-01 8.02295268e-01
-5.07437527e-01 2.11008877e-01 -4.63923424e-01 3.95039702e-03
1.18172562e+00 1.04570496e+00 -6.70328856e-01 -1.26545772e-01
-5.44400096e-01 6.29573226e-01 8.99340659e-02 4.90453780e-01
-4.59435612e-01 -3.98673952e-01 1.34239423e+00 1.45273730e-01
8.54865834e-02 4.76469100e-02 -2.94691443e-01 -1.06422162e+00
-1.35560334e-01 -4.09549236e-01 5.71893871e-01 -3.35929304e-01
-1.37078989e+00 2.97953814e-01 -2.89075226e-01 -9.49651599e-01
-5.46032377e-02 -1.50494829e-01 -8.02974284e-01 1.02702641e+00
-1.35543060e+00 -2.57500768e-01 1.92560717e-01 5.48457801e-01
-6.54063746e-02 -3.39743853e-01 4.47089046e-01 3.06915604e-02
-6.77534521e-01 6.63471103e-01 3.25949013e-01 1.89991891e-01
-8.75694957e-03 -1.08499551e+00 -2.82983989e-01 8.00854683e-01
-2.65921563e-01 5.95452845e-01 6.20391905e-01 -3.23356867e-01
-1.31384456e+00 -5.64849675e-01 4.47170168e-01 9.99557525e-02
5.89433849e-01 -1.32537887e-01 -2.86561131e-01 3.68194848e-01
1.29657999e-01 -2.06956699e-01 7.05277503e-01 3.22885700e-02
-4.66970205e-02 -4.97307032e-02 -1.38987136e+00 1.17686498e+00
1.47390819e+00 -2.45587602e-01 1.63086846e-01 -1.47696182e-01
3.92862260e-01 5.49630858e-02 -4.17317092e-01 -2.40152642e-01
4.35175776e-01 -1.20606458e+00 5.18783152e-01 -5.49118876e-01
-5.78965135e-02 -1.79272473e-01 -1.84258997e-01 -1.67405283e+00
-4.90579218e-01 -1.04357743e+00 6.24211788e-01 8.96564126e-01
5.19400895e-01 -1.14472353e+00 1.16347575e+00 1.09917700e+00
4.97151375e-01 -1.92293927e-01 -1.27429843e+00 -8.30242217e-01
5.15173018e-01 3.19619179e-01 7.27377713e-01 6.84342921e-01
5.66568375e-01 2.06334546e-01 -5.50674319e-01 7.48732910e-02
6.75818920e-01 2.38108218e-01 5.78271091e-01 -1.14899778e+00
-6.36083305e-01 -9.02272046e-01 3.78941000e-02 -9.79919970e-01
1.36869341e-01 -3.77091289e-01 -2.37205580e-01 -1.27500248e+00
6.65805459e-01 -9.79916871e-01 -1.98994935e-01 3.78518075e-01
1.01359472e-01 6.30043261e-03 6.63194537e-01 3.39785397e-01
-8.97908270e-01 2.06560671e-01 1.11149895e+00 -5.23755141e-03
-6.10986054e-01 4.15787607e-01 -1.40581405e+00 2.73913324e-01
7.79555261e-01 -1.45900860e-01 -5.98433495e-01 -2.67557073e-02
5.16978562e-01 6.82696044e-01 -8.23822841e-02 -3.22367013e-01
3.53373826e-01 -9.79962885e-01 -2.81686127e-01 3.40230286e-01
1.78526446e-01 -7.82951355e-01 5.29427886e-01 6.11368120e-01
-4.07872856e-01 -2.82796830e-01 -5.70637047e-01 4.68879044e-01
4.24508601e-01 -4.03318375e-01 7.76244402e-01 -5.21351278e-01
-2.46776074e-01 6.01488471e-01 -8.50217581e-01 2.48834521e-01
1.27344894e+00 -3.74456465e-01 -6.08624458e-01 -1.17841423e+00
-6.83421135e-01 5.87397099e-01 1.06658340e+00 -1.04300849e-01
3.37472036e-02 -9.30870354e-01 -6.64424658e-01 -1.75526723e-01
-2.98267305e-02 -3.30553025e-01 3.33094418e-01 4.64133710e-01
1.34156257e-01 6.28847852e-02 -5.25636315e-01 3.80685508e-01
-9.26640809e-01 5.61128855e-01 9.10371602e-01 -1.43842950e-01
3.25289033e-02 4.71933484e-01 6.79188073e-01 1.06552891e-01
6.29437044e-02 3.54551375e-01 -1.91340283e-01 1.50758743e-01
5.07103682e-01 4.66140419e-01 -4.67084497e-01 -5.01575708e-01
-2.72306770e-01 2.52979666e-01 -6.80965856e-02 -5.53397417e-01
1.09880257e+00 -5.66282392e-01 -6.99534267e-02 1.71028804e-02
3.54569793e-01 4.65855867e-01 -1.28668809e+00 -4.87334847e-01
-4.87344086e-01 -8.42098534e-01 -5.37796080e-01 -6.92685604e-01
-1.47682607e+00 1.65878937e-01 -1.13765016e-01 9.61619496e-01
1.02034652e+00 -1.60266131e-01 -5.24635948e-02 6.74318150e-02
1.14738417e+00 -1.05142689e+00 -2.56602932e-02 4.24607992e-01
1.85193643e-01 -7.46492743e-01 -2.48919100e-01 -3.67045403e-01
-7.20360100e-01 5.14304519e-01 5.77716172e-01 -6.45224601e-02
5.46957791e-01 8.71439502e-02 -2.14118794e-01 7.90686235e-02
-6.99133515e-01 -7.76155472e-01 -6.12118423e-01 7.45546579e-01
-5.33984937e-02 7.58079946e-01 -6.98687613e-01 1.01222074e+00
-3.18825990e-01 -1.22465432e-01 1.49545908e+00 5.96169770e-01
-7.30171978e-01 -1.18972945e+00 -1.02710046e-01 7.00780571e-01
-6.26957178e-01 -1.42417684e-01 -3.17036837e-01 5.47841609e-01
-8.55194107e-02 1.22321320e+00 3.30167949e-01 9.60362703e-02
3.55114847e-01 -4.65489179e-01 2.65839577e-01 -8.54793012e-01
-4.24843162e-01 2.83989985e-03 -2.06553228e-02 -8.05543587e-02
-2.72978365e-01 -4.54694360e-01 -9.53816593e-01 -9.92023766e-01
-2.47907326e-01 7.61820734e-01 -2.08115190e-01 6.48097217e-01
2.93945938e-01 9.21202675e-02 1.14801681e+00 -2.32344493e-01
-5.34826875e-01 -5.66942632e-01 -1.57489789e+00 2.03214735e-01
2.03820720e-01 -7.40674555e-01 -7.11648166e-01 -7.35134602e-01] | [4.296154499053955, 2.9792773723602295] |
f087b6d3-8c7f-4e74-8a93-ac17fccaac48 | developmental-reinforcement-learning-of | 2007.07793 | null | https://arxiv.org/abs/2007.07793v1 | https://arxiv.org/pdf/2007.07793v1.pdf | Developmental Reinforcement Learning of Control Policy of a Quadcopter UAV with Thrust Vectoring Rotors | In this paper, we present a novel developmental reinforcement learning-based controller for a quadcopter with thrust vectoring capabilities. This multirotor UAV design has tilt-enabled rotors. It utilizes the rotor force magnitude and direction to achieve the desired state during flight. The control policy of this robot is learned using the policy transfer from the learned controller of the quadcopter (comparatively simple UAV design without thrust vectoring). This approach allows learning a control policy for systems with multiple inputs and multiple outputs. The performance of the learned policy is evaluated by physics-based simulations for the tasks of hovering and way-point navigation. The flight simulations utilize a flight controller based on reinforcement learning without any additional PID components. The results show faster learning with the presented approach as opposed to learning the control policy from scratch for this new UAV design created by modifications in a conventional quadcopter, i.e., the addition of more degrees of freedom (4-actuators in conventional quadcopter to 8-actuators in tilt-rotor quadcopter). We demonstrate the robustness of our learned policy by showing the recovery of the tilt-rotor platform in the simulation from various non-static initial conditions in order to reach a desired state. The developmental policy for the tilt-rotor UAV also showed superior fault tolerance when compared with the policy learned from the scratch. The results show the ability of the presented approach to bootstrap the learned behavior from a simpler system (lower-dimensional action-space) to a more complex robot (comparatively higher-dimensional action-space) and reach better performance faster. | ['Ali A. Minai', 'Rumit Kumar', 'Aditya M. Deshpande', 'Manish Kumar'] | 2020-07-15 | null | null | null | null | ['developmental-learning', 'drone-controller'] | ['robots', 'robots'] | [-1.87193006e-01 4.54084873e-01 9.70238000e-02 3.40979040e-01
2.96468318e-01 -8.70093346e-01 4.21309948e-01 -2.32628897e-01
-5.47424197e-01 1.17458820e+00 -5.95105827e-01 -3.66759330e-01
-5.45181632e-01 -6.63671315e-01 -9.35087204e-01 -9.83617961e-01
-1.62321359e-01 4.40990627e-01 3.05846274e-01 -9.06363666e-01
-1.56922005e-02 7.35114276e-01 -1.63414311e+00 -5.33059180e-01
7.22968996e-01 4.98178571e-01 5.36768138e-01 8.94430280e-01
9.75920320e-01 2.85892338e-01 -5.42533755e-01 5.59303820e-01
7.79248238e-01 -2.69427568e-01 -3.11658204e-01 2.73221791e-01
3.00866902e-01 -2.89858580e-01 -1.17104901e-02 6.11417830e-01
3.51309448e-01 5.42513192e-01 5.41540146e-01 -9.69229996e-01
1.68373168e-01 1.77222073e-01 -8.16678721e-03 -4.16574568e-01
5.08704409e-02 5.26455402e-01 3.81313682e-01 -1.27136976e-01
4.44702059e-01 8.04288864e-01 6.91916406e-01 4.55512583e-01
-9.96746659e-01 -4.68771398e-01 -2.35293582e-01 -4.00804341e-01
-1.14959776e+00 2.72215545e-01 1.59568544e-02 -6.40915036e-01
1.00237870e+00 -1.05601147e-01 1.09359610e+00 7.87702858e-01
9.01000381e-01 -1.38041049e-01 1.29618776e+00 -3.41312557e-01
3.97509247e-01 8.97267908e-02 -5.57102382e-01 1.05196548e+00
6.77774727e-01 8.83226216e-01 3.83858740e-01 7.68779516e-02
9.52712536e-01 2.54725032e-02 -3.27779830e-01 -8.05828571e-01
-1.11015451e+00 6.73674941e-01 4.97041136e-01 2.66017854e-01
-6.35166585e-01 3.57129663e-01 1.77691579e-01 7.03929186e-01
-4.97767061e-01 1.23501372e+00 -9.80724931e-01 -1.88468426e-01
-5.28677046e-01 7.44048953e-01 1.00365508e+00 8.80522549e-01
5.13596535e-01 9.65785205e-01 3.01831543e-01 2.28275821e-01
-1.21033885e-01 9.19884503e-01 4.13959205e-01 -8.55849385e-01
3.44828963e-02 4.67945576e-01 5.37621856e-01 -6.22033238e-01
-7.32039750e-01 -5.30448794e-01 -4.76040363e-01 1.08891034e+00
1.85115755e-01 -1.19689167e+00 -8.74612629e-01 1.62104642e+00
4.43502516e-01 -3.26806605e-01 4.37739760e-01 1.05729043e+00
-2.81302392e-01 7.98157275e-01 -5.94544232e-01 -3.43789399e-01
1.03690743e+00 -6.32980347e-01 -3.20935220e-01 9.28368196e-02
5.25551081e-01 -3.19757879e-01 7.76250541e-01 8.10379803e-01
-7.04484344e-01 -7.59895444e-01 -1.58883262e+00 9.29468274e-01
-5.11282444e-01 3.61280441e-01 9.56994444e-02 3.43315542e-01
-1.01509154e+00 9.86651063e-01 -8.01084995e-01 -5.34572244e-01
-7.89124370e-01 7.13680267e-01 -5.93285441e-01 6.26454651e-01
-1.36238062e+00 1.34120905e+00 8.19741368e-01 -3.10190022e-01
-1.28652835e+00 -4.98846054e-01 -5.11464834e-01 -1.40695661e-01
8.22839379e-01 -1.01442409e+00 1.31449962e+00 -7.92626202e-01
-2.42077327e+00 -6.00063019e-02 7.83990026e-01 -6.75208747e-01
3.38330626e-01 -4.33274090e-01 1.85050324e-01 2.77369916e-01
-1.08557880e-01 4.79292870e-01 1.53901386e+00 -1.14149094e+00
-7.32595086e-01 1.83940813e-01 3.87944341e-01 3.62927616e-01
-1.62557021e-01 -8.50489616e-01 6.59912467e-01 -6.51318192e-01
-2.85761118e-01 -1.66046333e+00 -1.79253608e-01 -3.34206343e-01
1.46866530e-01 2.10919231e-01 8.58700156e-01 -2.11920589e-02
6.79038167e-01 -1.60049367e+00 7.44056344e-01 -6.13556579e-02
-4.29644912e-01 4.80644137e-01 1.76398486e-01 1.04562545e+00
-2.29114264e-01 -2.36537859e-01 -2.07110979e-02 5.23132145e-01
-4.43682045e-01 4.76384938e-01 -3.53804976e-01 3.22124094e-01
2.07713082e-01 2.97200650e-01 -9.46250618e-01 1.65315866e-01
5.86662665e-02 1.90423727e-01 -7.34586537e-01 3.63267332e-01
-4.77053702e-01 6.32696092e-01 -5.70464849e-01 3.04176331e-01
1.98466256e-01 5.77273726e-01 2.71899968e-01 -3.48170251e-02
-8.93695235e-01 -5.29205918e-01 -9.70656991e-01 1.21485639e+00
-8.64608109e-01 1.30096331e-01 6.20402694e-01 -9.72973526e-01
1.20622230e+00 3.58390957e-01 5.07891834e-01 -1.70052841e-01
4.78137016e-01 2.16017291e-01 3.36525053e-01 -5.47206759e-01
5.23192465e-01 -5.31555414e-01 -1.91410333e-01 5.43303601e-02
3.61649126e-01 -1.13744807e+00 -1.15271506e-03 -4.63824183e-01
1.08016431e+00 6.54048443e-01 3.69020313e-01 -5.79081714e-01
6.88403785e-01 5.68648517e-01 5.44147253e-01 3.76677990e-01
3.01589012e-01 -8.73344392e-03 3.96791458e-01 -3.17270219e-01
-1.09809685e+00 -6.55564964e-01 1.97527125e-01 6.13321662e-01
-4.90038618e-02 -2.46557042e-01 -7.08680511e-01 -3.00571263e-01
4.45079714e-01 6.45243347e-01 -4.65772718e-01 -4.66439247e-01
-7.92889178e-01 -2.65482485e-01 3.69996965e-01 8.27142373e-02
4.69960928e-01 -5.57607293e-01 -1.57679582e+00 2.59937882e-01
7.32393622e-01 -9.13045585e-01 2.08927706e-01 5.93641400e-01
-1.07938623e+00 -1.33102262e+00 -4.56851721e-01 -5.10789871e-01
5.43009877e-01 -5.61310872e-02 2.88145721e-01 -4.52792123e-02
-2.98014104e-01 9.09419537e-01 -4.91080105e-01 -4.25317466e-01
-4.45947647e-01 2.43766163e-03 7.73780048e-01 -5.18673718e-01
-9.02839541e-01 -4.37310785e-01 -2.39671156e-01 5.37430346e-01
-8.26341987e-01 1.68048199e-02 8.55091095e-01 1.34194887e+00
4.14932817e-01 4.67145354e-01 5.54046452e-01 -3.37635547e-01
7.34030128e-01 -2.44508862e-01 -1.34161115e+00 -2.15120882e-01
-6.87871516e-01 2.43265703e-01 1.37832415e+00 -5.81379235e-01
-6.91433609e-01 4.90518183e-01 2.37372234e-01 -5.91055274e-01
2.38523073e-02 1.95179090e-01 4.12740797e-01 -4.55896378e-01
5.99901795e-01 1.78668827e-01 6.13082528e-01 -2.53775656e-01
3.25196475e-01 2.43695170e-01 1.69506997e-01 -8.24779987e-01
1.20586216e+00 -7.80513734e-02 8.92564118e-01 -9.96246815e-01
-6.05376363e-02 1.17782399e-01 -6.95411384e-01 -3.11330229e-01
6.42760634e-01 -6.94238067e-01 -1.04752493e+00 3.76259089e-01
-8.17207813e-01 -6.55332685e-01 -4.53470290e-01 6.86207652e-01
-1.06522977e+00 -1.57582194e-01 -2.19565064e-01 -8.73559177e-01
-2.66374558e-01 -1.07452297e+00 6.76639676e-01 5.02991974e-01
2.06746534e-01 -8.13338280e-01 5.22368789e-01 -4.33111340e-01
4.98831272e-01 8.10490906e-01 8.21266174e-01 -4.13299575e-02
-2.37137944e-01 -1.83607697e-01 7.70026803e-01 4.88435835e-01
6.85851648e-02 2.35508472e-01 -3.05090129e-01 -1.18491542e+00
3.16618830e-01 -4.84912723e-01 3.82855803e-01 1.42866328e-01
1.36433542e-01 -3.54549348e-01 -3.65511835e-01 3.93948764e-01
1.87612176e+00 5.00381827e-01 -7.12838098e-02 5.70529819e-01
3.35002452e-01 5.70512474e-01 1.33129275e+00 4.48504984e-01
-2.33649969e-01 7.56307364e-01 1.07224488e+00 3.23063612e-01
3.08673620e-01 -2.45452687e-01 8.22603106e-01 3.81002575e-01
-4.97231185e-01 2.78984848e-02 -8.09744537e-01 4.04965788e-01
-1.73920059e+00 -4.22897637e-01 1.78584740e-01 2.54015398e+00
2.95167923e-01 2.25864295e-02 1.81667939e-01 9.71456096e-02
5.26694655e-01 -4.63202477e-01 -4.52806473e-01 -9.53726172e-01
2.58965880e-01 1.82624817e-01 9.82897460e-01 6.41243100e-01
-8.84902596e-01 8.17652524e-01 5.56379795e+00 2.01220244e-01
-1.70705950e+00 -4.01732862e-01 -5.49313605e-01 -4.86923456e-02
6.08675897e-01 3.88671458e-01 -8.20225358e-01 2.09134504e-01
1.09135890e+00 -1.38787031e-01 5.93698919e-01 9.71276581e-01
3.74069273e-01 -4.13257509e-01 -6.19863629e-01 3.89904499e-01
-1.94074333e-01 -8.25741947e-01 -1.32153735e-01 1.31432593e-01
8.65228653e-01 -3.56694758e-01 -1.43688112e-01 6.03496075e-01
-8.95141587e-02 -4.34749037e-01 6.53517425e-01 6.83036983e-01
7.61491716e-01 -8.16407800e-01 7.37468600e-01 8.14784825e-01
-9.10237491e-01 -8.51523221e-01 -4.01282400e-01 -4.28125232e-01
3.83824632e-02 -1.53443841e-02 -1.53499484e+00 8.59197021e-01
2.60356367e-01 3.75666976e-01 -3.10846686e-01 7.02350974e-01
-3.19338620e-01 2.20115095e-01 -3.79453421e-01 -6.55262232e-01
5.81540823e-01 -5.48523068e-01 1.15684640e+00 5.72101057e-01
6.80113196e-01 -5.22885807e-02 3.63353431e-01 3.88812006e-01
6.08134270e-01 -1.13890603e-01 -1.10383868e+00 1.11465529e-02
-1.29445165e-01 1.44312799e+00 -3.37528050e-01 -1.89275175e-01
2.53696531e-01 3.27004820e-01 1.15496311e-02 7.38817230e-02
-9.65407491e-01 -9.09461796e-01 6.49683595e-01 6.31829917e-01
6.01805449e-01 -7.02767670e-01 6.61625683e-01 -7.03594565e-01
-5.38118422e-01 -7.20145822e-01 -2.39183828e-01 -8.28155518e-01
-5.96874535e-01 5.36445737e-01 5.35364926e-01 -1.83201313e+00
-1.10780656e+00 -1.00970578e+00 -4.74939048e-01 7.12649941e-01
-1.10578728e+00 -7.72601306e-01 -1.21661201e-01 4.00791973e-01
2.49981835e-01 -5.29989123e-01 8.41673851e-01 -4.71366942e-01
-1.68336704e-01 -1.17049500e-01 1.82502151e-01 -6.23635888e-01
8.40315640e-01 -1.44208455e+00 -3.20935428e-01 6.06114507e-01
-8.29550922e-01 6.54504299e-01 1.02401412e+00 -8.28786552e-01
-2.02404261e+00 -9.46405888e-01 -1.27782539e-01 1.14109084e-01
6.95941091e-01 1.90453790e-02 -4.73763555e-01 3.15114319e-01
3.81433487e-01 -3.07629436e-01 -1.09509289e-01 -5.65774441e-01
5.69039524e-01 -3.83484602e-01 -1.18070495e+00 5.70754647e-01
3.20350736e-01 2.45581910e-01 -8.48931432e-01 2.41514161e-01
7.37190247e-01 -5.24991035e-01 -1.06093287e+00 7.50446558e-01
7.21952796e-01 -6.90444767e-01 3.60277921e-01 -6.26334488e-01
-6.81617334e-02 -7.85892129e-01 3.37803304e-01 -2.03380728e+00
-3.01194996e-01 -1.17585981e+00 -1.00260198e-01 5.64327538e-01
-2.25605723e-02 -9.27713752e-01 3.62833172e-01 -4.39248472e-01
-4.54449654e-01 -1.05365348e+00 -8.55507076e-01 -1.03409886e+00
2.94524610e-01 5.70318580e-01 1.65972244e-02 4.73293006e-01
2.16674849e-01 4.78947967e-01 -2.88614810e-01 5.94654560e-01
1.43364355e-01 1.41215608e-01 1.11894572e+00 -1.14151502e+00
-7.27448344e-01 2.27998063e-01 -2.54536450e-01 -5.99496067e-01
1.46528840e-01 -4.89588797e-01 1.87205911e-01 -1.11913109e+00
-1.04411948e+00 -3.44405413e-01 1.51561603e-01 5.60187876e-01
2.96656370e-01 -5.36633492e-01 3.05042595e-01 -8.24323744e-02
2.48769522e-01 5.05231202e-01 1.65010333e+00 2.00805217e-01
-4.77493286e-01 4.38125908e-01 -4.61263862e-03 5.71620345e-01
1.08361077e+00 -3.90427291e-01 -8.21952939e-01 -7.63753131e-02
2.90338427e-01 7.48994946e-01 2.43836313e-01 -1.77164328e+00
9.49617252e-02 -1.56190664e-01 2.46844009e-01 -1.54664963e-01
4.01134104e-01 -1.05623400e+00 2.11050600e-01 1.47727180e+00
3.82493794e-01 6.78852797e-01 6.14962220e-01 6.17955446e-01
-1.37987852e-01 -2.56822258e-01 9.26459491e-01 -1.98272094e-01
-4.38859403e-01 -2.52533257e-01 -8.28642845e-01 -5.33526063e-01
1.54114664e+00 -1.43741354e-01 -2.05430597e-01 -1.21044539e-01
-7.51088262e-01 2.10609555e-01 7.46397853e-01 3.42502415e-01
3.80566776e-01 -5.90654492e-01 -2.62681484e-01 4.27495778e-01
-3.95481914e-01 -2.01417014e-01 -1.56768113e-01 7.61906445e-01
-8.99162471e-01 5.65128684e-01 -1.18202913e+00 -5.17266393e-01
-1.02011263e+00 7.01642215e-01 7.85039008e-01 -4.13855724e-02
-4.33500767e-01 -9.97505058e-03 -2.16053873e-01 -4.69939619e-01
-4.39385235e-01 -7.91327834e-01 -2.24537924e-01 -1.63168088e-01
-1.70271903e-01 2.71063387e-01 1.40915126e-01 -1.04581356e-01
-7.14090466e-03 1.13718319e+00 4.65695679e-01 -4.22857672e-01
1.43251801e+00 4.40886974e-01 1.86519787e-01 2.37216398e-01
2.57329822e-01 1.09966956e-01 -1.44574404e+00 8.24876428e-01
-5.78321815e-01 -1.62912995e-01 -2.31320132e-02 -7.72876382e-01
-4.86086547e-01 5.34113765e-01 6.36304021e-01 2.44992703e-01
8.31337273e-01 -9.50751126e-01 1.70575827e-01 1.02844906e+00
9.46682751e-01 -1.15285158e+00 1.84934989e-01 1.02490830e+00
1.22313392e+00 -5.39897084e-01 2.48516709e-01 5.09868152e-02
-7.46293724e-01 1.58604228e+00 9.31670547e-01 -9.17924702e-01
4.26530302e-01 3.74875218e-01 1.83672868e-02 1.69965133e-01
-9.78283107e-01 -1.63218677e-01 -1.37085527e-01 8.74718785e-01
-2.01584697e-02 -1.01654232e-02 -5.14750600e-01 3.47197279e-02
-1.37675688e-01 -1.90721601e-01 1.04050469e+00 1.28598332e+00
-8.06902528e-01 -1.31191933e+00 -5.72416604e-01 -2.34722607e-02
1.27470884e-02 4.69824612e-01 -1.47201046e-01 1.51281404e+00
5.83493233e-01 4.87219036e-01 -1.82503924e-01 -5.44987381e-01
9.17939961e-01 -1.50498390e-01 7.56148040e-01 -8.55571568e-01
-1.05896449e+00 -1.16999723e-01 1.99488476e-02 -4.06153440e-01
-5.06510846e-02 -4.12124544e-01 -1.26984191e+00 2.16388389e-01
-3.01059932e-01 5.58245361e-01 8.52355599e-01 4.64928687e-01
3.88249427e-01 7.88414776e-01 7.26099968e-01 -1.11429930e+00
-1.23070073e+00 -9.77039337e-01 -5.42245209e-01 -3.32325339e-01
4.92888272e-01 -1.30985701e+00 -3.68679047e-01 -1.67528063e-01] | [4.804466247558594, 1.8025959730148315] |
7598a695-da87-45b3-9eb3-d6b0098f3057 | learning-utterance-level-representations | 2211.00523 | null | https://arxiv.org/abs/2211.00523v1 | https://arxiv.org/pdf/2211.00523v1.pdf | Learning utterance-level representations through token-level acoustic latents prediction for Expressive Speech Synthesis | This paper proposes an Expressive Speech Synthesis model that utilizes token-level latent prosodic variables in order to capture and control utterance-level attributes, such as character acting voice and speaking style. Current works aim to explicitly factorize such fine-grained and utterance-level speech attributes into different representations extracted by modules that operate in the corresponding level. We show that the fine-grained latent space also captures coarse-grained information, which is more evident as the dimension of latent space increases in order to capture diverse prosodic representations. Therefore, a trade-off arises between the diversity of the token-level and utterance-level representations and their disentanglement. We alleviate this issue by first capturing rich speech attributes into a token-level latent space and then, separately train a prior network that given the input text, learns utterance-level representations in order to predict the phoneme-level, posterior latents extracted during the previous step. Both qualitative and quantitative evaluations are used to demonstrate the effectiveness of the proposed approach. Audio samples are available in our demo page. | ['Pirros Tsiakoulis', 'Aimilios Chalamandaris', 'Spyros Raptis', 'Inchul Hwang', 'June Sig Sung', 'Georgia Maniati', 'Nikolaos Ellinas', 'Konstantinos Klapsas', 'Karolos Nikitaras'] | 2022-11-01 | null | null | null | null | ['expressive-speech-synthesis'] | ['speech'] | [ 1.73644304e-01 4.88227367e-01 -3.59364599e-01 -5.93558311e-01
-9.98184860e-01 -5.83160460e-01 5.81077695e-01 4.26111892e-02
1.53520554e-02 6.42774045e-01 9.72796559e-01 1.21689767e-01
-2.74846703e-03 -6.25372529e-01 -3.03191721e-01 -7.33146608e-01
1.71792842e-02 4.66171861e-01 -2.54199624e-01 4.74272668e-02
-2.05334667e-02 3.66048396e-01 -1.69572437e+00 5.52756548e-01
8.18710208e-01 8.72859180e-01 1.59790695e-01 7.45848775e-01
-1.93585292e-01 6.75128043e-01 -7.04727650e-01 1.30323082e-01
-1.03010528e-01 -3.71278644e-01 -6.19999766e-01 4.65079278e-01
1.90760836e-01 -1.70549124e-01 1.37139440e-01 9.61789548e-01
2.59229511e-01 2.59563744e-01 8.69018614e-01 -1.05698168e+00
-4.21113223e-01 1.06397617e+00 -8.57198164e-02 -2.12870851e-01
2.35992923e-01 1.73635051e-01 1.50656629e+00 -7.64276385e-01
2.68335015e-01 1.64118433e+00 2.01629713e-01 4.74524498e-01
-1.40511930e+00 -5.52523375e-01 3.34675789e-01 -2.87132472e-01
-1.00529456e+00 -7.38575041e-01 1.16045594e+00 -7.41303205e-01
8.66379440e-01 1.22548908e-01 3.82739425e-01 1.47923696e+00
-3.21930870e-02 7.81759977e-01 1.09771585e+00 -5.14168024e-01
3.08531880e-01 4.51980293e-01 3.13858688e-01 3.41586322e-01
-3.79426688e-01 1.08847208e-01 -5.77533960e-01 -1.27752140e-01
7.37780392e-01 -2.87889063e-01 -1.38004363e-01 -1.60363585e-01
-1.07131839e+00 9.11154032e-01 7.80545548e-02 5.88528574e-01
-6.80760443e-01 -1.12418100e-01 4.49513525e-01 1.16961718e-01
3.97967398e-01 5.62965870e-01 -5.01621008e-01 -3.58333051e-01
-9.57610786e-01 -5.54793067e-02 8.47716987e-01 9.01919842e-01
7.93053389e-01 4.44915742e-01 -3.91829938e-01 1.11854041e+00
5.35034776e-01 1.34143546e-01 7.88290560e-01 -1.01188862e+00
4.58782881e-01 4.53169793e-01 -1.63960382e-02 -7.73654521e-01
-2.64032513e-01 -3.51226032e-01 -8.09170544e-01 -1.04154935e-02
1.22569323e-01 -2.70188630e-01 -8.35547447e-01 2.16053438e+00
1.26464479e-02 -1.81355178e-01 2.73444235e-01 7.54082680e-01
5.99873066e-01 1.04867435e+00 4.53804404e-01 -4.93151635e-01
1.57429290e+00 -1.02836275e+00 -9.77698207e-01 -2.70178467e-01
2.84151554e-01 -5.86818576e-01 1.39269364e+00 2.18745694e-01
-1.37323141e+00 -8.20025027e-01 -1.09327269e+00 8.42179805e-02
-2.96319425e-01 3.66796821e-01 1.95695996e-01 6.51306570e-01
-8.75132382e-01 1.87361673e-01 -7.81573832e-01 -1.07395947e-01
-1.36315450e-01 2.91746378e-01 -2.39768773e-01 6.91203058e-01
-1.49220657e+00 5.40932298e-01 5.33096910e-01 -7.62699172e-02
-8.43025148e-01 -5.47675669e-01 -9.57801878e-01 6.17083132e-01
1.95021555e-01 -5.03853083e-01 1.28729475e+00 -7.58059084e-01
-2.15507674e+00 4.90494013e-01 -1.21415369e-01 -2.22592011e-01
9.41227004e-02 -1.05534457e-01 -1.44744501e-01 1.13503397e-01
-1.08276971e-01 6.58715665e-01 1.12420750e+00 -1.26732147e+00
-5.22281110e-01 -2.70847976e-01 7.94171318e-02 2.82336056e-01
-6.25200510e-01 5.22321984e-02 -9.86755565e-02 -8.01375508e-01
2.39074882e-02 -7.32288241e-01 -6.94754049e-02 -5.72257638e-01
-3.84803981e-01 -3.76239836e-01 6.53996289e-01 -4.72572178e-01
1.24646163e+00 -2.25796556e+00 6.92429781e-01 -1.05616853e-01
1.38824880e-01 -1.01385251e-01 -1.03347048e-01 3.90810162e-01
1.10166885e-01 2.57925510e-01 -6.62632361e-02 -8.87539446e-01
3.22941005e-01 3.33454907e-01 -4.37507570e-01 -2.01489925e-02
3.83757621e-01 5.63626289e-01 -7.23030031e-01 -3.61299545e-01
3.96911383e-01 5.99165797e-01 -6.87093735e-01 4.83519584e-01
-4.27585989e-01 7.55166829e-01 -4.87731010e-01 2.96905696e-01
3.02729845e-01 1.85431331e-01 1.64708897e-01 -2.24773258e-01
-2.17570364e-01 8.96648347e-01 -9.94974315e-01 1.53625369e+00
-9.67408657e-01 4.09429073e-01 4.42280412e-01 -6.88793600e-01
1.12790155e+00 8.62290859e-01 2.21961692e-01 -2.96838820e-01
1.03531159e-01 5.45287766e-02 -1.25648201e-01 -2.45944619e-01
7.20968366e-01 -5.10235608e-01 -6.56988561e-01 3.45635056e-01
3.54135573e-01 -1.98045269e-01 -3.09242368e-01 -1.95697963e-01
5.85674226e-01 -1.65223405e-01 4.60990936e-01 -3.54347795e-01
7.42751956e-01 -6.32528841e-01 6.12042308e-01 4.19669122e-01
-2.32712328e-01 3.72335106e-01 6.99814320e-01 -2.98741553e-02
-8.27048898e-01 -1.04554558e+00 -1.37491167e-01 1.45552564e+00
-3.00592542e-01 -3.76671702e-01 -8.21392953e-01 -4.80829716e-01
-2.25425005e-01 9.63131487e-01 -6.27127826e-01 -1.00424364e-01
-5.42012155e-01 -9.74635333e-02 5.20619631e-01 5.88148177e-01
-1.44019797e-02 -1.49458075e+00 -4.54507142e-01 3.83257985e-01
-2.62742579e-01 -1.17717898e+00 -3.32204163e-01 4.84523982e-01
-6.54221177e-01 -1.72083229e-01 -4.69942480e-01 -5.39035738e-01
2.35245362e-01 -5.28477550e-01 9.42865789e-01 -5.58763742e-01
2.40441829e-01 4.66510057e-02 -4.89238918e-01 -7.39388540e-02
-9.18033540e-01 4.44110125e-01 2.36245662e-01 2.38848254e-01
2.36268580e-01 -8.38882506e-01 -1.62821338e-01 8.53383988e-02
-7.83743083e-01 5.90942502e-02 5.75746715e-01 9.25127208e-01
3.59110922e-01 5.84894195e-02 9.73209262e-01 -7.10992634e-01
8.03609192e-01 -4.11511421e-01 -4.58114207e-01 -3.79731171e-02
-6.39588758e-02 3.29748213e-01 7.32801914e-01 -5.07154286e-01
-1.25247693e+00 -1.24736726e-02 -4.91626889e-01 -4.66244578e-01
-8.21729422e-01 5.79310536e-01 -7.74989545e-01 8.25743198e-01
2.46317193e-01 1.39727294e-01 -2.51563609e-01 -5.84981263e-01
6.94871664e-01 1.05972135e+00 5.23755729e-01 -8.35918367e-01
4.96814132e-01 -8.33009928e-02 -4.50620919e-01 -1.21501982e+00
-9.39489603e-01 -3.42701435e-01 -1.02342296e+00 5.50169535e-02
1.24096513e+00 -8.44250977e-01 -5.82072377e-01 2.95578063e-01
-1.27176559e+00 -1.81215838e-01 -4.96741742e-01 5.40360689e-01
-1.00452089e+00 -1.60459541e-02 -8.22778106e-01 -1.10069358e+00
-1.41787589e-01 -1.40975821e+00 1.36199093e+00 2.23293882e-02
-6.61771595e-01 -1.07965231e+00 1.08481735e-01 3.24923009e-01
2.76932508e-01 2.41275698e-01 1.28400052e+00 -8.47457170e-01
-4.07202750e-01 1.11573912e-01 1.53427318e-01 3.16551328e-01
4.80051279e-01 -1.15569621e-01 -1.53937685e+00 -1.22625977e-01
3.57816666e-01 -3.82429242e-01 5.84589243e-01 4.69648749e-01
8.17846775e-01 -7.72104323e-01 1.72421232e-01 4.22225356e-01
8.74233544e-01 2.08267197e-01 2.59317338e-01 -1.44988343e-01
7.49181151e-01 9.52577233e-01 3.97147775e-01 3.70363683e-01
3.25822145e-01 9.27488089e-01 6.91842213e-02 7.97495767e-02
-4.08199988e-02 -2.93607086e-01 5.77494025e-01 1.54496825e+00
1.99846253e-01 -1.24628238e-01 -5.67736983e-01 5.08533478e-01
-1.69289887e+00 -9.07496810e-01 5.65936983e-01 1.87954545e+00
1.11410213e+00 3.65250766e-01 2.33751729e-01 2.12694108e-01
4.95092243e-01 4.22627598e-01 -3.12072545e-01 -8.35162818e-01
1.50416940e-01 -1.52206063e-01 -1.50621176e-01 9.21488941e-01
-1.06334269e+00 1.26724625e+00 5.66462374e+00 6.84678972e-01
-1.21385550e+00 3.89478169e-02 4.23637509e-01 2.17834905e-01
-5.50835252e-01 -1.86155125e-01 -8.17530930e-01 2.90691972e-01
1.23538637e+00 -5.87597722e-03 1.81381404e-01 8.51655841e-01
4.26671475e-01 4.94051099e-01 -1.47535062e+00 5.14081001e-01
-1.49754643e-01 -1.09513891e+00 1.61054149e-01 2.83415258e-01
4.93044108e-01 -3.95855486e-01 2.88188905e-01 6.64230347e-01
2.01433092e-01 -1.05802429e+00 8.76494050e-01 4.64517891e-01
7.40231216e-01 -7.76102364e-01 4.84957159e-01 6.07504249e-01
-1.29961503e+00 -8.02200567e-03 -1.32106751e-01 -1.13168068e-01
3.14577013e-01 1.15584508e-01 -1.07657397e+00 3.65479499e-01
3.69186461e-01 4.33063477e-01 -8.82288218e-02 2.14085072e-01
-3.97350371e-01 7.61199713e-01 3.11463792e-02 1.00609146e-01
3.40033501e-01 -9.71171111e-02 7.34040141e-01 1.48630142e+00
1.33080631e-01 2.54788827e-02 4.29237127e-01 1.19515681e+00
2.07746979e-02 -1.29278265e-02 -4.20104831e-01 -4.18198496e-01
7.72559702e-01 1.22726309e+00 -4.18559790e-01 -3.35168153e-01
-2.61299998e-01 6.69374824e-01 2.96233535e-01 4.21178877e-01
-5.47884524e-01 -6.93216696e-02 1.00136554e+00 -5.23369461e-02
3.24550807e-01 -4.55487520e-01 -4.71667111e-01 -1.09716189e+00
-3.28723133e-01 -9.43240166e-01 -1.58697981e-02 -4.35907096e-01
-1.37178826e+00 1.06526303e+00 7.75989965e-02 -9.82558489e-01
-9.56207395e-01 -3.84501010e-01 -5.04023015e-01 1.25491977e+00
-1.24840820e+00 -1.21407163e+00 7.54121542e-02 3.35397929e-01
1.19914746e+00 -2.65708238e-01 1.23997426e+00 -1.13609225e-01
-6.05329454e-01 7.77866602e-01 -1.70424238e-01 1.78740755e-01
4.24383998e-01 -1.38309300e+00 2.15536773e-01 5.25563121e-01
2.00738475e-01 7.14121163e-01 7.66049385e-01 -3.34699690e-01
-9.03326571e-01 -8.44369471e-01 1.07307673e+00 -4.68692273e-01
7.51858234e-01 -7.57346034e-01 -1.01465964e+00 6.37348294e-01
2.88073778e-01 -4.63983774e-01 8.37239921e-01 4.45708513e-01
-4.02954310e-01 3.63497734e-02 -8.88170123e-01 4.87530500e-01
3.44786674e-01 -1.05960786e+00 -1.04810596e+00 -3.16719830e-01
1.20902336e+00 -2.56949682e-02 -1.07104588e+00 3.87024134e-01
5.96084058e-01 -8.73610735e-01 8.15927565e-01 -3.83878320e-01
3.67216259e-01 5.19686900e-02 -5.07885277e-01 -1.45505393e+00
-3.32716614e-01 -6.13676548e-01 -1.75384849e-01 1.71009123e+00
5.32754123e-01 -4.72054511e-01 5.97200394e-01 4.14717793e-01
-2.96205044e-01 -6.81243598e-01 -8.56769860e-01 -4.75409716e-01
3.44047964e-01 -3.00170422e-01 7.87884414e-01 8.05692315e-01
2.30396733e-01 9.12824810e-01 -4.65718061e-01 3.73844147e-01
2.98262239e-01 2.29208514e-01 5.37950754e-01 -1.39663696e+00
-4.64098662e-01 -6.06349468e-01 -3.39461505e-01 -1.10371661e+00
5.64682424e-01 -5.85456967e-01 3.87172818e-01 -1.09768879e+00
-1.28989518e-01 -3.84682626e-01 -4.40318912e-01 4.16143686e-01
-4.59893718e-02 -2.87594557e-01 3.75257820e-01 2.28197470e-01
-1.12608664e-01 9.13987637e-01 1.11721349e+00 -5.42142354e-02
-5.75969636e-01 2.00195596e-01 -5.86028337e-01 9.77813423e-01
7.61428237e-01 -3.15942019e-01 -5.99749148e-01 -1.86857104e-01
-3.93184990e-01 5.88993907e-01 -2.24389490e-02 -8.63596976e-01
3.36849764e-02 -2.57868379e-01 1.42958328e-01 -5.54581821e-01
9.87755418e-01 -8.00902903e-01 -2.99320757e-01 1.47598963e-02
-9.71473575e-01 -3.46525460e-01 1.00815468e-01 3.71833742e-01
-6.82323396e-01 -1.53082311e-01 7.96361685e-01 3.33248787e-02
-2.39191934e-01 9.24575254e-02 -7.06688404e-01 -2.35886678e-01
7.12693691e-01 -2.66564935e-01 3.34799327e-02 -6.12741888e-01
-1.23071218e+00 -2.53005158e-02 1.42071396e-01 7.20869660e-01
3.26790273e-01 -1.32246482e+00 -6.42528415e-01 6.81400061e-01
6.88450262e-02 -2.12168828e-01 2.36110896e-01 4.95339513e-01
2.23677158e-01 5.91148019e-01 -3.33606243e-01 -7.84223318e-01
-9.92699623e-01 5.22208333e-01 2.81993449e-02 -4.21614528e-01
-3.38316381e-01 8.05825949e-01 5.98472893e-01 -6.17100060e-01
5.53786933e-01 -7.00623751e-01 -4.88899559e-01 4.96185094e-01
2.78710395e-01 2.85083465e-02 -2.42155343e-01 -1.00660038e+00
-2.57904381e-01 4.07839566e-01 5.20513579e-02 -4.93781477e-01
1.17784178e+00 -5.26221216e-01 1.01025872e-01 1.07511020e+00
1.19954085e+00 3.64271939e-01 -1.44317937e+00 -1.80365071e-01
1.35372460e-01 -6.98203593e-02 -1.00393079e-01 -5.73778212e-01
-4.47064131e-01 1.37572682e+00 1.53556332e-01 4.79082197e-01
1.08308220e+00 2.57933110e-01 5.42358756e-01 4.09725979e-02
8.19701999e-02 -9.32640553e-01 2.82057941e-01 8.52998376e-01
1.04108262e+00 -8.43025446e-01 -6.35265529e-01 -4.30124432e-01
-8.31552267e-01 9.93477523e-01 5.66058457e-01 4.54460131e-03
5.68212926e-01 4.70764518e-01 1.89313367e-01 7.99112767e-02
-1.04292834e+00 -9.78061184e-03 4.61780369e-01 4.91528243e-01
7.28333235e-01 5.39582849e-01 1.42799765e-01 1.07471168e+00
-6.96167707e-01 -5.54012537e-01 2.69323766e-01 3.55564296e-01
-7.51087248e-01 -1.36929560e+00 -2.61580914e-01 -1.34604797e-01
-3.28791171e-01 -3.90033908e-02 -3.63611370e-01 4.12288547e-01
1.33161107e-02 1.03004360e+00 2.20814377e-01 -3.97436947e-01
2.03312635e-01 4.08428222e-01 9.49409157e-02 -1.05781555e+00
-5.21800101e-01 4.70775038e-01 2.74152249e-01 -3.63914490e-01
-2.69257009e-01 -4.85944360e-01 -1.09602237e+00 3.45034033e-01
-1.18226059e-01 4.00399357e-01 5.41503370e-01 1.01889741e+00
1.18258752e-01 8.59412491e-01 8.24030817e-01 -1.24246597e+00
-7.74350464e-01 -1.30739260e+00 -5.56614757e-01 2.14582860e-01
7.62521684e-01 -5.54071784e-01 -4.92521971e-01 1.48552701e-01] | [14.89428424835205, 6.637550354003906] |
b2a697ca-4751-4d74-931d-3bc94d6d331e | comformer-continual-learning-in-semantic-and | 2211.13999 | null | https://arxiv.org/abs/2211.13999v1 | https://arxiv.org/pdf/2211.13999v1.pdf | CoMFormer: Continual Learning in Semantic and Panoptic Segmentation | Continual learning for segmentation has recently seen increasing interest. However, all previous works focus on narrow semantic segmentation and disregard panoptic segmentation, an important task with real-world impacts. %a In this paper, we present the first continual learning model capable of operating on both semantic and panoptic segmentation. Inspired by recent transformer approaches that consider segmentation as a mask-classification problem, we design CoMFormer. Our method carefully exploits the properties of transformer architectures to learn new classes over time. Specifically, we propose a novel adaptive distillation loss along with a mask-based pseudo-labeling technique to effectively prevent forgetting. To evaluate our approach, we introduce a novel continual panoptic segmentation benchmark on the challenging ADE20K dataset. Our CoMFormer outperforms all the existing baselines by forgetting less old classes but also learning more effectively new classes. In addition, we also report an extensive evaluation in the large-scale continual semantic segmentation scenario showing that CoMFormer also significantly outperforms state-of-the-art methods. | ['Arthur Douillard', 'Matthieu Cord', 'Fabio Cermelli'] | 2022-11-25 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Cermelli_CoMFormer_Continual_Learning_in_Semantic_and_Panoptic_Segmentation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Cermelli_CoMFormer_Continual_Learning_in_Semantic_and_Panoptic_Segmentation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['panoptic-segmentation', 'continual-semantic-segmentation'] | ['computer-vision', 'computer-vision'] | [ 4.83869940e-01 -6.55144155e-02 -1.32414877e-01 -5.48859775e-01
-6.97923124e-01 -8.90844762e-01 6.12008989e-01 1.44353688e-01
-6.75476372e-01 5.52030742e-01 -2.38218278e-01 -3.80682588e-01
-4.17695306e-02 -7.98883200e-01 -9.51101363e-01 -5.98820925e-01
1.34600699e-01 7.11185515e-01 7.72932112e-01 1.05128184e-01
2.12108567e-02 3.01907361e-01 -1.36961842e+00 1.87169015e-01
1.21961308e+00 1.07036674e+00 3.40256453e-01 3.38745713e-01
-1.20115325e-01 4.17760700e-01 -5.18226206e-01 -3.92787308e-01
6.13350034e-01 -1.79491580e-01 -9.82528567e-01 4.79484081e-01
9.71432388e-01 -2.92086359e-02 -1.44928604e-01 1.02471268e+00
2.05236614e-01 3.16508621e-01 4.30019349e-01 -8.72747004e-01
-4.60630894e-01 7.94364810e-01 -7.91526675e-01 3.93901765e-01
-4.94097084e-01 3.80531788e-01 1.20272350e+00 -7.02802420e-01
4.87011224e-01 1.29168653e+00 8.80636454e-01 3.41101319e-01
-1.42131853e+00 -7.98604131e-01 9.01815295e-01 8.19170102e-02
-1.14679909e+00 -4.34809588e-02 6.66480422e-01 -1.54449135e-01
8.16244602e-01 2.21112832e-01 8.10218632e-01 9.65718150e-01
-5.29743433e-02 1.29288483e+00 1.38758934e+00 -2.90350825e-01
3.10965657e-01 1.26048438e-02 4.73710388e-01 7.69125760e-01
9.68352240e-03 1.17984638e-02 -2.17972919e-01 3.42901051e-01
3.40322614e-01 -4.34587896e-02 -5.54116070e-02 -4.70650405e-01
-9.40348625e-01 6.98685467e-01 5.77082515e-01 3.05243105e-01
-1.16804302e-01 3.26541275e-01 3.14369053e-01 1.94172621e-01
9.68360543e-01 5.07969558e-01 -9.24424887e-01 1.42069966e-01
-1.40361130e+00 1.78007469e-01 5.95297873e-01 6.99235678e-01
8.82558703e-01 1.26943246e-01 -3.99675369e-01 1.10873473e+00
4.14889902e-02 4.79825467e-01 3.01278651e-01 -9.11478996e-01
2.22378775e-01 4.05914426e-01 -1.69119477e-01 -3.59173059e-01
-4.05621797e-01 -8.83100092e-01 -4.91469562e-01 -1.11839874e-02
3.45111161e-01 -1.60331726e-01 -1.70032513e+00 1.70645034e+00
4.59637731e-01 7.18474150e-01 -1.85227424e-01 4.97582018e-01
3.24568212e-01 6.96505308e-01 3.72827470e-01 -5.12503125e-02
1.23232663e+00 -1.52002597e+00 -5.14590085e-01 -5.08488178e-01
1.25271901e-01 -7.39888906e-01 1.33079565e+00 8.01987588e-01
-7.01210439e-01 -5.90455532e-01 -9.75262463e-01 -3.12078726e-02
-6.06920421e-01 -8.80111009e-02 8.31816375e-01 7.64696121e-01
-1.13519478e+00 8.36813509e-01 -1.09762096e+00 -3.93126935e-01
6.62850976e-01 1.88547894e-01 2.51407504e-01 -9.17063057e-02
-1.06551909e+00 6.42964661e-01 7.79708207e-01 -9.48780477e-02
-1.02427769e+00 -1.00921834e+00 -6.28157377e-01 3.30368467e-02
8.59226584e-01 -5.64908981e-01 1.54590130e+00 -9.26611483e-01
-1.47845197e+00 7.32905865e-01 -1.15447398e-02 -1.03637183e+00
6.33572638e-01 -6.60185874e-01 -2.78094709e-01 1.48790061e-01
1.66445464e-01 1.20164502e+00 1.12953520e+00 -1.35676134e+00
-1.06928337e+00 -1.71704337e-01 2.63975650e-01 2.61596978e-01
-5.16408443e-01 -7.34933138e-01 -7.74959266e-01 -1.00366473e+00
1.20231844e-01 -9.65774596e-01 -3.32008481e-01 -4.39686552e-02
-4.45360452e-01 -3.23189437e-01 1.20408297e+00 -4.35322523e-01
1.13589597e+00 -2.14952612e+00 1.60549551e-01 -1.58772722e-01
1.74764663e-01 5.28454423e-01 -1.43502951e-01 -1.96066409e-01
3.78309414e-02 8.95630494e-02 -9.19028878e-01 -8.59257281e-01
7.45684505e-02 5.17952502e-01 -6.74606621e-01 1.83667704e-01
8.01143516e-03 9.29651737e-01 -8.96252930e-01 -2.56904900e-01
4.30743575e-01 3.47381890e-01 -5.72187543e-01 -1.31916448e-01
-8.22729528e-01 2.51486301e-01 -2.01564282e-01 7.46660888e-01
8.50513220e-01 -3.44919562e-01 -1.93537056e-01 -1.43298954e-01
-2.70970941e-01 -8.18346143e-02 -7.11747468e-01 2.04229593e+00
-5.16071677e-01 4.94889617e-01 -6.98208958e-02 -1.11509538e+00
5.56721985e-01 -1.08080611e-01 3.95532489e-01 -8.93080056e-01
-1.12339348e-01 3.19755822e-01 -3.55142534e-01 1.05241887e-01
6.67559624e-01 -1.98848918e-01 2.88417488e-02 2.35992238e-01
1.88392013e-01 -5.67080259e-01 4.67938840e-01 2.15977177e-01
7.75050938e-01 2.86245763e-01 -1.92032337e-01 -5.30932307e-01
3.56179863e-01 6.51252717e-02 8.17171991e-01 1.16129494e+00
-2.86748946e-01 7.10595608e-01 2.83624709e-01 -3.30369294e-01
-6.86947167e-01 -1.28685558e+00 -3.24100614e-01 1.13753152e+00
4.00652647e-01 -2.34139621e-01 -8.23240161e-01 -1.15400028e+00
1.73665866e-01 1.13704288e+00 -7.00403512e-01 5.28998021e-03
-7.05829620e-01 -1.33619261e+00 2.92535722e-01 4.77737755e-01
9.52153563e-01 -9.03738499e-01 -4.59602565e-01 3.72147411e-01
-5.49163334e-02 -1.10681760e+00 -6.24544561e-01 3.09036434e-01
-1.09055209e+00 -1.00422955e+00 -8.68802905e-01 -7.15701699e-01
1.89185500e-01 3.27932537e-01 1.18096817e+00 -3.71329337e-01
-5.18343270e-01 5.90369523e-01 -2.94997185e-01 -3.69556338e-01
4.11327375e-04 7.17911303e-01 -3.52306157e-01 2.36043945e-01
5.56960590e-02 -5.94937384e-01 -7.84015417e-01 1.31005943e-01
-1.06894326e+00 1.28713325e-01 5.45209587e-01 6.02362692e-01
9.44658220e-01 2.07759261e-01 5.19238472e-01 -1.54958546e+00
1.23869054e-01 -3.73385757e-01 -9.42610621e-01 4.46984291e-01
-1.05813313e+00 3.10938172e-02 5.29743731e-01 -3.98380041e-01
-1.29873669e+00 7.04178363e-02 -3.12495559e-01 -3.72357309e-01
2.62527727e-03 9.81465951e-02 8.05960670e-02 1.88559443e-02
3.68196160e-01 2.15366170e-01 -5.89074969e-01 -9.07068551e-01
7.15000868e-01 2.01397941e-01 5.45143008e-01 -6.92154646e-01
7.81099558e-01 8.02228272e-01 -3.07465911e-01 -8.32796097e-01
-1.64041352e+00 -6.19334996e-01 -6.64792120e-01 3.56840380e-02
8.44502985e-01 -9.42500234e-01 -1.72534347e-01 1.02512312e+00
-7.97279298e-01 -9.32255149e-01 -7.50715613e-01 2.21466515e-02
-3.73949200e-01 4.73502040e-01 -6.19632602e-01 -4.30750728e-01
-5.07003367e-01 -1.12116587e+00 1.22054791e+00 2.79207349e-01
2.37891853e-01 -1.09924471e+00 4.10866104e-02 3.71979624e-01
5.42654395e-01 1.49528190e-01 7.42772877e-01 -5.45672596e-01
-9.01062310e-01 3.35556060e-01 -3.86971354e-01 6.27039254e-01
2.27502242e-01 -3.05402517e-01 -1.12904835e+00 -5.05534649e-01
-4.98651192e-02 -3.92072916e-01 1.82933271e+00 3.96528989e-01
1.45020700e+00 -1.11435510e-01 -4.45228219e-01 9.59412515e-01
1.41112995e+00 1.96125776e-01 4.42040771e-01 3.40678543e-01
8.59597445e-01 3.49952012e-01 7.11458027e-01 1.15880147e-01
3.71205151e-01 3.46450210e-01 3.59633207e-01 -3.69998246e-01
-5.77713013e-01 -1.30472153e-01 -2.85210684e-02 6.39030814e-01
5.05249441e-01 -5.77724099e-01 -9.18028176e-01 8.43929410e-01
-1.87000620e+00 -4.00947750e-01 2.19528750e-01 1.92400455e+00
9.93327737e-01 4.94465947e-01 -1.92030862e-01 -3.29990536e-02
5.82423329e-01 6.29350483e-01 -9.59584653e-01 5.39722107e-02
-2.98857361e-01 8.27035844e-01 7.32446671e-01 7.05484629e-01
-1.72860026e+00 1.73122954e+00 6.09914684e+00 1.17361927e+00
-1.17682588e+00 4.86901343e-01 8.68955612e-01 -2.16249116e-02
-1.99456334e-01 1.73134029e-01 -9.70934331e-01 2.72585601e-01
5.79574943e-01 1.52887166e-01 2.83317506e-01 6.88475907e-01
-2.29172155e-01 -2.99202144e-01 -8.19525599e-01 6.67304516e-01
-6.12805150e-02 -1.19195807e+00 2.60512382e-01 -3.65785092e-01
1.00790691e+00 3.87667328e-01 4.61721420e-01 5.78324974e-01
4.93925363e-01 -6.59098625e-01 7.92953074e-01 2.96645612e-01
6.23244822e-01 -6.06240988e-01 2.55841225e-01 3.37914191e-03
-1.24879682e+00 -1.02560684e-01 -8.74067619e-02 3.40992481e-01
2.88162649e-01 9.30821180e-01 -6.24863029e-01 4.49907124e-01
7.72676289e-01 1.22147655e+00 -9.78858113e-01 1.31855953e+00
-4.64619875e-01 1.17114449e+00 -5.10431647e-01 5.38277745e-01
7.62272716e-01 -1.73037291e-01 4.97784764e-01 1.33048439e+00
3.90512645e-02 -2.15619773e-01 4.62792933e-01 7.60159135e-01
-2.35316887e-01 -7.64454827e-02 -3.19488943e-02 1.00955600e-02
2.28189662e-01 1.09619665e+00 -1.28156161e+00 -6.21129990e-01
-2.20197871e-01 1.24230170e+00 1.75827503e-01 4.27893102e-01
-8.94756973e-01 -2.05506027e-01 5.65402031e-01 -7.68905133e-02
7.86044300e-01 -1.44548506e-01 -3.64618123e-01 -1.23113453e+00
-2.67007113e-01 -5.78595400e-01 6.84108973e-01 -3.74053895e-01
-1.17628014e+00 4.71860170e-01 7.85675794e-02 -5.98439515e-01
4.08843070e-01 -4.34030265e-01 -3.49838853e-01 3.53881180e-01
-2.05790186e+00 -1.40682638e+00 -5.17215095e-02 3.92135620e-01
1.19923186e+00 1.66650638e-01 3.56091738e-01 4.55427319e-01
-5.37728488e-01 4.32802916e-01 2.26748109e-01 -2.52167344e-01
7.72853851e-01 -1.59300351e+00 7.41629422e-01 1.09201944e+00
3.98562431e-01 3.01377267e-01 6.54417038e-01 -6.66729212e-01
-6.63870335e-01 -1.59167743e+00 5.34198225e-01 -3.17254305e-01
6.73544824e-01 -5.53809285e-01 -1.05200756e+00 9.31015909e-01
2.51499146e-01 -5.03258072e-02 1.38369322e-01 1.78618193e-01
-4.51986909e-01 -5.95540047e-01 -9.69219804e-01 3.58706385e-01
1.16559529e+00 -2.03860417e-01 -6.53317511e-01 6.15975201e-01
1.29961336e+00 -4.16754723e-01 -3.03052962e-01 7.51594722e-01
1.98248118e-01 -8.14283967e-01 1.08932126e+00 -4.35984060e-02
-7.63455108e-02 -2.77255207e-01 1.30119711e-01 -1.30827987e+00
-1.42505541e-01 -6.44462764e-01 3.86531577e-02 1.20458734e+00
3.84893358e-01 -7.63067245e-01 7.96795428e-01 4.08810331e-03
-3.81665289e-01 -5.16217411e-01 -9.03455675e-01 -1.00458920e+00
3.22684318e-01 -4.70591128e-01 3.29582244e-01 7.70736933e-01
-8.41094255e-01 3.51203650e-01 -3.57931972e-01 1.04971461e-01
8.54269743e-01 4.11981761e-01 3.30643892e-01 -1.20271158e+00
-4.72330153e-01 -5.41155696e-01 2.42465958e-01 -1.42467368e+00
1.76735476e-01 -8.79663944e-01 2.38422737e-01 -1.35425341e+00
9.94222611e-02 -6.55208886e-01 -5.99666417e-01 7.44626284e-01
-2.76764989e-01 3.97596180e-01 2.13435814e-01 2.33386364e-02
-9.46760058e-01 8.21505606e-01 1.36985421e+00 -5.66364944e-01
-2.83774048e-01 1.44952387e-01 -5.61270237e-01 7.11004913e-01
8.17347765e-01 -5.49004853e-01 -7.27629721e-01 -7.26442575e-01
-1.87833235e-01 -6.58092618e-01 2.38377035e-01 -1.17976832e+00
1.27796948e-01 7.23461285e-02 -1.17331579e-01 -8.03990841e-01
3.84106517e-01 -5.67660809e-01 -2.05290884e-01 5.14573395e-01
-1.22880094e-01 -2.70639002e-01 6.04030073e-01 8.94870460e-01
-1.40091375e-01 -6.52226061e-02 1.11548936e+00 -3.21410656e-01
-1.15039766e+00 5.57906389e-01 -2.17057750e-01 4.46030259e-01
9.71440613e-01 1.50858417e-01 -2.23183662e-01 3.10380876e-01
-9.83574271e-01 6.20629489e-01 3.04951996e-01 5.64995229e-01
2.99929470e-01 -5.59353709e-01 -4.57284212e-01 -5.28528774e-03
-1.05275042e-01 3.79132509e-01 2.41127372e-01 5.93478799e-01
-5.79008281e-01 3.82462502e-01 1.70405760e-01 -7.84615159e-01
-9.89831805e-01 5.72909474e-01 3.68537873e-01 -5.25384486e-01
-8.40348184e-01 1.14386392e+00 6.44809544e-01 -6.12453461e-01
4.30399776e-01 -5.21316528e-01 -1.47073299e-01 3.83732468e-01
1.24275260e-01 2.35221207e-01 1.00518614e-01 -9.61588621e-02
-1.54896855e-01 5.55855811e-01 -5.61558068e-01 -1.91559643e-01
1.45389593e+00 -2.16582596e-01 -4.29324098e-02 6.88046455e-01
8.44577789e-01 -3.21569204e-01 -1.79328752e+00 -6.22245312e-01
3.04770350e-01 -2.98393846e-01 1.16081983e-01 -1.23835599e+00
-1.38493514e+00 1.05532503e+00 7.89145947e-01 -3.60343605e-02
1.19836318e+00 -1.35359600e-01 1.26836705e+00 2.97447383e-01
2.70888031e-01 -1.40796423e+00 2.52550066e-01 7.85487294e-01
5.39767861e-01 -1.20504594e+00 -9.67657119e-02 -5.80784142e-01
-2.94732571e-01 7.09360003e-01 6.94915414e-01 -3.72656211e-02
8.26317966e-01 7.99402669e-02 1.29304603e-01 -9.11929682e-02
-5.30755639e-01 -5.33084989e-01 2.13038176e-01 1.76036805e-01
-8.70688409e-02 2.25873888e-02 -2.28253618e-01 2.18380287e-01
-8.15479308e-02 -7.47562051e-02 2.36072347e-01 7.97237694e-01
-5.96761703e-01 -1.24387658e+00 -4.48906608e-02 3.88470083e-01
-3.40179324e-01 -3.93421948e-01 -2.12761030e-01 7.65463710e-01
3.64384770e-01 6.26894474e-01 2.58963019e-01 -3.31556611e-02
2.94256270e-01 2.17036113e-01 4.33300614e-01 -8.30826044e-01
-5.90828478e-01 2.12008879e-01 -2.13885963e-01 -4.15125728e-01
-5.06273329e-01 -8.25626373e-01 -1.00433266e+00 1.99602898e-02
-2.09370628e-01 4.88001034e-02 3.44730943e-01 9.82344031e-01
1.47747949e-01 7.80152261e-01 3.36732209e-01 -5.02635598e-01
-3.94224554e-01 -8.69342625e-01 -5.93022466e-01 3.39642674e-01
4.40452695e-01 -6.58711433e-01 -2.95058250e-01 1.03650078e-01] | [9.409417152404785, 1.880855679512024] |
bd62065c-97c9-49ee-8fdb-92a13c127313 | monocular-3d-object-detection-leveraging | 1904.01690 | null | http://arxiv.org/abs/1904.01690v1 | http://arxiv.org/pdf/1904.01690v1.pdf | Monocular 3D Object Detection Leveraging Accurate Proposals and Shape Reconstruction | We present MonoPSR, a monocular 3D object detection method that leverages
proposals and shape reconstruction. First, using the fundamental relations of a
pinhole camera model, detections from a mature 2D object detector are used to
generate a 3D proposal per object in a scene. The 3D location of these
proposals prove to be quite accurate, which greatly reduces the difficulty of
regressing the final 3D bounding box detection. Simultaneously, a point cloud
is predicted in an object centered coordinate system to learn local scale and
shape information. However, the key challenge is how to exploit shape
information to guide 3D localization. As such, we devise aggregate losses,
including a novel projection alignment loss, to jointly optimize these tasks in
the neural network to improve 3D localization accuracy. We validate our method
on the KITTI benchmark where we set new state-of-the-art results among
published monocular methods, including the harder pedestrian and cyclist
classes, while maintaining efficient run-time. | ['Steven L. Waslander', 'Jason Ku', 'Alex D. Pon'] | 2019-04-02 | monocular-3d-object-detection-leveraging-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Ku_Monocular_3D_Object_Detection_Leveraging_Accurate_Proposals_and_Shape_Reconstruction_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Ku_Monocular_3D_Object_Detection_Leveraging_Accurate_Proposals_and_Shape_Reconstruction_CVPR_2019_paper.pdf | cvpr-2019-6 | ['vehicle-pose-estimation'] | ['computer-vision'] | [-2.26994529e-01 -1.71195835e-01 -1.44203469e-01 -2.50486612e-01
-7.68140256e-01 -7.99369335e-01 6.22007370e-01 -9.01184883e-03
-6.33365154e-01 2.79495627e-01 -2.54920036e-01 -2.41844460e-01
5.76609850e-01 -3.25658113e-01 -1.03635192e+00 -4.59750533e-01
8.57195929e-02 5.48952222e-01 5.63622653e-01 2.44179308e-01
3.03920031e-01 1.01118529e+00 -1.31224775e+00 -8.43188092e-02
2.59257495e-01 1.15327990e+00 2.01524258e-01 6.04191601e-01
3.91860455e-01 3.28672230e-01 -2.27331758e-01 -3.02595615e-01
8.34646583e-01 1.93224445e-01 -7.04850778e-02 2.26412892e-01
9.09653008e-01 -5.71669459e-01 -4.54035342e-01 8.94384325e-01
5.95716476e-01 6.24708049e-02 8.05603802e-01 -1.16760039e+00
-3.95312190e-01 -3.92482895e-03 -1.05236471e+00 1.17461421e-01
2.92379975e-01 4.50126112e-01 9.19065118e-01 -1.44265890e+00
6.20790541e-01 1.36430120e+00 9.68878984e-01 3.12100112e-01
-1.27275181e+00 -5.83730459e-01 3.29617232e-01 -1.29356146e-01
-1.55317271e+00 -4.02658314e-01 7.83226550e-01 -4.40718919e-01
1.08972943e+00 -9.28039327e-02 7.34831154e-01 9.14514601e-01
1.49508923e-01 8.89271677e-01 9.01349008e-01 -4.11358863e-01
1.87549710e-01 1.97686493e-01 -8.84526223e-02 8.08900118e-01
5.05534351e-01 3.73553634e-01 -5.15340269e-01 -5.33428304e-02
7.90950239e-01 9.64628011e-02 1.22842826e-01 -1.24414635e+00
-1.00227141e+00 7.33842134e-01 8.72537553e-01 -4.30600435e-01
-2.21885636e-01 2.89574444e-01 3.96834910e-02 -2.28692785e-01
5.54371655e-01 2.07213357e-01 -2.75900692e-01 1.59774184e-01
-8.05796206e-01 5.10939717e-01 5.70300162e-01 1.07267594e+00
5.98765135e-01 -2.69307017e-01 -3.66350524e-02 5.49904585e-01
5.95934153e-01 9.37976360e-01 -3.69851679e-01 -1.08444190e+00
6.21829152e-01 8.13033223e-01 3.44652563e-01 -9.43374336e-01
-5.35447180e-01 -5.77538371e-01 -2.71018475e-01 5.62233746e-01
5.19744873e-01 5.36732338e-02 -9.54877138e-01 1.38136899e+00
7.92911708e-01 9.14518535e-02 -4.38835055e-01 1.23062468e+00
6.40843213e-01 3.80658865e-01 -2.22615704e-01 5.44491291e-01
1.24593091e+00 -9.00095463e-01 3.76834035e-01 -6.90686643e-01
5.26084661e-01 -7.77551413e-01 5.29234290e-01 1.44074783e-01
-1.06828558e+00 -4.28502500e-01 -1.05372989e+00 -3.29695046e-01
-2.06162736e-01 7.22870946e-01 4.48817223e-01 6.56938732e-01
-8.09285104e-01 6.63155690e-02 -9.95087028e-01 -3.49545836e-01
6.87809885e-01 3.01096022e-01 -4.64509875e-01 -1.45755857e-01
-3.73844564e-01 1.12206197e+00 2.68844754e-01 -5.33701815e-02
-8.68183613e-01 -9.17501092e-01 -9.36114907e-01 -7.42195547e-02
3.70379746e-01 -9.18087065e-01 1.24213374e+00 -1.03232481e-01
-1.10163999e+00 1.29451466e+00 -2.21646219e-01 -7.35444546e-01
8.37853849e-01 -2.73046285e-01 3.66207361e-01 1.12739727e-01
4.28370655e-01 1.31618893e+00 7.51228631e-01 -1.20423770e+00
-1.04605138e+00 -6.66721284e-01 4.51298580e-02 5.00503838e-01
3.19231361e-01 -9.00272653e-02 -8.38947296e-01 -1.91256061e-01
4.98552561e-01 -1.15974188e+00 -4.45772558e-01 6.79735243e-01
-6.01283669e-01 -1.34614930e-01 6.84186101e-01 -4.68367338e-01
3.22141111e-01 -2.13124871e+00 -1.06146552e-01 2.10576743e-01
3.72357190e-01 -6.04764372e-02 1.19313583e-01 -1.26757428e-01
2.47118816e-01 -2.08586395e-01 1.41592383e-01 -9.07605410e-01
1.32303521e-01 -2.40684718e-01 -2.61101604e-01 9.11781490e-01
4.45279837e-01 1.02225506e+00 -6.92809999e-01 -2.51301229e-01
5.02279222e-01 5.97536147e-01 -8.81328583e-01 -9.21915248e-02
-7.26388097e-02 1.63920462e-01 -4.37204808e-01 8.72582316e-01
9.08451557e-01 -4.15946722e-01 -5.30045092e-01 -2.72411764e-01
-2.46564344e-01 3.65207225e-01 -1.30708361e+00 1.68156219e+00
-2.27665439e-01 6.77039921e-01 5.38825244e-02 -7.34552145e-01
8.55628669e-01 -2.94674873e-01 4.36132878e-01 -3.78475964e-01
1.16262749e-01 2.19732136e-01 -3.28182399e-01 2.14990199e-01
5.71947455e-01 2.08436817e-01 1.04438206e-02 2.75503665e-01
-1.12031549e-01 -3.59917074e-01 -1.16325706e-01 1.74092367e-01
1.02833712e+00 3.77648085e-01 2.83446968e-01 1.05746388e-02
1.85975432e-01 2.86788493e-01 4.43900764e-01 1.02312422e+00
-4.44811165e-01 9.69275057e-01 3.86733413e-01 -5.34437120e-01
-1.46970463e+00 -1.28280425e+00 -2.38374189e-01 5.68351388e-01
4.11045223e-01 1.04487659e-02 -4.64456648e-01 -8.18335593e-01
5.03399432e-01 4.11162496e-01 -3.86995018e-01 1.18974261e-02
-7.04303265e-01 -6.19611859e-01 3.55381399e-01 6.95272744e-01
2.64532149e-01 -4.81740654e-01 -1.12254536e+00 1.24409735e-01
1.83931217e-01 -1.53662384e+00 -6.09892488e-01 3.65962386e-01
-6.07237220e-01 -1.14500034e+00 -7.41074204e-01 -5.52047372e-01
7.89285719e-01 9.21342015e-01 1.04575062e+00 -2.52907723e-01
-5.11774421e-01 4.73866642e-01 -1.23325393e-01 -4.00441468e-01
-8.47880766e-02 8.54166672e-02 2.50614077e-01 -3.15343767e-01
5.94851136e-01 -2.15382650e-01 -8.43985617e-01 3.93823087e-01
-2.79702283e-02 1.37418449e-01 6.35833025e-01 6.30493045e-01
7.77171373e-01 -3.63671154e-01 -3.41592124e-03 -2.23177761e-01
-1.76440194e-01 -2.05790594e-01 -1.21659291e+00 -4.49419692e-02
-2.31983140e-01 -2.12210119e-01 1.24083593e-01 -3.80049825e-01
-5.91031015e-01 8.19154501e-01 1.01313315e-01 -6.31783545e-01
-1.21096544e-01 -1.99147090e-01 8.97773579e-02 -5.71009457e-01
8.50969493e-01 1.37998924e-01 -5.08966409e-02 -4.94535744e-01
5.81718147e-01 3.75675291e-01 6.10370576e-01 -3.56272131e-01
1.10542333e+00 8.76215816e-01 2.70627439e-01 -7.41181314e-01
-9.82035279e-01 -8.53387654e-01 -9.78862941e-01 -1.69767469e-01
8.52326512e-01 -1.50527680e+00 -8.12683165e-01 2.22399324e-01
-1.48231065e+00 -4.06221440e-03 -1.99490964e-01 5.38567722e-01
-5.49561083e-01 2.85157859e-01 -3.72787774e-01 -9.35439289e-01
-2.37640306e-01 -1.15983140e+00 1.68108261e+00 4.20919433e-02
1.04075357e-01 -4.29297626e-01 -1.48286998e-01 2.43347570e-01
-3.76532041e-02 1.83289990e-01 3.80036205e-01 -2.90362746e-01
-1.20965731e+00 -5.36057293e-01 -6.32539988e-01 1.15630299e-01
-3.28910232e-01 -2.23773569e-01 -1.00499153e+00 -3.42606097e-01
-1.30007595e-01 -2.33837143e-01 1.04296350e+00 6.24714732e-01
7.22609460e-01 2.00335369e-01 -8.06502104e-01 7.88931906e-01
1.07987118e+00 -2.17365801e-01 -1.02592506e-01 2.57640600e-01
6.64831996e-01 2.82719433e-01 5.74677587e-01 4.14841354e-01
5.58307469e-01 9.61012721e-01 4.71721888e-01 -3.90759818e-02
-3.85207295e-01 -5.36613286e-01 2.70670116e-01 9.16941836e-02
2.93036103e-01 1.56847209e-01 -8.60098898e-01 4.53471899e-01
-1.72176790e+00 -6.27767205e-01 -4.96439496e-03 2.11723757e+00
3.44767600e-01 4.49842900e-01 3.46803278e-01 -2.36838460e-01
6.81735516e-01 -2.00944804e-02 -8.80980849e-01 4.23336953e-01
-1.91426501e-01 -1.94787517e-01 1.03228998e+00 4.30984527e-01
-1.58270204e+00 1.08507514e+00 6.29775524e+00 4.81548011e-01
-9.69589829e-01 3.12934816e-02 5.71621478e-01 -3.87764752e-01
2.89310306e-01 9.50690061e-02 -1.69100523e+00 2.35104963e-01
3.92978221e-01 2.32062891e-01 2.23931730e-01 1.31399393e+00
2.34815106e-01 -2.13787839e-01 -1.30721188e+00 1.22921240e+00
1.82795092e-01 -1.37625337e+00 -2.66115606e-01 1.31460786e-01
7.45730281e-01 5.94631255e-01 1.63093835e-01 1.40483677e-01
3.34431916e-01 -6.34410262e-01 1.15247333e+00 1.38139099e-01
4.72651601e-01 -6.18710041e-01 2.86380857e-01 5.44066489e-01
-1.32140195e+00 -1.56838879e-01 -6.52543247e-01 -5.99540994e-02
3.60195130e-01 5.52989781e-01 -1.23181784e+00 1.64980561e-01
7.68052042e-01 7.51195371e-01 -7.03792274e-01 1.60116458e+00
-3.23206812e-01 7.62196481e-02 -9.49064851e-01 -1.37378737e-01
2.97426254e-01 7.92760700e-02 8.97485673e-01 1.02102041e+00
3.52562487e-01 -5.75238690e-02 4.88765568e-01 1.31584167e+00
-8.19931030e-02 -2.38870293e-01 -6.67161345e-01 5.30120850e-01
6.92686975e-01 1.26229739e+00 -8.50488663e-01 -3.01829670e-02
-4.28499371e-01 6.66188478e-01 4.32352483e-01 1.30487770e-01
-7.98475325e-01 2.20541418e-01 6.75733805e-01 3.34231257e-01
6.96319818e-01 -5.35688758e-01 -4.91616428e-01 -1.27285671e+00
4.62066263e-01 -4.53046501e-01 7.97153916e-03 -7.54657269e-01
-1.33760190e+00 7.14844540e-02 -4.44755740e-02 -1.23562706e+00
-6.71471953e-02 -8.87100160e-01 -3.04956406e-01 8.24787736e-01
-1.52519941e+00 -1.30299008e+00 -3.03109765e-01 1.24649316e-01
3.93987656e-01 -9.07513872e-02 8.48364159e-02 3.36324155e-01
-2.31336132e-01 5.25207639e-01 -2.42051885e-01 1.32668421e-01
5.93410313e-01 -1.14863575e+00 1.01322341e+00 1.00563538e+00
3.47321272e-01 3.76511693e-01 4.18635756e-01 -7.67991483e-01
-1.58424342e+00 -1.18864810e+00 6.73055589e-01 -1.19120014e+00
3.84568334e-01 -8.63777637e-01 -4.30035293e-01 6.26565576e-01
-6.54367626e-01 4.31515515e-01 -1.25445992e-01 -8.07535574e-02
-4.98584449e-01 5.88873588e-03 -1.07960999e+00 6.36840224e-01
1.32310796e+00 -3.80528271e-01 -4.98829961e-01 3.68247151e-01
7.39226401e-01 -8.67508292e-01 -3.02235454e-01 3.24026495e-01
6.35600567e-01 -8.30455720e-01 1.55730069e+00 -3.84636462e-01
5.23241982e-02 -6.37331367e-01 -2.82598257e-01 -9.12616253e-01
-1.85000345e-01 -3.80140841e-01 -2.33852267e-01 6.49336874e-01
3.82902592e-01 -4.80303407e-01 1.28125429e+00 6.18543684e-01
-9.66042131e-02 -5.71020067e-01 -1.32329488e+00 -8.53463590e-01
-9.85810459e-02 -4.85583961e-01 2.71503508e-01 3.07412505e-01
-5.51745832e-01 3.83813202e-01 -2.23809168e-01 4.68982995e-01
1.16071773e+00 1.83595270e-01 1.21060002e+00 -1.21287799e+00
-2.24307388e-01 -5.47850072e-01 -6.70171022e-01 -1.90510213e+00
8.62458907e-03 -9.08814669e-01 7.87758827e-02 -1.09479773e+00
3.01548094e-01 -6.40871882e-01 9.76289213e-02 2.96105206e-01
-4.07371223e-02 5.51533699e-01 5.40028930e-01 1.42914236e-01
-7.38942564e-01 5.67635357e-01 9.08209980e-01 -5.38809337e-02
-2.72225678e-01 2.89053410e-01 -4.48055983e-01 8.78881276e-01
3.28765064e-01 -5.52013457e-01 5.00772148e-02 -4.53941494e-01
1.76004484e-01 -2.65423119e-01 1.11874080e+00 -1.07161129e+00
5.33580780e-01 2.27545887e-01 9.20527339e-01 -1.46045899e+00
9.34513569e-01 -7.96335518e-01 -3.76670539e-01 4.29881811e-01
-6.31952062e-02 -8.54662508e-02 -8.09397735e-03 6.45899653e-01
3.89161766e-01 3.44150439e-02 1.01654613e+00 -1.18661858e-01
-6.68070555e-01 5.13504863e-01 1.51627347e-01 4.11949493e-03
1.06935978e+00 -4.11157250e-01 -4.79224250e-02 -1.08357519e-01
-4.23250914e-01 3.05504501e-01 7.70807505e-01 5.19148886e-01
6.70433044e-01 -1.30997670e+00 -7.06655443e-01 2.91987538e-01
1.57968089e-01 3.27947855e-01 -1.37573108e-01 8.08781147e-01
-5.40498555e-01 7.15074837e-01 6.33417815e-02 -1.22330320e+00
-1.11170685e+00 4.20610279e-01 5.53581953e-01 2.24777445e-01
-9.88820016e-01 8.42144728e-01 3.52120131e-01 -6.66362584e-01
4.69460189e-01 -4.71157163e-01 2.77534992e-01 -3.36449742e-01
4.44106966e-01 2.71921962e-01 1.46756414e-02 -5.70364058e-01
-4.69514608e-01 8.36348593e-01 -4.44525182e-02 -9.91961882e-02
1.28190792e+00 -2.14568064e-01 3.62003028e-01 6.55794963e-02
1.06605506e+00 -2.52065092e-01 -1.94163418e+00 -4.32323813e-01
-1.38048545e-01 -7.04486370e-01 1.01659715e-01 -5.50652385e-01
-6.69014454e-01 8.50959420e-01 7.76209593e-01 -2.88775325e-01
3.43667686e-01 1.94490016e-01 6.45526826e-01 6.42347276e-01
5.37975550e-01 -8.24525774e-01 1.52706979e-02 6.18399262e-01
6.11952662e-01 -1.54672837e+00 2.70789146e-01 -4.25056159e-01
-1.84438854e-01 8.88062716e-01 5.77558458e-01 -4.47188109e-01
5.56542456e-01 2.55376518e-01 -2.15577051e-01 -9.51876864e-02
-3.93169314e-01 -1.32615104e-01 5.27170122e-01 5.58588028e-01
-1.71551138e-01 -1.18685411e-02 2.94329077e-01 2.60312945e-01
-1.54324949e-01 -3.70514572e-01 1.41501099e-01 7.06974089e-01
-6.10106826e-01 -7.58173168e-01 -5.89035332e-01 1.52869135e-01
-1.00161210e-01 8.79738256e-02 -5.15478432e-01 8.69820416e-01
2.01712996e-02 6.27463281e-01 5.48017472e-02 -3.20063382e-02
4.05396610e-01 -2.05400258e-01 6.05776191e-01 -6.85774207e-01
-7.58854523e-02 3.76672223e-02 -2.12864727e-01 -6.49841130e-01
-8.52284431e-02 -8.52762282e-01 -9.60940599e-01 -1.36593193e-01
-4.82959241e-01 -3.90668064e-01 9.34379280e-01 8.14182103e-01
5.23099005e-01 -2.14953840e-01 5.17162800e-01 -1.57688177e+00
-9.20376182e-01 -5.36540568e-01 -1.45722851e-01 4.93330769e-02
4.35939699e-01 -9.13029432e-01 -3.14099014e-01 -1.98362872e-01] | [7.716462135314941, -2.5901308059692383] |
395aa190-c9ac-405e-b4fe-944d54901484 | fine-grained-urban-flow-inference | 2002.02318 | null | https://arxiv.org/abs/2002.02318v1 | https://arxiv.org/pdf/2002.02318v1.pdf | Fine-Grained Urban Flow Inference | The ubiquitous deployment of monitoring devices in urban flow monitoring systems induces a significant cost for maintenance and operation. A technique is required to reduce the number of deployed devices, while preventing the degeneration of data accuracy and granularity. In this paper, we present an approach for inferring the real-time and fine-grained crowd flows throughout a city based on coarse-grained observations. This task exhibits two challenges: the spatial correlations between coarse- and fine-grained urban flows, and the complexities of external impacts. To tackle these issues, we develop a model entitled UrbanFM which consists of two major parts: 1) an inference network to generate fine-grained flow distributions from coarse-grained inputs that uses a feature extraction module and a novel distributional upsampling module; 2) a general fusion subnet to further boost the performance by considering the influence of different external factors. This structure provides outstanding effectiveness and efficiency for small scale upsampling. However, the single-pass upsampling used by UrbanFM is insufficient at higher upscaling rates. Therefore, we further present UrbanPy, a cascading model for progressive inference of fine-grained urban flows by decomposing the original tasks into multiple subtasks. Compared to UrbanFM, such an enhanced structure demonstrates favorable performance for larger-scale inference tasks. | ['Yu Zheng', 'Kun Ouyang', 'Zekun Tong', 'Yuxuan Liang', 'Ye Liu', 'Sijie Ruan', 'David S. Rosenblum'] | 2020-02-05 | null | null | null | null | ['fine-grained-urban-flow-inference'] | ['miscellaneous'] | [-2.21312076e-01 -2.66533583e-01 -1.15721188e-01 -3.87175769e-01
-5.75100839e-01 -1.57790929e-01 7.79982328e-01 1.65597424e-01
-1.14728130e-01 1.11136448e+00 5.32388449e-01 -5.47923565e-01
-1.83893397e-01 -1.61610425e+00 -4.26772892e-01 -4.42938954e-01
-2.45301172e-01 3.09088707e-01 5.89827538e-01 -1.53494716e-01
-1.37725979e-01 7.30911851e-01 -1.95846307e+00 3.55350733e-01
1.21582353e+00 8.99791598e-01 -5.85904531e-02 8.69544148e-01
-4.17634368e-01 1.18913102e+00 -8.82682741e-01 -1.65900588e-01
2.34720364e-01 -1.31689116e-01 -5.69743216e-01 -5.22905700e-02
5.88636816e-01 -7.50669479e-01 -4.05778497e-01 7.60571778e-01
5.03908575e-01 2.51666278e-01 6.02577686e-01 -1.50576627e+00
-3.21339101e-01 6.85715795e-01 -4.06256318e-01 5.44542253e-01
-9.69489198e-03 5.05713642e-01 7.96999395e-01 -4.86905903e-01
-7.27746710e-02 1.60308087e+00 6.79712057e-01 -1.06064811e-01
-9.96815681e-01 -8.26684296e-01 4.96593386e-01 -6.68402091e-02
-1.46714425e+00 -5.05753994e-01 4.07884777e-01 -7.75489628e-01
9.48459744e-01 2.60816604e-01 6.72697067e-01 7.07009614e-01
-1.23060681e-01 5.84373534e-01 7.34513044e-01 1.67412832e-01
3.33863527e-01 5.01063168e-02 5.68621233e-02 5.45070112e-01
5.65660357e-01 2.99114376e-01 -2.52900094e-01 -2.22965509e-01
7.85760999e-01 2.92743713e-01 1.12822221e-03 4.92569387e-01
-7.66861796e-01 7.96350718e-01 8.23829889e-01 1.88990727e-01
-3.85880053e-01 2.51366347e-01 3.15880686e-01 -1.17480345e-02
7.95078158e-01 -5.37319705e-02 -3.09271067e-01 -1.42419130e-01
-1.39431977e+00 5.26283681e-01 6.30368173e-01 8.67335200e-01
1.20378757e+00 3.12627375e-01 -5.62879264e-01 3.99776429e-01
3.54023248e-01 1.11951065e+00 6.97100982e-02 -9.09767807e-01
9.90827262e-01 7.84973979e-01 1.20980173e-01 -1.33189631e+00
-5.62716782e-01 -4.98617500e-01 -1.09004295e+00 9.37925577e-02
5.57765663e-01 -4.46562916e-01 -6.19047821e-01 1.64281452e+00
5.47161520e-01 5.93684912e-01 -3.73625696e-01 6.96054459e-01
7.52642334e-01 7.38710642e-01 4.58178997e-01 1.91743642e-01
1.24317849e+00 -7.16315806e-01 -5.85749388e-01 -2.28020415e-01
4.48683351e-01 -2.63026983e-01 9.55290794e-01 -3.23846608e-01
-9.15079892e-01 -7.41984546e-01 -7.88697124e-01 -9.94089618e-03
-8.18747759e-01 2.15484668e-02 6.54958665e-01 7.56734073e-01
-8.68868411e-01 4.49970961e-01 -7.17665374e-01 -9.91993099e-02
7.63904810e-01 1.44651502e-01 1.56316832e-01 1.57835081e-01
-1.35791481e+00 4.06056345e-01 1.76679835e-01 7.03534111e-04
-5.06613612e-01 -1.37345219e+00 -9.66845334e-01 4.49791640e-01
-3.59218977e-02 -1.03248882e+00 8.68868470e-01 -3.04666311e-01
-1.29051495e+00 3.00209790e-01 -4.15794551e-01 -4.66136843e-01
5.36711037e-01 2.61361338e-02 -5.06724000e-01 -4.96168062e-02
5.96633792e-01 4.47733521e-01 7.99666703e-01 -8.80453885e-01
-1.22571504e+00 -4.29418266e-01 2.04258263e-01 -2.24766955e-01
-2.36798555e-01 -2.84536123e-01 -2.86151040e-02 -7.58797467e-01
-7.22566783e-01 -5.13978481e-01 -3.78189236e-01 -2.21715450e-01
-3.05891901e-01 -1.36811882e-01 8.07907581e-01 -4.05390441e-01
1.72975159e+00 -1.96735275e+00 -6.59340143e-01 3.09528112e-01
5.38531482e-01 5.81237435e-01 9.64657031e-03 3.03919524e-01
3.69110972e-01 1.71601728e-01 -2.47010648e-01 -4.55310315e-01
1.33032903e-01 1.26194969e-01 -6.54531717e-01 2.17796594e-01
3.82252753e-01 1.17354870e+00 -1.14079511e+00 -5.03496468e-01
6.85359538e-01 3.73108536e-01 -6.05248392e-01 2.50320286e-01
-7.77537227e-02 3.59124959e-01 -6.75359488e-01 5.59386015e-01
1.01032400e+00 -3.67524177e-01 -2.72835195e-01 -5.68897761e-02
-2.26361245e-01 3.82529289e-01 -1.39350498e+00 1.06966770e+00
-7.28017330e-01 4.63588923e-01 1.48845956e-01 -5.94426036e-01
9.27696526e-01 1.75873227e-02 5.28164208e-01 -6.50867641e-01
1.29886210e-01 7.84388185e-02 -5.31885743e-01 -4.74387765e-01
5.73759735e-01 -1.58338472e-01 -3.35546762e-01 4.67766106e-01
-3.53373587e-01 4.65608686e-02 3.35316896e-01 8.84881243e-02
1.26689768e+00 -3.17621738e-01 3.18527281e-01 -2.71958023e-01
5.38405895e-01 -2.97144838e-02 5.63212693e-01 7.91340530e-01
-4.01731640e-01 1.73584983e-01 2.04388484e-01 -6.88925207e-01
-6.89567447e-01 -1.26684380e+00 -1.00473575e-01 1.06908107e+00
7.49800354e-02 -4.10010219e-01 -5.59245110e-01 -6.65759921e-01
6.98258519e-01 4.24544424e-01 -5.90651274e-01 1.75641939e-01
-6.48957372e-01 -1.23767412e+00 1.07540011e+00 7.19636917e-01
9.67324555e-01 -7.65307426e-01 -7.15341091e-01 1.63286403e-01
-4.04430002e-01 -1.18413866e+00 -3.73384744e-01 -3.75061214e-01
-7.94636488e-01 -9.80398834e-01 -2.70567596e-01 -4.99765463e-02
4.43602055e-01 3.77008915e-01 1.32574606e+00 4.90291864e-02
-9.91692171e-02 -2.12328702e-01 -9.04625803e-02 -4.39067721e-01
3.08306888e-03 4.39289421e-01 -1.71182767e-01 1.40797302e-01
5.14058173e-01 -8.47390831e-01 -6.08679950e-01 1.69095367e-01
-9.42326784e-01 -2.37807885e-01 3.67760122e-01 3.92441034e-01
1.59739822e-01 4.89938915e-01 9.96294081e-01 -9.40943718e-01
6.04654729e-01 -8.06170166e-01 -6.38445437e-01 -1.26256838e-01
-3.05052429e-01 9.90417302e-02 9.58887815e-01 -1.08397044e-01
-1.34086466e+00 -2.31443018e-01 -3.52209389e-01 -1.65481105e-01
-5.06161153e-01 1.99368782e-02 -3.46860290e-01 3.30427647e-01
7.01485455e-01 -2.40314335e-01 -3.39202195e-01 -3.92081857e-01
5.09094834e-01 1.01270139e+00 3.69180113e-01 -6.72278106e-01
9.51493144e-01 9.32881951e-01 5.05821928e-02 -1.00975800e+00
-9.45837021e-01 -4.68605161e-01 -5.80897570e-01 -1.94456518e-01
5.87270379e-01 -1.24621153e+00 -8.81145656e-01 4.45019126e-01
-1.01600480e+00 -5.51405907e-01 -7.08215952e-01 2.65610963e-01
-1.86100528e-01 -7.76286349e-02 -5.89216232e-01 -8.30244064e-01
-2.23100170e-01 -1.02164912e+00 1.45571852e+00 2.12247431e-01
-1.58591926e-01 -1.19994092e+00 1.18413456e-01 1.52312651e-01
7.87358880e-01 4.52246845e-01 5.27310789e-01 -4.53400314e-02
-8.65564942e-01 9.50484350e-02 -7.47014761e-01 -4.36890125e-02
3.44326645e-01 5.38930967e-02 -1.27912843e+00 -1.19888422e-03
-3.89311552e-01 2.32570872e-01 9.01323378e-01 5.40093482e-01
1.11029458e+00 -3.38808119e-01 -5.20934284e-01 7.80344546e-01
1.34787548e+00 -1.19906425e-01 5.79295397e-01 2.16772407e-01
8.57779920e-01 5.84660530e-01 3.38996291e-01 7.23638177e-01
8.42053771e-01 3.90836895e-01 3.44869763e-01 -2.26577505e-01
-4.43288147e-01 -5.52460551e-01 1.39324978e-01 4.18707639e-01
-2.97334194e-02 -3.37489069e-01 -7.22663045e-01 8.56909990e-01
-2.00264192e+00 -1.27273715e+00 -3.83368284e-01 1.95109165e+00
3.69162261e-01 4.08804789e-02 4.92290109e-01 2.11399481e-01
6.01373315e-01 5.06288707e-01 -5.39716661e-01 -1.75527990e-01
-1.62797663e-02 1.50550634e-01 7.86018431e-01 7.82736421e-01
-1.13109076e+00 9.42989945e-01 6.57912588e+00 1.02671862e+00
-1.07328033e+00 8.97050425e-02 5.32134175e-01 -9.77091640e-02
-4.11130190e-01 -3.56091380e-01 -1.24240077e+00 8.23558748e-01
1.16099167e+00 1.16802163e-01 3.06301028e-01 6.15537941e-01
6.81898832e-01 1.29231755e-02 -5.81767619e-01 7.95726776e-01
-4.44771439e-01 -1.53343725e+00 2.96651572e-01 1.58569649e-01
7.11402535e-01 2.76567996e-01 -1.93709135e-01 4.51919436e-01
8.47163916e-01 -8.44453931e-01 5.36345124e-01 5.33563018e-01
7.87653267e-01 -7.11818755e-01 7.03108907e-01 3.27432364e-01
-1.84925127e+00 -2.67853528e-01 -1.78791314e-01 -6.41192794e-01
6.03727341e-01 1.35553229e+00 -5.32876730e-01 5.81681669e-01
7.27741361e-01 8.91084075e-01 -4.81917411e-01 8.08237553e-01
-2.48634338e-01 6.17306530e-01 -5.27772605e-01 1.97271764e-01
1.33651361e-01 3.44859697e-02 2.60048509e-01 1.58272338e+00
3.68193567e-01 -8.88392851e-02 4.20007914e-01 9.40867186e-01
8.16029012e-02 -2.93425173e-01 -8.50466788e-01 4.93828923e-01
8.59492004e-01 1.17979586e+00 -3.63610655e-01 -7.15490460e-01
-4.27177727e-01 2.14247748e-01 4.18029428e-01 5.24239898e-01
-9.97602642e-01 -4.85691607e-01 1.28957057e+00 6.61883771e-01
3.25433850e-01 -5.67224361e-02 -6.63796723e-01 -1.17079413e+00
-1.26557112e-01 -4.38088179e-01 4.83806521e-01 -2.72593945e-01
-1.36676204e+00 5.83059549e-01 9.60071459e-02 -1.03578103e+00
-3.30230296e-01 -3.91287684e-01 -7.59383261e-01 1.00677168e+00
-1.95723271e+00 -1.08364284e+00 -6.43777788e-01 7.64325261e-01
2.92495191e-01 1.66551366e-01 4.20263886e-01 7.98709869e-01
-7.17951298e-01 3.82162094e-01 -3.15750003e-01 3.09772998e-01
3.91736507e-01 -1.16570592e+00 6.17926240e-01 1.02413821e+00
-3.56197506e-01 2.63798445e-01 3.50934982e-01 -7.41952300e-01
-6.64852500e-01 -2.07313895e+00 1.11136508e+00 -5.72350144e-01
7.42703855e-01 -3.20977032e-01 -5.74429572e-01 6.79679334e-01
-3.49497497e-01 3.11550230e-01 4.13809687e-01 1.15429154e-02
-1.76509500e-01 -5.60972214e-01 -1.35157466e+00 4.54463452e-01
1.25315750e+00 -4.88037109e-01 -3.18761349e-01 -5.21176904e-02
8.27619374e-01 1.90582052e-02 -1.05796492e+00 3.92358571e-01
2.92882800e-01 -1.16231585e+00 1.15648365e+00 -2.17498973e-01
2.13893071e-01 -7.02451408e-01 -1.16172016e-01 -1.33780396e+00
-5.29384255e-01 -4.15000677e-01 -3.31318289e-01 1.43207550e+00
1.60586208e-01 -9.72388327e-01 7.07899272e-01 5.13404667e-01
1.18699111e-01 -3.92217904e-01 -7.85266101e-01 -5.20688951e-01
8.92986655e-02 -6.50207639e-01 1.63170385e+00 6.02106452e-01
-1.27738848e-01 4.43098158e-01 -1.55839875e-01 4.93855953e-01
7.08243787e-01 3.54187965e-01 1.04842865e+00 -1.46019995e+00
-1.45283127e-02 -5.21668077e-01 -3.46128404e-01 -1.24806595e+00
1.22774690e-01 -6.87174976e-01 -1.17546812e-01 -1.46271122e+00
-2.17289016e-01 -8.62845361e-01 1.33331150e-01 2.66994596e-01
-3.86875838e-01 1.72101855e-01 9.55139175e-02 1.50426790e-01
-5.48534691e-01 6.20878935e-01 1.09927857e+00 -1.96913451e-01
-3.82447749e-01 2.23302037e-01 -7.72262931e-01 6.70026183e-01
9.65970576e-01 -3.49514544e-01 -5.23950458e-01 -5.97785234e-01
1.54831661e-02 -9.58300456e-02 5.06765664e-01 -1.26269805e+00
1.52412131e-01 -2.74763554e-01 3.15364748e-01 -7.05299914e-01
6.35990053e-02 -8.05054665e-01 1.21734023e-01 3.91897887e-01
3.03773340e-02 8.53418410e-02 2.80964285e-01 4.89756763e-01
-3.89449090e-01 5.23037255e-01 6.43163085e-01 -1.27480775e-01
-4.40262198e-01 8.05585563e-01 -6.40678525e-01 3.04451585e-01
9.20010746e-01 -4.55690473e-02 -7.14967191e-01 -1.76385388e-01
-3.42131078e-01 4.82510269e-01 7.37600699e-02 3.65613252e-01
2.27371141e-01 -1.39319515e+00 -7.47917175e-01 6.31329715e-01
-1.15402311e-01 3.43349427e-01 3.27816665e-01 5.72796881e-01
-3.44198763e-01 6.37633026e-01 8.54296088e-02 -5.02638519e-01
-6.05895042e-01 5.21499753e-01 3.72432679e-01 -7.55134642e-01
-5.12552023e-01 4.04957473e-01 3.96246761e-01 -7.78582156e-01
-1.41965672e-01 -8.02757323e-01 -2.44441837e-01 2.71047741e-01
9.75615382e-01 1.01362455e+00 1.63903505e-01 -7.44606555e-01
-2.58757383e-01 5.50503135e-01 6.22765183e-01 2.66034663e-01
1.17896163e+00 -5.34354270e-01 1.20852657e-01 3.00523102e-01
1.02971590e+00 1.04006611e-01 -1.41247869e+00 -3.25163454e-01
-3.91224563e-01 -7.94421256e-01 2.44775731e-02 -3.73745590e-01
-1.29421103e+00 9.33201313e-01 9.74857658e-02 6.12016201e-01
1.11968291e+00 -2.18352020e-01 1.18241680e+00 -1.06391020e-01
4.38011587e-01 -7.87227631e-01 -5.77321529e-01 5.34657717e-01
1.84109643e-01 -1.09687531e+00 -2.84010291e-01 -5.41985810e-01
-1.43284425e-01 6.25134349e-01 5.20968378e-01 -2.00521633e-01
1.18890452e+00 7.14756131e-01 -7.67483190e-02 -3.37465465e-01
-6.19301379e-01 -7.95110345e-01 8.03562924e-02 7.03212976e-01
3.15902010e-02 5.05582273e-01 3.37540478e-01 3.90332758e-01
-4.32485163e-01 3.51521850e-01 7.13861808e-02 4.32839572e-01
-6.28342390e-01 -6.56266928e-01 -5.16123652e-01 7.06621706e-01
-1.99758440e-01 -3.30106840e-02 2.43596107e-01 6.52583659e-01
5.17438948e-01 1.32435513e+00 4.89442676e-01 -3.94156873e-01
5.83848298e-01 -5.10717571e-01 -1.00016885e-01 -3.60183239e-01
-6.75789356e-01 -5.46927512e-01 1.16310120e-01 -8.26794565e-01
-4.28110957e-01 -5.84291697e-01 -1.03118408e+00 -1.00012863e+00
1.07390381e-01 9.98800918e-02 1.48024827e-01 1.03774834e+00
6.98826790e-01 8.94565463e-01 6.80620968e-01 -9.72090423e-01
-1.32210121e-01 -9.08527136e-01 -7.31283367e-01 4.45287526e-01
7.38991201e-01 -7.96403766e-01 -5.04975975e-01 -1.47701293e-01] | [6.4544596672058105, 2.1297760009765625] |
77b14756-c71a-4e99-ba40-c5efac173f40 | google-usm-scaling-automatic-speech | 2303.01037 | null | https://arxiv.org/abs/2303.01037v2 | https://arxiv.org/pdf/2303.01037v2.pdf | Google USM: Scaling Automatic Speech Recognition Beyond 100 Languages | We introduce the Universal Speech Model (USM), a single large model that performs automatic speech recognition (ASR) across 100+ languages. This is achieved by pre-training the encoder of the model on a large unlabeled multilingual dataset of 12 million (M) hours spanning over 300 languages, and fine-tuning on a smaller labeled dataset. We use multilingual pre-training with random-projection quantization and speech-text modality matching to achieve state-of-the-art performance on downstream multilingual ASR and speech-to-text translation tasks. We also demonstrate that despite using a labeled training set 1/7-th the size of that used for the Whisper model, our model exhibits comparable or better performance on both in-domain and out-of-domain speech recognition tasks across many languages. | ['Yonghui Wu', 'Françoise Beaufays', 'Johan Schalkwyk', 'Chung-Cheng Chiu', 'Pedro Moreno', 'Tara Sainath', 'Bhuvana Ramabhadran', 'Trevor Strohman', 'Hagen Soltau', 'Ginger Perng', 'Jason Riesa', 'Parisa Haghani', 'Daniel S. Park', 'Rohit Prabhavalkar', 'Andrew Rosenberg', 'Ke Hu', 'Zhong Meng', 'Gary Wang', 'Vera Axelrod', 'Bo Li', 'Nanxin Chen', 'Zhehuai Chen', 'Ankur Bapna', 'Yongqiang Wang', 'James Qin', 'Wei Han', 'Yu Zhang'] | 2023-03-02 | null | null | null | null | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 1.79064631e-01 3.76617283e-01 -4.76610690e-01 -5.18831313e-01
-1.78688896e+00 -7.38498867e-01 8.33047032e-01 -3.43225718e-01
-5.00765502e-01 6.52070165e-01 6.38389587e-01 -1.00101006e+00
6.55669212e-01 -1.62465423e-01 -9.14673567e-01 -1.71755895e-01
2.17652738e-01 9.33839381e-01 -1.21651471e-01 -3.02053064e-01
-4.08421397e-01 1.59901142e-01 -9.94707465e-01 5.17389834e-01
7.43205309e-01 5.61242282e-01 2.69383639e-01 8.63018751e-01
-1.76358163e-01 7.50818372e-01 -6.47073984e-01 -5.95832705e-01
2.28496253e-01 -2.42791593e-01 -9.90368247e-01 2.19882667e-01
5.29998779e-01 -2.06947640e-01 -8.16034019e-01 8.95452619e-01
7.28757203e-01 4.83402312e-02 4.65989888e-01 -6.64630890e-01
-9.94591057e-01 9.90305662e-01 -1.54022619e-01 2.66201526e-01
4.97196734e-01 7.47101158e-02 1.02238142e+00 -1.20558536e+00
7.47554660e-01 1.56910610e+00 3.00861806e-01 7.49337792e-01
-1.15873933e+00 -6.25116467e-01 -1.15374938e-01 -1.09641798e-01
-1.54065394e+00 -1.42631900e+00 1.22726142e-01 -1.75176546e-01
1.60008919e+00 1.14301302e-01 1.93533096e-02 1.38557482e+00
-2.77550578e-01 9.71067131e-01 8.71291578e-01 -6.72457635e-01
4.95256148e-02 1.75302267e-01 -2.14377940e-01 5.60587168e-01
-4.03079748e-01 1.38754055e-01 -8.18483889e-01 -1.20323092e-01
4.39184129e-01 -6.98278308e-01 -3.29959571e-01 2.63410777e-01
-1.53206027e+00 7.37359524e-01 -2.28229359e-01 2.42837802e-01
3.27009335e-02 -6.03394993e-02 5.73042691e-01 6.97194517e-01
5.66951156e-01 1.12071328e-01 -8.04202676e-01 -4.66697246e-01
-1.20952702e+00 -3.54678005e-01 8.76059055e-01 1.31970668e+00
6.74749076e-01 5.07364452e-01 -8.37522522e-02 1.24933672e+00
2.14171886e-01 1.29638577e+00 8.69586349e-01 -9.36231077e-01
1.03123689e+00 -7.49508068e-02 -2.66136706e-01 1.92480356e-01
8.42368528e-02 -3.53500158e-01 -5.53181469e-01 -4.29053754e-01
1.47584587e-01 -3.59938890e-01 -1.14738441e+00 1.83058774e+00
-8.81160982e-03 -4.33053598e-02 6.34846866e-01 6.63833559e-01
6.39048159e-01 1.02685523e+00 -2.08603926e-02 -1.93662718e-01
1.27077293e+00 -1.29731286e+00 -5.96130371e-01 -6.14451826e-01
8.76303852e-01 -1.01564324e+00 1.22419250e+00 -6.02669232e-02
-1.22112787e+00 -5.36523223e-01 -8.25479090e-01 -4.05378610e-01
-1.76019877e-01 4.61158693e-01 1.35152210e-02 7.38936484e-01
-1.49917877e+00 -1.27867162e-01 -8.77441704e-01 -6.84283376e-01
-6.19401522e-02 1.20905936e-01 -6.97135806e-01 -2.42512435e-01
-1.33193433e+00 1.05040240e+00 1.54115081e-01 -4.64099407e-01
-1.18684840e+00 -4.90591347e-01 -9.73503232e-01 -4.15999629e-02
9.34828371e-02 -4.44218993e-01 1.72513855e+00 -9.43379641e-01
-1.76565897e+00 1.19501328e+00 -6.83484912e-01 -5.60290933e-01
1.63236111e-01 -3.38488854e-02 -9.40461099e-01 8.69001374e-02
1.34644315e-01 7.12068677e-01 7.55254805e-01 -7.00404704e-01
-5.78839421e-01 -4.07160699e-01 -4.60006624e-01 4.56570387e-01
-2.96816021e-01 5.98823786e-01 -6.40818238e-01 -8.48868191e-01
-4.29794751e-02 -9.09915924e-01 1.92355029e-02 -5.95478475e-01
-4.29264635e-01 -1.87514007e-01 7.26490438e-01 -1.24041772e+00
1.15130949e+00 -2.09763074e+00 1.05938323e-01 -1.26722351e-01
-4.05123681e-01 3.86204690e-01 -6.09799087e-01 6.64833844e-01
-1.06791683e-01 8.71883780e-02 -1.41724482e-01 -9.04263616e-01
1.23584047e-01 4.29744005e-01 -5.83320498e-01 3.46612453e-01
1.36599749e-01 9.63626504e-01 -6.14781082e-01 -2.29283094e-01
1.45247996e-01 4.57845181e-01 -2.78613985e-01 3.72776896e-01
-1.09248325e-01 5.13745308e-01 6.90991953e-02 7.86667347e-01
3.13446164e-01 -6.35278672e-02 4.59130257e-01 3.01679134e-01
4.24083509e-02 1.20740259e+00 -7.45942712e-01 1.97817481e+00
-8.30082715e-01 7.19451427e-01 4.11661804e-01 -8.28438401e-01
6.78523004e-01 9.43382204e-01 6.37773424e-02 -8.31451774e-01
-1.80491880e-01 7.38724947e-01 -1.25738829e-01 -1.95526466e-01
4.39046800e-01 -1.48937508e-01 -2.54683018e-01 5.84679127e-01
5.44002414e-01 -1.10331759e-01 -4.85950932e-02 1.78354576e-01
9.29752111e-01 -2.53084958e-01 2.08316118e-01 -2.51620501e-01
5.00586808e-01 -1.39499456e-01 3.59641939e-01 6.25815570e-01
-8.94270241e-02 5.40755510e-01 3.37801389e-02 -8.06354955e-02
-1.37222934e+00 -1.18549109e+00 -1.04744166e-01 1.46553695e+00
-6.51170611e-01 -3.68222892e-01 -9.27728236e-01 -6.15970612e-01
-4.30018753e-02 8.57856989e-01 5.33453301e-02 1.22882500e-01
-7.44094908e-01 -4.98310655e-01 1.30739439e+00 3.50091338e-01
6.78656250e-02 -9.70426977e-01 5.43486774e-01 2.81468928e-01
-4.68462825e-01 -1.66842413e+00 -1.00510168e+00 3.63855153e-01
-4.51410294e-01 -5.51719248e-01 -1.00406826e+00 -1.29019070e+00
2.75432348e-01 1.50857523e-01 1.21655881e+00 -5.26208878e-01
2.06461459e-01 2.85239428e-01 -2.35753208e-01 1.45063192e-01
-1.24944007e+00 5.11003971e-01 5.38336217e-01 -1.35903850e-01
3.59635323e-01 -3.92092079e-01 5.70964590e-02 1.62179574e-01
-4.11597937e-01 -1.94085658e-01 5.57517648e-01 8.36695254e-01
4.08634424e-01 -7.09740937e-01 7.18046844e-01 -5.17817855e-01
4.87657458e-01 -3.60484779e-01 -5.86946189e-01 5.68046868e-01
-4.00854796e-01 1.62524864e-01 6.94177747e-01 -3.89211178e-01
-9.02818501e-01 -3.14899571e-02 -6.07451439e-01 -5.09006560e-01
-1.44070625e-01 3.16223592e-01 -4.49738681e-01 2.26880178e-01
5.56783319e-01 6.59960151e-01 -6.59529492e-02 -6.77696884e-01
8.76559019e-01 1.55036283e+00 7.38460064e-01 -5.24486244e-01
5.73435068e-01 4.59259050e-03 -8.22647572e-01 -1.04756987e+00
-5.47997475e-01 -5.78843057e-01 -4.75224674e-01 4.04444247e-01
8.91353607e-01 -1.49325919e+00 -3.80364865e-01 3.91119629e-01
-1.23974037e+00 -6.61261857e-01 -1.66108549e-01 4.58403587e-01
-6.04531467e-01 3.67271364e-01 -1.06869018e+00 -8.00330460e-01
-6.21896863e-01 -1.35066164e+00 1.51234663e+00 -5.58186352e-01
-7.96701610e-02 -8.86645555e-01 1.29180878e-01 6.58412158e-01
4.28233624e-01 -1.02285457e+00 8.31950665e-01 -1.02005708e+00
-4.01324779e-01 1.85874507e-01 6.54508942e-04 6.65291309e-01
8.85921493e-02 -4.06333476e-01 -1.17745447e+00 -6.76457942e-01
-4.08814400e-01 -6.31577790e-01 7.42969155e-01 1.58612505e-01
5.52348971e-01 -4.89020348e-01 -1.24163300e-01 7.00674891e-01
9.64573562e-01 1.75730512e-01 4.21202660e-01 1.81404352e-02
6.12931788e-01 4.18709219e-01 1.34438172e-01 5.90544678e-02
6.66636705e-01 7.10036933e-01 -3.24447095e-01 -9.68095809e-02
-5.60934305e-01 -6.14894032e-01 1.02502668e+00 1.54873478e+00
7.34859288e-01 -3.19590569e-01 -1.17708862e+00 7.59056389e-01
-1.47990048e+00 -7.48280942e-01 3.21520835e-01 2.30163717e+00
1.17381096e+00 -1.05903417e-01 1.66586593e-01 -2.81594932e-01
7.18781173e-01 1.98406726e-01 -2.78506607e-01 -5.20855904e-01
-5.88173985e-01 3.95213157e-01 7.87067890e-01 1.04692781e+00
-9.12934065e-01 1.61272025e+00 7.71385050e+00 1.02685332e+00
-1.33520246e+00 6.08349800e-01 7.84232318e-01 3.21494117e-02
-5.20989716e-01 9.36913770e-03 -1.05899811e+00 3.13273907e-01
1.79347014e+00 -2.32687756e-01 1.03870511e+00 6.96270943e-01
2.40942165e-01 4.67970103e-01 -1.09419143e+00 1.12499511e+00
2.39401907e-01 -1.19198298e+00 1.85383052e-01 7.22667249e-03
8.64116371e-01 1.01607800e+00 -2.49690954e-02 5.58688402e-01
7.08407283e-01 -1.07041323e+00 8.92337799e-01 -2.15687051e-01
1.68401694e+00 -5.39657295e-01 3.19814086e-01 5.52848876e-01
-1.08144665e+00 1.08929984e-01 -2.29317814e-01 4.13042992e-01
3.65150988e-01 3.55064362e-01 -1.26511133e+00 3.45907062e-01
2.98062116e-01 4.39493179e-01 -2.64914215e-01 2.86375999e-01
-3.40725631e-01 1.07730746e+00 -4.75645930e-01 2.35767111e-01
2.12988004e-01 5.91948852e-02 5.74999034e-01 1.54108906e+00
4.82921749e-01 -4.25411671e-01 3.15225214e-01 2.83910990e-01
-5.61001599e-01 2.71926969e-01 -6.21665716e-01 -4.56793785e-01
9.75446522e-01 7.13232636e-01 1.08552285e-01 -8.01839471e-01
-5.97887754e-01 1.29666090e+00 5.40489495e-01 6.34109616e-01
-3.62321824e-01 -2.23390594e-01 8.55509520e-01 -3.86893861e-02
3.25532526e-01 -4.76040959e-01 -1.15982533e-01 -1.50960553e+00
-9.51891113e-03 -1.35699213e+00 2.61104256e-01 -7.44397998e-01
-1.21561563e+00 9.18607712e-01 -6.20315671e-01 -9.06678081e-01
-8.29465151e-01 -5.34255326e-01 -1.18995801e-01 1.39510620e+00
-1.62967086e+00 -1.49478638e+00 6.54478848e-01 6.50367856e-01
8.20751250e-01 -6.93063080e-01 1.11688781e+00 6.24447107e-01
-3.75348270e-01 9.68199790e-01 4.63601172e-01 4.57074285e-01
1.01345372e+00 -1.16562867e+00 1.22268891e+00 8.66810977e-01
4.66408312e-01 6.17148459e-01 2.23715454e-01 -5.97435117e-01
-1.43226719e+00 -1.18495834e+00 1.56422913e+00 -5.96717775e-01
9.56835926e-01 -7.65680313e-01 -6.89797461e-01 1.25578868e+00
5.27750313e-01 -1.18050333e-02 7.39000201e-01 2.78636813e-01
-6.82489157e-01 -4.24084365e-02 -7.84281492e-01 5.66506028e-01
1.03299904e+00 -1.46641266e+00 -5.87437868e-01 4.22099143e-01
1.30187178e+00 -3.94993573e-01 -9.91913378e-01 -2.06361208e-02
3.47241700e-01 -2.05522597e-01 7.22973943e-01 -7.74619281e-01
-1.64520890e-02 -2.38748435e-02 -8.40896904e-01 -1.54886615e+00
-9.19396728e-02 -1.04620147e+00 1.20725498e-01 1.18791521e+00
8.83160591e-01 -6.86930597e-01 4.12508219e-01 -9.34608132e-02
-6.21220112e-01 -3.08332384e-01 -1.48254526e+00 -9.57873762e-01
4.50857937e-01 -6.39569104e-01 6.47253096e-01 1.07434523e+00
1.63867339e-01 6.76347136e-01 -2.96648592e-01 3.67437303e-01
4.31857675e-01 -3.20224762e-01 6.61164045e-01 -5.15305042e-01
-4.46794659e-01 -1.68398574e-01 -1.20130688e-01 -1.50340807e+00
6.18568838e-01 -1.38655698e+00 8.17816406e-02 -1.31471431e+00
-3.86687368e-02 -2.95955151e-01 -7.17145726e-02 6.68743134e-01
7.13622868e-02 8.09923857e-02 1.39211323e-02 2.96142489e-01
-5.54636657e-01 6.97371244e-01 8.22019100e-01 -2.63054907e-01
1.25867873e-01 -2.65053660e-01 -4.58121747e-01 1.33824527e-01
4.32613522e-01 -3.69728446e-01 -2.13709205e-01 -9.93116915e-01
-4.32326853e-01 5.50413311e-01 -1.74763843e-01 -8.85815561e-01
1.08740002e-01 3.46583761e-02 -6.41874894e-02 -4.71431434e-01
4.71191585e-01 -2.36276224e-01 -2.30066180e-01 1.33535549e-01
-4.95325178e-01 2.44825721e-01 1.18248440e-01 1.22830078e-01
-4.53655392e-01 1.36846438e-01 8.20411623e-01 -6.19468912e-02
-5.96724093e-01 2.70497859e-01 -6.87946856e-01 4.83748198e-01
2.79848576e-01 3.39369774e-01 -4.70380783e-01 -6.20745361e-01
-7.89412081e-01 1.14836037e-01 5.49649715e-01 7.32365191e-01
1.93320066e-01 -1.29942954e+00 -9.63958383e-01 5.29581606e-01
3.17149490e-01 -4.48018610e-01 -1.19558826e-01 3.89681786e-01
-1.61956221e-01 1.03186214e+00 4.90470201e-01 -5.36564469e-01
-9.73275721e-01 3.01627964e-01 5.88157713e-01 -2.14830041e-01
-2.78779864e-01 8.05395246e-01 -4.33751121e-02 -1.16182411e+00
3.11855078e-01 -1.31617591e-01 5.34021974e-01 -3.73971164e-01
5.40165782e-01 8.51163343e-02 5.25480151e-01 -1.06357849e+00
-5.11468232e-01 1.15069076e-01 1.67823173e-02 -9.32152092e-01
8.79843771e-01 -4.40059870e-01 1.00174867e-01 4.13122475e-01
1.46082163e+00 3.51535350e-01 -9.02634680e-01 -5.42439103e-01
6.40965551e-02 1.25178725e-01 3.76443937e-02 -9.32943225e-01
-6.57746077e-01 1.14356947e+00 3.58789831e-01 -3.18976760e-01
7.00740635e-01 3.14937681e-01 1.24402988e+00 6.48007333e-01
5.23104727e-01 -1.33672369e+00 -2.55998820e-01 1.12046134e+00
7.04010308e-01 -1.33522058e+00 -8.00635993e-01 -1.51177198e-01
-7.85356879e-01 7.95192599e-01 2.00711399e-01 3.92613888e-01
6.19679451e-01 5.50929666e-01 6.40927315e-01 3.61296326e-01
-1.01808941e+00 -2.84984827e-01 2.50278413e-01 6.32462561e-01
7.21160889e-01 4.88379657e-01 1.53527603e-01 4.72583234e-01
-5.28761864e-01 -1.97564498e-01 9.86739397e-02 6.88403547e-01
-4.36540604e-01 -1.36430669e+00 -2.74777234e-01 7.57542104e-02
-5.79992950e-01 -7.37569690e-01 -2.18968511e-01 2.94754058e-01
-4.20732290e-01 1.25861204e+00 2.10482940e-01 -4.22104120e-01
1.98929429e-01 5.36223412e-01 1.53925732e-01 -8.50260735e-01
-3.76903385e-01 3.20575505e-01 5.90477467e-01 -4.21498358e-01
1.43883705e-01 -5.97087443e-01 -1.16337335e+00 -2.39684075e-01
-1.85877886e-02 1.84517220e-01 8.86555672e-01 8.33507955e-01
5.92706621e-01 1.05722755e-01 7.19387949e-01 -4.18870836e-01
-1.03471327e+00 -1.30860531e+00 -4.11372483e-01 1.89613342e-01
5.61558187e-01 -1.35886427e-02 -3.96486849e-01 1.59814894e-01] | [14.391278266906738, 7.073164939880371] |
60305f8d-b5e5-4761-9ed8-265636e4563c | grammar-based-concept-alignment-for-domain | null | null | https://aclanthology.org/2021.cnl-1.2 | https://aclanthology.org/2021.cnl-1.2.pdf | Grammar-Based Concept Alignment for Domain-Specific Machine Translation | null | ['Aarne Ranta', 'Arianna Masciolini'] | null | null | null | null | cnl-2021-9 | ['concept-alignment'] | ['computer-vision'] | [-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] | [-7.387082099914551, 3.806807041168213] |
37b54a5b-9515-4ac9-8fd2-8c7fb27881d9 | masked-autoencoders-in-3d-point-cloud | 2207.01545 | null | https://arxiv.org/abs/2207.01545v1 | https://arxiv.org/pdf/2207.01545v1.pdf | Masked Autoencoders in 3D Point Cloud Representation Learning | Transformer-based Self-supervised Representation Learning methods learn generic features from unlabeled datasets for providing useful network initialization parameters for downstream tasks. Recently, self-supervised learning based upon masking local surface patches for 3D point cloud data has been under-explored. In this paper, we propose masked Autoencoders in 3D point cloud representation learning (abbreviated as MAE3D), a novel autoencoding paradigm for self-supervised learning. We first split the input point cloud into patches and mask a portion of them, then use our Patch Embedding Module to extract the features of unmasked patches. Secondly, we employ patch-wise MAE3D Transformers to learn both local features of point cloud patches and high-level contextual relationships between patches and complete the latent representations of masked patches. We use our Point Cloud Reconstruction Module with multi-task loss to complete the incomplete point cloud as a result. We conduct self-supervised pre-training on ShapeNet55 with the point cloud completion pre-text task and fine-tune the pre-trained model on ModelNet40 and ScanObjectNN (PB\_T50\_RS, the hardest variant). Comprehensive experiments demonstrate that the local features extracted by our MAE3D from point cloud patches are beneficial for downstream classification tasks, soundly outperforming state-of-the-art methods ($93.4\%$ and $86.2\%$ classification accuracy, respectively). | ['Meili Wang', 'Richard Dazeley', 'Lizhi Zhao', 'Xuequan Lu', 'Jincen Jiang'] | 2022-07-04 | null | null | null | null | ['point-cloud-completion', 'point-cloud-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 1.38300568e-01 1.53984666e-01 -5.50760925e-02 -4.14559513e-01
-8.84086430e-01 -4.34965014e-01 3.79485190e-01 -2.71110028e-01
1.19827725e-01 2.46038586e-01 -1.76008388e-01 -9.07794312e-02
-5.83541244e-02 -1.02976263e+00 -1.27987659e+00 -8.36734772e-01
-1.23619653e-01 4.17285204e-01 2.82455757e-02 -1.15005402e-02
-1.84824429e-02 9.20733392e-01 -1.73150623e+00 5.35131156e-01
6.23840809e-01 1.21678591e+00 3.47914189e-01 9.24391076e-02
-3.18823010e-01 5.08668661e-01 -4.59709644e-01 7.78409690e-02
5.04984140e-01 1.55886054e-01 -5.53897798e-01 3.62161756e-01
6.80681527e-01 -1.29240692e-01 -4.95505393e-01 7.07398891e-01
3.91269416e-01 6.54803738e-02 7.83689976e-01 -1.04859412e+00
-8.10476422e-01 2.71540850e-01 -6.40578210e-01 5.85998856e-02
-2.42018491e-01 3.05058420e-01 8.66728246e-01 -1.37532902e+00
3.27100337e-01 9.41824138e-01 8.06068063e-01 4.88027871e-01
-1.41575491e+00 -1.18033779e+00 1.80298209e-01 -1.11405753e-01
-1.70630562e+00 -3.26499641e-01 1.29223609e+00 -4.91211414e-01
1.38558090e+00 1.56026045e-02 5.52212894e-01 1.04710829e+00
-1.66946389e-02 7.45314002e-01 9.44223106e-01 -1.80388004e-01
1.28528282e-01 8.53752494e-02 -4.86385636e-02 7.92523861e-01
-1.95455775e-01 3.31704587e-01 -4.18245673e-01 -2.03867078e-01
1.21422696e+00 4.66379285e-01 -4.14660908e-02 -4.77173626e-01
-1.15211797e+00 8.21074069e-01 9.82352138e-01 2.61638314e-01
-5.62535822e-01 2.72275537e-01 -3.11370324e-02 5.10244846e-01
8.82844865e-01 2.24437788e-01 -5.99498868e-01 3.93794775e-01
-7.99616158e-01 3.43482792e-02 2.31347084e-01 1.22753608e+00
1.44774306e+00 3.86245459e-01 4.65228334e-02 9.84885812e-01
2.54963785e-01 6.64773047e-01 3.17552000e-01 -6.29717469e-01
5.68339944e-01 1.00050163e+00 -3.17915618e-01 -8.57233465e-01
-1.91989377e-01 -7.74572134e-01 -1.13044751e+00 4.98077869e-01
-2.83935606e-01 1.20001011e-01 -1.14960003e+00 1.55015624e+00
2.48022407e-01 7.28664637e-01 1.01773039e-01 6.86419845e-01
1.21155596e+00 8.21476936e-01 -1.84759200e-01 2.04759017e-01
1.06655383e+00 -8.42351675e-01 -3.75250243e-02 -4.19760466e-01
4.73714560e-01 -5.39481461e-01 1.04492819e+00 2.07096096e-02
-9.16461945e-01 -1.04631555e+00 -1.09926188e+00 1.78889129e-02
-3.07841152e-01 3.25870991e-01 5.49959958e-01 1.02773942e-01
-1.05904913e+00 6.80278778e-01 -1.04298103e+00 2.67540842e-01
9.75527585e-01 5.79438210e-01 -5.03949404e-01 -2.06337363e-01
-6.74916565e-01 4.64932978e-01 -3.27010602e-02 -1.20245904e-01
-1.20438147e+00 -1.05169511e+00 -1.02769768e+00 1.31269932e-01
-1.24335289e-02 -6.27009332e-01 8.03319395e-01 -7.72314489e-01
-1.24275553e+00 1.24774671e+00 -1.20472081e-01 -1.74251303e-01
-1.58282623e-01 -2.85542551e-02 -3.23655903e-01 2.02262327e-01
3.87464672e-01 1.04622734e+00 1.39597344e+00 -1.57776487e+00
-3.81443918e-01 -5.31558394e-01 -2.64198959e-01 2.44125754e-01
-1.54254630e-01 -4.08737034e-01 -3.74038279e-01 -8.85294199e-01
6.30554795e-01 -9.23091769e-01 -1.80343449e-01 1.09185144e-01
-2.84112871e-01 -4.44137305e-01 1.13524926e+00 -2.52190351e-01
4.14706886e-01 -2.49291992e+00 -1.12593099e-01 2.20903650e-01
3.85025114e-01 4.66019437e-02 -5.34161508e-01 1.53150171e-01
-5.51610291e-01 -1.48368701e-01 -4.92552459e-01 -6.74225450e-01
-2.14838147e-01 3.14225346e-01 -7.68772066e-01 5.71812749e-01
8.41060221e-01 1.15375817e+00 -5.13439059e-01 -1.47425488e-01
5.31037807e-01 7.08277881e-01 -4.60586637e-01 3.16376239e-01
-3.35225135e-01 5.37077546e-01 -5.08652866e-01 9.19468880e-01
9.35855567e-01 -5.19503653e-01 -6.25462949e-01 -1.98127285e-01
-4.92616557e-04 2.28468090e-01 -8.10823619e-01 2.17670417e+00
-7.66384065e-01 4.35934305e-01 -2.68362486e-03 -1.07389796e+00
1.39344418e+00 2.22072050e-01 7.69372404e-01 -4.34832782e-01
-2.67962050e-02 -5.81362993e-02 -4.00164217e-01 -3.08207750e-01
-3.75166014e-02 -2.93558031e-01 7.44721442e-02 4.97876644e-01
3.82901371e-01 -4.48617339e-01 -9.19161856e-01 -1.06449410e-01
1.23099840e+00 2.61433601e-01 -2.17308655e-01 -6.01867028e-02
3.22654009e-01 7.66014084e-02 2.17194125e-01 5.12095630e-01
2.02002600e-01 1.01107395e+00 -4.96853590e-02 -5.09405494e-01
-9.32468832e-01 -1.23231852e+00 -2.06342146e-01 7.96326995e-01
7.62264132e-02 -1.64918944e-01 -3.35807174e-01 -8.64897966e-01
3.47009510e-01 5.66413164e-01 -6.91101968e-01 -3.75439852e-01
-5.87134898e-01 -4.13864970e-01 3.46874416e-01 7.68241346e-01
5.28421938e-01 -1.23414791e+00 -2.43387073e-01 1.24433041e-01
2.97848254e-01 -1.10300243e+00 4.19111960e-02 5.37731171e-01
-1.26170731e+00 -8.08177531e-01 -5.78369737e-01 -1.16811025e+00
9.79415774e-01 6.47669792e-01 1.13021231e+00 1.42991722e-01
-1.77801043e-01 3.78082722e-01 -2.97481507e-01 -2.53598392e-01
2.55974345e-02 1.39941022e-01 -2.62255631e-02 1.30572051e-01
5.33034265e-01 -1.21756029e+00 -4.56631094e-01 4.08890694e-01
-7.39929795e-01 -1.50468767e-01 8.89634967e-01 9.12684739e-01
1.13714731e+00 6.90345615e-02 3.71440232e-01 -7.70554900e-01
-4.65483963e-02 -6.84807897e-01 -4.22812581e-01 -2.59644032e-01
-2.77051628e-01 -8.52485672e-02 4.35924411e-01 -4.26138550e-01
-6.77371204e-01 3.24157029e-01 -2.77998567e-01 -1.54614913e+00
-4.45419222e-01 2.60945946e-01 -3.78192097e-01 -3.86296064e-01
7.65325904e-01 5.24921060e-01 1.51162073e-01 -6.62976086e-01
2.76475459e-01 5.22254884e-01 4.41705197e-01 -5.93586266e-01
1.33741760e+00 8.74847829e-01 -7.50761777e-02 -7.49167621e-01
-8.08827221e-01 -5.18766344e-01 -7.63314724e-01 1.07499622e-01
7.98468947e-01 -1.39425457e+00 -4.99599963e-01 2.04222396e-01
-1.11459887e+00 -4.25334364e-01 -6.19580150e-01 2.40064532e-01
-6.11898601e-01 8.30417946e-02 -2.41110817e-01 -4.96392816e-01
-4.44351822e-01 -1.12003243e+00 1.50091660e+00 -2.17944086e-01
2.24953994e-01 -6.40984595e-01 -4.61026393e-02 2.63589889e-01
2.70823568e-01 2.87690729e-01 8.89365911e-01 -5.13249934e-01
-9.34174001e-01 -2.64288455e-01 -3.74280959e-01 6.96377993e-01
3.03294569e-01 -6.16554439e-01 -1.46744692e+00 -4.14000839e-01
3.38593215e-01 -3.13172340e-01 1.08178341e+00 3.48594517e-01
1.71115553e+00 -2.04877838e-01 -5.77983856e-01 1.10838020e+00
1.39542079e+00 -1.91614509e-01 5.27591884e-01 -2.12182719e-02
1.07510555e+00 4.32921290e-01 2.89033055e-01 2.44271472e-01
2.04358101e-01 3.66626054e-01 7.05006242e-01 -2.33239412e-01
-2.14649707e-01 -5.40404260e-01 1.50830269e-01 8.68375599e-01
3.33892182e-02 1.40523583e-01 -9.80268836e-01 6.55690432e-01
-1.60236204e+00 -5.37891448e-01 2.61080027e-01 1.79080486e+00
5.10544598e-01 1.42608792e-01 -3.83897811e-01 1.18363157e-01
6.59042656e-01 4.07938212e-01 -7.72014797e-01 4.03315395e-01
-2.86774039e-01 7.71199465e-01 3.31965894e-01 2.02951789e-01
-1.30479050e+00 1.23697627e+00 5.07341290e+00 9.07204568e-01
-1.35948026e+00 3.21363986e-01 4.57121253e-01 -1.06819302e-01
-3.43027771e-01 2.40911520e-03 -5.44557571e-01 2.84276038e-01
4.64605570e-01 5.76576829e-01 3.24935019e-01 1.07720637e+00
-2.71939695e-01 6.70808911e-01 -1.25572324e+00 1.36243701e+00
1.78991750e-01 -1.61618030e+00 2.24251430e-02 9.29638967e-02
8.07474911e-01 5.79124629e-01 3.03740710e-01 5.23727179e-01
1.77786112e-01 -1.15388048e+00 5.23195148e-01 2.87038118e-01
1.26452446e+00 -6.60841763e-01 3.75536770e-01 3.60132635e-01
-1.15134072e+00 7.02442275e-03 -7.28492916e-01 -3.49546112e-02
-1.64740399e-01 7.00088561e-01 -7.61947215e-01 4.79969114e-01
1.00821900e+00 1.16042280e+00 -3.31283569e-01 5.89198589e-01
-2.98809499e-01 7.35307634e-01 -4.91217047e-01 3.43541116e-01
2.36708969e-01 -1.08327046e-01 6.75487041e-01 7.49990046e-01
2.91800886e-01 2.30015963e-01 4.26374823e-02 1.45183957e+00
-2.03198239e-01 -1.95396990e-01 -8.68766963e-01 2.61558801e-01
4.87426072e-01 1.17861867e+00 -4.35859829e-01 -2.60066390e-01
-3.77641052e-01 9.39821541e-01 4.69657063e-01 4.13358569e-01
-5.35928488e-01 -2.05821097e-01 9.54780757e-01 1.84592843e-01
7.91636229e-01 -2.53527880e-01 -5.75807452e-01 -1.19550955e+00
1.43121779e-01 -5.94214082e-01 1.22146867e-01 -8.57595265e-01
-1.67341495e+00 7.42039442e-01 -1.37890399e-01 -1.75593889e+00
1.47173718e-01 -6.50068879e-01 -8.55920136e-01 8.64232719e-01
-1.60843849e+00 -1.58129680e+00 -5.61776638e-01 7.34784126e-01
4.87937152e-01 -5.66775501e-01 8.90670061e-01 2.18594059e-01
-1.99330419e-01 5.72286606e-01 -6.94438741e-02 2.31877640e-01
3.40754241e-01 -9.10127401e-01 6.91874743e-01 3.76204252e-01
4.07973439e-01 5.98636568e-01 -1.35029763e-01 -6.87611282e-01
-1.54631197e+00 -1.59328461e+00 3.46791387e-01 -6.32846951e-01
2.88507968e-01 -7.58980095e-01 -1.17734158e+00 7.98378468e-01
-2.13252425e-01 6.07734501e-01 6.02530122e-01 -4.60184552e-02
-5.92343390e-01 -3.36589605e-01 -1.16071880e+00 2.15556562e-01
1.39555097e+00 -8.56862843e-01 -6.74955189e-01 4.34179127e-01
1.12417924e+00 -3.92456472e-01 -8.88006091e-01 7.28880823e-01
-7.41333738e-02 -6.29511178e-01 1.34793770e+00 -4.64056224e-01
4.83947664e-01 -3.59910190e-01 -3.57259214e-01 -1.19041479e+00
-5.33139706e-01 -2.19169974e-01 -8.82237107e-02 1.02484572e+00
4.09659058e-01 -5.89774072e-01 1.33015335e+00 6.90270886e-02
-7.24742889e-01 -9.30499256e-01 -1.16771448e+00 -6.98306561e-01
3.24395627e-01 -7.07406402e-01 1.00336909e+00 1.23255920e+00
-5.82899988e-01 3.02182019e-01 6.26820624e-02 6.43540204e-01
7.84242153e-01 5.36191463e-01 8.41383636e-01 -1.20768142e+00
-1.96549654e-01 -2.49465317e-01 -5.51875532e-01 -1.26481354e+00
6.71472847e-01 -1.28638136e+00 -1.46376237e-01 -1.24044108e+00
-2.38644496e-01 -9.75231588e-01 -4.12882656e-01 8.97670388e-01
9.80190709e-02 4.51213181e-01 -8.04313123e-02 5.55420399e-01
-1.38435364e-01 1.00929725e+00 1.18217325e+00 -5.87621033e-01
-2.60296673e-01 2.02897444e-01 -4.26214993e-01 4.17810291e-01
5.95841527e-01 -5.80341995e-01 -3.29002649e-01 -8.17045987e-01
-8.03915858e-02 -1.95339397e-01 7.79642820e-01 -1.17168856e+00
1.71512842e-01 1.79335028e-02 7.41184175e-01 -1.05172896e+00
7.48206437e-01 -1.11835563e+00 2.98708305e-03 1.72910303e-01
1.06988112e-02 -2.12684587e-01 3.97958368e-01 6.26078784e-01
-2.94766575e-01 2.13493425e-02 6.25725448e-01 -2.88058698e-01
-5.46013176e-01 7.90023923e-01 1.44623652e-01 -3.66148740e-01
8.82864714e-01 -4.32306170e-01 -4.95985225e-02 -4.35270630e-02
-7.49955535e-01 5.32233976e-02 4.88177150e-01 5.34663916e-01
1.11670887e+00 -1.46986294e+00 -6.78068638e-01 8.60902786e-01
3.69524390e-01 9.17246819e-01 5.07923961e-01 2.93395936e-01
-3.67053062e-01 9.83944163e-02 -2.27418810e-01 -1.13100374e+00
-6.79864585e-01 6.19022191e-01 2.79948324e-01 2.71674812e-01
-1.03613245e+00 1.21959746e+00 5.09054184e-01 -8.67385924e-01
1.98375419e-01 -4.24397260e-01 7.17686117e-02 -2.93336213e-01
9.35251117e-02 -5.68688922e-02 3.15574080e-01 -6.04375601e-01
-5.09019494e-01 9.80560720e-01 7.84309879e-02 1.67274788e-01
1.87985897e+00 3.12546492e-01 -9.48256552e-02 2.93154925e-01
1.61661053e+00 -1.68846384e-01 -1.55828321e+00 -6.62875116e-01
-4.69898045e-01 -3.20964098e-01 3.81783634e-01 -4.73358780e-01
-1.41570652e+00 1.07180977e+00 7.64722705e-01 -2.75005013e-01
1.01763952e+00 4.03450757e-01 6.50345862e-01 3.95948350e-01
3.39419216e-01 -4.45321113e-01 1.55476883e-01 4.75982457e-01
1.10891187e+00 -1.31102514e+00 -6.40889704e-02 -5.12348890e-01
-3.82033467e-01 7.46406019e-01 8.34967256e-01 -7.58772016e-01
1.08743107e+00 1.87479168e-01 2.01781854e-01 -6.54228389e-01
-7.46603131e-01 -1.87955946e-01 4.57680941e-01 8.03809702e-01
-1.34754226e-01 -5.76555729e-02 7.64711440e-01 7.65790403e-01
-3.84794801e-01 -3.62250298e-01 -2.73818702e-01 9.11139667e-01
-1.91816270e-01 -8.53156984e-01 -3.22879136e-01 6.26756847e-01
1.63283542e-01 -1.08273923e-01 -3.35521758e-01 4.92986470e-01
4.09238219e-01 4.74522918e-01 5.73318303e-01 -8.47639143e-01
3.98144454e-01 9.40694299e-04 3.44349414e-01 -9.67373371e-01
-6.68416142e-01 5.05631976e-02 -4.84117836e-01 -4.05738890e-01
-2.28737786e-01 -4.61711168e-01 -1.20257092e+00 1.91161007e-01
-3.41432899e-01 -7.23264292e-02 6.86398864e-01 6.87530518e-01
8.68850946e-01 4.23597574e-01 1.09684312e+00 -1.32070339e+00
-4.09274936e-01 -1.06361020e+00 -3.60916466e-01 2.32435733e-01
5.70163846e-01 -9.43622828e-01 -5.00585914e-01 -1.43683553e-01] | [8.097677230834961, -3.4093072414398193] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.