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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
5afa9db3-590a-42fe-b434-e7de6320be35 | nn-copula-cd-a-copula-guided-interpretable | 2303.17448 | null | https://arxiv.org/abs/2303.17448v1 | https://arxiv.org/pdf/2303.17448v1.pdf | NN-Copula-CD: A Copula-Guided Interpretable Neural Network for Change Detection in Heterogeneous Remote Sensing Images | Change detection (CD) in heterogeneous remote sensing images is a practical and challenging issue for real-life emergencies. In the past decade, the heterogeneous CD problem has significantly benefited from the development of deep neural networks (DNN). However, the data-driven DNNs always perform like a black box where the lack of interpretability limits the trustworthiness and controllability of DNNs in most practical CD applications. As a strong knowledge-driven tool to measure correlation between random variables, Copula theory has been introduced into CD, yet it suffers from non-robust CD performance without manual prior selection for Copula functions. To address the above issues, we propose a knowledge-data-driven heterogeneous CD method (NN-Copula-CD) based on the Copula-guided interpretable neural network. In our NN-Copula-CD, the mathematical characteristics of Copula are designed as the losses to supervise a simple fully connected neural network to learn the correlation between bi-temporal image patches, and then the changed regions are identified via binary classification for the correlation coefficients of all image patch pairs of the bi-temporal images. We conduct in-depth experiments on three datasets with multimodal images (e.g., Optical, SAR, and NIR), where the quantitative results and visualized analysis demonstrate both the effectiveness and interpretability of the proposed NN-Copula-CD. | ['Gang Li', 'Xueqian Wang', 'Weiming Li'] | 2023-03-30 | null | null | null | null | ['change-detection'] | ['computer-vision'] | [ 1.39547512e-01 -4.85336691e-01 2.95245320e-01 -2.37534270e-01
-3.50440204e-01 -3.19867015e-01 3.54319930e-01 -1.25338897e-01
-1.62713736e-01 6.02557540e-01 -2.92108625e-01 -3.19586009e-01
-6.64993346e-01 -9.21779215e-01 -5.53444266e-01 -1.31134903e+00
-2.60131061e-01 6.87805340e-02 -4.50380659e-03 -2.82237113e-01
-2.18134835e-01 6.26304209e-01 -1.30186200e+00 -7.83969015e-02
1.29975128e+00 1.35795379e+00 3.43283981e-01 3.18559736e-01
2.10837856e-01 4.13317680e-01 -4.96125132e-01 -6.59110025e-02
2.94330150e-01 -4.58084494e-01 -8.31456631e-02 9.82738659e-02
-8.35730955e-02 -1.96109802e-01 -1.69157937e-01 1.28286946e+00
7.51301527e-01 5.22939721e-03 6.46285117e-01 -1.19346058e+00
-1.05118263e+00 2.22404778e-01 -7.63234615e-01 3.59121352e-01
-1.80807263e-01 2.42267817e-01 8.41160893e-01 -9.24701750e-01
3.07136178e-01 1.26428354e+00 7.88295865e-01 -7.68278316e-02
-1.04742014e+00 -5.63120425e-01 1.38909057e-01 3.45939487e-01
-1.42682314e+00 1.57531112e-01 9.28602755e-01 -5.21457970e-01
3.72723848e-01 2.67343879e-01 7.36455381e-01 6.43942058e-01
3.03387016e-01 6.85693681e-01 1.17188048e+00 -1.34448141e-01
7.97793269e-02 -1.13922842e-01 -1.55201420e-01 6.02974534e-01
3.99481058e-01 2.91935831e-01 2.49382913e-01 1.76465079e-01
9.50584531e-01 3.96657526e-01 -5.70395887e-01 -2.16347337e-01
-1.12004840e+00 7.96014547e-01 9.64265108e-01 1.84446484e-01
-4.97619092e-01 -1.37154713e-01 3.55507195e-01 1.50991991e-01
5.85024297e-01 2.12730944e-01 -2.19890937e-01 4.66990501e-01
-6.16967440e-01 1.33803235e-02 3.83754700e-01 6.09031796e-01
6.60532594e-01 7.29743093e-02 -1.72032043e-01 7.84158111e-01
2.00795427e-01 1.00133741e+00 2.94099361e-01 -4.21757609e-01
4.81095403e-01 7.58978248e-01 1.89133175e-02 -1.82005167e+00
-6.03194177e-01 -6.75070524e-01 -1.50182235e+00 1.35625497e-01
-5.01331650e-02 -4.44588184e-01 -7.18955278e-01 1.41241980e+00
3.79046798e-01 -8.19627494e-02 -2.37089377e-02 1.11073077e+00
7.00687766e-01 8.75604928e-01 -2.02983320e-01 -3.81001383e-01
1.02212226e+00 -6.51203871e-01 -8.36543739e-01 4.73881364e-02
-5.10544777e-02 -2.00307608e-01 9.27304745e-01 2.62074381e-01
-4.93500292e-01 -6.04898691e-01 -1.21822941e+00 3.84392619e-01
-3.61623079e-01 2.94358969e-01 4.52449262e-01 2.15146452e-01
-6.91403806e-01 3.43992710e-01 -7.78259158e-01 -1.22872241e-01
6.40464246e-01 1.24487475e-01 -2.62041897e-01 -1.40102923e-01
-1.27355850e+00 7.23318756e-01 3.98820758e-01 1.00958467e+00
-8.22208405e-01 -3.82665753e-01 -6.24002099e-01 1.30898416e-01
3.28563988e-01 -4.73826230e-01 5.99738061e-01 -1.39205015e+00
-1.18357027e+00 5.58908641e-01 3.18102181e-01 -9.33344886e-02
5.67352176e-01 -1.38975764e-02 -6.81197107e-01 3.93468529e-01
1.59859136e-02 3.18550527e-01 9.38019753e-01 -1.49541414e+00
-5.87352276e-01 -3.28234971e-01 -4.27341200e-02 1.18696488e-01
-2.29088813e-01 -2.71598455e-02 -1.47238165e-01 -6.12687051e-01
3.48740071e-01 -7.82266200e-01 4.76190597e-02 3.39690447e-01
-2.71834910e-01 9.57234278e-02 1.07762921e+00 -9.14272130e-01
1.19489563e+00 -2.16740179e+00 3.42391469e-02 2.87399352e-01
2.39292562e-01 4.04222459e-01 -1.91329509e-01 1.97249070e-01
-1.57361224e-01 9.93597209e-02 -7.46306062e-01 1.71155915e-01
-7.78730288e-02 2.86088824e-01 -2.10805640e-01 6.92438841e-01
6.26430750e-01 7.02566564e-01 -9.31462705e-01 -4.28214997e-01
1.59530744e-01 5.48177540e-01 -1.83741033e-01 2.64882565e-01
-2.10872963e-01 4.25451368e-01 -5.18629730e-01 7.01595426e-01
1.13790846e+00 -1.84648916e-01 -7.59018362e-02 -3.39784682e-01
-1.92742422e-01 -6.29733920e-01 -1.05852354e+00 9.48856533e-01
-4.03679162e-01 8.72722566e-01 2.15409353e-01 -1.29456580e+00
1.20138490e+00 1.17035508e-01 4.27196205e-01 -6.44187331e-01
1.52365834e-01 2.70094782e-01 1.06588475e-01 -1.05324292e+00
4.64410409e-02 -2.89515644e-01 1.90608263e-01 1.13012947e-01
-5.57219625e-01 -3.13219763e-02 -1.70128897e-01 -4.06470418e-01
5.95307648e-01 3.38269547e-02 2.78009564e-01 -2.91713297e-01
7.69761801e-01 1.66503757e-01 7.26836860e-01 3.36146057e-01
-3.22173148e-01 7.07322419e-01 3.44323248e-01 -6.33221090e-01
-6.94227338e-01 -9.99290586e-01 -4.46403950e-01 4.32388037e-01
5.48834205e-01 6.19192779e-01 -4.32605952e-01 -4.86330509e-01
-5.79700656e-02 2.90363908e-01 -5.50965905e-01 -2.25308299e-01
-3.95916820e-01 -1.18716407e+00 4.13355589e-01 4.53699291e-01
1.18127322e+00 -1.08029008e+00 -4.52899754e-01 8.97674188e-02
-3.42620701e-01 -9.95318115e-01 -1.76133975e-01 1.91209644e-01
-7.01039135e-01 -1.11711776e+00 -6.87908649e-01 -8.45993161e-01
9.36536551e-01 4.24378067e-01 7.33631015e-01 6.05424419e-02
-2.61569083e-01 2.37117425e-01 -3.50002736e-01 -3.92480642e-01
-1.00696601e-01 -4.42447484e-01 -1.91658452e-01 3.33081841e-01
2.68633425e-01 -5.90733767e-01 -8.46131444e-01 5.02319753e-01
-1.24417603e+00 -1.67962357e-01 7.05378950e-01 1.20179486e+00
6.75258040e-01 5.90315938e-01 7.40968525e-01 -3.10800403e-01
7.60119081e-01 -6.51691556e-01 -8.27366173e-01 3.78266752e-01
-3.88495773e-01 -3.40347409e-01 8.39490771e-01 -4.42209423e-01
-1.09846056e+00 -2.04941615e-01 1.28074557e-01 -4.42125082e-01
2.60252431e-02 1.17868376e+00 -4.17348266e-01 1.07397996e-02
6.38592184e-01 3.75720412e-01 1.21345654e-01 -8.64900723e-02
6.79863989e-02 8.53539824e-01 6.75881088e-01 -4.40762758e-01
8.82418573e-01 6.52182937e-01 5.99282980e-02 -8.39052558e-01
-7.48987317e-01 -1.95964530e-01 -5.21978438e-01 -2.97749937e-01
9.88915801e-01 -1.05110812e+00 -6.69850171e-01 7.87768841e-01
-1.21560609e+00 -1.98173285e-01 1.00926816e-01 4.72232431e-01
-2.36429214e-01 4.58995461e-01 -1.97103247e-01 -8.88352394e-01
-5.94214499e-01 -9.77842808e-01 9.12958741e-01 4.31367487e-01
5.16757905e-01 -1.12808478e+00 -6.86511456e-04 1.49298996e-01
4.15573895e-01 8.27988148e-01 1.09965885e+00 -2.91735500e-01
-5.92498779e-01 -2.63001025e-01 -6.10232353e-01 7.17584848e-01
4.03359890e-01 2.79930741e-01 -8.96324277e-01 -2.25634515e-01
1.96947694e-01 4.45634052e-02 6.44055426e-01 6.37181342e-01
1.23052585e+00 -4.08630908e-01 -1.69675946e-01 8.11956227e-01
1.77734160e+00 5.15340686e-01 6.07900500e-01 4.67147708e-01
6.11121297e-01 5.08096397e-01 5.59807479e-01 5.18544376e-01
4.77660477e-01 2.87108958e-01 7.80928969e-01 -4.01269555e-01
3.03165317e-01 9.07339435e-03 3.32736850e-01 7.94675946e-01
-2.44874954e-01 -4.34936613e-01 -9.23075855e-01 6.45831227e-01
-1.83590627e+00 -1.05908966e+00 -2.49584213e-01 1.96157110e+00
6.62144423e-01 -2.08410710e-01 -1.96835086e-01 -5.11911474e-02
1.15234196e+00 1.34530097e-01 -7.70596206e-01 1.05112977e-01
-5.80316067e-01 -2.15564221e-01 1.54583961e-01 4.93557155e-02
-1.35772038e+00 3.02386940e-01 5.21627140e+00 7.57944882e-01
-1.44898200e+00 -1.12807594e-01 6.75348043e-01 2.97971517e-01
-2.27545694e-01 -2.37674430e-01 -3.29708219e-01 6.62876427e-01
1.97768077e-01 4.69073839e-02 3.43755811e-01 7.49734700e-01
6.41504884e-01 -1.02905013e-01 -6.72370136e-01 1.04304063e+00
-2.24895440e-02 -1.08839619e+00 -1.67130213e-02 -2.20959559e-01
1.03583348e+00 1.03162691e-01 2.34556958e-01 9.27971974e-02
7.60545433e-02 -9.80192482e-01 4.78800654e-01 7.45376885e-01
6.53800070e-01 -4.46086735e-01 1.12321842e+00 4.03719753e-01
-1.10134208e+00 -3.33603173e-01 -6.92947090e-01 9.13822651e-02
-3.35686319e-02 8.40297937e-01 -3.52813125e-01 8.93513441e-01
8.02122951e-01 9.51000035e-01 -4.13274199e-01 1.12746108e+00
-3.48936409e-01 4.62164819e-01 -3.80568653e-01 -8.53192881e-02
2.60170460e-01 -7.09845185e-01 6.10572457e-01 1.06849933e+00
5.18557012e-01 1.65666655e-01 -1.56148644e-02 1.09919238e+00
2.32274011e-01 -7.75885358e-02 -4.03583348e-01 1.66907255e-03
3.52154493e-01 1.35584474e+00 -6.55959070e-01 -3.05355657e-02
-1.03780620e-01 6.64470077e-01 -6.20153844e-02 5.11654258e-01
-8.43576372e-01 -5.37152290e-01 4.29074764e-01 -1.83886603e-01
5.33500791e-01 -2.76950270e-01 -2.82016397e-01 -1.14172018e+00
3.94936383e-01 -8.17850053e-01 3.57972175e-01 -1.00766194e+00
-1.68320930e+00 7.54447162e-01 7.31301084e-02 -1.58039629e+00
3.59189004e-01 -6.50640368e-01 -1.11389720e+00 9.12494063e-01
-1.85352433e+00 -1.13602698e+00 -9.71589327e-01 7.12614894e-01
1.75005393e-04 -5.65346591e-02 5.48380017e-01 3.66044760e-01
-9.01041806e-01 2.12631121e-01 5.26775181e-01 2.80802161e-01
3.13302517e-01 -9.69664216e-01 -1.11098990e-01 1.01772785e+00
-4.42840427e-01 4.29296285e-01 4.98765945e-01 -5.84703982e-01
-1.23265648e+00 -1.26564002e+00 2.03920916e-01 3.52433436e-02
7.36193836e-01 -3.17504518e-02 -8.81616473e-01 3.18270713e-01
6.16113916e-02 3.78741026e-01 2.76398897e-01 -3.88874590e-01
1.22719910e-02 -7.35056043e-01 -1.15205824e+00 3.98952186e-01
7.24963486e-01 -3.96018535e-01 -4.85830516e-01 4.14256662e-01
7.64206827e-01 -1.43003419e-01 -8.52102220e-01 5.11926830e-01
3.82847488e-01 -9.40088451e-01 8.01170647e-01 -2.23299414e-01
6.54047668e-01 -7.88915694e-01 -1.30821064e-01 -1.52617085e+00
-1.50590479e-01 -1.95582107e-01 3.45641971e-01 1.08736587e+00
1.25700206e-01 -9.39406216e-01 1.74845353e-01 1.38926804e-01
-2.09290907e-01 -7.66894042e-01 -8.63329768e-01 -8.52721989e-01
-3.36932461e-03 -2.87960917e-01 7.26983368e-01 9.24649835e-01
-3.87848616e-01 6.26337752e-02 -1.39067769e-01 7.07381785e-01
4.11745816e-01 3.10937375e-01 5.75318754e-01 -1.24708319e+00
-1.90101400e-01 -4.75839734e-01 -4.84030396e-01 -1.04218137e+00
-1.43944025e-01 -6.63362443e-01 2.76853949e-01 -1.75249875e+00
9.38954055e-02 -5.72976053e-01 -1.54373065e-01 3.72114837e-01
-3.32872927e-01 -1.97061643e-01 5.42170415e-03 3.05320650e-01
-2.04375550e-01 9.98194993e-01 1.64543235e+00 -4.98266041e-01
-1.39105305e-01 1.41491413e-01 -4.76582915e-01 5.37346721e-01
6.82786167e-01 -2.97452569e-01 -4.91278380e-01 -5.96850395e-01
3.68239582e-01 1.98665056e-02 6.90078974e-01 -1.05470252e+00
2.16989651e-01 -2.77967244e-01 3.49442780e-01 -7.99006045e-01
-3.12459916e-02 -1.30066693e+00 2.46372312e-01 3.97719741e-01
1.27816662e-01 5.87734804e-02 1.59513637e-01 7.03858674e-01
-6.74776912e-01 1.00666538e-01 8.02501142e-01 2.12656640e-04
-5.71355224e-01 5.69048882e-01 -1.26060024e-01 -9.50104892e-02
1.03638172e+00 -3.36836785e-01 -6.46328509e-01 -3.69491816e-01
-4.43701535e-01 4.59763825e-01 -1.04752686e-02 1.30802214e-01
6.80849969e-01 -1.29829466e+00 -7.02777803e-01 3.40609998e-01
1.27268136e-01 6.11529648e-01 6.04792714e-01 1.13499534e+00
-9.34787214e-01 8.08240920e-02 -3.39696497e-01 -9.40005243e-01
-7.08807647e-01 5.51909089e-01 8.21554005e-01 -7.51778781e-02
-5.33056855e-01 4.04878229e-01 3.02233696e-01 -5.58866680e-01
-1.23042561e-01 -6.31831765e-01 -3.01213235e-01 1.49211586e-01
3.09386998e-01 3.18404138e-01 -5.81725221e-03 -3.98784250e-01
-2.86936790e-01 7.43879437e-01 4.29335773e-01 2.87257880e-01
1.71841097e+00 -2.48552859e-01 -4.03715789e-01 2.06610173e-01
1.20997763e+00 -4.21355516e-01 -1.63057196e+00 -2.97931373e-01
-3.13330233e-01 -2.79718250e-01 8.76281261e-02 -8.63766015e-01
-1.41175795e+00 1.00725496e+00 8.48374188e-01 4.82101709e-01
1.56519926e+00 -4.01238918e-01 5.58904886e-01 5.44833660e-01
7.61782601e-02 -1.01969910e+00 6.57139271e-02 3.67459863e-01
1.18576300e+00 -1.48012447e+00 1.23172395e-01 -1.54508680e-01
-6.99873447e-01 1.37907648e+00 5.29690802e-01 -1.34249955e-01
8.75874281e-01 -1.40966475e-01 1.78080246e-01 -4.88318115e-01
-1.82856083e-01 -8.04486498e-02 2.14109316e-01 7.62467265e-01
-2.56724983e-01 1.80383399e-01 -1.26718864e-01 7.62580156e-01
1.72845960e-01 -1.20269910e-01 3.76631498e-01 5.96077204e-01
-3.34446311e-01 -2.95270592e-01 -6.18930936e-01 2.70657182e-01
-3.82238701e-02 -1.65875256e-02 -2.74435543e-02 9.54949558e-01
3.68262798e-01 9.47407246e-01 1.04268648e-01 -3.72961819e-01
2.93461651e-01 -5.75610936e-01 -2.18054913e-02 -1.18854441e-01
-3.55279177e-01 1.04443152e-02 -3.86118799e-01 4.26533073e-02
-7.98565388e-01 -6.82830930e-01 -1.21668804e+00 -1.42105862e-01
-4.76167560e-01 -1.47573158e-01 5.48817515e-01 1.06861639e+00
3.34427536e-01 4.71933067e-01 1.16872668e+00 -7.13792861e-01
-5.06462216e-01 -9.96619821e-01 -8.04459691e-01 2.09988758e-01
5.23587644e-01 -7.97519863e-01 -5.94497025e-01 8.87938868e-03] | [9.974052429199219, -1.5468181371688843] |
ea88a203-8f4a-4367-b61a-2f780700616a | hyperspectral-pansharpening-based-on-improved | 2107.02630 | null | https://arxiv.org/abs/2107.02630v1 | https://arxiv.org/pdf/2107.02630v1.pdf | Hyperspectral Pansharpening Based on Improved Deep Image Prior and Residual Reconstruction | Hyperspectral pansharpening aims to synthesize a low-resolution hyperspectral image (LR-HSI) with a registered panchromatic image (PAN) to generate an enhanced HSI with high spectral and spatial resolution. Recently proposed HS pansharpening methods have obtained remarkable results using deep convolutional networks (ConvNets), which typically consist of three steps: (1) up-sampling the LR-HSI, (2) predicting the residual image via a ConvNet, and (3) obtaining the final fused HSI by adding the outputs from first and second steps. Recent methods have leveraged Deep Image Prior (DIP) to up-sample the LR-HSI due to its excellent ability to preserve both spatial and spectral information, without learning from large data sets. However, we observed that the quality of up-sampled HSIs can be further improved by introducing an additional spatial-domain constraint to the conventional spectral-domain energy function. We define our spatial-domain constraint as the $L_1$ distance between the predicted PAN image and the actual PAN image. To estimate the PAN image of the up-sampled HSI, we also propose a learnable spectral response function (SRF). Moreover, we noticed that the residual image between the up-sampled HSI and the reference HSI mainly consists of edge information and very fine structures. In order to accurately estimate fine information, we propose a novel over-complete network, called HyperKite, which focuses on learning high-level features by constraining the receptive from increasing in the deep layers. We perform experiments on three HSI datasets to demonstrate the superiority of our DIP-HyperKite over the state-of-the-art pansharpening methods. The deployment codes, pre-trained models, and final fusion outputs of our DIP-HyperKite and the methods used for the comparisons will be publicly made available at https://github.com/wgcban/DIP-HyperKite.git. | ['Vishal M. Patel', 'Jeya Maria Jose Valanarasu', 'Wele Gedara Chaminda Bandara'] | 2021-07-06 | null | null | null | null | ['pansharpening'] | ['computer-vision'] | [ 9.14022088e-01 -2.63785958e-01 2.01610595e-01 -3.32280070e-01
-1.03246844e+00 -4.00423437e-01 2.94138908e-01 -4.93615657e-01
-1.48884058e-01 5.99168718e-01 6.51959181e-02 -5.24534248e-02
-2.44522557e-01 -1.28429008e+00 -7.73839414e-01 -1.18222737e+00
2.00084582e-01 -1.56996086e-01 -1.18379351e-02 -2.63446629e-01
-1.51629895e-01 5.90413272e-01 -1.57310534e+00 1.27295613e-01
1.43328214e+00 1.34906459e+00 4.63604689e-01 5.33085167e-01
4.27317232e-01 4.61816996e-01 -1.24341957e-01 8.18544626e-02
8.09781551e-01 -6.10570490e-01 -3.65169853e-01 1.96255922e-01
5.41640520e-01 -4.19921100e-01 -6.13029003e-01 1.52171934e+00
3.06315601e-01 5.13398945e-01 4.16948289e-01 -9.27533388e-01
-8.85170519e-01 4.72593457e-01 -1.02345467e+00 -3.43929678e-01
-1.05825715e-01 3.52391422e-01 8.79258096e-01 -7.36735940e-01
2.61286467e-01 6.42838955e-01 7.39353418e-01 7.91723207e-02
-1.12498808e+00 -7.97380745e-01 -2.73438215e-01 2.39233971e-01
-1.44371581e+00 -2.49988794e-01 1.12159240e+00 -6.76352456e-02
4.54894453e-01 4.32653904e-01 6.86755359e-01 5.44556677e-01
-6.43525124e-02 4.04666096e-01 1.37924755e+00 -4.55440730e-01
-5.45193106e-02 -2.97250301e-01 3.67864296e-02 6.53376162e-01
1.12821385e-01 5.00531673e-01 -5.89994565e-02 1.33998424e-01
7.38105536e-01 2.44166195e-01 -8.66848052e-01 2.36940961e-02
-9.16818857e-01 5.77328503e-01 1.19815588e+00 3.52179497e-01
-6.70041323e-01 -1.92374110e-01 -1.34682119e-01 5.64421490e-02
4.15353030e-01 3.93963128e-01 -2.62653798e-01 6.80567384e-01
-1.10608160e+00 -1.69271708e-01 2.14275479e-01 5.95485687e-01
1.38327551e+00 1.13405086e-01 -3.10108718e-02 1.01489162e+00
1.54971793e-01 6.32259011e-01 3.00043523e-01 -1.01402938e+00
2.79878646e-01 6.45245790e-01 4.67232913e-02 -8.68483782e-01
-3.45818698e-01 -5.16651452e-01 -1.23404706e+00 2.71337509e-01
1.01342000e-01 -3.60431761e-01 -1.22825897e+00 1.71301055e+00
2.06709579e-01 2.83309519e-01 1.32636562e-01 1.19327140e+00
7.03704298e-01 1.18200684e+00 -2.47239158e-01 -2.61850446e-01
1.01681519e+00 -1.21063757e+00 -3.75762224e-01 -4.43426400e-01
1.16063438e-01 -6.13624454e-01 1.02247286e+00 1.75217152e-01
-9.75825548e-01 -7.77490914e-01 -1.23908436e+00 -1.58371598e-01
-4.28097159e-01 5.81541717e-01 5.55239022e-01 4.68121558e-01
-1.07912958e+00 8.31366122e-01 -6.64174139e-01 -1.35329500e-01
3.77802402e-01 6.71272129e-02 -3.23191315e-01 -4.12620872e-01
-1.21099496e+00 4.50664014e-01 7.85322249e-01 5.12468219e-01
-7.47812092e-01 -8.83170664e-01 -8.04153025e-01 2.64934570e-01
4.23386723e-01 -4.55682218e-01 6.27419055e-01 -1.25828528e+00
-1.58357775e+00 6.40837729e-01 6.11922704e-02 1.44436210e-02
2.03326613e-01 -7.61010824e-03 -7.16523051e-01 3.84497166e-01
-2.71340627e-02 5.78933239e-01 9.59577620e-01 -1.40403259e+00
-6.12867594e-01 -4.02351767e-01 -2.62755826e-02 3.93405616e-01
-1.99709132e-01 -3.80915254e-01 -3.29458416e-01 -5.85190415e-01
3.95163506e-01 -8.02277863e-01 -1.79533347e-01 1.81275457e-02
-5.75887442e-01 3.51882577e-01 8.81785214e-01 -1.02599788e+00
6.75104201e-01 -2.29085875e+00 2.19373889e-02 1.70578703e-01
4.15221928e-03 5.54385781e-01 -5.45993984e-01 1.75133213e-01
-4.64603633e-01 -5.92931882e-02 -9.91261899e-01 -3.73055646e-03
-1.53067604e-01 -1.96756199e-01 -3.20801854e-01 5.93226969e-01
4.07546163e-02 8.63746405e-01 -7.49643981e-01 -6.61487281e-02
4.76617277e-01 7.27041900e-01 -5.01461811e-02 3.48151982e-01
-2.31746837e-01 3.15332681e-01 -1.21066734e-01 7.12958097e-01
1.26896632e+00 -9.18509513e-02 5.88356666e-02 -8.34968150e-01
-4.64035928e-01 -7.19648227e-02 -1.06841195e+00 1.61656213e+00
-2.79321820e-01 5.11299372e-01 4.17054564e-01 -9.79339778e-01
1.00133872e+00 7.41732866e-02 5.43957770e-01 -8.50311816e-01
1.44404862e-02 3.33481550e-01 -2.64662057e-01 -6.93322793e-02
5.51930487e-01 -2.82738209e-01 3.72490048e-01 2.59294897e-01
-1.28452271e-01 -5.15849233e-01 2.00245939e-02 -2.42918149e-01
3.34254861e-01 1.28308177e-01 9.35563967e-02 -2.77150780e-01
6.82441473e-01 7.96748921e-02 5.30429721e-01 6.34437263e-01
-3.31413746e-02 7.82781720e-01 1.03317700e-01 -1.85822502e-01
-1.02314889e+00 -8.57827485e-01 -2.23179042e-01 7.77655721e-01
3.58520001e-01 1.68166250e-01 -7.23082364e-01 -3.22813720e-01
-3.19904685e-01 8.80946815e-01 -5.25848866e-01 -1.90010384e-01
-4.04604971e-01 -1.19684470e+00 3.32912743e-01 2.78247416e-01
1.25061154e+00 -9.23790038e-01 -3.74907047e-01 8.69323835e-02
-3.28784257e-01 -9.54231262e-01 -4.92282271e-01 1.77119285e-01
-6.03786826e-01 -1.14374483e+00 -5.61785638e-01 -6.06259167e-01
4.71533895e-01 6.16994619e-01 7.26860046e-01 -1.16036780e-01
-2.61666417e-01 -7.37021193e-02 -3.39611232e-01 -1.13631651e-01
-1.67116091e-01 4.54182969e-03 -4.07726914e-01 3.22509080e-01
5.38840331e-02 -7.17235684e-01 -8.56950581e-01 2.17736527e-01
-1.45941174e+00 5.46101987e-01 8.20611417e-01 8.54333162e-01
8.09669197e-01 6.76459491e-01 2.69637167e-01 -6.22205555e-01
7.17289895e-02 -2.69006968e-01 -8.95853043e-01 4.15881783e-01
-4.93802637e-01 -1.87825561e-01 6.94686472e-01 -1.29070804e-01
-1.51783192e+00 1.41820967e-01 -3.49065036e-01 -2.66150385e-01
-3.30690384e-01 7.19579995e-01 -3.09978485e-01 -4.34870452e-01
6.35946035e-01 5.63792825e-01 -3.25507611e-01 -3.94929349e-01
3.62840861e-01 6.37915671e-01 8.55663478e-01 -2.59468794e-01
1.17937434e+00 7.42093146e-01 1.22126341e-01 -1.05873084e+00
-1.18992484e+00 -3.87696952e-01 -5.06209850e-01 -9.27214772e-02
8.28704476e-01 -1.05615866e+00 -3.34641427e-01 8.86695504e-01
-6.51772261e-01 -7.87203908e-01 -3.68355244e-01 4.27135527e-01
-3.33980411e-01 5.67075789e-01 -6.56403303e-01 -5.25853038e-01
-5.96675158e-01 -1.00439274e+00 1.03274345e+00 6.37060821e-01
7.68585563e-01 -7.53963590e-01 6.08882047e-02 4.62337524e-01
5.95888317e-01 4.85316187e-01 8.37675214e-01 7.98578411e-02
-6.42904162e-01 -2.52699312e-02 -8.85557652e-01 6.10668719e-01
3.41818988e-01 7.04656690e-02 -1.16960526e+00 -3.99042040e-01
2.79448420e-01 -4.46841747e-01 1.34513879e+00 6.95893168e-01
1.38403165e+00 -1.20899953e-01 -5.97527549e-02 1.44734788e+00
2.00345612e+00 1.39189571e-01 9.23601866e-01 2.72316366e-01
8.10727358e-01 4.92977709e-01 5.39074063e-01 7.62519911e-02
-6.35259319e-03 3.83866608e-01 6.84817374e-01 -6.23916030e-01
-2.75456041e-01 -2.44581640e-01 2.66839832e-01 3.28045905e-01
-3.28531176e-01 -2.35936552e-01 -5.10678649e-01 3.78623933e-01
-1.60279179e+00 -9.32225049e-01 -1.32392839e-01 2.04408646e+00
7.99898624e-01 -3.87269408e-01 -3.56885433e-01 4.71216924e-02
8.66102397e-01 6.75259292e-01 -8.62154484e-01 3.89308155e-01
-5.40922225e-01 3.55762303e-01 7.52075672e-01 6.90582633e-01
-1.25229883e+00 8.60682726e-01 5.03781080e+00 9.00385320e-01
-1.42183459e+00 -6.82647154e-02 6.62104309e-01 2.83006281e-01
-2.95945257e-01 -3.89994420e-02 -3.80228132e-01 1.88746899e-01
6.56046689e-01 -4.87319641e-02 8.14960659e-01 4.76042330e-01
8.88860226e-02 -1.14197001e-01 -5.21531701e-01 9.61295307e-01
-1.83915403e-02 -1.12811458e+00 -7.58736581e-02 -8.72753561e-02
1.09679258e+00 4.36293989e-01 6.17518015e-02 6.14916235e-02
3.11487347e-01 -9.80807245e-01 3.96266162e-01 6.56791449e-01
1.25154960e+00 -6.09573483e-01 6.05911255e-01 2.15744063e-01
-1.50217640e+00 -2.49765024e-01 -7.03179896e-01 2.78570890e-01
-2.78988704e-02 9.85731721e-01 -1.92425832e-01 1.09880519e+00
6.67501330e-01 8.72293115e-01 -3.82445991e-01 9.04826224e-01
-5.63394725e-01 5.42047381e-01 -3.11566502e-01 7.56451190e-01
3.05119187e-01 -9.40772176e-01 4.31082278e-01 1.07466590e+00
6.84722960e-01 4.78517145e-01 1.82315782e-01 1.37346220e+00
-2.66081750e-01 -2.98206180e-01 -3.41987014e-01 -8.71745124e-02
6.82983324e-02 1.65535271e+00 -4.66965914e-01 -2.78589517e-01
-3.26882690e-01 9.08737123e-01 -7.37748249e-03 7.31211424e-01
-8.43240857e-01 -5.23425937e-01 5.66529632e-01 -1.77869752e-01
2.19432741e-01 1.01886352e-03 -1.38067469e-01 -1.33811629e+00
-1.47074595e-01 -7.85738468e-01 3.79533499e-01 -1.29469025e+00
-1.13858926e+00 7.20284045e-01 -2.58128852e-01 -1.03017938e+00
2.76255548e-01 -5.86816251e-01 -6.31858110e-01 1.28117609e+00
-2.03002024e+00 -1.34541750e+00 -1.06417155e+00 7.24919140e-01
2.97569573e-01 3.55822176e-01 6.36846960e-01 1.53726563e-01
-7.29441583e-01 8.01658705e-02 2.99281746e-01 1.28573794e-02
5.65979779e-01 -1.04874802e+00 8.86609331e-02 1.26898575e+00
-2.67248273e-01 -7.75905140e-03 4.59328949e-01 -5.51364899e-01
-1.22248483e+00 -1.50953758e+00 1.99585006e-01 2.51685977e-01
5.03150761e-01 1.01582915e-01 -1.09800398e+00 6.62504971e-01
1.87315971e-01 1.54622942e-01 3.41422945e-01 -5.94821930e-01
-2.60763377e-01 -6.07324660e-01 -1.21031702e+00 4.56173807e-01
8.56476843e-01 -6.00897908e-01 -2.77737260e-01 2.97894090e-01
6.99984908e-01 -2.99748123e-01 -7.23446488e-01 8.03310931e-01
3.81180257e-01 -1.35138011e+00 9.70040739e-01 1.64595127e-01
6.33648574e-01 -5.27293742e-01 -4.10857767e-01 -1.55274320e+00
-6.09225869e-01 -2.51038134e-01 2.97601163e-01 8.98697495e-01
2.70056814e-01 -7.37060785e-01 6.69609189e-01 4.37914014e-01
-4.64037418e-01 -2.89747775e-01 -3.25350910e-01 -4.30654347e-01
-8.26208517e-02 -2.01428786e-01 8.14515591e-01 1.15446639e+00
-6.11301303e-01 -7.75627121e-02 -3.05121541e-01 8.63720357e-01
9.33241367e-01 7.38737166e-01 2.85153151e-01 -1.08611119e+00
-2.56355077e-01 -4.28551555e-01 3.45817775e-01 -8.33006561e-01
1.13576792e-01 -8.24500084e-01 2.40381688e-01 -1.67492008e+00
2.59470761e-01 -3.29547018e-01 -4.54518229e-01 6.96777582e-01
-3.35538566e-01 5.98715186e-01 7.45756701e-02 2.34557092e-01
4.89096306e-02 7.46308208e-01 1.45116019e+00 -3.84359568e-01
-3.35369378e-01 -2.38512501e-01 -6.76764667e-01 6.24900579e-01
1.00932288e+00 -2.91800559e-01 -1.79434448e-01 -4.81640279e-01
9.75552872e-02 2.20997497e-01 5.66118419e-01 -1.24829149e+00
2.21576005e-01 -3.98528606e-01 5.99770963e-01 -5.50492883e-01
4.39395219e-01 -8.82945716e-01 5.55620432e-01 2.70481795e-01
-1.07328422e-01 -8.53322506e-01 2.15936780e-01 2.09790871e-01
-3.38306278e-01 -1.87431738e-01 1.18989587e+00 -1.53039858e-01
-7.74222374e-01 6.42082214e-01 1.66980281e-01 -5.42174578e-01
7.88473248e-01 -1.49359226e-01 -6.36230409e-01 -3.26255560e-01
-2.89274395e-01 -8.88473019e-02 7.78125882e-01 -1.68463036e-01
5.76188505e-01 -1.02105975e+00 -7.74792194e-01 5.04702270e-01
-2.72115301e-02 1.84097335e-01 6.61235154e-01 7.23981917e-01
-7.64397442e-01 1.61171108e-01 -3.52468759e-01 -2.91072905e-01
-8.28870058e-01 4.14133161e-01 7.87736714e-01 -3.02583367e-01
-5.86293817e-01 6.08726025e-01 4.25685912e-01 -7.15484560e-01
-2.80093908e-01 -1.89362854e-01 -4.43796664e-02 -1.97731555e-01
4.86325592e-01 3.83830518e-01 3.09105273e-02 -6.31344736e-01
3.69530581e-02 6.91777527e-01 4.12802637e-01 1.13664918e-01
1.70203674e+00 -3.85047086e-02 -5.25182366e-01 -7.40507469e-02
1.18678987e+00 -6.69142678e-02 -1.79450655e+00 -3.95456880e-01
-5.52959979e-01 -5.96117735e-01 6.49791598e-01 -8.89327347e-01
-1.65161443e+00 8.01441312e-01 6.41086340e-01 1.66691557e-01
1.92345643e+00 -3.34124655e-01 7.76762128e-01 2.81381816e-01
5.19728400e-02 -8.74982238e-01 -2.54691631e-01 3.08264226e-01
8.48167837e-01 -1.22072005e+00 4.83360179e-02 -4.45665181e-01
-3.68845701e-01 1.28247321e+00 5.23345530e-01 -5.04960641e-02
4.78261262e-01 5.57741476e-03 2.85759792e-02 -1.86931670e-01
-1.20830357e-01 -3.12494576e-01 4.82170463e-01 6.06196225e-01
1.47933066e-01 2.71456391e-01 1.30896732e-01 3.52210253e-01
-8.21382925e-02 4.80403453e-02 4.81164306e-01 3.51397514e-01
-4.91502374e-01 -6.74921691e-01 -4.85381246e-01 4.63624418e-01
2.59262323e-02 -3.17235678e-01 -1.27544165e-01 5.43692410e-01
1.99690744e-01 8.93372953e-01 -1.02708086e-01 -3.44374120e-01
2.66308993e-01 -1.55385718e-01 1.81077048e-01 -2.73856372e-01
-1.73979223e-01 1.75834328e-01 -2.51962870e-01 -5.63459396e-01
-6.73514009e-01 -3.34056050e-01 -8.49129379e-01 -2.73746043e-01
-4.11980182e-01 1.07336663e-01 5.09149432e-01 6.80667222e-01
7.27764517e-02 4.49961752e-01 8.51753175e-01 -1.28144133e+00
-3.19053799e-01 -1.01840842e+00 -1.17041278e+00 2.08442822e-01
5.02908707e-01 -1.18640676e-01 -6.97282016e-01 -1.11876115e-01] | [10.210897445678711, -1.9791988134384155] |
8b291889-ca4b-4b15-a36a-252bff3746e1 | is-risk-sensitive-reinforcement-learning | 2307.00547 | null | https://arxiv.org/abs/2307.00547v1 | https://arxiv.org/pdf/2307.00547v1.pdf | Is Risk-Sensitive Reinforcement Learning Properly Resolved? | Due to the nature of risk management in learning applicable policies, risk-sensitive reinforcement learning (RSRL) has been realized as an important direction. RSRL is usually achieved by learning risk-sensitive objectives characterized by various risk measures, under the framework of distributional reinforcement learning. However, it remains unclear if the distributional Bellman operator properly optimizes the RSRL objective in the sense of risk measures. In this paper, we prove that the existing RSRL methods do not achieve unbiased optimization and can not guarantee optimality or even improvements regarding risk measures over accumulated return distributions. To remedy this issue, we further propose a novel algorithm, namely Trajectory Q-Learning (TQL), for RSRL problems with provable convergence to the optimal policy. Based on our new learning architecture, we are free to introduce a general and practical implementation for different risk measures to learn disparate risk-sensitive policies. In the experiments, we verify the learnability of our algorithm and show how our method effectively achieves better performances toward risk-sensitive objectives. | ['Dongsheng Li', 'Weinan Zhang', 'Xufang Luo', 'Kan Ren', 'Minghuan Liu', 'Ruiwen Zhou'] | 2023-07-02 | null | null | null | null | ['q-learning', 'distributional-reinforcement-learning', 'management'] | ['methodology', 'methodology', 'miscellaneous'] | [-0.25563127 0.05381701 -0.40415555 -0.3425316 -1.13269 -0.5166749
0.21181712 0.21696894 -0.55161244 1.0222887 0.07470082 -0.50763685
-0.857947 -0.95018387 -0.59449756 -0.72467846 -0.32043287 0.2717622
0.01793152 -0.1662441 0.45398498 0.43934125 -1.3411688 -0.3447203
1.2634718 1.0044807 -0.07288059 0.43763956 0.01379536 0.7128074
-0.5827922 -0.47292072 0.36632973 -0.29063523 -0.6374526 -0.41510174
-0.02800848 -0.39399132 0.18200801 1.3352929 0.44257024 0.35249528
0.81294155 -1.3817493 -0.42970446 0.7130603 -0.45737895 0.14829107
-0.04522844 0.08149338 1.2024915 -0.45661324 0.09031065 1.3842372
0.5165166 0.6827452 -0.88790864 -0.43599415 0.38815746 0.03675539
-0.91846997 0.22468051 0.55880624 -0.3280213 0.42335674 0.32814264
0.4839962 0.8700234 0.601597 0.8599974 1.3910015 -0.21185459
0.70249915 0.25862893 -0.01207401 0.33474204 0.46230692 0.68041533
0.01736211 -0.34776783 0.52724695 0.11820246 -0.03860125 -0.57058203
-0.72471654 1.069944 0.14423847 0.06454277 -0.34989184 0.26008788
0.3778549 0.4659154 0.49993488 0.564774 -0.2660139 -0.10891312
-0.5318582 0.60257715 0.715752 0.7370883 0.46561667 0.3138615
-0.60742885 0.37139416 0.3499042 0.57879335 0.3347829 -1.028492
0.56464195 0.34357733 0.6042761 -0.76614916 -0.37387225 -0.5251286
-0.4847181 0.5291296 0.42922646 -0.5044442 -0.4072284 1.8357866
0.22376099 -0.08694134 0.3714693 0.7971072 -0.1476125 0.67196774
0.32079318 -0.7658919 0.8171729 -0.4419741 -0.8216993 0.2525339
0.60221106 -0.21708508 1.5056787 0.4451074 -1.0171419 -0.10878947
-0.9581624 0.6590669 -0.12866734 -0.404983 0.49510834 1.0160468
-0.67660594 0.9163932 -0.58379936 -0.03131137 0.4812247 0.10202706
0.3387251 0.30839482 -1.500067 0.9693011 0.48540786 0.0114602
-1.2366347 -0.8680892 -0.5236419 0.13070826 0.8858736 -0.21464327
1.4678681 -0.85906196 -1.7565904 0.2712678 0.5793494 -0.7360717
0.91887105 -0.3530987 -0.30978402 0.06111706 -0.15722556 -0.24591875
0.97096187 -1.162158 -0.8559009 -0.34130406 0.32249576 0.2524276
-0.44234297 0.17136651 0.6238822 -0.7422884 -0.5883133 -0.67071444
-0.5728575 -0.41660157 -0.18721485 -0.40682113 0.31292272 -0.3342795
1.3571576 -2.0460668 -0.04837322 0.4264297 -0.18418795 0.3430999
0.10027377 0.44609612 0.08355072 0.2456626 -0.609208 0.04511894
0.5515072 0.1731004 -0.70231324 0.47325054 0.11672447 0.741029
-1.0731382 -0.27728492 0.10950855 -0.02117284 -0.54297787 0.53692317
-0.43565673 0.27055287 -0.8135976 0.41871086 0.6123879 0.30961868
0.03508542 0.28024012 -0.19325691 -0.30659354 -1.2276344 1.0686314
-0.51860535 -0.09324019 -0.09671845 -1.3424144 1.0420555 0.15974121
0.5765406 -0.5873887 -0.03657148 0.3186108 -0.30354762 -0.5026072
0.31627643 -0.6650992 -0.27324405 0.652707 -0.20372263 -0.02431791
-0.1669868 -0.24294147 0.6732429 0.18941282 0.3304922 -0.62743366
0.547938 -0.27307272 0.7466089 0.77637815 -0.48320788 0.01071108
0.9147003 -0.3105914 -0.6316082 -1.4146252 -0.08633575 1.1385626
0.01605586 -0.01239733 -0.70725214 -1.175774 0.27086058 1.147161
-0.47342446 -0.46208534 -0.4801327 -1.0560559 0.28045982 0.44281247
0.336678 -0.9662147 -0.8303041 0.19453897 0.34184176 -0.34582597
-0.39018926 0.09194983 -0.8455718 -1.0707959 -0.7680695 -0.11879517
0.40014082 -0.19082023 0.8778291 -0.4837304 0.16098519 0.52635217
-0.2563641 -0.8437246 -0.34220514 -0.10817298 0.22539279 0.2536053
0.08438563 -0.4251514 -0.6993605 0.27440196 -1.0262005 -0.80565125
0.35312715 0.6824355 0.60633695 0.4152105 1.3437423 -0.94053644
1.1262171 -0.5177143 -1.1939824 0.5957058 -1.1084152 0.56059587
0.8258093 -0.13943684 -1.2759539 -0.35783312 -0.08245768 -0.27961156
0.13800724 0.40370393 -0.3140855 0.13795863 0.5188739 0.01395843
0.12455659 -0.36054763 0.5054939 0.5018918 0.21986996 -0.8890887
0.77833194 0.332575 0.17957619 -0.19504666 -1.0662326 -0.06846742
-0.08847591 -0.30665302 0.75908726 -0.5444151 -1.0814916 0.13466744
-0.50143576 -0.2914491 -0.68774843 0.6411707 -1.2559512 0.23689088
-0.20182432 -1.4852809 -0.38028044 -1.011357 0.4724317 0.17669117
0.3075014 -1.2493461 0.3033865 -0.14345941 0.48950145 0.43730438
1.1129454 -0.59227264 -0.3667462 0.04442098 0.09999986 0.5568833
0.09015554 -0.07179578 -0.7494711 -0.63121504 0.34492996 -0.4212405
0.83055216 0.4353175 1.3828081 -0.7016075 0.20115215 0.5433521
1.7456013 0.43871132 0.46494603 0.5080571 0.31847322 1.0211028
1.303638 0.6897727 0.2104532 0.35619873 0.8177828 0.3158499
0.8004783 -0.43999463 0.58369607 0.17156565 -0.095978 0.09660993
-0.7452207 0.27225965 -2.0826733 -1.1824243 0.34707698 2.6037996
0.8584446 0.1523937 0.56807274 -0.04973875 0.5917175 0.08141223
-0.86246485 -0.7701769 0.17588127 -0.02338999 0.76100767 0.6038049
-0.9810608 0.6345277 6.79612 1.054393 -0.85367036 -0.04148761
0.6501591 0.00854317 -0.7348645 -0.15315266 -0.72371244 0.48566553
1.0502505 -0.6827785 0.18957959 0.98307973 0.40716165 0.151116
-0.8254949 0.6428201 -0.44253385 -0.9384386 -0.06277648 -0.05542659
0.71432084 -0.71047485 0.3883812 0.71195024 0.56963336 -1.2345976
0.60923415 0.8069056 0.4089854 -1.5793691 0.71483874 0.52151513
-0.6968741 -0.63592416 -0.5978222 0.04696633 -0.06143164 0.5166048
-0.49873286 0.78427815 0.44579327 0.54687655 -0.21916714 0.84400815
-0.22655635 0.6911972 0.10756584 -0.24719131 0.42199752 -0.34220406
0.54684657 0.81765527 0.43651062 -0.19799775 0.15989201 0.9490216
0.36667436 0.25881574 -0.65403336 0.07523978 0.27578115 1.006992
-0.2436976 0.10460157 -0.20467617 0.24904595 0.2763478 0.3632083
-0.848574 -0.3910347 0.7969381 -0.16460682 0.01417591 0.0177777
-0.08561419 -0.8656042 0.05380638 -0.8476083 0.81314653 -0.02872965
-1.6340954 0.30572867 0.25239655 -1.3644153 -0.37885213 -0.5707952
-0.5450657 0.7208566 -1.7607136 -0.5635712 0.40781116 0.7077902
0.10263205 -0.33527288 0.4765391 -0.10578538 -0.64782685 0.68692404
0.61222005 -0.3729809 0.6504765 -1.7508025 -0.1918776 0.66827303
-0.35270342 0.24387255 0.8080456 -0.47456175 -1.2529299 -1.3341168
0.25074726 -0.36306867 0.8398405 -0.03166262 -0.873244 0.3877146
-0.05590796 -0.2204156 0.548396 -0.09835667 -0.16092685 -0.6364334
-1.3840052 0.71031976 0.7988957 -0.22575596 -0.7214496 0.27604088
1.0430369 -0.02244936 -0.985084 0.53495693 0.24399282 -0.99045396
0.9124381 -1.002445 0.08480693 -0.02975441 -0.20211717 -1.5644172
-0.04255497 -0.8991555 -0.19503829 1.1186827 0.16531342 -0.92344266
0.50611705 0.4951437 0.04982243 -0.97074956 -1.1424627 -1.3323221
0.7570801 -0.47850066 0.86432314 0.5334719 0.13938808 -0.30811396
-0.53748524 0.14532542 1.1034454 0.35751063 0.3262804 -0.90153664
-0.3229783 -0.66275364 -0.03919106 -0.35568377 0.47368395 -0.6967117
0.13539505 -1.2154166 0.04819325 -0.52597654 -0.87631184 0.11519223
-0.4491411 -0.5148064 0.15137587 -0.11789756 -0.6274313 0.9353892
1.263426 0.02858949 -0.24789952 0.6418097 -0.8883652 0.67298883
1.2263402 -0.44104156 -0.8449412 0.22604226 0.35844973 0.3149416
0.27656573 -0.71003014 -0.23149902 -0.8965408 -0.19805244 -0.4132378
-0.31661224 -0.7609262 -0.14308985 0.7310308 -0.74870056 0.11369313
-0.15007117 0.770229 -0.10990236 -0.5354485 0.8948458 -0.01014894
-0.41587895 0.5895239 -0.3521904 0.40528736 1.2713374 0.3101937
-0.12477117 -0.38420817 -0.398772 0.5293315 0.03786289 0.22854361
0.67652315 -1.4035201 -0.7316743 -0.18074739 0.01732855 -0.26634455
0.18332478 0.5338853 -0.07527982 0.46469742 -0.23090151 -0.12362468
-0.61908025 0.87975466 0.701668 -0.57197404 -0.3572655 0.32928035
0.15587509 -0.3869283 0.40824333 -0.31526423 -0.3282931 0.16967535
0.59091145 0.6887705 -0.13504691 -0.09033035 -0.08559857 0.492326
0.12432088 -0.2946124 1.1262832 -0.16529128 0.23682404 0.5286957
0.9113866 -0.14442086 -1.6934056 -0.03714357 0.52874136 -0.5189307
-0.18181998 -0.6926085 -0.896046 0.82560605 0.7214099 0.34245834
1.1986604 -0.48449662 0.50958186 0.3601801 0.66127545 -1.422035
0.09651251 0.33003405 0.91482246 -1.2383598 -0.22501199 0.08820897
-0.88102 0.96198475 0.53405756 -0.33553287 0.7275599 0.13112313
0.04402487 0.2488515 -0.7316021 -0.24749427 0.04782191 0.8008261
-0.05334031 0.3925517 -0.63416183 0.69205195 -0.03427527 -0.1594795
0.44035867 0.7047163 -0.66356623 -1.0879292 -0.26319364 0.30336702
-0.67534614 0.32707486 0.17968087 0.72807735 -0.32455555 0.93697846
-0.22737445 -0.08013187 0.52828777 -0.09840407 0.29551196 -0.26065412
-0.42818236 -0.21882755 -0.27624112 -0.66384166 -0.3271774 -0.582195
-1.252288 -0.1733494 0.0967699 0.45099092 0.22686735 0.9501981
0.01726582 0.3006 1.2770494 -0.06786361 -1.8469117 -0.4550141
-1.0281748 0.22049421 0.44346878 -0.6979957 -0.54569286 -0.640586 ] | [4.238212585449219, 2.571080446243286] |
123e64e6-be8a-43f5-ae72-394689ee42e1 | occlusion-net-2d3d-occluded-keypoint | null | null | http://openaccess.thecvf.com/content_CVPR_2019/html/Reddy_Occlusion-Net_2D3D_Occluded_Keypoint_Localization_Using_Graph_Networks_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Reddy_Occlusion-Net_2D3D_Occluded_Keypoint_Localization_Using_Graph_Networks_CVPR_2019_paper.pdf | Occlusion-Net: 2D/3D Occluded Keypoint Localization Using Graph Networks | We present Occlusion-Net, a framework to predict 2D and 3D locations of occluded keypoints for objects, in a largely self-supervised manner. We use an off-the-shelf detector as input (like MaskRCNN) that is trained only on visible key point annotations. This is the only supervision used in this work. A graph encoder network then explicitly classifies invisible edges and a graph decoder network corrects the occluded keypoint locations from the initial detector. Central to this work is a trifocal tensor loss that provides indirect self-supervision for occluded keypoint locations that are visible in other views of the object. The 2D keypoints are then passed into a 3D graph network that estimates the 3D shape and camera pose using the self-supervised re-projection loss. At test time, our approach successfully localizes keypoints in a single view under a diverse set of severe occlusion settings. We demonstrate and evaluate our approach on synthetic CAD data as well as a large image set capturing vehicles at many busy city intersections. As an interesting aside, we compare the accuracy of human labels of invisible keypoints against those obtained from geometric trifocal-tensor loss.
| [' Srinivasa G. Narasimhan', ' Minh Vo', 'N. Dinesh Reddy'] | 2019-06-01 | null | null | null | cvpr-2019-6 | ['3d-car-instance-understanding', '3d-object-reconstruction-from-a-single-image', 'vehicle-pose-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-2.72838563e-01 3.89359027e-01 -2.63422906e-01 -4.59006488e-01
-8.59920740e-01 -7.62209713e-01 5.57342052e-01 -1.74950525e-01
-1.50941968e-01 2.27767885e-01 -9.85172391e-02 -2.22324401e-01
4.18362528e-01 -3.90882671e-01 -1.21693206e+00 -2.67259389e-01
7.23201856e-02 7.99431443e-01 5.07330775e-01 9.69428718e-02
1.48990065e-01 1.02170062e+00 -1.43198097e+00 7.85431080e-03
2.71919698e-01 8.52805912e-01 3.64753231e-02 7.07238972e-01
2.16108277e-01 7.20654428e-01 -2.48351574e-01 -5.39503336e-01
6.92878544e-01 3.59180778e-01 -2.92019069e-01 3.59105617e-01
1.52762258e+00 -6.04362667e-01 -7.53499627e-01 8.57144475e-01
1.20054312e-01 -5.20816930e-02 5.17295122e-01 -1.58295369e+00
-4.98916298e-01 -1.17402017e-01 -7.22954750e-01 1.58963367e-01
4.76591796e-01 3.33858073e-01 1.12331820e+00 -1.34444201e+00
9.72950459e-01 1.44436240e+00 8.10995221e-01 6.76311031e-02
-1.44457150e+00 -6.84446573e-01 3.46750528e-01 4.81484039e-03
-1.59628522e+00 -6.31562650e-01 1.09988225e+00 -6.07921183e-01
8.98917019e-01 -1.84648946e-01 5.14455795e-01 1.08920395e+00
2.80810863e-01 7.40869999e-01 5.76228142e-01 -1.98503345e-01
7.45390058e-02 1.35216430e-01 4.35468554e-02 1.22763383e+00
2.03944802e-01 3.04486454e-01 -3.97802204e-01 -1.70983344e-01
8.35015059e-01 3.28741193e-01 5.11620268e-02 -1.29479361e+00
-1.14548886e+00 8.10777247e-01 7.34635830e-01 -3.71219933e-01
-9.51696038e-02 4.72529948e-01 2.99605206e-02 1.52216330e-01
6.22150064e-01 6.32435754e-02 -5.48047245e-01 4.01577741e-01
-5.83561242e-01 2.65377641e-01 6.34343922e-01 1.22030008e+00
1.16643584e+00 2.27902960e-02 1.59598336e-01 2.78690189e-01
4.32079077e-01 9.89332020e-01 -3.42738122e-01 -1.22211778e+00
6.57065511e-01 6.93864286e-01 2.34836832e-01 -1.27642310e+00
-3.93165916e-01 -4.83753562e-01 -2.59408385e-01 7.74213374e-01
2.74969071e-01 1.02977835e-01 -1.20332372e+00 1.54382718e+00
6.24545336e-01 3.12481403e-01 -2.40297526e-01 1.00106239e+00
7.12287784e-01 2.97847599e-01 -3.06193501e-01 4.29789633e-01
1.04749882e+00 -1.16591489e+00 -3.34385455e-01 -5.71936190e-01
5.20825028e-01 -8.59585464e-01 8.28033745e-01 4.75034490e-02
-9.65826750e-01 -6.14005446e-01 -8.84862065e-01 -4.65957254e-01
-3.07153165e-01 4.24304575e-01 5.42699158e-01 2.60487109e-01
-1.19756949e+00 1.36054799e-01 -9.37591016e-01 -3.41495603e-01
4.32190746e-01 2.11077586e-01 -7.43991911e-01 -2.63008863e-01
-4.45525408e-01 1.07645953e+00 -7.23028928e-02 -4.81226034e-02
-1.16307175e+00 -7.45695591e-01 -1.28726983e+00 -1.62370235e-01
6.10817254e-01 -7.50701189e-01 1.05574203e+00 -6.04855299e-01
-1.14172196e+00 1.17798185e+00 -3.40348631e-01 -4.24149364e-01
7.84603775e-01 -3.50167453e-01 -1.48850933e-01 2.62708396e-01
6.46985590e-01 1.01531696e+00 9.74916458e-01 -1.59344840e+00
-6.59632325e-01 -5.95260859e-01 1.95725426e-01 1.85193554e-01
4.73208845e-01 -3.59017372e-01 -8.57579350e-01 -4.61183727e-01
4.11076009e-01 -1.19534504e+00 -1.15716584e-01 4.74902183e-01
-6.70834303e-01 -2.01460928e-01 1.26515007e+00 -4.80193824e-01
2.96824366e-01 -2.09391403e+00 8.21564719e-03 2.80834794e-01
3.94873619e-01 -8.13572332e-02 -2.51794845e-01 1.57975584e-01
-1.65267780e-01 -2.68205136e-01 2.36950815e-01 -9.03016031e-01
-9.80515927e-02 1.62124142e-01 -4.66107279e-01 8.93814743e-01
3.33285242e-01 1.15315700e+00 -8.38314712e-01 -2.99187809e-01
8.39907825e-01 6.23457253e-01 -6.06611609e-01 2.35931560e-01
-2.89331883e-01 3.09126794e-01 -3.63084614e-01 8.18157315e-01
7.82201588e-01 -3.49119037e-01 -5.06986558e-01 -4.81016308e-01
-4.19857614e-02 4.09977660e-02 -1.20148420e+00 1.94395745e+00
-3.14599484e-01 7.82484770e-01 -2.91479863e-02 -5.54215491e-01
7.91298270e-01 -5.28145842e-02 4.68366534e-01 -4.48288947e-01
-2.06946186e-03 -4.14003171e-02 -5.36081076e-01 -3.30877066e-01
3.27729672e-01 4.31696713e-01 1.04125738e-01 2.89590985e-01
1.28065154e-01 -2.02452317e-01 -1.59626782e-01 3.54215533e-01
1.20375478e+00 2.99704373e-01 1.47258677e-02 1.09689593e-01
3.02834362e-01 1.29408583e-01 3.67220610e-01 4.48309392e-01
-6.62833601e-02 1.03419948e+00 3.35995913e-01 -9.09011006e-01
-1.31276441e+00 -1.29911661e+00 1.99675970e-02 7.09687412e-01
4.71931279e-01 -3.57496947e-01 -3.19851428e-01 -1.03325605e+00
3.74523252e-01 5.50104082e-01 -6.30672038e-01 6.65017888e-02
-5.19469857e-01 2.30439425e-01 1.92677200e-01 4.82567728e-01
2.33035371e-01 -6.10629320e-01 -5.62830567e-01 -1.78813800e-01
-3.65592465e-02 -1.55475140e+00 -7.20344782e-01 1.98414579e-01
-5.07105350e-01 -1.37432456e+00 -3.70185673e-01 -8.10522139e-01
1.03222895e+00 7.85864770e-01 1.02949345e+00 2.62599345e-02
-2.62167245e-01 6.28154457e-01 5.12653254e-02 -1.32381514e-01
-2.10896954e-01 -1.01204589e-01 4.49970514e-02 1.72404572e-01
3.08203995e-01 -3.74526262e-01 -6.20141149e-01 5.12092233e-01
-1.76473156e-01 3.74093987e-02 5.43331504e-01 3.71103466e-01
8.47222865e-01 -2.22266868e-01 -2.03140929e-01 -5.84742308e-01
-2.30105236e-01 -2.18303129e-01 -1.06940961e+00 8.66734982e-02
-2.23664135e-01 8.92838314e-02 3.49000245e-01 -2.30562195e-01
-6.46378040e-01 7.43390739e-01 2.31302559e-01 -1.22591412e+00
-2.26786748e-01 -1.87479302e-01 -1.39993727e-01 -3.72059733e-01
5.60983539e-01 -2.06307650e-01 -6.11446686e-02 -4.81077909e-01
6.67363346e-01 1.81138024e-01 7.01617241e-01 -2.90127546e-01
1.19671655e+00 9.35926259e-01 7.20001534e-02 -6.25674784e-01
-1.10973465e+00 -6.64079785e-01 -9.21888411e-01 -1.92415446e-01
9.83807087e-01 -1.31385624e+00 -6.99872077e-01 2.46200599e-02
-1.50729072e+00 -3.07451010e-01 -3.57895613e-01 4.32876080e-01
-6.14810050e-01 1.07510932e-01 -2.59936392e-01 -5.84099472e-01
4.34470288e-02 -1.32128704e+00 1.91097689e+00 -2.91498542e-01
-4.10775095e-03 -8.57840419e-01 -7.69105926e-02 3.30442250e-01
-2.07473174e-01 3.69267553e-01 4.47253078e-01 -2.10838646e-01
-1.21092212e+00 -5.03114700e-01 -3.69362444e-01 2.62345046e-01
-1.62458614e-01 -4.46777977e-02 -1.11586833e+00 -3.72050941e-01
-2.72343725e-01 -2.79770404e-01 7.92412281e-01 4.99644995e-01
7.76612878e-01 2.24938672e-02 -6.96977735e-01 8.71795535e-01
1.24082756e+00 -2.08989888e-01 2.38485813e-01 1.03481308e-01
1.26159298e+00 5.11579335e-01 4.68313575e-01 -1.77834690e-01
7.66411901e-01 8.40860009e-01 7.35955060e-01 -3.92486751e-01
-4.16222155e-01 -7.86740720e-01 2.04253346e-01 1.48920089e-01
2.97547787e-01 -1.50375128e-01 -9.21760321e-01 5.00061691e-01
-1.85366714e+00 -7.78229535e-01 -1.91317067e-01 2.20431828e+00
2.22536638e-01 3.39115173e-01 -1.17917970e-01 -2.77730674e-01
6.30075514e-01 2.43842840e-01 -8.15179586e-01 1.69256181e-01
-1.41176935e-02 -3.31605613e-01 9.15160000e-01 9.88821566e-01
-1.34273314e+00 1.22541714e+00 6.26829195e+00 2.73610741e-01
-1.04249585e+00 5.38554937e-02 5.82346320e-01 3.36542539e-02
-1.24715261e-01 2.79315919e-01 -9.20952320e-01 1.51684940e-01
3.42975080e-01 3.79267216e-01 2.75266051e-01 1.18612897e+00
7.79680982e-02 -4.96537015e-02 -1.38756585e+00 1.15195012e+00
4.52385783e-01 -1.41916275e+00 -1.73149154e-01 3.01292360e-01
8.51621091e-01 6.67270958e-01 1.30964592e-01 -1.02279022e-01
5.51375568e-01 -6.32530510e-01 9.71269071e-01 4.29482967e-01
8.30019116e-01 -5.04580855e-01 4.19967979e-01 1.55231908e-01
-1.22736216e+00 1.29216030e-01 -3.26650500e-01 8.77215266e-02
3.38794112e-01 4.96210396e-01 -1.02871919e+00 1.35563344e-01
7.14379966e-01 9.39132035e-01 -8.05976510e-01 9.13007379e-01
-6.11197889e-01 1.21454850e-01 -5.89244783e-01 6.28731728e-01
9.39240903e-02 2.67887823e-02 7.79287815e-01 8.34057093e-01
7.07525536e-02 -2.14509144e-01 5.66840649e-01 9.49192822e-01
-1.25138730e-01 -4.48839128e-01 -1.10742009e+00 6.01690590e-01
5.55025816e-01 1.33236063e+00 -7.92180538e-01 -3.30872595e-01
-4.12163734e-01 9.74717975e-01 4.86980826e-01 7.03352571e-01
-8.86702061e-01 6.33618906e-02 7.17398167e-01 4.32257324e-01
6.26977980e-01 -5.25635064e-01 -9.01261792e-02 -1.31561637e+00
3.17200989e-01 -3.02064657e-01 4.22132947e-02 -1.19588387e+00
-1.19191086e+00 3.67711097e-01 3.22130620e-02 -1.34783626e+00
-1.54866114e-01 -5.94463766e-01 -3.85213137e-01 6.53437138e-01
-1.52644181e+00 -1.50944281e+00 -4.72986907e-01 5.01734972e-01
4.84498799e-01 -7.46345073e-02 3.75676006e-01 1.90429851e-01
-2.58075774e-01 3.84632528e-01 -2.77204573e-01 2.17223898e-01
6.59287691e-01 -1.21781731e+00 9.34186876e-01 9.09631014e-01
5.07373691e-01 4.42103595e-01 5.09316087e-01 -6.92071974e-01
-1.47030485e+00 -1.49253345e+00 8.09890270e-01 -1.17198479e+00
5.56628942e-01 -7.85740852e-01 -5.49749494e-01 1.24588549e+00
-1.85352340e-01 7.30261087e-01 -1.60313129e-01 -8.80918875e-02
-5.81678033e-01 -1.41746745e-01 -9.28077698e-01 6.04387045e-01
1.31656611e+00 -6.86961353e-01 -4.54066455e-01 6.46084189e-01
9.51956511e-01 -8.44694138e-01 -3.62416238e-01 2.65599638e-01
4.20576572e-01 -8.61589134e-01 1.33268774e+00 -3.90146792e-01
1.15466632e-01 -6.03678346e-01 -2.19663158e-01 -1.08216071e+00
-1.07786037e-01 -5.76725423e-01 -9.76266935e-02 7.60084391e-01
1.85377359e-01 -5.23942649e-01 1.05302513e+00 6.26486123e-01
-1.98234349e-01 -4.31630760e-01 -8.66345108e-01 -6.22557878e-01
-5.23108840e-01 -6.38613522e-01 3.23658705e-01 6.22056901e-01
-6.98527753e-01 6.63398802e-01 -2.51723498e-01 6.18168533e-01
1.06962287e+00 1.67381108e-01 1.50086474e+00 -1.24244356e+00
6.74244985e-02 5.50062954e-02 -1.03282058e+00 -1.39495838e+00
5.25997758e-01 -8.98045838e-01 -5.30249961e-02 -1.25458276e+00
-5.22533320e-02 -5.21109223e-01 2.03640312e-01 5.06716907e-01
1.89056307e-01 5.65128922e-01 1.76102415e-01 2.70180315e-01
-7.77956903e-01 4.14841413e-01 1.15118074e+00 -2.69766212e-01
-1.91885427e-01 -6.85539544e-02 -2.43879184e-01 1.11871302e+00
3.12078387e-01 -5.27385592e-01 -4.44693238e-01 -8.10026407e-01
1.05694763e-01 -1.97690390e-02 1.02317953e+00 -9.10008073e-01
3.73305053e-01 1.08537255e-02 6.40144527e-01 -1.15258777e+00
6.40845776e-01 -1.15450740e+00 4.71245907e-02 2.98756640e-02
-1.93355039e-01 3.62441748e-01 2.31792349e-02 8.83260250e-01
1.88398123e-01 3.33143979e-01 4.33196157e-01 3.23973782e-02
-7.22453892e-01 7.17612505e-01 2.74057746e-01 3.26236375e-02
1.22867393e+00 -4.17429060e-01 -2.01981038e-01 -3.85236084e-01
-7.30265677e-01 3.30568820e-01 9.02695775e-01 7.35008299e-01
8.88525486e-01 -1.32994342e+00 -5.79233885e-01 6.59474969e-01
3.43686342e-01 4.02169555e-01 -1.14885814e-01 6.90154970e-01
-5.88598430e-01 5.42667150e-01 1.46712825e-01 -1.26224136e+00
-1.27844810e+00 6.58016324e-01 4.40175235e-01 1.89407825e-01
-1.18114126e+00 6.75770938e-01 5.79830289e-01 -6.52829707e-01
4.74941254e-01 -5.82129538e-01 3.14342141e-01 -3.83110374e-01
3.47080231e-01 2.08124235e-01 1.45110354e-01 -1.10006130e+00
-4.67049956e-01 9.49939311e-01 -6.91060871e-02 -1.85592949e-01
1.19265020e+00 -3.32644552e-01 1.65581107e-01 3.43086958e-01
1.66618848e+00 1.00467190e-01 -1.90597403e+00 -3.59467268e-01
-3.30486894e-01 -7.53276885e-01 1.21287532e-01 -4.63220268e-01
-1.17260432e+00 8.07694376e-01 6.53406620e-01 -2.50127852e-01
3.96466345e-01 3.64896744e-01 5.07277608e-01 5.83410561e-01
4.95031923e-01 -6.63534343e-01 2.23224059e-01 3.75985444e-01
8.98102641e-01 -1.54004645e+00 1.11537285e-01 -4.98605788e-01
-4.28821146e-01 9.03048217e-01 5.96938014e-01 -5.17367542e-01
5.95164895e-01 1.87522694e-01 2.04552844e-01 -5.80163419e-01
-5.23824632e-01 -1.67975485e-01 3.91592085e-01 7.38366663e-01
-1.32653490e-01 -9.59562734e-02 5.91960669e-01 -3.93886596e-01
-2.27455571e-01 -2.70087302e-01 2.66619682e-01 6.93816066e-01
-2.42484674e-01 -6.17940187e-01 -4.42995310e-01 6.53308034e-02
4.86711180e-03 1.80335552e-01 -4.41484839e-01 8.43602002e-01
2.04277396e-01 7.95087397e-01 2.36765593e-01 -3.57193172e-01
4.81304079e-01 -2.25753501e-01 4.26657200e-01 -7.15644658e-01
6.86374009e-02 1.42990044e-04 -1.13707244e-01 -1.02430224e+00
-3.08826089e-01 -5.51965773e-01 -1.16709054e+00 -4.34986725e-02
-3.42295736e-01 -2.43048996e-01 8.01743388e-01 7.06422448e-01
5.50560057e-01 2.33088583e-01 7.17374682e-01 -1.37116098e+00
-2.36800924e-01 -4.52342689e-01 -1.89214915e-01 4.78387028e-01
9.30177093e-01 -9.32732344e-01 -5.39599359e-01 1.90190196e-01] | [7.707833290100098, -2.555600166320801] |
9579963c-1013-4343-9c4c-3e6b02bfa692 | a-comparison-of-audio-preprocessing | 2212.05335 | null | https://arxiv.org/abs/2212.05335v1 | https://arxiv.org/pdf/2212.05335v1.pdf | A Comparison of Audio Preprocessing Techniques and Deep Learning Algorithms for Raga Recognition | Ragas form the foundation for Indian Classical Music. The task of Raga Recognition has gained traction in the Music Information Retrieval community in the recent past, which can be attributed to the nuances of Indian Classical Music that have resulted in a plethora of research problems in Computing. In this work, we used two different digital audio signal processing techniques to preprocess audio samples of Carnatic classical ragas that were then processed by various Deep Learning models. Their results were compared in order to infer which DASP technique is better suited to the task of raga recognition. We obtained state of the art results, with our best model reaching a testing accuracy of 98.1%. We also compared each model ability to distinguish between similar ragas. | ['Vandana Jagtap', 'Devayani Hebbar'] | 2022-12-10 | null | null | null | null | ['audio-signal-processing', 'music-information-retrieval'] | ['audio', 'music'] | [ 2.38270894e-01 -4.16373491e-01 1.08200088e-01 1.53639540e-02
-9.01031494e-01 -6.48131728e-01 6.11031175e-01 1.98708653e-01
-2.85124600e-01 2.44165391e-01 2.97493100e-01 1.12203456e-01
-4.72025096e-01 -8.85054529e-01 -3.02818567e-01 -8.07262361e-01
-2.37029552e-01 3.58487487e-01 3.28356326e-02 -4.72377509e-01
5.13982177e-01 6.53528035e-01 -1.87267888e+00 6.83840752e-01
3.61763090e-01 1.22636867e+00 1.30829429e-02 7.67837286e-01
2.08074544e-02 6.09802365e-01 -6.24569952e-01 -3.54478389e-01
2.91521609e-01 -6.50858998e-01 -7.52319157e-01 -1.75609410e-01
4.25592393e-01 9.39103514e-02 -4.43700671e-01 8.75747859e-01
7.30395377e-01 1.54354006e-01 5.84964573e-01 -6.51237905e-01
-2.69976765e-01 1.24170923e+00 -5.28315842e-01 3.08790058e-01
4.97477561e-01 -3.80300432e-01 1.46604121e+00 -7.08263874e-01
2.59745836e-01 1.13193130e+00 6.68762207e-01 8.05089921e-02
-1.30892909e+00 -7.22000480e-01 -6.90530300e-01 6.61876500e-01
-1.58236289e+00 -3.78984809e-01 1.09503889e+00 -4.46857810e-01
4.99463290e-01 4.21659291e-01 1.05511582e+00 7.99389362e-01
9.09785777e-02 8.87903035e-01 1.40423954e+00 -5.97025871e-01
1.66276202e-01 -4.91758585e-01 1.60063475e-01 8.88277814e-02
-2.11284399e-01 4.74870317e-02 -6.91695869e-01 1.59322187e-01
6.67361379e-01 -2.99098432e-01 -1.28849939e-01 2.32152104e-01
-1.20121050e+00 9.56881821e-01 4.45636481e-01 7.98349380e-01
-4.80766952e-01 -6.84131458e-02 4.95500088e-01 6.35014951e-01
-4.21884619e-02 7.45603740e-01 -5.24782669e-03 -3.43460947e-01
-1.25860357e+00 6.29269898e-01 5.91191530e-01 -4.72633988e-02
2.04262346e-01 2.27574572e-01 6.93059936e-02 1.25167096e+00
1.11736275e-01 1.93157956e-01 7.20678091e-01 -7.87337065e-01
-7.70966848e-03 4.26887512e-01 -3.79098982e-01 -1.19823420e+00
-4.40977186e-01 -5.71474016e-01 -1.00133693e+00 2.44577050e-01
6.93675220e-01 4.58174855e-01 -7.17225671e-01 1.26112783e+00
-2.20157057e-01 -5.13773039e-02 -1.71690047e-01 9.31652725e-01
5.51737368e-01 7.83684790e-01 -3.56673419e-01 3.04675382e-02
1.24527776e+00 -3.43145043e-01 -5.75128257e-01 1.34207949e-01
3.83436345e-02 -1.44776905e+00 1.10034263e+00 1.17059159e+00
-1.00599754e+00 -8.56070757e-01 -1.44080770e+00 2.11695075e-01
-1.66938052e-01 1.98592663e-01 6.99003518e-01 8.19997072e-01
-7.64662802e-01 1.08330786e+00 -5.56244612e-01 -2.58998781e-01
5.75569451e-01 3.17142904e-01 -1.07451402e-01 2.28111312e-01
-1.08351684e+00 4.62708950e-01 5.91853559e-01 1.55675784e-01
-7.26978123e-01 -2.75762111e-01 -2.24381819e-01 -1.01168253e-01
2.37621710e-01 -3.00274968e-01 1.20271409e+00 -1.11466527e+00
-1.62393463e+00 1.26958621e+00 4.54212755e-01 -9.19531643e-01
3.92150789e-01 -4.12116140e-01 -6.73737824e-01 1.12882324e-01
-2.99335122e-01 1.47481769e-01 1.03551257e+00 -7.18337595e-01
-3.78524750e-01 -5.63852489e-01 -1.81502953e-01 -1.35709375e-01
6.25794474e-03 2.74666905e-01 -2.12129295e-01 -1.25554764e+00
6.12895310e-01 -1.07036734e+00 2.25663379e-01 -5.82398057e-01
-3.23817611e-01 -2.18779579e-01 4.18586552e-01 -7.77866006e-01
1.34778559e+00 -2.35616064e+00 4.51815337e-01 3.12392682e-01
-4.12332974e-02 3.10800850e-01 -1.49267986e-01 6.40150964e-01
-2.43362814e-01 -1.09730355e-01 -3.35312635e-02 2.33365282e-01
5.49881756e-02 -1.39990002e-01 -8.32723022e-01 3.27809572e-01
-9.07383114e-02 5.34614623e-01 -3.69178683e-01 -3.08705747e-01
2.46618614e-02 6.01286173e-01 -3.74914259e-01 2.28224769e-01
-1.43617451e-01 2.26413950e-01 -2.43274897e-01 7.71447062e-01
4.14417982e-01 2.92250782e-01 1.01104733e-02 -3.52149427e-01
-4.51281630e-02 5.51272452e-01 -1.38915241e+00 1.75124109e+00
-6.03828356e-02 7.87123442e-01 3.32456362e-03 -1.11877549e+00
1.10090852e+00 2.77072310e-01 5.61083555e-01 -7.78168559e-01
4.48657125e-01 4.60701495e-01 4.10697252e-01 -2.33992010e-01
7.52962112e-01 -4.46320146e-01 -1.12480722e-01 4.76303309e-01
-4.82468940e-02 -3.74168575e-01 2.48119101e-01 -1.62070125e-01
8.16658318e-01 5.64218536e-02 3.23725969e-01 -5.78494333e-02
6.24748826e-01 -1.42777205e-01 1.37729034e-01 6.82624936e-01
-4.01593670e-02 8.33747983e-01 2.28979722e-01 -5.32601416e-01
-7.91779459e-01 -1.34092510e+00 -1.20214656e-01 1.27453589e+00
-3.40731561e-01 -4.91172433e-01 -6.68331206e-01 1.07890613e-01
-2.76589870e-01 3.65641057e-01 -4.02687639e-01 -1.10342391e-01
-8.18213284e-01 -7.00779915e-01 9.51786697e-01 3.55225176e-01
5.19830227e-01 -1.42513037e+00 -5.55889785e-01 5.32842219e-01
-3.06159966e-02 -7.31092930e-01 -8.97491872e-02 1.82760969e-01
-9.00455952e-01 -1.11664331e+00 -6.81905210e-01 -5.63619912e-01
-5.51089346e-01 8.31343308e-02 1.07657659e+00 -2.60738641e-01
-5.14247537e-01 2.24807233e-01 -5.61067164e-01 -4.31929410e-01
-6.45681262e-01 2.61354893e-01 1.01191074e-01 2.00822622e-01
3.22528571e-01 -9.87298608e-01 -3.03585023e-01 9.19090807e-02
-1.06699443e+00 -1.99727297e-01 7.94069886e-01 5.02879918e-01
7.11854160e-01 1.93296462e-01 7.23957539e-01 -5.94449341e-01
7.89336979e-01 -7.88385645e-02 -4.76574510e-01 -2.60274500e-01
-3.05285186e-01 -1.17748246e-01 4.33919281e-01 -5.25304794e-01
-3.91031355e-01 -3.53644490e-02 -5.22649646e-01 -1.81661561e-01
-2.88709044e-01 7.17132270e-01 -1.33356556e-01 8.09425041e-02
7.27602959e-01 2.81047523e-01 -3.25988501e-01 -8.17648351e-01
2.72892505e-01 9.06281114e-01 8.38794231e-01 -5.01146972e-01
6.30177319e-01 2.05955207e-01 1.27739251e-01 -1.20038593e+00
-9.17037785e-01 -3.31956774e-01 -6.29769862e-01 -3.16633254e-01
8.09203148e-01 -6.55359685e-01 -6.74782991e-01 7.43976414e-01
-7.98692524e-01 1.10393129e-01 -3.43469024e-01 5.33729970e-01
-7.37603128e-01 3.07696640e-01 -7.06963718e-01 -8.58074427e-01
-4.62323517e-01 -7.58228779e-01 6.26343310e-01 1.13251433e-01
-5.54461241e-01 -3.39757085e-01 4.50419724e-01 4.76131499e-01
2.55771101e-01 2.61660874e-01 1.06868649e+00 -6.48431718e-01
-4.38853055e-01 -4.11246479e-01 6.59834146e-02 3.29624176e-01
1.27162663e-02 -2.19759136e-01 -1.35931802e+00 -1.17834702e-01
1.98356323e-02 -4.62725222e-01 1.03068089e+00 2.44213492e-01
1.13360870e+00 1.63811222e-01 2.48460978e-01 4.38011080e-01
1.13925755e+00 3.03477675e-01 9.03909504e-01 4.99527067e-01
4.60629791e-01 3.37450594e-01 6.45630062e-01 3.85714829e-01
-2.89668798e-01 9.38248277e-01 3.96119982e-01 3.58142912e-01
-4.27177846e-01 -1.94996879e-01 4.13715869e-01 1.18701839e+00
-5.01760781e-01 1.19918540e-01 -9.19389129e-01 5.07085800e-01
-1.50315845e+00 -1.22237980e+00 -2.08943374e-02 2.26458168e+00
8.70101333e-01 6.94987699e-02 7.00112283e-01 1.16361344e+00
3.94410998e-01 3.82450670e-01 1.58031303e-02 -2.97820121e-01
-4.55480754e-01 8.35217237e-01 -7.39703253e-02 5.02816327e-02
-1.27759063e+00 7.87631094e-01 6.46334696e+00 9.92564976e-01
-1.37113631e+00 -2.59355724e-01 1.03719942e-01 -8.90779123e-02
1.53379738e-01 -2.98920035e-01 -1.59026638e-01 1.61085382e-01
9.64847803e-01 -1.01285391e-01 7.49649167e-01 3.97869378e-01
1.11069083e-01 1.77075356e-01 -1.03149176e+00 1.46977413e+00
8.80660862e-02 -1.27589583e+00 2.27635249e-01 -3.03061325e-02
5.11817813e-01 7.45130032e-02 1.92803472e-01 1.45844221e-01
-2.19192505e-01 -1.23424196e+00 7.60226727e-01 5.66003203e-01
2.90290207e-01 -9.63507175e-01 5.38914025e-01 2.35953648e-02
-1.16810775e+00 -1.21654436e-01 -1.20904058e-01 -3.26927632e-01
-2.15082705e-01 3.76044005e-01 -8.35727930e-01 6.10806465e-01
7.84132063e-01 7.44637430e-01 -5.94536960e-01 1.31660330e+00
-5.09046093e-02 1.04577112e+00 -2.84269303e-01 1.18673272e-01
3.39890569e-02 -3.93565565e-01 7.05691814e-01 1.01486635e+00
5.13879716e-01 -2.04200715e-01 -4.38327678e-02 6.93957746e-01
-9.96503830e-02 5.98433256e-01 -1.93833962e-01 -6.38164997e-01
1.51729494e-01 9.67093706e-01 -9.11494970e-01 -6.83621988e-02
1.04939982e-01 7.18924165e-01 -1.60763949e-01 -8.73758346e-02
-4.52888578e-01 -3.33691806e-01 3.92873973e-01 4.56898302e-01
1.83286533e-01 -4.21415627e-01 -2.58765727e-01 -8.32271814e-01
-2.41615474e-01 -1.19162846e+00 6.27961040e-01 -6.98703527e-01
-1.27013564e+00 5.91638744e-01 -3.46722603e-01 -1.36290145e+00
-2.15768024e-01 -5.61583638e-01 -5.95512629e-01 7.03188956e-01
-1.01325047e+00 -1.07463253e+00 -7.50228539e-02 6.61186397e-01
5.07929802e-01 -4.81347293e-01 1.26398718e+00 5.75353920e-01
-2.31216829e-02 2.93606400e-01 1.51790127e-01 2.25530520e-01
8.79320502e-01 -1.17060053e+00 3.73531915e-02 6.42934561e-01
1.12860298e+00 2.36956984e-01 8.51562798e-01 -7.49588013e-02
-1.43628752e+00 -6.36211872e-01 7.17056632e-01 -2.74795592e-02
1.00205958e+00 5.43600768e-02 -1.02307844e+00 3.59218240e-01
1.65658310e-01 -6.85274839e-01 8.97173643e-01 3.23181421e-01
-5.94292283e-01 -4.36115712e-01 -6.42489910e-01 4.04491156e-01
6.08174980e-01 -7.55374372e-01 -1.04202497e+00 -2.20750198e-01
6.25402182e-02 1.50818191e-02 -7.41275012e-01 1.36147082e-01
8.87718737e-01 -1.10067296e+00 1.02547622e+00 -3.09219569e-01
3.61342072e-01 -4.06737328e-01 -4.78707492e-01 -1.16471553e+00
-2.49778610e-02 -5.59563875e-01 2.48875394e-02 1.09680557e+00
-4.67668250e-02 -1.93772942e-01 5.80254912e-01 -4.33480889e-01
5.39427297e-03 -2.04526529e-01 -8.23654175e-01 -5.64968407e-01
3.43620181e-02 -8.13917100e-01 3.72038841e-01 9.91548955e-01
-2.21397474e-01 6.00714743e-01 -3.91118288e-01 -2.80305386e-01
5.03816068e-01 7.65200615e-01 7.62272775e-01 -1.77896523e+00
-5.83060622e-01 -9.33428943e-01 -7.74901927e-01 -5.01720965e-01
-8.21873322e-02 -1.24279261e+00 -2.58643448e-01 -1.14700937e+00
-1.07627131e-01 -2.38061875e-01 -7.02204108e-01 4.28639412e-01
2.25152075e-01 1.03637147e+00 4.56198007e-01 2.17565283e-01
-1.96870267e-01 2.30248258e-01 8.48054528e-01 -4.11121637e-01
-5.10442197e-01 3.30326945e-01 -6.04599714e-01 8.29922259e-01
1.08652365e+00 -4.02013063e-01 -7.73387402e-02 -8.32014754e-02
5.94845295e-01 -2.33723479e-03 3.11775267e-01 -1.55156803e+00
-1.47465318e-01 2.84149945e-01 2.73680478e-01 -7.97419608e-01
6.34503782e-01 -5.21618426e-01 4.07830805e-01 3.94176573e-01
-3.73852938e-01 -2.30196908e-01 3.30128580e-01 3.22725803e-01
-6.29232287e-01 -3.28926928e-02 1.04180110e+00 1.78417966e-01
-7.27910459e-01 -1.96221009e-01 -4.39378291e-01 -2.26267904e-01
4.35776323e-01 -9.61011872e-02 4.28353667e-01 -3.54533792e-01
-1.15269923e+00 -7.81381845e-01 9.83858854e-02 5.49218595e-01
5.57059705e-01 -1.29740322e+00 -9.97380793e-01 1.99726403e-01
2.12899130e-03 -6.67161405e-01 2.60529011e-01 8.04214895e-01
-6.22909009e-01 1.82932466e-01 -6.35773838e-01 -6.57064140e-01
-1.71949911e+00 3.86271268e-01 1.44188136e-01 -6.23736195e-02
-8.14175725e-01 5.76603413e-01 -4.67640489e-01 3.16354120e-03
2.43122175e-01 -3.74875218e-02 -4.38609630e-01 5.09134650e-01
6.78750813e-01 4.88017261e-01 1.63929105e-01 -7.93520868e-01
-1.75361782e-01 8.38273883e-01 -1.05142780e-02 -2.02741221e-01
1.56803179e+00 2.93206900e-01 -2.18895465e-01 1.01349664e+00
9.16795433e-01 4.13345069e-01 -3.46442968e-01 -1.51668325e-01
2.99245298e-01 -5.34138381e-01 3.45638931e-01 -9.31743860e-01
-9.16866779e-01 1.06355941e+00 9.21549678e-01 7.39674747e-01
1.39620495e+00 -9.08292308e-02 6.78986132e-01 4.87361640e-01
3.69070113e-01 -9.48949277e-01 -2.67400909e-02 4.77108568e-01
1.08806479e+00 -8.09232354e-01 1.11467741e-01 6.11233572e-03
-3.00628573e-01 1.38705933e+00 -4.84475702e-01 -5.71905732e-01
5.23552537e-01 5.11351041e-02 1.95590198e-01 -1.12933807e-01
-1.16482139e-01 -2.96374500e-01 5.39708436e-01 3.31272542e-01
8.28419268e-01 2.28546679e-01 -5.84683716e-01 9.81002629e-01
-8.12424541e-01 1.13207869e-01 4.34264481e-01 5.12786210e-01
-6.26810074e-01 -1.30570936e+00 -7.29511023e-01 2.90440261e-01
-8.40614855e-01 8.90238024e-03 -8.40324402e-01 7.39748538e-01
-4.48211730e-02 9.39829290e-01 -3.03307474e-01 -6.09240532e-01
2.28810728e-01 1.99827373e-01 8.53605270e-01 -2.03683645e-01
-9.06155586e-01 5.87536752e-01 -1.33321704e-02 -4.75051463e-01
-5.40131450e-01 -6.43455088e-01 -1.18503118e+00 -3.06778193e-01
6.60534203e-02 2.86653131e-01 6.47853911e-01 7.87347555e-01
-1.23811774e-01 6.34762883e-01 6.14519715e-01 -1.00324285e+00
-4.43257451e-01 -1.17355776e+00 -1.11622679e+00 4.65899199e-01
6.68943971e-02 -2.72771657e-01 -1.71449214e-01 2.43871465e-01] | [15.895161628723145, 5.208682537078857] |
7c123ab5-055e-42e1-bb46-f63098abbfed | texpose-neural-texture-learning-for-self | 2212.12902 | null | https://arxiv.org/abs/2212.12902v2 | https://arxiv.org/pdf/2212.12902v2.pdf | TexPose: Neural Texture Learning for Self-Supervised 6D Object Pose Estimation | In this paper, we introduce neural texture learning for 6D object pose estimation from synthetic data and a few unlabelled real images. Our major contribution is a novel learning scheme which removes the drawbacks of previous works, namely the strong dependency on co-modalities or additional refinement. These have been previously necessary to provide training signals for convergence. We formulate such a scheme as two sub-optimisation problems on texture learning and pose learning. We separately learn to predict realistic texture of objects from real image collections and learn pose estimation from pixel-perfect synthetic data. Combining these two capabilities allows then to synthesise photorealistic novel views to supervise the pose estimator with accurate geometry. To alleviate pose noise and segmentation imperfection present during the texture learning phase, we propose a surfel-based adversarial training loss together with texture regularisation from synthetic data. We demonstrate that the proposed approach significantly outperforms the recent state-of-the-art methods without ground-truth pose annotations and demonstrates substantial generalisation improvements towards unseen scenes. Remarkably, our scheme improves the adopted pose estimators substantially even when initialised with much inferior performance. | ['Benjamin Busam', 'Nassir Navab', 'Fabian Manhardt', 'Hanzhi Chen'] | 2022-12-25 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Chen_TexPose_Neural_Texture_Learning_for_Self-Supervised_6D_Object_Pose_Estimation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Chen_TexPose_Neural_Texture_Learning_for_Self-Supervised_6D_Object_Pose_Estimation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['6d-pose-estimation'] | ['computer-vision'] | [ 7.00767756e-01 6.56556845e-01 3.69692683e-01 -3.69028658e-01
-1.33439481e+00 -6.57236397e-01 8.17726851e-01 -3.08248401e-01
-5.60286283e-01 7.47817636e-01 -3.35912257e-01 1.60662830e-01
-8.91152769e-02 -6.16176605e-01 -1.29779029e+00 -9.02742147e-01
3.05978090e-01 1.00750542e+00 3.63391340e-01 -6.80997670e-02
2.30135694e-01 7.27673292e-01 -1.64016795e+00 1.61542326e-01
6.64258122e-01 1.30180240e+00 1.94466636e-01 7.21578062e-01
2.86164314e-01 5.45478284e-01 -4.52355653e-01 -6.12483978e-01
6.06936753e-01 -6.88446611e-02 -7.30275810e-01 7.20589936e-01
9.93373036e-01 -3.51763546e-01 -3.24147083e-02 6.53792381e-01
5.57164907e-01 -1.03539102e-01 7.59347141e-01 -8.49767327e-01
3.26880068e-02 -1.27351329e-01 -5.68043351e-01 -4.05116081e-01
3.25249970e-01 1.56267315e-01 6.21250570e-01 -8.77024531e-01
9.06675100e-01 1.09721994e+00 9.13409948e-01 4.60742295e-01
-1.56458259e+00 -2.21357718e-01 -1.85246632e-01 -2.80194730e-01
-1.16445529e+00 -6.76967204e-01 8.60643923e-01 -4.29016411e-01
5.45688450e-01 4.05268490e-01 4.30103481e-01 1.51859462e+00
-9.37025845e-02 7.59279013e-01 1.58661163e+00 -6.14200771e-01
2.55873263e-01 1.64520770e-01 -8.81300569e-01 7.16799080e-01
-6.14570454e-02 1.70197353e-01 -4.61084276e-01 1.03559336e-02
1.17259729e+00 -3.70387077e-01 -1.91187516e-01 -1.25952029e+00
-1.32228589e+00 4.45654005e-01 3.87452394e-01 -2.05671936e-01
-3.68722081e-01 3.19373131e-01 2.71945387e-01 2.64316380e-01
7.40292370e-01 4.81956303e-01 -5.55734932e-01 9.19779763e-02
-8.12243044e-01 3.74676257e-01 6.64788842e-01 9.20645654e-01
7.79542208e-01 1.44710079e-01 9.68950316e-02 8.37638617e-01
2.91569561e-01 8.38599801e-01 3.67149431e-03 -1.16797650e+00
5.70501328e-01 2.96908468e-01 3.65908563e-01 -8.77032042e-01
-1.10492639e-01 -6.53232634e-01 -5.32597899e-01 3.40596378e-01
8.21351886e-01 7.59020597e-02 -1.18998003e+00 1.39624190e+00
6.63810551e-01 1.43998697e-01 -2.60200184e-02 8.87223721e-01
4.88097042e-01 1.00655921e-01 -1.92301854e-01 -3.83207649e-02
8.43733728e-01 -6.82668924e-01 -2.71099478e-01 -3.48468751e-01
2.54204780e-01 -1.03371930e+00 7.07604945e-01 6.79429471e-01
-1.14064646e+00 -3.78248692e-01 -9.27844048e-01 1.15203641e-01
-9.49988738e-02 3.72333229e-01 5.81818581e-01 7.30492353e-01
-9.55923557e-01 7.12501884e-01 -8.58053207e-01 -2.89800495e-01
5.99854589e-01 5.63599706e-01 -5.21847963e-01 5.58426566e-02
-7.26864636e-01 9.40259635e-01 4.56313491e-01 3.43679249e-01
-1.05599499e+00 -5.67326844e-01 -9.56565619e-01 -6.38402879e-01
8.67520571e-01 -8.14095259e-01 1.20500231e+00 -1.21634495e+00
-1.72149992e+00 1.30735993e+00 3.95947635e-01 -4.29755658e-01
1.21107471e+00 -2.83705533e-01 1.93329409e-01 1.48678556e-01
-5.82288988e-02 7.59294987e-01 1.33870268e+00 -1.78929067e+00
-2.56014943e-01 -4.69544470e-01 -4.58855275e-03 3.92357856e-01
2.39155307e-01 -4.81206775e-01 -7.94754207e-01 -6.51968658e-01
2.83414960e-01 -1.07933199e+00 -3.61524433e-01 2.12888032e-01
-3.56556922e-01 4.26292211e-01 7.25843072e-01 -5.89875519e-01
2.19715104e-01 -1.76005268e+00 6.22571051e-01 2.64082313e-01
8.30356404e-03 1.17403969e-01 -1.68186352e-01 2.80653596e-01
6.80040419e-02 -4.80963409e-01 -3.47757727e-01 -7.07497358e-01
7.21809687e-03 4.89340484e-01 -7.86695555e-02 8.28335762e-01
4.44263756e-01 9.58721340e-01 -6.64817274e-01 -3.46578151e-01
6.08974040e-01 6.55370593e-01 -5.05148113e-01 2.95781016e-01
-6.16223812e-01 1.18580544e+00 -3.75660092e-01 5.57347476e-01
8.06313097e-01 6.54684529e-02 9.44369882e-02 -3.24669152e-01
9.61771905e-02 -9.02626589e-02 -1.34750330e+00 2.05667257e+00
-6.22990489e-01 1.32132053e-01 3.68814498e-01 -1.21669292e+00
9.52773511e-01 2.55552202e-01 4.47824508e-01 -5.08336961e-01
2.88622081e-01 5.69192290e-01 -4.04708713e-01 -5.37524521e-01
1.51277944e-01 -3.88777107e-01 -7.52711669e-02 5.37925512e-02
3.56374413e-01 -9.55110967e-01 -3.30972522e-01 -4.02137697e-01
6.46038651e-01 9.36183929e-01 -6.06548321e-03 -2.64908727e-02
7.32100010e-01 -2.38707542e-01 3.18075866e-01 5.44984579e-01
1.95923418e-01 1.17131150e+00 4.27315623e-01 -6.06705308e-01
-1.52298176e+00 -9.41067517e-01 -1.90094098e-01 7.26423800e-01
-1.32741760e-02 9.51429829e-02 -8.46704841e-01 -9.26056564e-01
-1.83276013e-02 2.57077098e-01 -8.72743607e-01 2.39656463e-01
-7.18280494e-01 -7.25168347e-01 3.87954175e-01 2.86260128e-01
4.21978086e-01 -9.16368544e-01 -5.25236547e-01 3.77370976e-02
-1.01311076e-02 -1.26361465e+00 -2.49626394e-02 4.05397892e-01
-7.48207510e-01 -1.01485384e+00 -1.00905526e+00 -5.55460215e-01
6.35429263e-01 -3.53099465e-01 1.20565867e+00 -2.87045479e-01
-3.44634980e-01 6.33533180e-01 -1.94533050e-01 -3.97376984e-01
-4.67821479e-01 9.44780484e-02 -3.27740125e-02 3.57474148e-01
-3.01544905e-01 -6.21710837e-01 -6.17669940e-01 4.68413442e-01
-9.70367551e-01 1.07761778e-01 8.85493577e-01 9.68473792e-01
8.85194123e-01 -2.27552548e-01 1.52356595e-01 -1.10823226e+00
-1.57275960e-01 6.93400577e-02 -6.74231887e-01 1.66586533e-01
-2.49553695e-01 3.05542678e-01 4.05192196e-01 -2.61751205e-01
-1.31959105e+00 6.77380979e-01 -2.94841796e-01 -5.96077561e-01
-4.50030178e-01 -7.33314157e-02 -3.36501092e-01 -6.94274366e-01
7.04664707e-01 2.65774041e-01 1.47021130e-01 -4.97906804e-01
2.59151518e-01 2.82077104e-01 6.05470896e-01 -9.64422107e-01
1.08568919e+00 7.06092179e-01 3.20131391e-01 -7.30284989e-01
-9.83264983e-01 -3.74642819e-01 -1.05519664e+00 -2.68475801e-01
7.45696723e-01 -9.27101016e-01 -5.95894217e-01 6.38322771e-01
-1.04399264e+00 -5.03950179e-01 -3.43919754e-01 2.85949796e-01
-1.23376560e+00 3.92850041e-01 -2.55143166e-01 -7.55984306e-01
-1.28466878e-02 -1.14306450e+00 1.93276000e+00 -2.90469259e-01
8.17157254e-02 -9.45838034e-01 -4.12844978e-02 8.55652511e-01
1.99069530e-01 8.67012382e-01 1.83185086e-01 -8.75291601e-02
-9.37026381e-01 -1.47125006e-01 1.16446624e-02 5.36678493e-01
-2.68160999e-01 -2.83076823e-01 -1.29184878e+00 -3.36403161e-01
1.26172394e-01 -8.21105421e-01 7.23829150e-01 1.88070089e-01
1.18666828e+00 -1.77143067e-01 -5.77613674e-02 9.40812886e-01
1.57715797e+00 -3.15140456e-01 5.59718013e-01 5.04443944e-01
9.00405943e-01 9.60866928e-01 8.16090763e-01 3.14071655e-01
-5.39936796e-02 1.00069165e+00 8.10039580e-01 -1.68821320e-01
-1.31285846e-01 -2.26740912e-01 1.01435088e-01 3.07179272e-01
-4.83192116e-01 -1.64994001e-01 -9.23754275e-01 2.68169671e-01
-1.68005204e+00 -4.91596371e-01 -2.51413435e-01 2.31883359e+00
6.29930139e-01 2.64581949e-01 -6.69098496e-02 2.63980776e-01
2.36823425e-01 1.57717854e-01 -4.41436231e-01 -5.65950871e-02
-1.47995755e-01 5.03735483e-01 8.26263070e-01 5.67484081e-01
-1.41017294e+00 9.50815916e-01 5.36193562e+00 9.76287246e-01
-1.13389385e+00 6.25902042e-02 7.14247465e-01 1.46211103e-01
-2.53642827e-01 -2.81861395e-01 -3.66042078e-01 -1.45098614e-02
6.24701738e-01 7.44376481e-01 3.33997488e-01 6.74104035e-01
-1.86476521e-02 -2.00157627e-01 -1.04237604e+00 9.44975853e-01
2.88012207e-01 -1.00325251e+00 -4.89650704e-02 7.82559514e-02
8.55334163e-01 -1.12527385e-01 4.22528327e-01 -9.33046266e-03
-1.50737958e-02 -1.13750398e+00 9.73577201e-01 7.04117596e-01
1.02129936e+00 -6.94630980e-01 7.84188449e-01 3.96072537e-01
-7.84082949e-01 1.95396006e-01 -2.07052559e-01 1.69706702e-01
1.29755273e-01 4.59472030e-01 -8.84193301e-01 1.01812434e+00
4.41135138e-01 5.72569549e-01 -6.72176301e-01 7.86005795e-01
-2.04456478e-01 2.26952508e-01 -4.86986279e-01 3.88009578e-01
3.22675586e-01 -1.90438434e-01 6.02114499e-01 8.62489760e-01
1.35863960e-01 -3.53090763e-01 9.40413177e-02 7.13458538e-01
1.87343031e-01 -9.92521346e-02 -4.53381449e-01 3.80245179e-01
-1.08559936e-01 1.21377778e+00 -9.26350713e-01 1.15748597e-02
-1.13146499e-01 1.41530049e+00 3.37772191e-01 2.43648872e-01
-6.83021367e-01 1.85063928e-01 1.15800403e-01 4.35047209e-01
6.21125102e-01 -1.26038074e-01 -2.76137680e-01 -1.21040893e+00
2.78546512e-01 -9.46654737e-01 -1.26476824e-01 -7.45804727e-01
-1.05906987e+00 5.45591772e-01 -4.21417765e-02 -1.29838824e+00
-3.72534633e-01 -8.27319920e-01 -5.13663255e-02 8.33292842e-01
-1.28015387e+00 -1.76358759e+00 -2.06376836e-01 2.40411595e-01
5.19858539e-01 -3.30252722e-02 7.84100354e-01 1.11936375e-01
-2.66565494e-02 4.35511500e-01 8.72178152e-02 -1.48630172e-01
8.01015735e-01 -1.52023280e+00 2.97517389e-01 5.11162817e-01
1.67370424e-01 7.67537802e-02 9.16544318e-01 -4.19292241e-01
-1.31964481e+00 -1.05383480e+00 4.52177346e-01 -9.76312995e-01
2.82514960e-01 -7.44786620e-01 -5.63669741e-01 5.44569433e-01
-1.81922063e-01 3.34805042e-01 -2.51982128e-03 -2.74805546e-01
-6.73799366e-02 -6.98719472e-02 -1.14436364e+00 3.18409860e-01
9.91901696e-01 -4.36462551e-01 -4.77958649e-01 2.94120073e-01
2.28424415e-01 -9.23934639e-01 -9.01330233e-01 8.16626847e-01
6.49903893e-01 -1.17979276e+00 1.24972844e+00 -1.42002091e-01
3.24788660e-01 -3.74798536e-01 -8.52062702e-02 -1.08355165e+00
3.52483809e-01 -7.54435837e-01 7.67173916e-02 1.06798637e+00
2.12226272e-01 -3.60664904e-01 1.23504651e+00 2.50616074e-01
-7.00504705e-02 -7.60876298e-01 -9.46443141e-01 -7.31139302e-01
3.63112874e-02 -4.16209102e-01 1.72113910e-01 7.42339015e-01
-9.04053330e-01 2.26636216e-01 -6.68165267e-01 2.15545401e-01
1.05430448e+00 3.22944969e-02 1.16264832e+00 -1.07817018e+00
-7.33023047e-01 -1.20452298e-02 -6.06789052e-01 -1.15965569e+00
1.67174950e-01 -5.67420661e-01 1.63366124e-01 -1.02957821e+00
-2.59214807e-02 -5.81954300e-01 2.37847984e-01 1.88434035e-01
1.23847879e-01 8.67509246e-01 -1.17232643e-01 2.09182780e-02
-6.24189734e-01 7.23097682e-01 1.68058395e+00 1.24355815e-01
1.02364972e-01 1.83290094e-01 -1.23776615e-01 7.70283461e-01
4.53309476e-01 -3.56376678e-01 -4.24158484e-01 -4.77547765e-01
2.76571602e-01 1.32172599e-01 8.55442047e-01 -1.00924063e+00
-7.75134936e-02 1.33217365e-01 6.82958484e-01 -4.10358101e-01
6.91940904e-01 -1.01277328e+00 4.28368747e-01 1.65825024e-01
-2.16397643e-01 -5.09756923e-01 1.72656417e-01 8.20699632e-01
-1.85412839e-02 -1.25626341e-01 7.70961761e-01 -3.40739846e-01
-4.63900387e-01 3.53412241e-01 7.08850697e-02 1.87109578e-02
1.05823195e+00 -4.18738157e-01 3.63435626e-01 -2.50129968e-01
-1.03820467e+00 -1.62567005e-01 9.10257101e-01 2.05248848e-01
4.03288931e-01 -1.13963151e+00 -7.37338066e-01 3.76317322e-01
2.29518175e-01 6.12992227e-01 1.98088810e-01 9.53440905e-01
-9.20028329e-01 3.35455179e-01 -2.35050052e-01 -9.47841465e-01
-1.17269611e+00 3.69609207e-01 6.60156250e-01 -2.11188376e-01
-4.77432162e-01 1.01320434e+00 2.41720155e-01 -1.05022299e+00
3.15806448e-01 -2.56221533e-01 1.26450181e-01 -1.63108990e-01
-1.33456275e-01 1.04957730e-01 3.94910663e-01 -8.42987895e-01
-1.35955572e-01 9.18999076e-01 3.49624008e-02 -4.14228439e-01
1.59871590e+00 -1.25776738e-01 1.94214925e-01 3.59323472e-01
1.07611048e+00 7.96715319e-02 -2.00983286e+00 -1.13025732e-01
-7.41834119e-02 -6.23206794e-01 -6.84678331e-02 -8.30390036e-01
-9.73325968e-01 7.42984235e-01 5.90126097e-01 -3.29577386e-01
9.06302094e-01 9.92926881e-02 4.76576686e-01 2.17355803e-01
6.97103858e-01 -1.05554509e+00 2.00029418e-01 2.55564749e-01
1.09050369e+00 -1.44826031e+00 2.53763162e-02 -7.76547909e-01
-4.36338484e-01 1.16070068e+00 4.40070361e-01 -4.50988561e-01
2.26297483e-01 2.18233660e-01 9.07714814e-02 -2.23502845e-01
-3.44868481e-01 -1.71178162e-01 6.81664586e-01 7.49819398e-01
2.39256963e-01 -2.20126644e-01 -2.56345216e-02 -5.34563847e-02
-1.34428069e-01 -2.61136800e-01 1.58722982e-01 8.25863600e-01
2.06478462e-02 -1.34512413e+00 -6.05523646e-01 1.58708036e-01
-5.93353689e-01 1.89745232e-01 -5.19440949e-01 9.96575058e-01
4.04372752e-01 3.44988227e-01 -1.06032327e-01 -1.34267509e-01
3.78381193e-01 -2.39058752e-02 1.16884613e+00 -5.66197455e-01
-4.43241864e-01 3.64923716e-01 1.94584429e-01 -8.08672249e-01
-7.92987347e-01 -8.16378534e-01 -4.40970719e-01 1.19766973e-01
-3.50869298e-01 -3.17075610e-01 5.94237745e-01 9.36555445e-01
-1.38005495e-01 4.99835640e-01 4.46338117e-01 -1.53765631e+00
-6.40954077e-01 -8.26500058e-01 -5.12203276e-01 5.17758369e-01
4.18305874e-01 -8.35422754e-01 -3.69395167e-01 2.95961648e-01] | [8.052486419677734, -2.7173879146575928] |
9e4218fa-ab1c-432c-b7c7-77c95ef28a05 | a-dataset-for-research-on-short-text | null | null | https://aclanthology.org/D13-1096 | https://aclanthology.org/D13-1096.pdf | A Dataset for Research on Short-Text Conversations | null | ['Hang Li', 'Hao Wang', 'Enhong Chen', 'Zhengdong Lu'] | 2013-10-01 | null | null | null | emnlp-2013-10 | ['short-text-conversation'] | ['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] | [-7.3513875007629395, 3.76395583152771] |
d1f6cc24-a660-4c2b-a39d-26f2002d6335 | semi-equivariant-gnn-architectures-for-jet | 2202.06941 | null | https://arxiv.org/abs/2202.06941v1 | https://arxiv.org/pdf/2202.06941v1.pdf | Semi-Equivariant GNN Architectures for Jet Tagging | Composing Graph Neural Networks (GNNs) of operations that respect physical symmetries has been suggested to give better model performance with a smaller number of learnable parameters. However, real-world applications, such as in high energy physics have not born this out. We present the novel architecture VecNet that combines both symmetry-respecting and unconstrained operations to study and tune the degree of physics-informed GNNs. We introduce a novel metric, the \textit{ant factor}, to quantify the resource-efficiency of each configuration in the search-space. We find that a generalized architecture such as ours can deliver optimal performance in resource-constrained applications. | ['Jason Wong', 'Savannah Thais', 'Daniel Murnane'] | 2022-02-14 | null | null | null | null | ['jet-tagging'] | ['graphs'] | [ 3.19885314e-02 1.32626981e-01 -3.46264154e-01 -1.73211619e-01
1.08235665e-01 -5.32058358e-01 5.52538574e-01 -8.16775709e-02
-3.95310014e-01 7.03010142e-01 -4.97176573e-02 -7.13680208e-01
-5.87549388e-01 -9.99110818e-01 -8.86239350e-01 -6.80373132e-01
-3.60745698e-01 5.81655800e-01 2.03711048e-01 -5.13284326e-01
3.96518826e-01 8.44874620e-01 -1.17620277e+00 -2.64258504e-01
8.09564650e-01 7.84474373e-01 2.46864960e-01 4.07743514e-01
8.48645568e-02 4.15493935e-01 -4.24021691e-01 -2.32850984e-01
7.87023902e-01 -4.93895322e-01 -6.78545713e-01 -7.62743890e-01
2.34979555e-01 1.44925177e-01 -9.10095692e-01 9.46794152e-01
5.04338443e-01 6.36290848e-01 3.31897706e-01 -1.27927864e+00
-8.09891224e-01 1.08661342e+00 -3.60065579e-01 6.26200497e-01
-4.05103981e-01 4.64718550e-01 1.11356568e+00 -2.23321021e-01
7.38798738e-01 1.07768798e+00 2.64112473e-01 4.44389224e-01
-1.07193387e+00 -6.06958210e-01 -3.53669450e-02 3.36459428e-01
-1.40386653e+00 -2.93158263e-01 8.51820946e-01 1.67269558e-01
1.41864073e+00 3.72681350e-01 6.58623278e-01 1.07814670e+00
5.87547421e-01 6.54627606e-02 6.80534363e-01 -2.80719548e-01
4.93130773e-01 -3.88048232e-01 -8.19916278e-02 8.34644735e-01
8.49456728e-01 4.56273288e-01 -8.60139012e-01 4.03806791e-02
9.37433779e-01 -1.99629143e-01 -2.18987584e-01 -6.88420355e-01
-1.18469512e+00 8.81728053e-01 1.05569184e+00 3.20714712e-01
-3.35081428e-01 6.64273202e-01 1.67401522e-01 1.81751549e-01
1.06624961e-01 1.47712529e+00 -6.29319847e-01 -9.35683846e-02
-5.96955836e-01 1.13217101e-01 8.85807335e-01 8.85581553e-01
6.10149086e-01 4.43417132e-01 -1.66985974e-01 4.25961882e-01
-2.04832822e-01 2.79499412e-01 3.34225446e-01 -9.35109138e-01
4.14569452e-02 8.22910428e-01 -4.52745020e-01 -9.88419533e-01
-7.52806246e-01 -8.76484692e-01 -1.15724409e+00 1.01228163e-01
1.45709619e-01 -2.29095861e-01 -8.98251712e-01 1.87532544e+00
2.25960284e-01 1.37393236e-01 -1.85362235e-01 1.00582457e+00
6.08347118e-01 6.43293858e-01 -1.31707266e-01 1.38952345e-01
1.05345750e+00 -1.06195760e+00 -1.89576060e-01 -4.97070074e-01
4.78573978e-01 -2.31719851e-01 9.34960842e-01 1.69595882e-01
-1.00903499e+00 -8.11524615e-02 -1.40752602e+00 3.54261808e-02
-4.02423561e-01 -4.40008789e-01 1.40686107e+00 8.00568759e-01
-1.29736233e+00 1.12844336e+00 -9.91045654e-01 -4.13306713e-01
3.40797603e-01 8.01381409e-01 -2.21031219e-01 1.81673303e-01
-1.03733265e+00 9.16324258e-01 8.96242678e-01 -4.13731821e-02
-7.83387065e-01 -8.07188570e-01 -4.98727739e-01 4.90423828e-01
8.40270877e-01 -1.04363942e+00 7.57644415e-01 -4.82988596e-01
-1.67004883e+00 3.30265194e-01 4.17814404e-01 -5.39021373e-01
-2.73717325e-02 3.99733156e-01 -2.13862479e-01 -2.05467179e-01
-6.21811688e-01 6.61600769e-01 4.82393682e-01 -8.45600963e-01
-5.98831549e-02 -1.04756318e-01 5.18739760e-01 2.70453215e-01
-5.70337772e-01 -2.09763244e-01 -5.11337459e-01 -2.86898077e-01
4.10072505e-01 -1.45328641e+00 -5.88727295e-01 -1.52658314e-01
-6.13629103e-01 -6.40197396e-02 3.28419030e-01 -4.47925851e-02
8.91667128e-01 -1.69239724e+00 6.07225001e-01 4.52430248e-01
6.40015185e-01 -5.10381395e-03 -4.00186926e-01 5.40137768e-01
3.93785350e-02 4.17204559e-01 1.94557413e-01 3.32265049e-01
1.54972017e-01 4.27955627e-01 3.20926011e-02 3.80804777e-01
2.92692371e-02 1.16295087e+00 -7.11188793e-01 -1.85502786e-02
-7.80915022e-02 3.21017176e-01 -9.94231582e-01 -1.58404619e-01
-4.66349691e-01 2.47063518e-01 -6.44197464e-01 2.97530204e-01
5.50091028e-01 -7.88322806e-01 5.77653468e-01 -3.94613110e-02
1.28973067e-01 2.48723239e-01 -8.10441315e-01 1.66059148e+00
-4.43893790e-01 6.09592080e-01 -1.57857299e-01 -8.59319448e-01
6.44782305e-01 -5.51215947e-01 3.82892907e-01 -1.08349073e+00
3.58915836e-01 -1.52834579e-01 6.67933702e-01 1.13814957e-01
7.62340784e-01 3.16633731e-01 -1.34391695e-01 4.04390305e-01
1.57629073e-01 -4.93552536e-02 4.13207203e-01 3.33400160e-01
1.70586872e+00 -1.81096643e-01 3.91169041e-01 -7.58664310e-01
9.19409990e-02 -1.12827376e-01 4.62679327e-01 1.21819997e+00
2.49307845e-02 3.81108336e-02 6.86167836e-01 -6.05083108e-01
-1.26965463e+00 -8.87220323e-01 2.94540823e-01 1.37891197e+00
3.10376823e-01 -5.11998951e-01 -6.87127173e-01 -3.11293066e-01
-2.00301871e-01 6.69348955e-01 -4.08989847e-01 -5.99388003e-01
-8.61632168e-01 -8.73251736e-01 3.92302364e-01 2.21392363e-01
4.37093258e-01 -1.19285798e+00 -6.91687703e-01 -2.41616935e-01
3.20096970e-01 -1.17149723e+00 -5.52771807e-01 5.01904666e-01
-7.80624568e-01 -8.37773025e-01 -1.26287341e-01 -4.88508046e-01
6.10243142e-01 1.33589342e-01 1.40406656e+00 5.93211949e-01
-3.23776007e-01 -1.88526586e-01 -3.19859356e-01 -2.17659831e-01
-3.83503020e-01 6.89487934e-01 3.06874305e-01 -5.83432972e-01
-2.34449301e-02 -9.51970279e-01 -7.03393042e-01 1.32022768e-01
-8.61907125e-01 1.26123860e-01 8.76752973e-01 7.04581261e-01
3.91614079e-01 2.34065652e-01 7.00933263e-02 -6.16573930e-01
6.83057487e-01 -3.00929278e-01 -8.92573297e-01 2.02193514e-01
-8.18228006e-01 8.44464302e-01 7.85802722e-01 -5.15657306e-01
-4.68558848e-01 -2.79696047e-01 2.08388418e-01 -4.77241158e-01
3.14401984e-01 6.71919286e-01 -4.38816510e-02 -8.10360849e-01
9.05334949e-01 1.66547492e-01 -3.72880638e-01 -1.25609219e-01
4.94479656e-01 -2.62251079e-01 3.66506100e-01 -8.76678169e-01
1.05179322e+00 -5.27084479e-03 8.24817955e-01 -8.26614916e-01
-5.05531430e-01 7.34944046e-02 -1.80421486e-01 6.60071895e-02
5.84673703e-01 -4.52487975e-01 -1.21365082e+00 1.88148245e-01
-1.09941173e+00 -4.71561134e-01 -3.27276051e-01 4.65284824e-01
-3.82409602e-01 1.33694708e-01 -3.36957127e-01 -4.92910951e-01
-4.49248970e-01 -1.21809363e+00 5.22569835e-01 5.08051991e-01
4.99359183e-02 -7.13266432e-01 7.23269358e-02 -1.42132774e-01
9.32687521e-01 1.46586567e-01 1.42177188e+00 -7.29152441e-01
-1.24630737e+00 1.09509595e-01 -2.96555847e-01 -2.13581726e-01
-3.58351111e-01 -1.56727046e-01 -6.00160062e-01 -4.62094635e-01
-1.04558825e-01 -6.02987744e-02 8.75909984e-01 4.46148038e-01
1.59573507e+00 -5.71268260e-01 -3.71046066e-01 1.19568396e+00
1.44984913e+00 1.76374793e-01 7.13343501e-01 2.68557787e-01
8.70696902e-01 5.64372279e-02 -2.51124501e-01 2.77929246e-01
-1.95421830e-01 6.73941076e-01 7.66384900e-01 8.19758251e-02
-7.57828355e-02 -1.74002796e-01 -8.05033837e-03 1.03250742e+00
-3.26511800e-01 -1.04154551e+00 -1.03280127e+00 -1.01138009e-02
-1.66172171e+00 -5.45428574e-01 4.83028382e-01 1.96169484e+00
4.94240046e-01 5.73564351e-01 -2.10962996e-01 -3.52676660e-01
4.22789872e-01 3.64919573e-01 -1.00902998e+00 -4.65028077e-01
-4.98132184e-02 6.63298428e-01 1.18857694e+00 1.89653188e-01
-5.59881747e-01 9.81369615e-01 6.86917496e+00 1.06775558e+00
-1.06787407e+00 3.73040251e-02 4.72044021e-01 -4.17094916e-01
-4.39565718e-01 1.72304600e-01 -5.19997597e-01 2.05272540e-01
1.26319623e+00 -5.03827751e-01 1.19266212e+00 8.72622967e-01
-1.50948703e-01 1.43899873e-01 -1.23286426e+00 8.15532923e-01
-6.62550032e-02 -1.86700189e+00 3.62954646e-01 3.91108721e-01
6.60092533e-01 3.84030014e-01 1.40991643e-01 3.88058186e-01
5.71714103e-01 -1.18649805e+00 4.64721978e-01 2.92544842e-01
7.76657522e-01 -8.56138885e-01 4.04251575e-01 1.96839944e-01
-8.99689138e-01 -8.88273790e-02 -5.31825483e-01 -5.97665086e-02
-9.45783928e-02 3.85694593e-01 -8.48778963e-01 5.91622889e-01
6.69065356e-01 9.60913971e-02 -5.98833501e-01 1.00449443e+00
-5.41045750e-03 3.48746777e-01 -6.36250496e-01 -7.53573656e-01
3.09317917e-01 -3.69525909e-01 7.87721932e-01 4.86838818e-01
4.20646518e-01 2.67157972e-01 1.40272722e-01 1.16032243e+00
-5.87117493e-01 -5.75641878e-02 -5.91358066e-01 -5.78062117e-01
6.16716981e-01 1.24157691e+00 -1.21012163e+00 2.22326517e-01
2.75797814e-01 6.29936695e-01 5.67740858e-01 1.90866813e-01
-9.26394463e-01 -2.41489992e-01 7.42710173e-01 4.62868847e-02
1.37620091e-01 -7.03414679e-01 -2.69232124e-01 -8.61628354e-01
-1.36066794e-01 -8.37581158e-01 1.65986955e-01 -6.42504632e-01
-9.34340417e-01 6.65211141e-01 3.57216671e-02 -5.40204704e-01
4.96361144e-02 -9.53180909e-01 -7.31896818e-01 5.41983604e-01
-1.13925290e+00 -7.31396437e-01 -2.72690237e-01 1.06638230e-01
2.48619601e-01 -5.27686477e-01 7.02882349e-01 5.85169755e-02
-9.48535860e-01 8.99109364e-01 1.09638922e-01 -3.43278170e-01
1.18187986e-01 -1.18103719e+00 8.18396449e-01 8.51820052e-01
1.38656765e-01 8.50994587e-01 1.04340613e+00 -5.64171791e-01
-2.18347859e+00 -7.43909538e-01 3.48862559e-01 -3.13103586e-01
7.81934142e-01 -5.28167129e-01 -5.60356617e-01 4.98883396e-01
3.39610279e-01 -9.68551338e-02 1.88662767e-01 4.42281514e-01
-3.79116058e-01 -3.66783924e-02 -8.60396802e-01 9.67988551e-01
1.79126394e+00 -1.98595360e-01 9.15060043e-02 6.67501032e-01
1.33270061e+00 -6.79465532e-01 -6.89999640e-01 4.86912757e-01
2.30727360e-01 -7.05048859e-01 1.22623146e+00 -8.95963073e-01
4.56508458e-01 2.38494016e-02 -7.59969326e-03 -1.42946160e+00
-7.79982150e-01 -1.05676913e+00 -2.29035035e-01 3.06323886e-01
5.90459645e-01 -8.60477686e-01 1.09391224e+00 2.70762652e-01
-3.50236595e-01 -8.82336617e-01 -1.06930935e+00 -1.12247825e+00
1.80934608e-01 7.14529753e-02 9.55721796e-01 8.02254021e-01
-2.16174155e-01 4.76825684e-01 -2.53168046e-01 1.85935751e-01
5.25123000e-01 -1.31478891e-01 4.71203715e-01 -1.18776405e+00
-5.06019473e-01 -8.82642269e-01 -4.82465267e-01 -6.84133410e-01
2.05361575e-01 -1.26804626e+00 -2.11063504e-01 -1.19951653e+00
2.43201897e-01 -5.42485237e-01 -5.67049444e-01 5.22294521e-01
2.25735053e-01 3.65504399e-02 8.45466703e-02 -3.59867699e-02
-6.01229608e-01 6.10908091e-01 1.22262537e+00 -2.24317908e-02
-1.41754866e-01 -6.17868304e-01 -8.40129018e-01 4.44677591e-01
8.90984297e-01 -6.89594090e-01 -5.75844646e-01 -4.02263850e-01
7.73275137e-01 -1.56223342e-01 1.62223384e-01 -1.24810565e+00
5.44834137e-01 -6.88789487e-01 2.91387379e-01 -2.24706963e-01
2.95275539e-01 -5.42772233e-01 4.33166713e-01 6.39645100e-01
-2.92969108e-01 4.52216864e-01 1.89753190e-01 5.06523371e-01
6.33403182e-01 -1.28469184e-01 5.91185451e-01 -2.17514321e-01
-6.53907776e-01 4.78893369e-01 6.56322092e-02 8.68545845e-02
8.75652552e-01 -1.93135161e-02 -7.67605186e-01 -1.59976229e-01
-2.35998735e-01 1.41075701e-01 4.72674757e-01 3.78750026e-01
2.02541679e-01 -1.12838507e+00 -4.42370117e-01 1.77387059e-01
-1.58805355e-01 -2.82603234e-01 3.62080246e-01 5.50746083e-01
-9.70566630e-01 5.85374355e-01 -5.16421556e-01 -3.60024631e-01
-7.07988739e-01 6.36691570e-01 4.63380337e-01 -5.84280789e-01
-5.71915030e-01 9.47114527e-01 2.87106540e-02 -5.56180000e-01
-1.05083376e-01 -7.81446844e-02 1.79929063e-01 -4.96244252e-01
2.13297028e-02 4.62389231e-01 3.46265823e-01 -8.39505941e-02
-4.62020010e-01 1.64573580e-01 -5.27362116e-02 2.32534841e-01
1.38895953e+00 5.05533755e-01 -3.16608727e-01 -1.88882515e-01
8.34122419e-01 -2.49923676e-01 -7.79116094e-01 -8.97940472e-02
-8.91548023e-02 -1.77830175e-01 4.88748372e-01 -7.38395631e-01
-1.36575985e+00 3.47863913e-01 4.79430795e-01 2.47907147e-01
8.99427354e-01 -1.84606642e-01 7.39106476e-01 1.17478478e+00
6.01188898e-01 -8.89850259e-01 7.25219324e-02 7.83387244e-01
7.56493628e-01 -8.21614921e-01 3.02743226e-01 -2.73091793e-01
-1.15937911e-01 9.16940928e-01 9.26566243e-01 -2.99945325e-01
6.33215070e-01 1.80373847e-01 -5.56453824e-01 -5.21406829e-01
-8.01128149e-01 -3.86775620e-02 3.73458624e-01 3.60004067e-01
-9.78525132e-02 1.90944031e-01 -1.74510509e-01 2.98795819e-01
-6.58710063e-01 -4.26233381e-01 6.70519769e-01 7.40265310e-01
-3.54158580e-01 -9.88969505e-01 2.40424454e-01 8.00040364e-01
-1.03546269e-01 -2.41608098e-01 -4.59697545e-01 8.64715159e-01
-2.04080656e-01 5.26125550e-01 5.84345050e-02 -6.47589087e-01
2.29648337e-01 -1.44748390e-01 6.58743441e-01 -5.31860113e-01
-6.02032959e-01 -5.65578103e-01 1.89942122e-01 -8.50125134e-01
8.85569826e-02 -8.74759182e-02 -1.04922819e+00 -9.70750630e-01
-3.12388003e-01 3.82960849e-02 7.83001304e-01 7.29440033e-01
6.87863469e-01 9.76559520e-01 3.39405119e-01 -1.05548191e+00
-6.04186773e-01 -5.23669839e-01 -5.43856382e-01 -1.15937248e-01
4.38730940e-02 -6.60481095e-01 -4.49234128e-01 -9.31355655e-01] | [6.553464412689209, 5.917568206787109] |
ec36c1a4-420f-4a08-b11e-a2908ac3eacc | learning-to-simulate-tree-branch-dynamics-for | 2306.03410 | null | https://arxiv.org/abs/2306.03410v1 | https://arxiv.org/pdf/2306.03410v1.pdf | Learning to Simulate Tree-Branch Dynamics for Manipulation | We propose to use a simulation driven inverse inference approach to model the joint dynamics of tree branches under manipulation. Learning branch dynamics and gaining the ability to manipulate deformable vegetation can help with occlusion-prone tasks, such as fruit picking in dense foliage, as well as moving overhanging vines and branches for navigation in dense vegetation. The underlying deformable tree geometry is encapsulated as coarse spring abstractions executed on parallel, non-differentiable simulators. The implicit statistical model defined by the simulator, reference trajectories obtained by actively probing the ground truth, and the Bayesian formalism, together guide the spring parameter posterior density estimation. Our non-parametric inference algorithm, based on Stein Variational Gradient Descent, incorporates biologically motivated assumptions into the inference process as neural network driven learnt joint priors; moreover, it leverages the finite difference scheme for gradient approximations. Real and simulated experiments confirm that our model can predict deformation trajectories, quantify the estimation uncertainty, and it can perform better when base-lined against other inference algorithms, particularly from the Monte Carlo family. The model displays strong robustness properties in the presence of heteroscedastic sensor noise; furthermore, it can generalise to unseen grasp locations. | ['Fabio Ramos', 'Paulo Borges', 'Jason Williams', 'Tirthankar Bandyopadhyay', 'Jayadeep Jacob'] | 2023-06-06 | null | null | null | null | ['density-estimation'] | ['methodology'] | [ 3.42961311e-01 2.60047823e-01 -1.25884786e-01 2.36537695e-01
-4.57492977e-01 -9.81319666e-01 4.23942119e-01 -7.07089826e-02
-9.85970050e-02 8.26922357e-01 -5.93611225e-02 -1.11096062e-01
-5.24211764e-01 -9.86243427e-01 -1.18972552e+00 -7.80092001e-01
-2.91594267e-01 8.84321868e-01 1.81241304e-01 -3.38039249e-01
6.50278181e-02 7.00654745e-01 -1.33926427e+00 -3.81352842e-01
9.54153240e-01 7.58832574e-01 3.75612617e-01 9.29582596e-01
4.27016765e-01 3.72102618e-01 8.05788562e-02 -2.34957784e-01
7.89717510e-02 1.76307067e-01 -5.58812439e-01 7.07230866e-02
7.62411430e-02 -7.15256631e-01 -8.09699818e-02 7.46975720e-01
6.80192411e-02 1.01695962e-01 1.05909395e+00 -8.54886472e-01
-5.18432677e-01 7.25111663e-01 -4.51637119e-01 -1.57476619e-01
3.28061767e-02 5.80600977e-01 9.18966353e-01 -5.51163495e-01
8.02852392e-01 1.29711759e+00 9.68253791e-01 3.94556910e-01
-1.81540775e+00 -1.38239432e-02 3.86906534e-01 -4.99457955e-01
-1.26673961e+00 -7.98268765e-02 5.50051272e-01 -6.68532610e-01
1.96295589e-01 1.71339914e-01 1.06684101e+00 1.40027690e+00
4.98958647e-01 9.12246525e-01 6.73710883e-01 1.27832010e-01
6.68931067e-01 -5.21283805e-01 -2.05647051e-01 8.65088344e-01
4.44970727e-01 5.16848683e-01 -3.84458899e-01 -5.08517623e-01
1.23730659e+00 -1.02409557e-01 -3.93780738e-01 -7.63145387e-01
-1.04821777e+00 5.97069800e-01 5.04548490e-01 -3.30349982e-01
-6.27090394e-01 6.46138370e-01 6.08452484e-02 -3.17528903e-01
4.33173418e-01 1.61359802e-01 -5.60679913e-01 -2.73913443e-01
-7.55870283e-01 8.29280436e-01 1.26613510e+00 8.56433034e-01
4.51328039e-01 1.00224920e-01 1.66478455e-02 1.46942168e-01
7.84681678e-01 1.12446284e+00 -5.35419464e-01 -1.51016641e+00
3.26828361e-02 1.92761034e-01 6.45905375e-01 -8.27373862e-01
-1.69489831e-01 -3.63409817e-01 -7.23632812e-01 5.10872543e-01
7.73508966e-01 -1.25051692e-01 -1.10041785e+00 1.88162088e+00
5.22682369e-01 2.50601947e-01 -2.88615614e-01 9.67798710e-01
1.32880583e-01 3.95738512e-01 1.09655499e-01 3.70301679e-02
8.52018774e-01 2.94765122e-02 -2.10929587e-01 -2.29599066e-02
1.82430699e-01 -2.38990039e-01 8.95052493e-01 4.18382168e-01
-1.32675183e+00 -5.20359799e-02 -7.34814346e-01 1.25018269e-01
-7.09644333e-02 7.10207224e-02 7.14688420e-01 2.63747573e-01
-8.68164837e-01 1.28548563e+00 -1.59920919e+00 -2.75604635e-01
5.85033298e-01 1.58730984e-01 2.65647888e-01 2.73140788e-01
-5.80850720e-01 8.53968740e-01 2.09612455e-02 6.16949022e-01
-1.50162530e+00 -1.00624859e+00 -5.76475203e-01 -1.23907123e-02
4.56500441e-01 -8.79007816e-01 9.27940786e-01 -2.92633772e-01
-1.93665302e+00 6.65602505e-01 6.89880028e-02 -2.59191900e-01
9.84947026e-01 -4.41115677e-01 5.75563848e-01 -6.80529475e-02
-1.89271986e-01 7.29274511e-01 9.60811257e-01 -1.46643674e+00
1.46736994e-01 -6.42951071e-01 -6.56037256e-02 1.66875601e-01
4.81357455e-01 -1.11372638e+00 1.33546680e-01 -5.02675414e-01
4.28840935e-01 -1.30602062e+00 -4.05843556e-01 8.87451470e-01
-4.65685964e-01 2.81848609e-01 8.30670297e-01 -7.91524291e-01
4.06463236e-01 -1.80794466e+00 8.60891402e-01 5.03791273e-01
1.74102068e-01 -2.39467829e-01 1.56533942e-01 4.69758779e-01
7.37731993e-01 8.26075077e-02 -5.66046119e-01 2.03482419e-01
2.42027909e-01 5.58995426e-01 -6.50441945e-01 5.30613780e-01
3.30868036e-01 1.02526832e+00 -1.13501179e+00 -2.70973265e-01
2.98449218e-01 5.57221472e-01 -4.48365092e-01 -4.68502007e-02
-8.30163956e-01 7.53156960e-01 -1.03775656e+00 8.83427262e-01
6.16205871e-01 -3.15966576e-01 1.75739899e-01 -1.31457478e-01
-1.15383148e-01 -1.86534941e-01 -1.11973011e+00 1.76327753e+00
-4.16753769e-01 4.26125452e-02 8.85902822e-01 -7.01946735e-01
8.37583780e-01 -3.39773074e-02 2.87150532e-01 4.30320710e-01
1.21128730e-01 2.28971258e-01 -1.83931485e-01 -3.49660993e-01
3.15582216e-01 2.09586084e-01 2.23163992e-01 1.88445553e-01
-1.54972613e-01 -1.05505919e+00 -2.75063753e-01 7.41389096e-02
1.04758954e+00 1.26310313e+00 -2.42381573e-01 -5.89955568e-01
-8.43529478e-02 6.40839934e-02 2.49620214e-01 7.11923540e-01
6.41498640e-02 1.77731752e-01 2.82167464e-01 -2.26883739e-01
-9.46459770e-01 -1.76100349e+00 -4.29078758e-01 7.76844442e-01
4.41730559e-01 1.44005433e-01 -6.41271055e-01 3.13348509e-02
7.43607402e-01 5.33870637e-01 -7.26921141e-01 -1.78894088e-01
-3.40192884e-01 -6.35622323e-01 3.58660698e-01 6.66355193e-01
2.65405089e-01 -8.42933297e-01 -9.57333684e-01 6.36291623e-01
5.26834046e-05 -8.07134211e-01 9.17327851e-02 1.78216264e-01
-1.23543525e+00 -9.80428338e-01 -7.12909698e-01 -1.09787449e-01
2.81962663e-01 -4.84107107e-01 1.09916711e+00 -6.40962124e-02
-6.52423203e-01 8.14768910e-01 -7.46320412e-02 -2.84519941e-01
-4.88872439e-01 1.07698523e-01 -1.90941781e-01 -4.22762275e-01
-1.81809545e-01 -1.24501169e+00 -5.96150219e-01 2.17394248e-01
-4.96845156e-01 -2.58320775e-02 3.23237449e-01 8.16014588e-01
7.30656087e-01 -5.20946145e-01 2.17763949e-02 -2.99225241e-01
3.18652093e-01 -3.80251229e-01 -8.77801895e-01 3.45451325e-01
-8.21722299e-02 4.06751364e-01 1.91672727e-01 -6.49006188e-01
-9.12389636e-01 1.60381630e-01 1.86285362e-01 -3.36594790e-01
5.46658374e-02 5.25033414e-01 1.99896097e-02 -2.93808818e-01
6.46372318e-01 -3.80288959e-02 2.82023191e-01 -2.94143051e-01
6.65556967e-01 4.83266376e-02 7.54616380e-01 -1.34339476e+00
8.33457828e-01 9.09409165e-01 6.32994413e-01 -9.41698790e-01
-4.97790843e-01 4.76960450e-01 -6.80196106e-01 -4.64789957e-01
7.26842105e-01 -6.19419038e-01 -1.52396131e+00 6.38091326e-01
-9.08692420e-01 -9.62189734e-01 -6.05131865e-01 3.82487267e-01
-9.68318224e-01 1.76179186e-01 -6.63511992e-01 -1.29004788e+00
-1.99799687e-01 -1.01967847e+00 1.71030378e+00 1.55314699e-01
-1.88613698e-01 -1.07656908e+00 -2.02632491e-02 8.13699269e-04
3.72125685e-01 1.05007243e+00 8.21593225e-01 1.59802675e-01
-1.21681845e+00 -1.23932190e-01 1.53049812e-01 -1.58139005e-01
-2.45293722e-01 6.97307467e-01 -7.28508949e-01 -2.13160247e-01
-3.98597091e-01 -4.47489083e-01 8.75038385e-01 1.04620767e+00
1.10928643e+00 5.05573787e-02 -8.38102102e-01 5.34532607e-01
1.18183017e+00 -4.71112132e-01 2.84762055e-01 -2.99084574e-01
5.78326225e-01 6.35450423e-01 2.91933179e-01 7.37447917e-01
3.95724446e-01 3.26124400e-01 1.09885085e+00 3.49837780e-01
3.71999979e-01 -3.79866838e-01 8.14921260e-02 4.00824368e-01
-5.12224078e-01 -3.55352700e-01 -8.40616584e-01 4.13020223e-01
-1.80644143e+00 -9.14442062e-01 -1.31404653e-01 2.25633717e+00
7.31840551e-01 9.40947160e-02 -4.79542501e-02 -2.27020189e-01
4.67789412e-01 -2.44600810e-02 -1.34376657e+00 2.86677643e-03
2.34633479e-02 3.99081796e-01 7.69964218e-01 7.50134647e-01
-7.64795661e-01 8.50018501e-01 6.60249329e+00 4.16224003e-01
-8.24562907e-01 -2.29963690e-01 3.48026544e-01 2.02689096e-01
-5.99853516e-01 3.44172299e-01 -6.22404754e-01 2.83614278e-01
6.43816769e-01 9.65837985e-02 7.86694169e-01 6.50938392e-01
4.10161793e-01 -4.32127953e-01 -1.04440916e+00 2.35418022e-01
-6.19187772e-01 -1.40452397e+00 -2.75214940e-01 3.09372783e-01
3.48680526e-01 1.87994838e-01 4.24691476e-02 -4.22181673e-02
1.23029816e+00 -8.46234024e-01 9.04744387e-01 9.38126981e-01
5.85982561e-01 -2.65695125e-01 4.31158394e-02 7.01010883e-01
-1.18446362e+00 2.49579892e-01 -1.83046460e-01 -1.90332875e-01
4.45837498e-01 7.09550202e-01 -4.26047534e-01 2.86379695e-01
6.23900771e-01 8.34628463e-01 6.94867074e-02 5.12908280e-01
-1.10625796e-01 8.14670265e-01 -1.13116956e+00 -1.78759247e-01
-1.14476517e-01 -7.13050246e-01 1.01005721e+00 4.28967088e-01
1.00687012e-01 2.40776800e-02 4.53881770e-01 1.59718311e+00
2.52730072e-01 -5.51726401e-01 -5.26039898e-01 -2.13918149e-01
5.93750894e-01 9.27192986e-01 -8.31151545e-01 1.91824794e-01
5.01199424e-01 6.38717890e-01 2.86771804e-01 4.92329419e-01
-8.15701246e-01 1.74587369e-01 5.59622407e-01 1.85367107e-01
4.78750736e-01 -5.14115810e-01 -3.01173836e-01 -1.14119768e+00
1.27003297e-01 -2.51931638e-01 -1.85757607e-01 -1.00472307e+00
-1.38985229e+00 -2.37549737e-01 3.03250700e-01 -6.11162543e-01
-3.12619925e-01 -8.19989800e-01 -5.74224114e-01 8.82171750e-01
-1.07137728e+00 -1.35287857e+00 -4.35559452e-01 -2.33299769e-02
2.23166775e-02 6.54866695e-01 8.10030520e-01 -6.02559865e-01
-2.73663789e-01 -3.90733927e-02 3.31607401e-01 -2.10297674e-01
-6.70161322e-02 -1.19059455e+00 3.23104143e-01 5.67488611e-01
-3.39888215e-01 3.20751756e-01 9.49267626e-01 -1.07484710e+00
-1.81215954e+00 -7.80057669e-01 -2.94279426e-01 -5.52959442e-01
8.59985471e-01 -2.73642778e-01 -1.02226138e+00 6.12115145e-01
-5.64096689e-01 3.38929206e-01 -1.53367192e-01 5.29297441e-02
8.31665024e-02 1.92961216e-01 -1.28120494e+00 5.15807271e-01
1.35230625e+00 -2.68585443e-01 -1.83572933e-01 3.03174764e-01
2.87721962e-01 -9.55746114e-01 -1.17791927e+00 6.34866655e-01
1.07281387e+00 -3.89372021e-01 1.05261374e+00 -4.35088724e-01
5.59261322e-01 -1.64587274e-01 -3.07076067e-01 -1.05327344e+00
4.39297110e-02 -9.31904256e-01 -4.48490620e-01 1.06844962e+00
1.67290032e-01 -3.89499277e-01 9.67594385e-01 6.02901042e-01
1.56711921e-01 -8.26719224e-01 -9.12661791e-01 -7.89973557e-01
4.08521891e-01 -2.63331324e-01 3.16013366e-01 3.21706176e-01
-5.44997036e-01 -3.05245250e-01 1.15196137e-02 5.57409406e-01
1.18695438e+00 1.43851399e-01 7.14993298e-01 -1.72738051e+00
-7.01285124e-01 -2.51827538e-01 -1.91008955e-01 -1.31379580e+00
2.67793208e-01 -6.56950593e-01 4.02855873e-01 -1.18798387e+00
-1.53374463e-01 -5.65305233e-01 6.02706075e-01 1.23192936e-01
-1.14124842e-01 -5.01902223e-01 -1.31865725e-01 1.30883068e-01
2.82542139e-01 8.44519436e-01 1.46392012e+00 9.72186550e-02
-2.85781115e-01 1.60154969e-01 1.23426698e-01 9.23431933e-01
3.71405900e-01 -3.56039703e-01 -3.05125356e-01 -4.21425551e-01
2.26127401e-01 3.83431673e-01 1.08091891e+00 -7.41156161e-01
-1.68903410e-01 -4.07192022e-01 2.97148347e-01 -5.12501001e-01
5.27246594e-01 -7.94824064e-01 3.98685068e-01 6.25222087e-01
-4.52783972e-01 -3.67853105e-01 2.16552198e-01 1.18808520e+00
6.68076396e-01 -1.75592899e-01 7.05573916e-01 -2.90158719e-01
-1.54864714e-01 3.62184107e-01 -4.80625302e-01 3.61710310e-01
5.74123561e-01 -3.18852514e-01 -1.08454123e-01 -2.94842005e-01
-1.10893238e+00 3.36377501e-01 7.16049016e-01 -1.44742936e-01
2.32039467e-01 -6.16391599e-01 -8.02652717e-01 1.50897466e-02
-3.29422444e-01 5.24613023e-01 -2.68052835e-02 8.27928185e-01
-7.27835596e-01 -5.04124761e-01 7.67647624e-02 -1.17938054e+00
-5.47912478e-01 -2.39630248e-02 4.72790062e-01 -2.13127494e-01
-7.65439212e-01 8.82682264e-01 -2.26826519e-01 -4.72556293e-01
1.17934234e-01 -8.95466626e-01 4.34396684e-01 -4.03211892e-01
-1.92173705e-01 5.07028401e-01 -3.83105218e-01 -1.15940563e-01
-1.01003915e-01 6.66139781e-01 3.51493478e-01 -3.42911720e-01
1.50086451e+00 -5.95056191e-02 2.09289696e-02 6.69856548e-01
2.65775710e-01 -3.80284965e-01 -2.12510085e+00 3.12957019e-02
-1.31882831e-01 -2.62407124e-01 2.28831381e-01 -7.41407573e-01
-6.90551341e-01 7.91804790e-01 3.75878513e-01 1.88291669e-01
4.41706359e-01 -4.99643274e-02 4.87165153e-01 4.90880907e-01
6.96693957e-01 -7.81273723e-01 -2.70161361e-01 3.10798943e-01
8.61423075e-01 -8.31538796e-01 1.80613995e-01 -6.51194155e-01
-1.06100440e-01 1.16786599e+00 3.19872260e-01 -7.17613697e-01
1.06523693e+00 8.58148456e-01 -4.72613066e-01 -9.87079591e-02
-6.52818501e-01 -4.47986424e-02 -5.75672872e-02 7.77257562e-01
-3.15085016e-02 1.91801414e-01 2.40773708e-01 -3.00784619e-03
-6.46500383e-03 2.59745002e-01 3.42800230e-01 1.22749984e+00
-4.92473125e-01 -6.77755475e-01 -2.37543866e-01 5.17684698e-01
1.09679379e-01 1.39629558e-01 -2.31889531e-01 9.14637566e-01
-1.85209960e-01 3.45790029e-01 8.37322418e-03 2.02087402e-01
1.76677197e-01 -7.99101144e-02 9.21355724e-01 -2.66113192e-01
-2.16275021e-01 -2.53705289e-02 -2.68074363e-01 -6.22091353e-01
-3.55719596e-01 -1.02863681e+00 -1.08006525e+00 -4.71619755e-01
-6.26074910e-01 -2.66499221e-01 6.91247404e-01 1.06577694e+00
3.83688748e-01 3.59553874e-01 -2.33024806e-02 -1.48637021e+00
-1.18243825e+00 -7.54965246e-01 -6.24437094e-01 2.55328000e-01
2.52015829e-01 -1.02385247e+00 -4.94003236e-01 2.18868047e-01] | [5.503587245941162, -0.3750949203968048] |
768e7794-661c-44b8-b412-77094f4ff6d5 | hinted-networks | 1812.06297 | null | http://arxiv.org/abs/1812.06297v1 | http://arxiv.org/pdf/1812.06297v1.pdf | Hinted Networks | We present Hinted Networks: a collection of architectural transformations for
improving the accuracies of neural network models for regression tasks, through
the injection of a prior for the output prediction (i.e. a hint). We ground our
investigations within the camera relocalization domain, and propose two
variants, namely the Hinted Embedding and Hinted Residual networks, both
applied to the PoseNet base model for regressing camera pose from an image. Our
evaluations show practical improvements in localization accuracy for standard
outdoor and indoor localization datasets, without using additional information.
We further assess the range of accuracy gains within an aerial-view
localization setup, simulated across vast areas at different times of the year. | ['Joel Lamy-Poirier', 'Anqi Xu'] | 2018-12-15 | null | null | null | null | ['camera-relocalization'] | ['computer-vision'] | [ 1.48366034e-01 2.05610901e-01 -3.24432760e-01 -5.35047650e-01
-5.78754961e-01 -7.10195839e-01 7.13641107e-01 -3.13055694e-01
-3.88500154e-01 6.18966103e-01 2.36989185e-01 -2.92614430e-01
-3.36925417e-01 -3.13859969e-01 -1.07322788e+00 -5.69532692e-01
-2.34283552e-01 5.77425100e-02 -9.77985635e-02 -2.67399281e-01
-1.53209016e-01 7.78342128e-01 -1.08597493e+00 -2.77441293e-02
2.32145816e-01 9.41146076e-01 -6.08714670e-03 9.16677713e-01
6.83775723e-01 1.16319060e+00 -5.25944173e-01 -2.15406507e-01
5.53897023e-01 6.19525909e-02 -5.10402977e-01 -1.26821086e-01
9.70717251e-01 -4.06117529e-01 -7.40685165e-01 6.51585162e-01
4.06554163e-01 1.95671901e-01 5.16990721e-01 -1.21859074e+00
-6.48370326e-01 5.38171113e-01 -1.00643136e-01 1.42702982e-01
4.24503982e-01 5.56291267e-02 9.41999674e-01 -1.02384865e+00
5.47635317e-01 7.92001009e-01 1.41260695e+00 2.83021927e-01
-1.31970370e+00 -4.53686446e-01 2.51789570e-01 1.91496580e-03
-1.62647557e+00 -6.07358158e-01 5.67886949e-01 -5.58963418e-01
1.26410925e+00 2.05016155e-02 3.38761002e-01 1.48628354e+00
3.23246300e-01 3.34837019e-01 6.82844937e-01 -3.14445436e-01
9.26416442e-02 2.66497433e-01 -1.50461063e-01 8.13943326e-01
3.17124099e-01 4.57655758e-01 -5.87754190e-01 -4.98513021e-02
8.96243632e-01 2.50663310e-01 -5.13194621e-01 -8.41333985e-01
-1.30628312e+00 7.57413685e-01 1.13347101e+00 5.26714809e-02
-4.31093484e-01 6.15315259e-01 1.60440654e-02 9.75237936e-02
5.94742060e-01 7.49366045e-01 -8.66119146e-01 3.30033511e-01
-1.15800166e+00 -2.65881419e-02 7.67694712e-01 1.14060330e+00
9.68790293e-01 2.17794850e-01 -2.30369787e-03 5.55399537e-01
4.03614223e-01 6.41745925e-01 4.30938959e-01 -9.47919309e-01
5.70277572e-01 4.20634180e-01 2.71035850e-01 -1.36079895e+00
-5.70727170e-01 -6.64787948e-01 -8.31217527e-01 -8.67935345e-02
7.62484223e-02 -4.20400381e-01 -8.38847280e-01 1.77179754e+00
-6.95398748e-02 5.18699467e-01 7.01330453e-02 9.27458942e-01
6.18184745e-01 5.75442135e-01 -2.12784156e-01 4.61043924e-01
7.65193403e-01 -1.29099405e+00 -2.24627495e-01 -6.83319449e-01
7.23012328e-01 -4.25015867e-01 7.02154219e-01 2.06776202e-01
-5.80912292e-01 -8.18893194e-01 -1.36168635e+00 -5.46901748e-02
-7.44575441e-01 7.05663621e-01 5.05780935e-01 4.15055782e-01
-1.48633492e+00 8.14621389e-01 -1.04476476e+00 -6.80156171e-01
-5.19690774e-02 9.01577592e-01 -8.75447750e-01 2.53207944e-02
-8.69779706e-01 1.06033468e+00 8.43875781e-02 5.51482916e-01
-9.77734625e-01 -6.04770899e-01 -1.26092374e+00 -2.21349932e-02
5.94325848e-02 -7.74501026e-01 1.08693326e+00 -8.14917862e-01
-1.50942099e+00 5.06228387e-01 1.23545723e-02 -7.97952831e-01
2.46149436e-01 -5.28154254e-01 -2.62304097e-01 -3.70042890e-01
6.95978180e-02 1.10563314e+00 9.75585282e-01 -1.22628677e+00
-5.53769886e-01 -2.74117678e-01 2.84932882e-01 1.48223981e-01
-1.57673791e-01 -5.84372878e-01 -2.52270937e-01 -4.79254603e-01
2.02895492e-01 -1.35515499e+00 -6.85711682e-01 -2.27518931e-01
-4.75315899e-01 4.52411115e-01 7.73244321e-01 -7.29963660e-01
7.13562429e-01 -2.01267195e+00 3.38037878e-01 1.72224864e-01
2.25779369e-01 -5.27515300e-02 -2.72875190e-01 4.19655383e-01
-4.61025536e-01 1.07479997e-01 9.83940586e-02 -5.64214528e-01
3.57483067e-02 3.33771259e-01 -5.39337993e-01 8.72155726e-01
4.72559333e-02 9.02323008e-01 -7.51240015e-01 1.64267123e-01
6.55151486e-01 7.20399797e-01 -6.64296865e-01 4.11661983e-01
7.12192580e-02 5.23599625e-01 -9.64096785e-02 5.34603119e-01
4.40605789e-01 -1.75080806e-01 -1.01265624e-01 -4.68142211e-01
-3.90055701e-02 2.53080755e-01 -9.56392407e-01 1.95765054e+00
-8.30363214e-01 1.12661231e+00 -8.66216943e-02 -4.91553038e-01
9.24296498e-01 9.46705416e-02 4.41136032e-01 -3.46955180e-01
7.37027079e-02 -1.31493196e-01 -3.03636104e-01 -1.01161636e-01
8.60020220e-01 4.41596329e-01 -2.93130308e-01 -3.32104973e-02
7.42971241e-01 -9.21168625e-02 -3.83360207e-01 -2.42576763e-01
1.31248808e+00 5.06883621e-01 3.07247192e-01 -2.10842267e-01
4.69566524e-01 1.14206471e-01 8.34943280e-02 9.42579031e-01
2.61724778e-02 7.46803761e-01 1.08885080e-01 -9.26249027e-01
-1.00126660e+00 -8.22671056e-01 1.74088299e-01 1.23209679e+00
2.02258617e-01 -6.72412872e-01 -5.54578543e-01 -8.23477209e-01
6.62130117e-02 4.72307742e-01 -6.37840688e-01 -3.23195606e-01
-5.21673143e-01 -5.84742904e-01 8.56122077e-01 7.81543791e-01
5.29983401e-01 -5.28001964e-01 -6.86976075e-01 -1.16562881e-01
5.25268875e-02 -1.40130746e+00 -1.78615190e-02 1.02172279e+00
-8.04144442e-01 -1.02251601e+00 -5.59747517e-01 -6.75949216e-01
6.88604414e-01 2.26757377e-01 1.16411138e+00 -1.92870095e-01
1.32989101e-02 7.11847484e-01 -1.39610544e-01 -1.70222506e-01
1.70302987e-01 5.17790377e-01 4.41913813e-01 -3.66977900e-01
1.47681445e-01 -3.97919416e-01 -5.35093307e-01 5.10262668e-01
-3.52209270e-01 -2.74096370e-01 7.24294662e-01 9.90755200e-01
4.39111292e-01 -1.69134527e-01 -4.73360926e-01 -5.72664678e-01
2.73038328e-01 -4.12168175e-01 -9.55436110e-01 1.60522163e-02
-4.67019141e-01 1.76952034e-01 8.33284020e-01 -2.77115703e-01
-4.71995264e-01 5.77016771e-01 -1.16861619e-01 -6.01189137e-01
-3.22653204e-01 3.05745482e-01 1.83986858e-01 -7.50018954e-01
8.42069089e-01 -1.71035886e-01 -4.59674418e-01 -1.88029960e-01
6.71739995e-01 3.79807025e-01 6.73970520e-01 -3.45803320e-01
1.07928026e+00 1.29630953e-01 1.94539711e-01 -9.06043828e-01
-7.98631191e-01 -4.23387229e-01 -1.14719903e+00 -7.36087039e-02
6.82023644e-01 -1.28427172e+00 -5.26088834e-01 -3.96120995e-02
-1.10472250e+00 -6.74840331e-01 -2.47715056e-01 7.70516396e-01
-4.40002561e-01 -2.48928130e-01 -4.69983965e-01 -2.16803238e-01
1.29879937e-01 -1.36772585e+00 1.35550880e+00 1.73445761e-01
-1.53369874e-01 -1.27743447e+00 3.53129804e-01 -1.10628277e-01
4.59995836e-01 2.69989222e-01 4.03232217e-01 -6.06131911e-01
-8.13689470e-01 -4.19352353e-01 -5.72343394e-02 3.96714568e-01
-4.10522521e-02 -8.62894729e-02 -1.31611514e+00 -4.43865478e-01
-2.56608546e-01 -2.80128270e-01 8.46079767e-01 4.95801538e-01
8.75867546e-01 -4.94010985e-01 -4.75501835e-01 1.28052974e+00
1.53654993e+00 -2.42093444e-01 4.89907712e-01 4.94627446e-01
8.98564279e-01 4.60583605e-02 2.96152741e-01 2.19196439e-01
2.35785291e-01 8.71058106e-01 8.33677888e-01 -3.72950435e-01
-1.45792058e-02 -6.68000400e-01 3.47363442e-01 4.95117515e-01
-2.51650333e-01 -4.87522215e-01 -9.80714917e-01 2.57915467e-01
-1.78845108e+00 -4.52684492e-01 1.20877095e-01 2.09122014e+00
-1.46425627e-02 -1.30492002e-01 -4.92487550e-01 -3.09548259e-01
3.81343484e-01 6.03272378e-01 -4.07696337e-01 -1.37813151e-01
2.20703870e-01 1.18966721e-01 1.53902233e+00 7.19162226e-01
-1.47540486e+00 1.21492636e+00 7.78445578e+00 7.26575181e-02
-1.37160289e+00 7.44606107e-02 3.58754873e-01 3.04648519e-01
3.53777766e-01 6.14874065e-02 -9.18184161e-01 -7.89310709e-02
1.24613440e+00 6.49995267e-01 6.07264161e-01 1.45352149e+00
-1.11656033e-01 8.34397227e-02 -1.54795146e+00 9.86811161e-01
3.77810180e-01 -1.37007499e+00 -2.30187491e-01 1.52024880e-01
6.56659126e-01 5.53494930e-01 3.34703654e-01 5.33403158e-01
4.78819579e-01 -1.07975531e+00 6.94943607e-01 4.58432287e-01
7.91811109e-01 -5.08579552e-01 1.21616399e+00 2.77794927e-01
-1.07374430e+00 -2.07728803e-01 -5.38417161e-01 -2.81153947e-01
-1.69437394e-01 -1.24920923e-02 -1.26243818e+00 4.02544916e-01
7.63641000e-01 1.04381561e+00 -1.25138915e+00 7.36038327e-01
-5.61342061e-01 5.61971486e-01 -3.76091719e-01 2.28133991e-01
4.31008816e-01 8.78323242e-03 3.21848840e-01 9.79205728e-01
4.25794661e-01 -4.87382531e-01 8.18476453e-02 7.77508140e-01
-7.40303993e-02 -5.19881487e-01 -1.26314485e+00 5.67246854e-01
3.59849304e-01 1.29533684e+00 -4.56968486e-01 3.54422599e-01
-2.01282665e-01 1.17087197e+00 3.30766469e-01 5.67862213e-01
-1.20903063e+00 -2.11455300e-01 6.24168634e-01 3.35043781e-02
4.40616965e-01 -5.10313392e-01 3.60546038e-02 -1.04320884e+00
-2.48795003e-01 -5.49604118e-01 -1.49785951e-01 -1.28741276e+00
-8.23744476e-01 6.24093711e-01 1.28003508e-01 -1.23162293e+00
-6.27775192e-01 -1.14934349e+00 -2.95477480e-01 6.14547014e-01
-1.24603701e+00 -1.47919476e+00 -6.20603740e-01 5.86485684e-01
3.33497047e-01 -3.46498489e-01 1.02076638e+00 2.90837735e-01
-6.42551661e-01 5.30398488e-01 -3.43312882e-02 3.03101659e-01
7.96893895e-01 -1.33440185e+00 4.93210733e-01 9.72749054e-01
6.20845795e-01 9.89401281e-01 8.38920116e-01 -1.21658467e-01
-1.56655061e+00 -1.46027732e+00 5.12723804e-01 -9.91255462e-01
5.88376760e-01 -7.46803105e-01 -2.65632346e-02 1.56727684e+00
2.95202315e-01 3.44145179e-01 3.56618434e-01 4.36253220e-01
-1.21465281e-01 -2.69965768e-01 -9.12253916e-01 6.57625198e-01
7.59760499e-01 -7.26813614e-01 -3.42092127e-01 3.28495055e-01
9.96807218e-01 -7.07712471e-01 -8.78155351e-01 6.44158363e-01
5.43958783e-01 -1.01359653e+00 1.42173338e+00 -2.92241901e-01
1.59389660e-01 -3.29052389e-01 -7.36801803e-01 -1.52904677e+00
-7.06718385e-01 -4.04683471e-01 3.71138863e-02 9.42185104e-01
6.08614922e-01 -5.35979867e-01 1.02566397e+00 1.48429692e-01
-2.38388956e-01 -4.61965740e-01 -8.47653925e-01 -5.53830564e-01
-3.52712214e-01 -3.56475592e-01 5.10315120e-01 6.35510623e-01
-5.81984341e-01 5.75567722e-01 -7.34421432e-01 9.17200446e-01
1.75776705e-01 -6.56517625e-01 1.18772209e+00 -1.06482887e+00
-2.16923922e-01 1.05987996e-01 -9.63996589e-01 -1.16402435e+00
3.44589651e-01 -5.82406342e-01 2.75742203e-01 -1.24723446e+00
-3.75899822e-01 -1.01838849e-01 -2.03209773e-01 5.30809820e-01
4.78915691e-01 4.65865701e-01 1.70919187e-02 2.84008920e-01
-6.81749225e-01 4.62722510e-01 3.49078119e-01 -2.50222862e-01
-1.40779912e-01 1.29373059e-01 -3.19341600e-01 1.11746657e+00
5.86501956e-01 -5.54228723e-01 -5.19194007e-01 -7.73468912e-01
2.92628646e-01 -5.68219870e-02 8.01557302e-01 -1.67153943e+00
4.27720398e-01 2.49399588e-01 9.14600432e-01 -6.40735269e-01
6.57770991e-01 -1.45510483e+00 1.27623022e-01 3.73141319e-01
-4.62745667e-01 6.57918215e-01 2.54070491e-01 6.51022255e-01
-1.51345164e-01 1.58994317e-01 4.47492898e-01 -6.72685774e-03
-1.16156340e+00 1.35688663e-01 -1.23361796e-01 -4.95951414e-01
8.56454849e-01 -1.40574291e-01 -3.57739478e-01 -4.89373207e-01
-8.70349705e-01 -9.54402164e-02 5.23948371e-01 4.61030304e-01
5.46648681e-01 -1.38306224e+00 -5.05061522e-02 6.14749253e-01
3.25301677e-01 -1.16763696e-01 -4.18620586e-01 8.77136767e-01
-8.71058643e-01 9.44798350e-01 -1.30727947e-01 -9.20913100e-01
-8.98180127e-01 4.56623435e-01 7.75784552e-01 -3.07626992e-01
-3.45598012e-01 8.33121121e-01 -1.46404132e-01 -1.12199092e+00
5.53923070e-01 -8.57432902e-01 1.70653649e-02 -3.52287620e-01
-8.39196052e-03 1.87059551e-01 2.73542821e-01 -7.41743743e-01
-5.36780953e-01 6.73126280e-01 4.05068040e-01 -9.75098312e-02
1.28084993e+00 -1.21860139e-01 1.36599943e-01 5.10949254e-01
1.12543154e+00 4.11847718e-02 -1.39227247e+00 1.81378976e-01
3.51235121e-02 -1.54765919e-01 3.41922611e-01 -8.28767657e-01
-8.88991058e-01 6.22067869e-01 8.79882514e-01 -2.38875598e-01
7.60854185e-01 -1.48745745e-01 1.58947498e-01 1.04288816e+00
6.93675578e-01 -5.70950508e-01 -1.03750087e-01 6.85358524e-01
6.60906255e-01 -1.24873209e+00 1.67494193e-01 6.13105111e-02
-4.77616906e-01 1.04790282e+00 6.79717720e-01 -5.84746063e-01
6.88388228e-01 3.38179916e-01 9.01955962e-02 -1.18461788e-01
-4.54252154e-01 -2.63527017e-02 4.29033995e-01 7.55534291e-01
4.32990074e-01 -1.76111743e-01 8.11659813e-01 4.13700014e-01
-3.43785822e-01 -2.36508116e-01 2.03360766e-01 7.33692169e-01
-1.21092565e-01 -6.88088834e-01 -3.96997273e-01 1.26648676e-02
2.90805846e-03 -1.52607709e-01 -7.92561352e-01 1.38447618e+00
4.71141189e-01 8.32056761e-01 -8.64760280e-02 -1.02706873e+00
4.30977553e-01 -1.67595327e-01 3.27301234e-01 -5.61544061e-01
-4.90875244e-01 -3.93102229e-01 -1.60190836e-01 -9.80065107e-01
-4.06484038e-01 -1.72233865e-01 -4.93349552e-01 -1.54957727e-01
-5.86310565e-01 -4.22613462e-04 8.08221996e-01 6.58611238e-01
3.38969648e-01 7.27139413e-01 5.58322966e-01 -1.57077277e+00
-4.05436546e-01 -1.02898848e+00 -3.02581787e-01 -3.21364403e-01
8.24628353e-01 -6.69296086e-01 -4.71624285e-01 8.10244083e-02] | [7.666461944580078, -2.142548084259033] |
b1621242-1939-4523-bb62-4943427e69b5 | infer-intermediate-representations-for-future | 1903.10641 | null | http://arxiv.org/abs/1903.10641v1 | http://arxiv.org/pdf/1903.10641v1.pdf | INFER: INtermediate representations for FuturE pRediction | In urban driving scenarios, forecasting future trajectories of surrounding
vehicles is of paramount importance. While several approaches for the problem
have been proposed, the best-performing ones tend to require extremely detailed
input representations (eg. image sequences). But, such methods do not
generalize to datasets they have not been trained on. We propose intermediate
representations that are particularly well-suited for future prediction. As
opposed to using texture (color) information, we rely on semantics and train an
autoregressive model to accurately predict future trajectories of traffic
participants (vehicles) (see fig. above). We demonstrate that using semantics
provides a significant boost over techniques that operate over raw pixel
intensities/disparities. Uncharacteristic of state-of-the-art approaches, our
representations and models generalize to completely different datasets,
collected across several cities, and also across countries where people drive
on opposite sides of the road (left-handed vs right-handed driving).
Additionally, we demonstrate an application of our approach in multi-object
tracking (data association). To foster further research in transferrable
representations and ensure reproducibility, we release all our code and data. | ['Madhava Krishna K', 'Krishna Murthy J.', 'Junaid Ahmed Ansari', 'Shashank Srikanth', 'Karnik Ram R', 'Sarthak Sharma'] | 2019-03-26 | null | null | null | null | ['activity-prediction', 'activity-prediction'] | ['computer-vision', 'time-series'] | [ 1.76303804e-01 -6.27243370e-02 -1.88388735e-01 -6.28458858e-01
-5.07675767e-01 -5.46750486e-01 1.09861827e+00 -1.03279755e-01
-4.13216680e-01 6.72972143e-01 3.65630895e-01 -5.65061808e-01
6.65050000e-02 -9.52897906e-01 -8.80760550e-01 -4.56463754e-01
-2.08768457e-01 4.23874646e-01 3.97086859e-01 -5.33463240e-01
1.08898416e-01 7.25606978e-01 -2.10750961e+00 4.50575829e-01
5.24814308e-01 7.24740803e-01 1.73541680e-01 7.60378480e-01
-7.39201233e-02 7.88745999e-01 -1.99033275e-01 -5.22699654e-01
3.06322336e-01 -3.65631580e-02 -5.99457681e-01 -1.16186310e-02
8.83686185e-01 -2.16779813e-01 -7.92659760e-01 7.02497721e-01
8.37277249e-02 1.95198968e-01 8.43263805e-01 -1.52039409e+00
-7.87148774e-01 1.13462575e-01 -3.46347332e-01 3.49548608e-01
1.13004856e-01 1.93580732e-01 7.34517992e-01 -6.59810305e-01
6.78285360e-01 1.39041972e+00 7.48417616e-01 6.79201901e-01
-1.24679816e+00 -7.22081959e-01 6.46617770e-01 5.43232143e-01
-1.15848494e+00 -7.10030615e-01 4.74997789e-01 -6.79921627e-01
6.73636913e-01 3.50274920e-01 4.83159840e-01 1.25564778e+00
1.45037755e-01 9.66600835e-01 1.06647825e+00 -8.28771964e-02
-6.25919923e-02 2.43079320e-01 1.44050028e-02 3.95822316e-01
2.91340172e-01 4.44040686e-01 -4.22331274e-01 5.33412620e-02
3.74853998e-01 2.93244928e-01 5.70913143e-02 -5.39864302e-01
-1.38074946e+00 7.72499859e-01 4.51490223e-01 1.67983949e-01
-4.18779343e-01 4.90559161e-01 2.45711595e-01 1.83440924e-01
6.71686172e-01 -1.69538811e-01 -2.33534917e-01 -2.47914433e-01
-1.12100399e+00 6.43831968e-01 5.20472288e-01 1.02479172e+00
1.02785301e+00 -1.59493294e-02 -1.84155419e-01 3.78795713e-01
2.20545813e-01 6.23938143e-01 2.47544169e-01 -8.78813505e-01
4.99024540e-01 2.66462564e-01 3.46308082e-01 -1.15548396e+00
-4.63902682e-01 -1.40997142e-01 -5.48874199e-01 5.38644731e-01
6.31170809e-01 -1.63005039e-01 -1.01716018e+00 1.73785174e+00
4.99543063e-02 5.14010012e-01 3.45248543e-02 9.57916617e-01
5.81236541e-01 5.47641098e-01 4.06962723e-01 2.29643017e-01
1.20731854e+00 -8.46101344e-01 -4.31373566e-01 -4.94164556e-01
5.56431532e-01 -4.54118609e-01 5.77958167e-01 1.88233152e-01
-9.25620914e-01 -7.72601128e-01 -7.80334234e-01 -5.76055869e-02
-9.26581264e-01 -3.57976370e-02 6.94135487e-01 8.20768952e-01
-1.30324733e+00 5.33485353e-01 -7.48999119e-01 -7.10521460e-01
4.57536161e-01 2.35148951e-01 -2.78919727e-01 -8.25911760e-02
-1.14837825e+00 1.28426647e+00 -2.78432923e-03 1.03206351e-01
-9.00445104e-01 -6.48848474e-01 -7.02545404e-01 -2.87109643e-01
2.90488359e-02 -5.86060941e-01 9.91727471e-01 -1.09904540e+00
-9.74118710e-01 9.06463861e-01 -4.47798550e-01 -7.94993222e-01
8.18938315e-01 -2.38114089e-01 -7.77135730e-01 -1.76128939e-01
1.31336138e-01 1.10016024e+00 7.49132872e-01 -1.26443708e+00
-9.35171485e-01 -2.79837728e-01 1.62227869e-01 -6.35161847e-02
-1.88551128e-01 -1.12267456e-03 -2.52862066e-01 -6.34097576e-01
-2.16273353e-01 -1.13025057e+00 -4.47173953e-01 2.78761029e-01
1.34933519e-03 -1.05987594e-01 1.15186322e+00 -6.11846924e-01
1.06352985e+00 -2.06412196e+00 -1.72073469e-01 8.00982341e-02
-2.13873256e-02 1.96036190e-01 -2.00637370e-01 3.36203903e-01
3.85154672e-02 -3.75679247e-02 -1.79437965e-01 -1.57747090e-01
1.18738584e-01 3.99254084e-01 -5.02017736e-01 6.14912570e-01
3.24423790e-01 1.03055644e+00 -9.38741744e-01 -3.10992718e-01
4.15926993e-01 6.87900484e-01 -4.50638682e-01 -2.12197945e-01
-2.41805226e-01 4.84793305e-01 -5.20674586e-01 2.76993245e-01
7.37771094e-01 -1.22390315e-01 8.59704167e-02 1.06862113e-01
-4.36360717e-01 1.58512831e-01 -1.09665442e+00 1.51874650e+00
-5.39903104e-01 1.04786706e+00 -2.35334545e-01 -1.04915094e+00
8.74652326e-01 9.48136598e-02 4.55183685e-01 -9.06317413e-01
-3.17779064e-01 2.62300298e-02 -1.53085351e-01 -4.97849643e-01
9.24363911e-01 -1.18811443e-01 -1.21892586e-01 3.53810310e-01
-3.83733332e-01 2.76910126e-01 1.72537729e-01 4.33914550e-02
8.79734397e-01 6.01157725e-01 -1.23997681e-01 -2.35309660e-01
4.89467323e-01 3.60240787e-01 3.39116931e-01 7.35027969e-01
-4.78272349e-01 6.69934511e-01 2.62410194e-01 -8.79089117e-01
-1.16496491e+00 -1.14200330e+00 -1.38340861e-01 1.34661376e+00
2.06325814e-01 -8.88383016e-02 -4.00519639e-01 -5.87117195e-01
2.96595544e-01 8.87519002e-01 -8.28539193e-01 8.05493817e-02
-8.11994553e-01 -5.90800703e-01 5.06812930e-01 7.85309494e-01
2.11315811e-01 -8.01225364e-01 -9.28187668e-01 2.33517036e-01
-1.12859800e-01 -1.06571877e+00 -1.43413678e-01 -3.71482491e-01
-6.56925678e-01 -9.38218832e-01 -8.33470166e-01 -3.83574963e-01
4.00328785e-01 5.56525767e-01 1.25875425e+00 5.26055507e-02
-2.03637689e-01 6.44938052e-01 -1.14131957e-01 -5.19479692e-01
-2.84489721e-01 -8.60373378e-02 1.37544153e-02 2.95370132e-01
6.01935506e-01 -2.94841856e-01 -7.45095015e-01 3.16585153e-01
-4.79440153e-01 1.70049921e-01 5.37551761e-01 4.21314985e-01
2.89286911e-01 -3.25318813e-01 4.39971179e-01 -8.43232274e-01
2.10035965e-01 -7.94921637e-01 -5.86334348e-01 2.56271929e-01
-3.15364689e-01 1.00540914e-01 4.51428622e-01 -3.53216261e-01
-1.10835218e+00 1.97684407e-01 -1.80080943e-02 -4.96506393e-01
-5.94203472e-01 -3.27779651e-02 3.55268121e-01 1.52300283e-01
5.90073287e-01 1.75524071e-01 3.17680240e-02 -2.62787253e-01
7.34716177e-01 5.08613944e-01 5.16021848e-01 -3.98085445e-01
7.98629165e-01 9.93204057e-01 6.95334300e-02 -8.08383524e-01
-4.07091618e-01 -3.01914454e-01 -7.78347254e-01 -5.16983688e-01
1.00898898e+00 -1.04243279e+00 -8.32948625e-01 4.24388386e-02
-9.80512857e-01 -2.92188793e-01 -4.69447440e-03 4.92570043e-01
-7.43490160e-01 1.01812050e-01 -7.69503042e-02 -1.10956490e+00
3.29560459e-01 -9.69338894e-01 1.17833686e+00 6.41940744e-04
-2.14320838e-01 -1.06476605e+00 -3.82006192e-03 2.63029128e-01
7.28685021e-01 4.31464046e-01 5.77861965e-01 -3.10585260e-01
-8.48687053e-01 -3.17071676e-02 -3.68313462e-01 -2.32836246e-01
4.71943952e-02 1.26984380e-02 -1.16153729e+00 -1.21926129e-01
-6.43805742e-01 -6.87988549e-02 1.24250519e+00 4.11977857e-01
1.19202769e+00 -1.92352772e-01 -8.37878287e-01 2.58080661e-01
1.11869419e+00 2.12950669e-02 7.55436301e-01 5.31275749e-01
6.14584088e-01 1.12863433e+00 7.08852053e-01 2.92183787e-01
9.51210618e-01 9.32072163e-01 4.72022355e-01 -1.26340419e-01
-3.60570312e-01 -2.27768689e-01 4.73135322e-01 -5.52270934e-03
-3.83912474e-01 -2.29257673e-01 -1.06878138e+00 8.77726436e-01
-2.05414581e+00 -1.43451810e+00 -5.90943694e-01 2.06752753e+00
2.20294535e-01 -5.61385602e-03 5.42029858e-01 -2.39583880e-01
7.20006943e-01 3.59213740e-01 -5.11719465e-01 -3.86650860e-01
-6.29852060e-03 -9.65500548e-02 8.12310517e-01 3.86958212e-01
-1.22550142e+00 8.81347239e-01 7.19075775e+00 4.70348120e-01
-1.15332127e+00 2.11163357e-01 7.42984116e-01 -2.28827670e-01
-5.28071702e-01 2.25595497e-02 -8.70280147e-01 4.69722092e-01
1.27318811e+00 6.25992566e-02 4.42698091e-01 7.60755122e-01
2.66775578e-01 3.17141367e-03 -1.08291399e+00 7.83030033e-01
2.23964583e-02 -1.42878747e+00 -1.37976572e-01 1.63688406e-01
6.66692853e-01 2.49850884e-01 3.00307900e-01 4.55219716e-01
5.13607383e-01 -1.19120991e+00 1.14173901e+00 9.86623108e-01
5.45289993e-01 -5.52437305e-01 3.06080669e-01 2.57358372e-01
-1.35424685e+00 -2.46184736e-01 -3.08462232e-01 -1.70889214e-01
2.85400718e-01 2.54226744e-01 -5.67265689e-01 4.85280454e-01
8.79297435e-01 1.03592098e+00 -7.86335886e-01 7.41673946e-01
1.83892995e-01 3.82743508e-01 -7.91778117e-02 6.63019111e-03
4.40180361e-01 -1.62997916e-01 2.24243417e-01 1.48344183e+00
3.98755640e-01 -1.78599298e-01 1.01102240e-01 8.06491792e-01
3.20625871e-01 -1.86091393e-01 -1.13178861e+00 1.45408824e-01
1.64271578e-01 1.11099470e+00 -4.75888371e-01 -5.20314455e-01
-9.49993134e-01 6.85526192e-01 3.06931555e-01 5.80051422e-01
-1.02119243e+00 7.41692707e-02 1.25824404e+00 4.80138034e-01
4.24666017e-01 -3.81558746e-01 -2.61936963e-01 -8.89206946e-01
-1.19671218e-01 -4.69745755e-01 3.04772228e-01 -7.62269974e-01
-1.24963856e+00 3.97736788e-01 8.09760690e-02 -1.44917572e+00
-3.91756296e-01 -7.82716095e-01 -5.48923194e-01 8.33373904e-01
-1.93065619e+00 -1.46053350e+00 -2.12732673e-01 5.66897571e-01
5.41508496e-01 -8.30294192e-02 5.06164193e-01 3.80350828e-01
-2.09289685e-01 1.16010755e-01 -2.72466689e-02 5.01944236e-02
4.92244095e-01 -9.27485466e-01 8.90293956e-01 9.02302444e-01
2.42162451e-01 3.76005501e-01 7.85304904e-01 -6.15014076e-01
-1.47384894e+00 -1.36962736e+00 9.80179965e-01 -7.24914074e-01
7.83198953e-01 -4.10043180e-01 -7.53626525e-01 1.09245276e+00
1.59133941e-01 2.59666413e-01 3.20762873e-01 2.25169629e-01
-2.70878464e-01 -1.93286508e-01 -8.27711344e-01 7.57727325e-01
1.25600255e+00 -4.54032719e-01 -4.49784100e-01 1.76083148e-01
2.33761519e-01 -2.36917749e-01 -4.25019681e-01 1.86561078e-01
8.19486380e-01 -1.21610475e+00 1.23357511e+00 -8.23793769e-01
1.70880824e-01 -4.89442408e-01 -2.18888283e-01 -1.04160500e+00
-5.51249325e-01 -1.97779819e-01 -7.36994494e-04 9.47169483e-01
4.85105902e-01 -6.75610542e-01 8.87751937e-01 7.59242237e-01
-1.34219363e-01 -4.99456882e-01 -8.97175312e-01 -7.97223747e-01
3.36070031e-01 -8.26741755e-01 8.37894738e-01 6.72309041e-01
-3.62845749e-01 -3.02679569e-01 -6.43253446e-01 2.88640380e-01
5.11083007e-01 1.91296607e-01 1.01187694e+00 -1.38285959e+00
1.22715376e-01 -5.63685775e-01 -7.46733665e-01 -1.13859594e+00
3.84286880e-01 -8.95941436e-01 -9.65004880e-03 -1.68384004e+00
7.76375681e-02 -7.83667743e-01 -3.25683892e-01 3.99254888e-01
4.42339629e-02 4.56732213e-01 2.51617908e-01 1.42393291e-01
-5.53186953e-01 3.96994919e-01 9.68051851e-01 -2.66011566e-01
1.71765536e-01 9.05811712e-02 -6.64624333e-01 5.96460998e-01
7.31898487e-01 -4.00349379e-01 -3.79164636e-01 -5.24504900e-01
1.15863755e-01 -7.18993619e-02 1.01675498e+00 -1.15766060e+00
2.72136509e-01 -3.95420849e-01 6.02355897e-01 -7.72979856e-01
6.16026163e-01 -7.73742437e-01 3.66892606e-01 3.99732292e-01
-2.86596239e-01 3.18905890e-01 2.56786376e-01 8.32817554e-01
-4.36305404e-02 1.75033480e-01 4.26323891e-01 -2.19325140e-01
-1.28546500e+00 4.86399293e-01 -6.26846254e-01 -6.49511963e-02
1.28579688e+00 -5.02425194e-01 -3.99756551e-01 -5.24844944e-01
-7.19213605e-01 2.68192023e-01 5.19057453e-01 9.78952467e-01
5.26304543e-01 -1.35909581e+00 -8.11437547e-01 3.04528981e-01
2.47105792e-01 -6.93611264e-01 1.86002567e-01 8.09846997e-01
-2.28696525e-01 6.02249801e-01 -3.89172345e-01 -6.84322715e-01
-1.01026642e+00 7.65568316e-01 2.50217825e-01 1.48986250e-01
-8.29670310e-01 3.59423578e-01 3.81221980e-01 -4.08864737e-01
-9.51892436e-02 -2.06053257e-01 -3.49208862e-01 2.38823742e-01
5.08647025e-01 4.73119140e-01 -2.17255894e-02 -1.13509345e+00
-5.56677818e-01 6.15293562e-01 1.04143448e-01 -1.42698199e-01
1.30470002e+00 -3.04762363e-01 4.53900427e-01 4.84682769e-01
9.02161300e-01 -1.75783619e-01 -1.63416886e+00 7.02247322e-02
4.66866083e-02 -6.12670958e-01 -8.10006931e-02 -4.89778519e-01
-1.03891289e+00 1.18380272e+00 8.54161441e-01 2.46854231e-01
8.58901799e-01 6.02571480e-02 6.55727983e-01 1.38654023e-01
5.39628088e-01 -9.51680839e-01 -3.18331689e-01 3.88406336e-01
7.00463831e-01 -1.43665314e+00 -1.09238818e-01 -1.41768411e-01
-9.13769245e-01 1.17240310e+00 4.09152955e-01 -5.48282340e-02
7.17462778e-01 4.67589684e-02 2.32629314e-01 -1.32968158e-01
-8.86209548e-01 -7.22613335e-01 3.25882226e-01 8.63797128e-01
4.77243543e-01 2.08089530e-01 1.88556001e-01 -1.92938056e-02
-1.63252547e-01 4.23746556e-02 4.14392650e-01 7.35686481e-01
-3.77460122e-01 -9.90659475e-01 -4.39427793e-01 4.40164149e-01
-1.08208813e-01 9.85040814e-02 -1.94967128e-02 9.66924667e-01
2.73898482e-01 1.12145805e+00 4.94222313e-01 -2.85903633e-01
3.94640654e-01 2.96985637e-02 3.26205403e-01 -2.56189704e-01
-3.65085930e-01 -5.46848655e-01 2.52051145e-01 -6.61167085e-01
-7.33321965e-01 -1.05155849e+00 -9.87534285e-01 -6.43289149e-01
2.80904055e-01 -3.44996601e-01 6.64746642e-01 8.78233910e-01
4.73493040e-01 3.51777285e-01 3.86146486e-01 -1.10616469e+00
1.85871858e-03 -5.47868371e-01 -3.28136265e-01 6.61178589e-01
6.12917542e-01 -9.93923128e-01 2.09085848e-02 1.83465809e-01] | [6.266190528869629, 0.6481020450592041] |
9182651b-7430-4de8-9063-b84cdddaa78e | dissect-disentangled-simultaneous | 2105.15164 | null | https://arxiv.org/abs/2105.15164v4 | https://arxiv.org/pdf/2105.15164v4.pdf | DISSECT: Disentangled Simultaneous Explanations via Concept Traversals | Explaining deep learning model inferences is a promising venue for scientific understanding, improving safety, uncovering hidden biases, evaluating fairness, and beyond, as argued by many scholars. One of the principal benefits of counterfactual explanations is allowing users to explore "what-if" scenarios through what does not and cannot exist in the data, a quality that many other forms of explanation such as heatmaps and influence functions are inherently incapable of doing. However, most previous work on generative explainability cannot disentangle important concepts effectively, produces unrealistic examples, or fails to retain relevant information. We propose a novel approach, DISSECT, that jointly trains a generator, a discriminator, and a concept disentangler to overcome such challenges using little supervision. DISSECT generates Concept Traversals (CTs), defined as a sequence of generated examples with increasing degrees of concepts that influence a classifier's decision. By training a generative model from a classifier's signal, DISSECT offers a way to discover a classifier's inherent "notion" of distinct concepts automatically rather than rely on user-predefined concepts. We show that DISSECT produces CTs that (1) disentangle several concepts, (2) are influential to a classifier's decision and are coupled to its reasoning due to joint training (3), are realistic, (4) preserve relevant information, and (5) are stable across similar inputs. We validate DISSECT on several challenging synthetic and realistic datasets where previous methods fall short of satisfying desirable criteria for interpretability and show that it performs consistently well and better than existing methods. Finally, we present experiments showing applications of DISSECT for detecting potential biases of a classifier and identifying spurious artifacts that impact predictions. | ['Rosalind W. Picard', 'Brian Eoff', 'Brendan Jou', 'Chun-Liang Li', 'Been Kim', 'Asma Ghandeharioun'] | 2021-05-31 | dissect-disentangled-simultaneous-1 | https://openreview.net/forum?id=qY79G8jGsep | https://openreview.net/pdf?id=qY79G8jGsep | iclr-2022-4 | ['interpretability-techniques-for-deep-learning'] | ['miscellaneous'] | [ 3.83671045e-01 4.50029671e-01 -5.31160295e-01 -5.14768600e-01
-5.85668266e-01 -7.52789259e-01 9.82553959e-01 2.64587283e-01
-6.17225282e-02 1.02736712e+00 4.45001692e-01 -9.21072721e-01
-2.56728262e-01 -8.12971950e-01 -7.45908976e-01 -5.40640056e-01
8.53239670e-02 5.12074471e-01 -2.28542596e-01 1.78918213e-01
4.67456579e-01 2.31131613e-01 -1.58826983e+00 4.69815880e-01
1.13921201e+00 3.12516093e-01 -4.00437027e-01 3.36485356e-01
3.67131643e-02 8.78019750e-01 -7.19948471e-01 -6.47062302e-01
2.03803241e-01 -8.97849143e-01 -6.80351436e-01 -2.17648372e-01
3.65958691e-01 -7.84823522e-02 2.28604272e-01 1.02311051e+00
1.33171678e-01 -1.25471175e-01 9.06383157e-01 -1.58581614e+00
-8.25149417e-01 1.07573211e+00 -4.30634290e-01 7.23498091e-02
-2.75673587e-02 5.30685663e-01 1.24150455e+00 -4.85146910e-01
4.78225589e-01 1.50461304e+00 6.31833494e-01 8.93616319e-01
-1.79804158e+00 -1.12888646e+00 3.16574663e-01 -7.84916505e-02
-1.00482738e+00 -3.94482344e-01 7.04736471e-01 -6.36659741e-01
4.80913401e-01 7.74299383e-01 6.74127221e-01 1.62423539e+00
3.19469333e-01 5.96211553e-01 1.42843759e+00 -1.90779820e-01
7.13921070e-01 3.73807788e-01 2.81823337e-01 5.24143696e-01
7.30770350e-01 4.88437057e-01 -4.62872028e-01 -4.98319983e-01
3.22257549e-01 4.79733795e-02 -3.35153133e-01 -3.37943643e-01
-1.11217940e+00 1.26504314e+00 4.35142934e-01 7.04213232e-03
-1.72770336e-01 2.03607321e-01 -2.54217051e-02 1.43418580e-01
5.12633681e-01 1.06755137e+00 -3.59557033e-01 5.27857840e-02
-9.01913524e-01 5.58520317e-01 8.71644557e-01 6.45348370e-01
5.78557193e-01 5.45016117e-02 -2.72537798e-01 2.40274459e-01
7.54350200e-02 3.41133922e-01 5.06436110e-01 -8.46035898e-01
1.32291675e-01 7.54583418e-01 3.55684422e-02 -8.35116506e-01
-1.56703919e-01 -7.28557408e-01 -7.39264846e-01 3.53732228e-01
4.17883694e-01 -1.72628880e-01 -9.30338681e-01 2.14127183e+00
-4.22378965e-02 -1.13183059e-01 5.65309599e-02 9.22626615e-01
4.40154761e-01 2.83162355e-01 3.86426419e-01 -1.09664224e-01
1.23614979e+00 -2.95709521e-01 -3.90070140e-01 -4.82783258e-01
5.78420281e-01 -2.74842739e-01 1.19986904e+00 1.96204022e-01
-8.17963779e-01 -3.57078850e-01 -1.15257847e+00 1.19357742e-01
-3.19857508e-01 -3.72029245e-01 1.17880070e+00 9.34760392e-01
-7.82909572e-01 6.06849968e-01 -5.73108613e-01 4.93296757e-02
8.98623049e-01 2.18706861e-01 -9.50202271e-02 5.54805100e-02
-1.22868979e+00 7.92451262e-01 3.25691521e-01 -3.24326932e-01
-1.14358294e+00 -1.01357317e+00 -6.31997943e-01 5.37379265e-01
4.51899976e-01 -1.13974237e+00 9.20004010e-01 -1.20788908e+00
-8.18250835e-01 5.45976818e-01 -2.91868031e-01 -5.67299426e-01
6.54433727e-01 -6.66325446e-03 -3.55165422e-01 -4.40276563e-01
4.37863737e-01 8.48613024e-01 5.57233214e-01 -1.45651650e+00
-4.95978236e-01 -4.11701381e-01 1.64045930e-01 2.63882130e-02
1.64290033e-02 -3.39121312e-01 5.22653937e-01 -5.01572311e-01
2.10471656e-02 -8.09834301e-01 -3.42158437e-01 -1.36215582e-01
-8.85678709e-01 3.54888774e-02 7.43131280e-01 -1.10434845e-01
9.37435985e-01 -1.77049077e+00 -2.76490629e-01 2.96547472e-01
7.11152673e-01 6.16015345e-02 1.30455479e-01 1.73180014e-01
-4.17957366e-01 9.34261441e-01 -4.08073962e-01 -2.42141709e-02
6.68798387e-02 1.62650377e-01 -6.20296597e-01 2.33165026e-01
3.20207953e-01 9.75602984e-01 -1.08508003e+00 -1.27482712e-01
-2.81822532e-02 1.52099594e-01 -8.41866791e-01 -9.63744894e-02
-3.02757651e-01 2.94051647e-01 -2.98917383e-01 2.97340512e-01
4.58353758e-01 -2.62065262e-01 3.59784186e-01 2.21269965e-01
4.68737893e-02 7.12315261e-01 -9.42520440e-01 8.55930805e-01
-1.93660989e-01 8.13436627e-01 -5.44696212e-01 -9.68111873e-01
8.65937412e-01 7.30970800e-02 -6.55151159e-02 -2.64876962e-01
-1.42266080e-02 1.69679746e-01 3.47302884e-01 -2.48850271e-01
2.30793759e-01 -6.94373608e-01 4.52919938e-02 9.08189058e-01
-1.74515426e-01 4.56886664e-02 -1.80069417e-01 3.15826237e-01
1.07752430e+00 -2.38351405e-01 5.40563226e-01 -5.57655573e-01
-4.34855418e-03 -8.23759101e-03 7.34209895e-01 1.13704634e+00
-1.71036310e-02 4.86360580e-01 1.04669142e+00 -5.50616026e-01
-9.20438111e-01 -1.29216003e+00 -5.67725115e-02 6.90006852e-01
-1.06290197e-02 -1.12330787e-01 -5.25896430e-01 -9.98076975e-01
3.53346050e-01 1.59754336e+00 -1.11401856e+00 -4.71796662e-01
-2.34275255e-02 -1.03566849e+00 5.55803478e-01 5.11360288e-01
2.91617960e-01 -6.41583681e-01 -6.96239233e-01 -1.06624678e-01
-3.23924839e-01 -5.26259720e-01 -2.42646426e-01 3.08391243e-01
-9.94424164e-01 -1.28110993e+00 4.07889821e-02 -2.37260927e-02
9.52145815e-01 3.01217079e-01 1.28413856e+00 6.40962645e-02
-7.21127465e-02 -2.98641622e-01 1.15302444e-01 -8.25488329e-01
-7.87790954e-01 -8.42435360e-02 -7.51978979e-02 -2.02445626e-01
7.50368834e-01 -8.18276167e-01 -4.78565753e-01 2.74850428e-01
-7.88250148e-01 3.80921811e-01 6.66771412e-01 9.86137748e-01
1.73531398e-01 -7.62849674e-02 8.84173930e-01 -1.56621361e+00
8.82166088e-01 -7.26389349e-01 -4.95343864e-01 1.73706979e-01
-1.28614342e+00 5.27188063e-01 5.11750340e-01 -4.88908708e-01
-1.07123494e+00 -2.67823547e-01 3.66598666e-01 -2.17619643e-01
-2.35299975e-01 3.12469482e-01 -3.29972863e-01 5.39281905e-01
1.28860378e+00 5.29796164e-03 3.41166146e-02 -1.49752617e-01
5.39016843e-01 4.70065862e-01 3.12160730e-01 -6.11002564e-01
8.47916663e-01 5.31302869e-01 -7.56115317e-02 -1.42381102e-01
-1.02639484e+00 -5.57467118e-02 -3.47080469e-01 6.29432052e-02
6.20369315e-01 -7.70817816e-01 -7.87608802e-01 -3.72839272e-01
-1.00913262e+00 -7.62885436e-02 -4.11688060e-01 4.62676883e-01
-2.46952444e-01 -1.98917702e-01 -9.48877782e-02 -9.48608756e-01
1.20215155e-02 -9.98326480e-01 6.30867362e-01 2.37313434e-01
-8.24494541e-01 -1.15835071e+00 -1.40060365e-01 3.15504104e-01
3.71264100e-01 6.10968530e-01 1.37028360e+00 -1.07002640e+00
-7.32125342e-01 -1.27089441e-01 -1.16367184e-01 6.50300533e-02
2.40081012e-01 -2.55348533e-02 -1.31025326e+00 -6.57772645e-02
-3.24958153e-02 -1.40592083e-01 1.07175219e+00 2.93652654e-01
1.24253809e+00 -8.86857331e-01 -5.92357099e-01 3.38221312e-01
1.21499348e+00 2.56996483e-01 6.02190375e-01 1.86900962e-02
4.74146873e-01 8.01334441e-01 -9.76365060e-02 1.62080467e-01
2.94094533e-01 2.44616643e-01 3.73599142e-01 -2.50603169e-01
1.66896120e-01 -7.60978878e-01 2.04979628e-01 1.17962724e-02
-4.57954146e-02 -1.80584177e-01 -7.54471481e-01 5.91885984e-01
-1.67554498e+00 -1.37384582e+00 -3.02015901e-01 2.31181741e+00
8.92087042e-01 4.42746729e-01 2.46281363e-02 1.23101391e-01
6.19196534e-01 -5.60965687e-02 -7.95941174e-01 -6.95408165e-01
-1.27274960e-01 2.38349259e-01 3.29182595e-01 6.64921105e-01
-6.83885276e-01 6.27450466e-01 6.35213232e+00 4.09458607e-01
-8.71782899e-01 -2.40364373e-02 1.10491025e+00 -1.07979126e-01
-1.26878810e+00 5.97940683e-01 -5.08003175e-01 4.48769331e-01
7.38874435e-01 -5.20293951e-01 1.42976031e-01 9.65619981e-01
2.57005662e-01 -5.51585928e-02 -1.72036767e+00 2.90389925e-01
-4.88067754e-02 -1.51961553e+00 5.38012087e-01 3.18673909e-01
8.55491817e-01 -4.63064671e-01 2.01686546e-01 2.99559176e-01
9.21737969e-01 -1.51991665e+00 8.31744075e-01 3.04541141e-01
6.35086119e-01 -7.08464682e-01 6.77855790e-01 4.59317833e-01
-3.16103280e-01 -2.06517458e-01 -3.13933402e-01 -4.15254623e-01
-4.99308735e-01 8.61331105e-01 -1.29241490e+00 3.29053730e-01
1.40515268e-01 3.56862217e-01 -6.00867450e-01 6.12396657e-01
-7.39063740e-01 1.00499320e+00 1.57094091e-01 -2.81307429e-01
8.27124417e-02 2.39807442e-01 6.38346553e-01 1.19184899e+00
4.83612716e-02 1.60530403e-01 -2.68883586e-01 1.79131341e+00
-1.00821130e-01 -3.10662895e-01 -7.81846285e-01 -1.26495659e-01
6.94562733e-01 8.94312263e-01 -6.74386978e-01 -4.40142483e-01
-2.56871358e-02 4.43794221e-01 -2.02358644e-02 5.28057575e-01
-6.97607934e-01 -1.56430304e-01 1.03948975e+00 1.52328715e-01
-2.09183022e-01 2.98309147e-01 -7.98659801e-01 -1.17004192e+00
-1.51114941e-01 -1.03388536e+00 4.35377270e-01 -7.29828119e-01
-1.28865886e+00 2.47603685e-01 -7.55371377e-02 -1.06572378e+00
-4.86505508e-01 -3.50827366e-01 -8.50294232e-01 1.11778200e+00
-1.11373746e+00 -9.82045174e-01 -1.67222142e-01 1.72716260e-01
4.35159028e-01 3.34916003e-02 9.47520494e-01 -4.12179083e-01
-2.81719059e-01 6.16413414e-01 -2.71191031e-01 5.62997796e-02
4.35048878e-01 -1.52258492e+00 5.38706660e-01 9.17234421e-01
3.49501193e-01 1.10681784e+00 9.32424307e-01 -6.68209016e-01
-7.97640681e-01 -1.03428984e+00 1.02347827e+00 -9.30800915e-01
3.74696374e-01 -6.84080362e-01 -9.76213157e-01 9.50348973e-01
-8.37835670e-02 -4.41424578e-01 1.05210662e+00 7.40800381e-01
-7.95120358e-01 9.66958329e-03 -1.08880115e+00 8.16472232e-01
1.06759548e+00 -3.99554193e-01 -7.97989488e-01 1.24591202e-01
6.24450088e-01 -1.53032124e-01 -2.17323810e-01 1.90709978e-01
7.02382088e-01 -1.21104801e+00 7.07299113e-01 -1.21058857e+00
8.92982662e-01 -1.43210948e-01 4.47156020e-02 -1.61384058e+00
-4.78566557e-01 -6.16460860e-01 2.14042351e-01 1.07211959e+00
9.87761378e-01 -8.98199677e-01 6.83122277e-01 9.90978420e-01
6.21781573e-02 -6.05317831e-01 -5.06724000e-01 -6.07070208e-01
2.67244488e-01 -5.57731628e-01 1.03375614e+00 1.36243784e+00
2.08135992e-01 6.03595853e-01 -1.25201821e-01 1.44854277e-01
7.94547379e-01 4.85826731e-01 7.83337831e-01 -1.47526407e+00
-3.42538357e-01 -7.07193136e-01 -3.74072790e-01 -5.09846389e-01
1.19998589e-01 -1.09875977e+00 -2.02819079e-01 -1.27514696e+00
7.79541016e-01 -5.04209757e-01 -1.92282274e-01 7.48163998e-01
-6.74753129e-01 -2.43986189e-01 1.32668018e-01 2.25351408e-01
-1.22979596e-01 2.63288379e-01 1.02538097e+00 -1.52905196e-01
-4.37282883e-02 -2.64947675e-02 -1.83891749e+00 6.95880532e-01
7.79444873e-01 -7.33250141e-01 -7.74456620e-01 -1.36374190e-01
3.20373148e-01 -1.39158547e-01 8.72529566e-01 -6.71810508e-01
-7.86261708e-02 -5.44353545e-01 7.95729339e-01 1.71886049e-02
-1.36312738e-01 -2.35090882e-01 3.81033808e-01 6.97494745e-01
-1.05465209e+00 -8.11203048e-02 1.45391852e-01 5.58361948e-01
1.64838895e-01 2.34706663e-02 6.19672239e-01 -2.41825074e-01
-2.07843065e-01 -3.53109017e-02 -3.64373475e-01 3.53826374e-01
7.62683153e-01 5.69463102e-03 -7.71792829e-01 -5.04170775e-01
-4.22428280e-01 1.83217436e-01 4.23461616e-01 5.16520560e-01
5.78229368e-01 -1.11517179e+00 -8.87643695e-01 2.50458509e-01
1.28330857e-01 -3.08170766e-01 -1.45014673e-01 3.64172161e-01
1.46034807e-01 4.41634387e-01 -1.36903480e-01 -4.58011121e-01
-9.45563257e-01 5.47804356e-01 3.84822011e-01 -1.34311885e-01
-4.29877192e-01 7.87183404e-01 8.53936553e-01 -2.77402192e-01
-2.17714563e-01 -3.23835820e-01 2.48960312e-02 -1.35836467e-01
4.89299685e-01 5.37156984e-02 -2.55818188e-01 -4.80791479e-02
-2.78634667e-01 -2.11136431e-01 -5.36677726e-02 -2.41507456e-01
1.12331462e+00 2.36586198e-01 1.26258194e-01 5.26281118e-01
7.97340691e-01 8.61251503e-02 -1.14817989e+00 1.69484720e-01
-2.90147588e-03 -7.77849078e-01 -1.44264877e-01 -1.27453756e+00
-7.40522683e-01 1.05373752e+00 2.51836300e-01 3.45834047e-01
7.46877968e-01 -4.46879119e-02 9.12017077e-02 3.90860587e-02
1.32440969e-01 -5.96647978e-01 -2.81039216e-02 -2.19526842e-01
8.62443209e-01 -1.14574301e+00 2.99893692e-02 -2.98296005e-01
-7.37246931e-01 8.49951327e-01 5.68676353e-01 2.12362453e-01
2.59704679e-01 1.58423305e-01 1.64306797e-02 -2.44047165e-01
-1.26413238e+00 -7.25907683e-02 3.34665328e-01 6.25135660e-01
5.06771803e-01 5.13868153e-01 -3.02135199e-01 8.33465815e-01
-6.68666780e-01 -1.11685231e-01 7.10230172e-01 3.75850171e-01
-2.58950025e-01 -7.15471148e-01 -2.68099785e-01 6.97297454e-01
-3.67695808e-01 -3.04077268e-01 -7.85436332e-01 9.96913075e-01
2.94399112e-01 9.59291041e-01 4.85006981e-02 -6.01401925e-01
7.29335174e-02 2.30566680e-01 3.45676541e-02 -7.21568048e-01
-3.55656654e-01 -3.83412927e-01 2.51641482e-01 -5.20095766e-01
-3.76790427e-02 -7.47107267e-01 -1.00238073e+00 -5.21431804e-01
-3.18267494e-01 4.24806446e-01 4.89498377e-01 1.11232841e+00
5.37211597e-01 4.09624785e-01 6.48797393e-01 1.73301983e-03
-7.35995173e-01 -7.34001517e-01 -2.43487224e-01 5.45469224e-01
3.87948513e-01 -6.64180636e-01 -7.64841199e-01 1.21446084e-02] | [8.724835395812988, 5.591401100158691] |
a9330ef3-f778-4cf6-9998-4a7dc448f09b | music-auto-tagging-using-cnns-and-mel | 1911.04824 | null | https://arxiv.org/abs/1911.04824v3 | https://arxiv.org/pdf/1911.04824v3.pdf | How Low Can You Go? Reducing Frequency and Time Resolution in Current CNN Architectures for Music Auto-tagging | Automatic tagging of music is an important research topic in Music Information Retrieval and audio analysis algorithms proposed for this task have achieved improvements with advances in deep learning. In particular, many state-of-the-art systems use Convolutional Neural Networks and operate on mel-spectrogram representations of the audio. In this paper, we compare commonly used mel-spectrogram representations and evaluate model performances that can be achieved by reducing the input size in terms of both lesser amount of frequency bands and larger frame rates. We use the MagnaTagaTune dataset for comprehensive performance comparisons and then compare selected configurations on the larger Million Song Dataset. The results of this study can serve researchers and practitioners in their trade-off decision between accuracy of the models, data storage size and training and inference times. | ['Dmitry Bogdanov', 'Jay Ho Jeon', 'Andres Ferraro', 'Xavier Serra', 'Jason Yoon'] | 2019-11-12 | null | null | null | null | ['music-auto-tagging'] | ['music'] | [-1.39952414e-02 -7.25228667e-01 -1.20618105e-01 -6.04786798e-02
-7.75305688e-01 -7.48359263e-01 7.75699764e-02 2.79728800e-01
-6.41516805e-01 4.55938339e-01 1.62441537e-01 5.91973215e-02
-4.36704129e-01 -6.06811762e-01 -2.61670172e-01 -3.72145057e-01
-3.62529427e-01 2.10396945e-01 1.52483806e-01 -7.51123354e-02
3.27091694e-01 2.94147968e-01 -1.91033661e+00 5.38792968e-01
1.58472478e-01 1.23801816e+00 6.63289875e-02 1.07874978e+00
3.58520783e-02 5.83288252e-01 -1.02919590e+00 -1.15993991e-01
2.74219722e-01 -2.11052760e-01 -8.15253377e-01 -5.39838493e-01
5.89953721e-01 -1.14568233e-01 -3.81697923e-01 8.39537024e-01
1.06748676e+00 3.89001876e-01 2.20058933e-01 -9.85877573e-01
-1.90081313e-01 1.13110256e+00 -2.50495344e-01 6.66426301e-01
2.57836014e-01 -5.22710197e-02 1.24186850e+00 -4.99272704e-01
1.24807626e-01 9.72133219e-01 1.03180265e+00 1.55166373e-01
-1.00166059e+00 -8.32999468e-01 -3.75303060e-01 5.34434974e-01
-1.57152355e+00 -5.15799522e-01 8.11956644e-01 -4.09086943e-01
1.28824270e+00 4.94078904e-01 7.18389809e-01 5.38832903e-01
-1.84947047e-02 4.86326247e-01 5.35652995e-01 -6.47510231e-01
1.28532156e-01 -2.69278884e-01 9.83545482e-02 2.95537233e-01
-8.44293758e-02 4.66999086e-03 -9.35558856e-01 -1.40777662e-01
8.19934011e-01 -3.76995593e-01 -1.27533108e-01 2.34178513e-01
-1.19154060e+00 7.37203121e-01 2.31671557e-01 5.55376291e-01
-2.54503548e-01 5.98031640e-01 1.03605223e+00 4.92499471e-01
4.00891691e-01 7.77258396e-01 -6.10754967e-01 -7.60286391e-01
-1.52360642e+00 4.38286185e-01 7.77811289e-01 5.29851735e-01
1.76528439e-01 4.69919652e-01 -2.78652877e-01 1.06742036e+00
-1.39874071e-01 3.22758518e-02 8.85670841e-01 -1.07299304e+00
3.51711273e-01 2.21523210e-01 -1.76362008e-01 -8.77953231e-01
-5.55343032e-01 -6.79456651e-01 -7.13490129e-01 -7.45429471e-02
6.19155765e-01 -1.15736462e-01 -6.91666722e-01 1.52080059e+00
2.75815316e-02 4.52553034e-01 -2.37590864e-01 1.06579781e+00
7.86241353e-01 5.53592801e-01 -6.74050152e-02 8.92984048e-02
1.49783671e+00 -7.86794841e-01 -6.69760227e-01 1.55414432e-01
4.11804318e-01 -1.30167747e+00 1.12120044e+00 7.55160034e-01
-1.21073878e+00 -1.10240948e+00 -1.20610845e+00 -2.29177639e-01
-3.56817245e-01 6.14755273e-01 8.59720349e-01 7.23055840e-01
-8.66959095e-01 1.18169475e+00 -7.36866534e-01 -1.66341007e-01
2.99717426e-01 6.81522310e-01 2.19154105e-01 6.22158527e-01
-1.29172289e+00 2.66721904e-01 6.92328095e-01 -1.14676012e-02
-6.52666330e-01 -9.15602565e-01 -2.49714762e-01 5.24299920e-01
7.20749050e-02 -3.68930250e-01 1.65032196e+00 -8.62367988e-01
-1.59259689e+00 6.64072812e-01 4.05552477e-01 -8.64562392e-01
8.22498798e-02 -5.33385217e-01 -6.03705347e-01 3.21701467e-02
-3.48881423e-01 6.45310998e-01 6.73011839e-01 -1.73790500e-01
-8.61799777e-01 -7.88212419e-02 2.60139436e-01 -1.52164185e-02
-4.74490941e-01 2.35704735e-01 -2.02899948e-01 -1.11290920e+00
-2.40667671e-01 -1.13126612e+00 1.10802889e-01 -3.31859618e-01
-1.94331080e-01 -2.48765215e-01 6.17459178e-01 -5.79794228e-01
1.63893211e+00 -2.32655120e+00 3.07276379e-02 -1.22137479e-01
-2.64276952e-01 5.14953077e-01 -7.82903060e-02 3.18850368e-01
-1.71800479e-01 -1.41424071e-02 3.22607428e-01 -5.16253375e-02
7.14431852e-02 -1.41464382e-01 -4.95586187e-01 2.52964079e-01
-1.67735606e-01 5.19038081e-01 -5.03318071e-01 -3.79805535e-01
2.92917550e-01 5.67843139e-01 -7.70817697e-01 1.37738019e-01
-1.13899574e-01 1.71138510e-01 -4.69681993e-02 4.50798661e-01
1.42399728e-01 6.99358806e-02 1.27971023e-01 -3.82635772e-01
-1.84391692e-01 9.66365516e-01 -1.34870720e+00 2.23408866e+00
-5.62580943e-01 1.01969659e+00 -2.15755075e-01 -8.42248142e-01
8.02524090e-01 5.88886440e-01 7.00762570e-01 -4.13985014e-01
3.56407702e-01 2.43662208e-01 3.87530923e-01 -2.13319466e-01
8.32663238e-01 -4.75486778e-02 -4.61848453e-03 5.54997623e-01
2.96467781e-01 1.51191249e-01 3.51742297e-01 -3.54099602e-01
7.88513184e-01 8.89512971e-02 7.14611262e-02 -2.70975411e-01
2.80080646e-01 -1.92378551e-01 3.29303414e-01 6.66523278e-01
-2.67613064e-02 4.93847668e-01 6.90524951e-02 -7.38298714e-01
-1.13002372e+00 -6.63399935e-01 -2.09869012e-01 1.56174672e+00
-3.64945531e-01 -9.14597631e-01 -7.64553368e-01 -6.28436208e-02
-7.69258616e-03 2.73583144e-01 -3.22213292e-01 -1.19454928e-01
-6.89434528e-01 -5.89926422e-01 1.23237324e+00 6.49271190e-01
2.56727636e-01 -1.38843167e+00 -1.14514828e+00 4.72734720e-01
-7.63019838e-04 -6.92945719e-01 -3.55586499e-01 5.05744517e-01
-9.61154461e-01 -8.95971060e-01 -6.39398634e-01 -7.78737545e-01
-4.48482096e-01 2.32743323e-02 1.33241701e+00 4.85754646e-02
-6.40020609e-01 7.16947839e-02 -5.70483088e-01 -7.60261774e-01
-1.06990054e-01 7.72854030e-01 3.02656025e-01 -2.75629789e-01
5.46929538e-01 -8.22228789e-01 -6.90947771e-01 6.28835261e-02
-8.98460746e-01 -2.69271642e-01 1.91922173e-01 5.13368845e-01
6.07511997e-01 2.18179643e-01 6.37756884e-01 -5.50432742e-01
8.10536265e-01 -1.30045667e-01 -5.78377485e-01 -9.65730771e-02
-5.38512945e-01 7.53285587e-02 5.70040822e-01 -7.89034426e-01
-3.49410504e-01 5.90834692e-02 -2.67414510e-01 -5.98896921e-01
-1.12317339e-01 5.84364474e-01 4.00801420e-01 1.85311228e-01
7.22159505e-01 -9.82366279e-02 -4.80075628e-01 -9.64082360e-01
2.50786155e-01 1.02115786e+00 5.68867326e-01 -5.73738754e-01
1.92011684e-01 1.32004812e-01 -1.36361599e-01 -7.31962621e-01
-9.21486139e-01 -5.53478777e-01 -6.12164319e-01 -3.90769929e-01
5.25268078e-01 -8.34749162e-01 -9.33152735e-01 3.46058249e-01
-8.25230956e-01 -1.22444164e-02 -5.54111362e-01 7.29386628e-01
-6.28303826e-01 1.89843308e-02 -6.99092925e-01 -8.83045614e-01
-7.42230296e-01 -8.85881364e-01 1.10257649e+00 2.26056054e-01
-6.19097650e-01 -5.80694854e-01 3.05179536e-01 -1.59903392e-02
5.84267795e-01 -1.38324067e-01 8.96973252e-01 -6.05819583e-01
-1.12917304e-01 -1.82875901e-01 1.03130169e-01 3.69392961e-01
-3.73453647e-02 -9.01716799e-02 -1.50150347e+00 -4.04419452e-01
-3.89162898e-01 -4.96284783e-01 9.68463361e-01 7.72070527e-01
1.64970648e+00 -4.67801616e-02 2.21123412e-01 7.06559241e-01
1.20780635e+00 2.77739018e-01 5.61952293e-01 4.13581491e-01
4.27032530e-01 3.01409006e-01 6.10000014e-01 6.62378907e-01
-3.62503022e-01 9.92881119e-01 1.31632671e-01 2.31316566e-01
-4.84365851e-01 -1.95837498e-01 1.55082181e-01 1.17658174e+00
-2.87479311e-01 -5.53007685e-02 -7.62834013e-01 6.46174073e-01
-1.85583973e+00 -1.04421508e+00 1.84219658e-01 2.28996515e+00
9.47117746e-01 1.45415440e-01 5.40646136e-01 1.03165507e+00
6.47023380e-01 1.16337672e-01 -3.51976961e-01 -5.09456575e-01
1.03360824e-01 7.43178785e-01 1.85670748e-01 2.99169421e-02
-1.32802069e+00 7.56572068e-01 7.17924786e+00 1.17614591e+00
-1.49408758e+00 1.20773464e-01 2.45859742e-01 -9.37282264e-01
4.86667275e-01 -2.42664337e-01 -6.05986595e-01 4.09680754e-01
1.62137747e+00 1.97487581e-03 7.27042198e-01 8.51292789e-01
4.85037155e-02 1.55880228e-01 -1.22781670e+00 1.46987379e+00
-1.91363439e-01 -1.55407345e+00 -1.07413210e-01 3.91350500e-03
3.83742243e-01 1.08359292e-01 7.25472122e-02 3.77237171e-01
-2.45948911e-01 -9.97046471e-01 9.92677450e-01 5.13389945e-01
1.01551557e+00 -1.05791521e+00 8.05459201e-01 -1.45530522e-01
-1.52149951e+00 -2.48794064e-01 -4.42813516e-01 -4.82483000e-01
-1.95513945e-02 4.54647034e-01 -7.64091909e-01 3.26151729e-01
1.15969491e+00 5.74120045e-01 -5.57871878e-01 1.34774232e+00
3.17888618e-01 1.09643853e+00 -4.76084173e-01 -8.16259906e-02
-6.08385950e-02 6.44994229e-02 3.86560172e-01 1.51955628e+00
3.81588310e-01 -3.29695612e-01 3.53442170e-02 4.89772975e-01
-1.90955520e-01 2.19887376e-01 -6.70056492e-02 -4.02077377e-01
5.63883960e-01 1.16663885e+00 -6.80564344e-01 -3.04168403e-01
-3.25827040e-02 4.80363071e-01 2.15015382e-01 -1.09093055e-01
-7.95496464e-01 -6.42261147e-01 9.00700212e-01 1.89384088e-01
3.88166904e-01 -1.77873611e-01 -2.15556607e-01 -8.84076834e-01
-3.14130038e-01 -8.65426064e-01 5.31152964e-01 -7.58560300e-01
-9.32645023e-01 6.25957668e-01 -2.57044882e-01 -1.42668867e+00
-4.09432769e-01 -5.27829051e-01 -2.96117485e-01 6.47775292e-01
-1.24038124e+00 -7.74653256e-01 1.50849754e-02 4.61817384e-01
6.84530020e-01 -5.14438570e-01 1.23004651e+00 8.45755696e-01
-2.98972487e-01 8.16069186e-01 7.88229406e-02 2.22551033e-01
7.39503443e-01 -1.15035927e+00 2.88211793e-01 2.98599422e-01
1.09243917e+00 4.62486863e-01 5.45946062e-01 -1.03985801e-01
-1.10415399e+00 -8.69299114e-01 6.57549858e-01 -1.07315689e-01
5.86191297e-01 -2.39531875e-01 -5.61159074e-01 2.91492969e-01
2.48974904e-01 -3.19426805e-01 1.05420434e+00 6.72029138e-01
-3.53715628e-01 -3.68432432e-01 -6.03872895e-01 1.38974890e-01
6.95898116e-01 -7.92400837e-01 -4.46055263e-01 -4.38188612e-02
4.47412014e-01 -4.61000443e-01 -1.13310552e+00 2.25589290e-01
1.10462010e+00 -7.81858563e-01 9.89084482e-01 -7.08101511e-01
2.04646111e-01 -2.05744162e-01 -2.27805525e-01 -1.06669223e+00
-4.82852221e-01 -6.52844012e-01 -2.35569671e-01 1.23563743e+00
2.44942456e-01 2.61276752e-01 7.50556886e-01 -1.74343407e-01
-1.26923144e-01 -6.45033777e-01 -1.12149537e+00 -6.80647552e-01
-1.44999668e-01 -8.17060590e-01 6.32257342e-01 7.36448765e-01
7.30861276e-02 4.60746020e-01 -5.18805921e-01 -3.40193927e-01
6.33351505e-02 4.29073989e-01 7.34452546e-01 -1.58449471e+00
-5.91010094e-01 -5.31269670e-01 -6.40560210e-01 -5.73141158e-01
-2.21322581e-01 -7.91372657e-01 -1.99851885e-01 -9.41321194e-01
-4.01260108e-02 -3.14672470e-01 -7.90949821e-01 6.48881257e-01
2.73686945e-01 8.25349748e-01 4.38777298e-01 2.90711910e-01
-5.26151896e-01 9.16346237e-02 7.26769567e-01 -1.66897207e-01
-3.64958972e-01 2.52413098e-02 -2.01826990e-01 7.99470484e-01
1.02367806e+00 -5.60536444e-01 -2.28645027e-01 -4.84841138e-01
5.35883784e-01 -1.26328841e-01 1.60597280e-01 -1.70014513e+00
4.65264060e-02 2.86284178e-01 4.71636653e-01 -6.92111313e-01
6.17119312e-01 -6.17970407e-01 3.69132221e-01 2.73344964e-01
-6.27312958e-01 1.04171678e-01 5.65688968e-01 2.91387647e-01
-4.09994841e-01 -3.32789809e-01 7.24883914e-01 5.17387092e-02
-4.56502348e-01 -8.21815878e-02 -2.29488686e-01 -1.28263935e-01
4.13587809e-01 -1.16393529e-01 5.81973717e-02 -4.01121140e-01
-8.07150543e-01 -5.32205999e-01 -6.92092925e-02 6.53455794e-01
1.24977410e-01 -1.46353352e+00 -6.06816649e-01 6.57573342e-02
-1.20635614e-01 -5.37761569e-01 2.30859742e-01 5.66436827e-01
-7.65932858e-01 7.31618941e-01 -4.55403298e-01 -4.74518508e-01
-1.58622575e+00 3.18745762e-01 2.76261002e-01 -2.55409747e-01
-4.37523037e-01 9.54701662e-01 -4.89380568e-01 8.57441500e-03
6.84825778e-01 -5.87001503e-01 -2.72966474e-01 3.61265957e-01
6.77272201e-01 4.93168354e-01 3.58119041e-01 -4.09363568e-01
-2.83968121e-01 4.87642556e-01 9.64318439e-02 -6.98124170e-02
1.40681016e+00 2.72390991e-01 1.83433428e-01 8.32783937e-01
9.09752727e-01 5.35773598e-02 -8.15023005e-01 -2.75487844e-02
5.06952517e-02 -4.60208535e-01 6.03443980e-01 -8.96617472e-01
-1.31163013e+00 1.10471106e+00 1.17041934e+00 6.40244007e-01
1.33213496e+00 -3.03507209e-01 9.71079946e-01 2.83165187e-01
2.66901016e-01 -1.36196756e+00 -1.68291599e-01 4.63651657e-01
5.95506012e-01 -5.86668551e-01 1.15442201e-01 8.11497644e-02
-1.46665365e-01 1.35197031e+00 2.69874424e-01 -4.57582146e-01
7.51276612e-01 3.73202503e-01 1.30413368e-01 -1.51545197e-01
-6.54986918e-01 -2.56269544e-01 6.00781679e-01 1.88283339e-01
1.11515892e+00 6.33825958e-02 -3.96568507e-01 8.87183309e-01
-9.29957032e-01 1.44934669e-01 1.15040764e-01 7.40032971e-01
-4.11835283e-01 -1.22349572e+00 -4.62765008e-01 5.59608757e-01
-9.66818631e-01 -2.33368784e-01 -4.41031188e-01 5.52486360e-01
4.98138279e-01 9.59079981e-01 3.47192198e-01 -7.03049481e-01
2.52926648e-01 2.34623000e-01 6.18312836e-01 -4.87468392e-01
-1.21617091e+00 5.18586576e-01 1.31722346e-01 -3.77624452e-01
-6.05164647e-01 -3.09823662e-01 -1.02527881e+00 -3.66668493e-01
-5.99824190e-01 4.35275465e-01 8.03379774e-01 6.38648212e-01
5.04950762e-01 1.15926218e+00 1.70271456e-01 -1.10939479e+00
-5.15699506e-01 -1.54843080e+00 -8.40494454e-01 1.70287639e-01
1.87525645e-01 -5.04871845e-01 -4.09191363e-02 2.23249763e-01] | [15.757858276367188, 5.256225109100342] |
96417aa4-a0c3-48ad-8f67-96c2de69f1fe | grassmannian-learning-mutual-subspace-method | 2111.04352 | null | https://arxiv.org/abs/2111.04352v1 | https://arxiv.org/pdf/2111.04352v1.pdf | Grassmannian learning mutual subspace method for image set recognition | This paper addresses the problem of object recognition given a set of images as input (e.g., multiple camera sources and video frames). Convolutional neural network (CNN)-based frameworks do not exploit these sets effectively, processing a pattern as observed, not capturing the underlying feature distribution as it does not consider the variance of images in the set. To address this issue, we propose the Grassmannian learning mutual subspace method (G-LMSM), a NN layer embedded on top of CNNs as a classifier, that can process image sets more effectively and can be trained in an end-to-end manner. The image set is represented by a low-dimensional input subspace; and this input subspace is matched with reference subspaces by a similarity of their canonical angles, an interpretable and easy to compute metric. The key idea of G-LMSM is that the reference subspaces are learned as points on the Grassmann manifold, optimized with Riemannian stochastic gradient descent. This learning is stable, efficient and theoretically well-grounded. We demonstrate the effectiveness of our proposed method on hand shape recognition, face identification, and facial emotion recognition. | ['Kazuhiro Fukui', 'Takumi Kobayashi', 'Bernardo B. Gatto', 'Naoya Sogi', 'Lincon S. Souza'] | 2021-11-08 | null | null | null | null | ['facial-emotion-recognition', 'face-identification'] | ['computer-vision', 'computer-vision'] | [ 6.01794291e-03 -1.73439398e-01 -1.28916368e-01 -5.82499981e-01
-1.34768724e-01 -6.63115382e-01 4.76685375e-01 -5.70188165e-01
-3.65142137e-01 -1.16151735e-01 8.63649771e-02 2.07122475e-01
-3.94170374e-01 -4.00269747e-01 -5.84134042e-01 -8.39448869e-01
8.53774622e-02 1.86005041e-01 -5.60290456e-01 6.36959672e-02
2.18987331e-01 9.22367573e-01 -1.64991784e+00 2.81803608e-01
3.26333523e-01 1.19447052e+00 -9.55489203e-02 6.05027914e-01
8.93127918e-02 6.24734282e-01 -6.19835500e-03 -5.48861384e-01
4.41248059e-01 -1.89934865e-01 -7.82974422e-01 6.16952002e-01
6.99066281e-01 -1.37848228e-01 -4.60277349e-01 1.22994268e+00
2.47744843e-01 1.80818751e-01 9.36804116e-01 -1.42818236e+00
-8.05731535e-01 2.61266697e-02 -4.39062446e-01 -2.81721681e-01
1.64855689e-01 -5.49645312e-02 1.02373552e+00 -1.61975765e+00
6.41397655e-01 1.67084980e+00 2.82937258e-01 7.53056943e-01
-1.39070821e+00 -3.63579482e-01 -1.15927551e-02 2.60105044e-01
-1.59689021e+00 -5.43239355e-01 9.93758917e-01 -5.45276880e-01
5.82691848e-01 3.66919130e-01 5.22811413e-01 1.06766701e+00
-7.44939223e-02 8.70134830e-01 7.54717588e-01 -4.41855073e-01
2.57633090e-01 2.33806506e-01 2.30316266e-01 6.91335380e-01
-6.10734150e-02 1.92268938e-02 -7.55851567e-01 -4.87921275e-02
7.89697647e-01 4.79387432e-01 -3.86826634e-01 -8.51488113e-01
-9.92480159e-01 7.28123188e-01 4.53639030e-01 3.56638342e-01
-4.34252083e-01 -6.98528215e-02 1.29592791e-01 3.12330008e-01
3.23381335e-01 5.36443666e-02 -2.65449613e-01 1.37101069e-01
-8.46532881e-01 -1.24703996e-01 8.29150736e-01 9.46940303e-01
8.14383566e-01 4.40636352e-02 4.70098890e-02 6.44465685e-01
8.09424102e-01 7.24967897e-01 4.39655811e-01 -9.28737462e-01
2.76720077e-01 8.47961605e-01 -2.41430223e-01 -1.43943310e+00
-3.38331401e-01 -2.64821261e-01 -1.13484406e+00 3.60395402e-01
3.95822555e-01 8.95071924e-02 -4.53147203e-01 1.99544501e+00
2.61384428e-01 1.91991895e-01 2.71247804e-01 1.10135138e+00
5.08953869e-01 2.20823094e-01 -3.09500784e-01 -9.10898075e-02
1.21244371e+00 -6.88081443e-01 -5.87146699e-01 1.19706415e-01
5.02054453e-01 -6.83787942e-01 9.69286263e-01 3.85356545e-01
-8.91735613e-01 -5.16399801e-01 -9.69353437e-01 1.21560432e-01
-3.74195904e-01 5.77753067e-01 2.23866291e-02 6.88709319e-01
-1.14324808e+00 8.94149840e-01 -7.19607711e-01 -5.23134649e-01
5.58259964e-01 3.14786106e-01 -8.00563812e-01 -9.19691846e-02
-5.50897181e-01 6.13898039e-01 2.18901679e-01 5.25593042e-01
-7.53709257e-01 -3.07920903e-01 -9.38970029e-01 -9.86852683e-03
2.38620177e-01 -4.64804918e-01 6.76687419e-01 -1.35711229e+00
-1.55433834e+00 1.03452575e+00 -2.67190784e-01 6.66072173e-03
3.03889513e-01 -3.75923276e-01 -3.00659776e-01 2.40031049e-01
-3.14336300e-01 4.61393833e-01 1.51139438e+00 -1.16246665e+00
-2.35956050e-02 -1.03007877e+00 -2.11667359e-01 2.34245330e-01
-8.01598907e-01 1.56128824e-01 -4.00414109e-01 -3.98177564e-01
4.69690442e-01 -1.14700079e+00 -5.08026639e-03 1.32335976e-01
-3.74430746e-01 -1.01904005e-01 1.04529619e+00 -5.60628474e-01
9.47692215e-01 -2.35488033e+00 6.72820508e-01 4.70693767e-01
2.18161464e-01 2.27588877e-01 -4.89294708e-01 2.11603537e-01
-4.04548287e-01 -6.54048100e-02 -2.86202520e-01 -4.42249477e-01
9.43191648e-02 7.50490576e-02 -3.28002959e-01 7.69570053e-01
5.45336366e-01 7.37700939e-01 -8.28653336e-01 -1.14100121e-01
1.98055670e-01 6.24342501e-01 -4.08790559e-01 4.51361418e-01
2.24169552e-01 4.99941170e-01 -2.66197085e-01 5.27485192e-01
9.17387962e-01 -1.48340121e-01 1.40592933e-01 -5.94008029e-01
1.04676262e-01 -3.80779415e-01 -1.58751619e+00 1.78811979e+00
-3.17944109e-01 5.94255567e-01 1.88129041e-02 -9.85330999e-01
1.12942004e+00 3.58444542e-01 4.61745888e-01 -7.69048780e-02
3.27989340e-01 2.02471122e-01 9.11867339e-03 -5.52268744e-01
3.05143110e-02 2.27800533e-01 4.77921754e-01 5.04859686e-01
6.18744791e-01 2.26118162e-01 -7.33061582e-02 6.71403408e-02
7.04132259e-01 7.50535429e-02 1.40823320e-01 -4.80534136e-01
8.70188177e-01 -7.52151012e-01 3.69693339e-01 3.14240247e-01
-3.00592352e-02 7.98964322e-01 3.66512060e-01 -5.62900662e-01
-9.09216046e-01 -9.04197454e-01 -9.59983692e-02 8.04854035e-01
-1.96761236e-01 -4.78898346e-01 -1.03258610e+00 -7.20044374e-01
-6.85300380e-02 2.67280519e-01 -7.05794394e-01 -3.98958951e-01
-2.16394857e-01 -3.82271707e-01 3.29018235e-01 5.03097415e-01
4.25649971e-01 -1.00800288e+00 -5.75871766e-01 -8.49435627e-02
5.12391292e-02 -1.15352285e+00 -6.86237812e-01 -1.61011413e-01
-9.38735008e-01 -1.31363142e+00 -6.11559212e-01 -6.15280867e-01
9.73236799e-01 4.34304059e-01 9.71834362e-01 -1.09655805e-01
-4.19304401e-01 1.08606672e+00 -2.11189583e-01 -2.66849339e-01
-1.54621396e-02 -4.69869643e-01 5.15054524e-01 1.18144572e+00
6.81460440e-01 -6.48710430e-01 -5.04932642e-01 3.60522836e-01
-1.06816578e+00 -1.68325961e-01 5.35268843e-01 9.55092609e-01
3.82312894e-01 -3.28528732e-01 6.82055503e-02 -3.18948805e-01
2.99876928e-01 -3.51539969e-01 -4.62178856e-01 4.34817642e-01
-3.39990705e-01 1.82341054e-01 4.62722480e-01 -4.08740640e-01
-6.12400711e-01 4.27273571e-01 2.27472261e-01 -1.22574484e+00
-3.22509289e-01 3.76273006e-01 -5.73443830e-01 -1.25149950e-01
6.04638636e-01 2.66739756e-01 3.43757331e-01 -5.76730430e-01
6.13670528e-01 5.74269235e-01 4.21000957e-01 -3.75500768e-01
8.64098608e-01 7.24602342e-01 2.25453973e-01 -1.24572325e+00
-6.94684684e-01 -7.53478348e-01 -1.37954330e+00 -4.76103038e-01
8.59365463e-01 -6.60922766e-01 -8.66263330e-01 5.59020877e-01
-1.35904312e+00 7.16125444e-02 -9.41390395e-02 7.28956938e-01
-6.72614872e-01 4.94625777e-01 -2.30593696e-01 -1.13534868e+00
-3.52462560e-01 -9.70181584e-01 1.15476036e+00 7.15622082e-02
-1.10862121e-01 -9.32463586e-01 -1.66012466e-01 -2.19454449e-02
3.14745848e-04 1.32946610e-01 7.13251889e-01 -6.97759509e-01
-5.02506733e-01 -2.43552193e-01 -1.67376116e-01 7.19755769e-01
2.23319411e-01 1.48208722e-01 -1.28705013e+00 -5.76858163e-01
5.13420463e-01 -2.76329070e-01 7.08435774e-01 2.36591563e-01
1.27352655e+00 -5.30962825e-01 2.72449013e-02 7.71013141e-01
1.44672835e+00 -7.07444251e-02 4.89147395e-01 8.01910982e-02
9.11226153e-01 1.04994428e+00 2.90803730e-01 4.33854610e-01
-9.33060236e-03 6.28322840e-01 5.34089923e-01 4.81023081e-02
2.92428821e-01 -7.28819072e-02 5.86670816e-01 8.75126779e-01
-1.76633582e-01 4.74288985e-02 -6.96080983e-01 2.81655937e-01
-2.06200695e+00 -9.69046116e-01 1.74660191e-01 2.45794964e+00
1.08471997e-01 -4.16070461e-01 7.79478922e-02 3.18882793e-01
7.30863750e-01 -1.76007785e-02 -5.93665659e-01 -1.74257502e-01
-3.43834668e-01 2.74937730e-02 2.47734282e-02 1.87368169e-01
-1.08240080e+00 7.59998798e-01 5.18955231e+00 5.70663989e-01
-1.18659353e+00 -1.71663344e-01 6.12979293e-01 8.24196413e-02
7.90804923e-02 -1.42011583e-01 -5.76449335e-01 1.28924295e-01
7.25934684e-01 6.07441887e-02 6.60958350e-01 1.04384816e+00
3.31741571e-02 4.79122579e-01 -1.68780541e+00 1.76283753e+00
5.11963546e-01 -1.08016443e+00 3.05706114e-01 1.97445393e-01
4.22685593e-01 -2.24328786e-01 4.21774745e-01 8.75252299e-03
-2.30026349e-01 -1.15385091e+00 8.31410706e-01 8.50469589e-01
5.90896010e-01 -6.82466269e-01 4.64702547e-01 2.92771995e-01
-1.06995916e+00 -1.35703072e-01 -6.17681801e-01 1.40446410e-01
-3.92979026e-01 3.02606106e-01 -5.58228195e-01 5.79945147e-01
8.24812889e-01 1.07306015e+00 -5.95055819e-01 7.24607289e-01
1.58846945e-01 3.88282120e-01 -1.96302250e-01 8.96611437e-02
2.75216460e-01 -8.32824707e-01 8.93914640e-01 1.11929047e+00
4.51709867e-01 1.30725890e-01 1.00984879e-01 1.03447318e+00
-1.79611653e-01 3.94359291e-01 -9.86955225e-01 -8.96331891e-02
-9.67104174e-03 1.67783248e+00 -6.28683031e-01 -1.19665712e-01
-4.08357710e-01 1.08527863e+00 3.58535737e-01 4.88314837e-01
-3.42279553e-01 -2.47848451e-01 9.21622753e-01 -5.02873957e-01
1.71541899e-01 -2.76580036e-01 4.17995416e-02 -1.41860986e+00
3.55343223e-01 -9.59939063e-01 2.70375937e-01 -6.23408437e-01
-1.44850194e+00 7.63693213e-01 -3.56406718e-01 -1.51316714e+00
-2.65799046e-01 -1.19819736e+00 -5.61979413e-01 7.23766088e-01
-1.08804560e+00 -1.06551647e+00 -4.32922930e-01 1.06366265e+00
3.11568141e-01 -5.57388663e-01 9.23536062e-01 1.34098217e-01
-5.63647211e-01 5.98505139e-01 3.82016093e-01 5.51563799e-01
5.11343777e-01 -1.31923223e+00 -8.81925374e-02 6.57561600e-01
6.60359621e-01 7.29948223e-01 3.18806559e-01 -2.39686184e-02
-1.85837269e+00 -1.28661513e+00 4.46265787e-01 -7.08802342e-01
5.26921690e-01 -5.11714458e-01 -9.20642316e-01 5.79986811e-01
5.75022101e-02 1.34266287e-01 8.05349708e-01 9.10389051e-02
-7.86018491e-01 -3.07085723e-01 -9.93191540e-01 7.04246700e-01
1.07614100e+00 -8.49480093e-01 -2.76932091e-01 3.58367115e-01
2.88182110e-01 -7.30505213e-02 -7.91818678e-01 3.34503412e-01
6.68755412e-01 -9.88391101e-01 8.31563175e-01 -1.07397091e+00
2.55189061e-01 -2.17756212e-01 -6.11041069e-01 -1.25135875e+00
-2.29685783e-01 -6.81618154e-01 -1.53885499e-01 1.16986537e+00
1.54181913e-01 -4.07817990e-01 7.49892533e-01 6.09450698e-01
1.64382532e-01 -7.67462194e-01 -8.95642281e-01 -9.83334482e-01
-1.95852712e-01 -5.66462874e-01 6.21724069e-01 9.11541343e-01
-1.19608805e-01 3.58314961e-01 -5.03089130e-01 2.83464402e-01
9.85613823e-01 1.53215900e-01 7.51820147e-01 -1.28866434e+00
5.68987466e-02 -5.61355114e-01 -9.10761833e-01 -8.13189507e-01
4.11409259e-01 -1.07698333e+00 -1.50142983e-01 -7.26037443e-01
3.06298077e-01 1.34008482e-01 -4.21016306e-01 1.86099589e-01
2.44429603e-01 8.79843608e-02 4.77502882e-01 3.72368783e-01
-5.11892855e-01 8.44219625e-01 9.31997895e-01 -2.38238245e-01
-9.69043374e-02 -7.39507526e-02 -3.53395253e-01 9.46466863e-01
5.08528590e-01 -1.43567130e-01 -1.85039401e-01 -3.59134763e-01
-1.37486294e-01 -2.54467309e-01 6.08751714e-01 -9.76058841e-01
1.43038854e-01 2.11436823e-02 6.27932072e-01 -3.89876276e-01
4.14038569e-01 -1.17918968e+00 5.68487458e-02 1.57242924e-01
-3.67664456e-01 5.62617108e-02 -1.32236093e-01 6.19556010e-01
-2.77591258e-01 -2.45385081e-01 9.21653926e-01 9.90128741e-02
-4.97665346e-01 8.17538321e-01 -2.57984642e-02 -3.22905302e-01
8.54945362e-01 -2.08037049e-01 3.48143905e-01 -3.36177230e-01
-9.11091626e-01 -1.41899303e-01 3.09170514e-01 6.38422787e-01
9.98600006e-01 -1.87247574e+00 -5.82652152e-01 4.53895926e-01
2.80767173e-01 -1.94312453e-01 2.61094362e-01 7.98232079e-01
-2.50531416e-02 3.71423215e-01 -3.15244973e-01 -1.20046341e+00
-1.43274903e+00 6.82669938e-01 4.93916720e-01 1.54533550e-01
-3.90412986e-01 7.41016388e-01 4.72271502e-01 -7.06406891e-01
3.90522510e-01 -1.13281116e-01 -4.41210151e-01 1.38982832e-01
7.12722719e-01 3.44330311e-01 2.75543779e-02 -1.16047788e+00
-3.86729479e-01 9.84954298e-01 1.51917085e-01 -2.52240598e-01
1.31517088e+00 -1.16689943e-01 -4.99624342e-01 5.26054263e-01
1.85424721e+00 -3.55562478e-01 -1.19269550e+00 -5.52330434e-01
2.90058583e-01 -6.80136859e-01 -1.48897946e-01 -9.89307389e-02
-1.25081873e+00 1.11521649e+00 1.00392520e+00 -7.23353326e-02
1.15084851e+00 -1.27322212e-01 1.19540341e-01 7.47815549e-01
1.80297613e-01 -1.06222188e+00 6.05084240e-01 5.69838762e-01
1.36046326e+00 -1.27059710e+00 -5.36892056e-01 -2.71294355e-01
-5.46561599e-01 1.68533051e+00 5.10383248e-01 -1.87094569e-01
1.04534900e+00 -4.58271146e-01 8.74379557e-03 -3.43388110e-01
-4.70740139e-01 -2.57859111e-01 6.62890077e-01 4.87027645e-01
1.56028315e-01 -2.02108789e-02 2.73765773e-01 4.85273600e-01
1.48482658e-02 -2.59839147e-01 1.67207614e-01 6.41095579e-01
-1.46698371e-01 -8.10514510e-01 -5.95624208e-01 1.94987774e-01
-1.73490256e-01 1.68822512e-01 -3.94861549e-01 4.92140710e-01
-3.79096232e-02 8.70958924e-01 1.71186373e-01 -6.59216523e-01
3.77402484e-01 2.40497559e-01 6.43535495e-01 -4.76081818e-01
1.42653480e-01 -3.04379109e-02 -6.38634145e-01 -7.98925817e-01
-5.56997836e-01 -7.56502092e-01 -7.40541816e-01 1.54523812e-02
-2.13778868e-01 9.09850001e-02 7.68575966e-01 9.46599424e-01
4.35306937e-01 -1.17110707e-01 1.22631478e+00 -1.10905874e+00
-8.73348832e-01 -8.31545353e-01 -8.59154761e-01 8.88280451e-01
3.11329097e-01 -7.26425409e-01 -4.10969883e-01 2.20102534e-01] | [8.052536964416504, 3.866126537322998] |
6b81e60d-dd41-4bca-8d1d-c0b1d62b25e0 | training-a-deep-learning-model-for-single | null | null | https://www.nature.com/articles/s41598-021-03299-4 | https://www.nature.com/articles/s41598-021-03299-4 | Training a deep learning model for single-cell segmentation without manual annotation | Advances in the artificial neural network have made machine learning techniques increasingly more important in image analysis tasks. Recently, convolutional neural networks (CNN) have been applied to the problem of cell segmentation from microscopy images. However, previous methods used a supervised training paradigm in order to create an accurate segmentation model. This strategy requires a large amount of manually labeled cellular images, in which accurate segmentations at pixel level were produced by human operators. Generating training data is expensive and a major hindrance in the wider adoption of machine learning based methods for cell segmentation. Here we present an alternative strategy that trains CNNs without any human-labeled data. We show that our method is able to produce accurate segmentation models, and is applicable to both fluorescence and bright-field images, and requires little to no prior knowledge of the signal characteristics. | ['Ji Yu', 'Nizam Ud Din'] | 2021-12-14 | null | null | null | scientific-reports-article-number-23995-2021 | ['cell-segmentation'] | ['medical'] | [ 6.10632896e-01 -6.40213937e-02 4.00792688e-01 -3.94151926e-01
-6.46285713e-01 -6.14985645e-01 2.00201824e-01 4.29310530e-01
-1.06855357e+00 1.11136675e+00 -7.27530718e-01 -3.78704697e-01
3.55302483e-01 -7.54778802e-01 -6.70888960e-01 -1.07171571e+00
3.10461909e-01 5.12439072e-01 5.45024097e-01 5.68913408e-02
3.41696918e-01 8.43284905e-01 -1.17338741e+00 1.64741993e-01
7.20993519e-01 8.75708997e-01 4.52172577e-01 9.33340549e-01
-3.64820361e-01 6.16502523e-01 -6.53934896e-01 -2.56740358e-02
1.46185398e-01 -5.60500085e-01 -1.03342378e+00 3.49171400e-01
3.44124474e-02 3.59502211e-02 1.77307039e-01 1.03971863e+00
5.15490592e-01 -9.14063379e-02 7.14437962e-01 -7.41693854e-01
-6.01958297e-02 1.79326668e-01 -3.99935961e-01 3.47650498e-01
-1.56689763e-01 1.11463591e-01 3.58353823e-01 -6.33595288e-01
8.11552644e-01 6.48994923e-01 6.74873412e-01 6.26927376e-01
-1.75255370e+00 -3.01819652e-01 -3.00951213e-01 -1.90466821e-01
-1.18316066e+00 -2.48165458e-01 4.63205904e-01 -6.60484970e-01
7.32858360e-01 1.64245181e-02 9.22906995e-01 5.46399713e-01
1.31752267e-01 5.82279444e-01 1.55036211e+00 -6.75141990e-01
5.07908940e-01 1.40452877e-01 1.31710425e-01 4.96378928e-01
2.20326751e-01 -1.58966467e-01 1.21545322e-01 7.65838176e-02
1.34868836e+00 -3.20113115e-02 -2.91246623e-01 -2.00308368e-01
-1.31465757e+00 7.88028300e-01 2.48235315e-01 8.37311149e-01
-2.49528080e-01 1.77233443e-01 3.17690045e-01 3.84757258e-02
3.18965137e-01 7.71560550e-01 -4.91824895e-01 -3.26526277e-02
-1.22630847e+00 1.24488257e-01 6.15514278e-01 3.91231537e-01
1.00000167e+00 -9.85531956e-02 2.40403205e-01 6.47777021e-01
-3.92364040e-02 2.05043092e-01 4.37861443e-01 -1.30131030e+00
-3.24810237e-01 7.65088201e-01 1.75049499e-01 -7.22229660e-01
-5.72564065e-01 -1.49855614e-01 -8.37365627e-01 6.16564810e-01
1.22004962e+00 -2.12388113e-01 -1.21887755e+00 1.25212991e+00
1.67384163e-01 -1.21519923e-01 2.44933218e-02 7.56308436e-01
3.68600070e-01 4.66614038e-01 -1.43864021e-01 -3.37358952e-01
1.00230300e+00 -5.34490585e-01 -7.08483100e-01 7.05413446e-02
7.58647442e-01 -7.41476238e-01 8.89854372e-01 5.94181180e-01
-9.27817523e-01 -4.00296539e-01 -8.40232134e-01 -9.94686410e-02
-5.97716510e-01 3.36917825e-02 6.18159354e-01 5.66879570e-01
-1.21016693e+00 7.42808640e-01 -1.00169909e+00 -6.51920617e-01
8.84156227e-01 8.49803329e-01 -5.78301847e-01 1.83880508e-01
-6.27770364e-01 7.69481242e-01 5.25132656e-01 2.08980605e-01
-6.10206902e-01 -2.56644309e-01 -5.24309695e-01 -9.19709727e-02
2.90854871e-02 -4.32823777e-01 1.23554969e+00 -1.33674777e+00
-1.69691479e+00 1.32389712e+00 -1.51851416e-01 -5.80504835e-01
3.73445392e-01 3.61806750e-01 2.59266824e-01 3.95631701e-01
-3.07477295e-01 1.03080010e+00 4.57697868e-01 -1.50193799e+00
-6.86865330e-01 -2.65304655e-01 -1.37950778e-01 -4.46002185e-01
-1.03065381e-02 4.87885810e-02 -1.62243962e-01 -3.18795145e-01
7.60106668e-02 -8.30911100e-01 -5.77934563e-01 1.11486390e-01
-2.23313883e-01 -3.43248161e-04 8.11658025e-01 -4.15348649e-01
5.00206947e-01 -1.89146900e+00 -5.09344041e-02 2.50845820e-01
2.95521498e-01 5.65155983e-01 3.43884490e-02 2.87187964e-01
1.50684282e-01 1.16089068e-01 -4.96260107e-01 -2.28958756e-01
-5.01118362e-01 4.01016802e-01 1.08491525e-01 4.84006971e-01
1.93964958e-01 9.22432840e-01 -8.25930357e-01 -7.69897103e-01
3.03463489e-01 6.10140681e-01 -3.21561247e-01 2.70482630e-01
-1.58301130e-01 1.03239262e+00 -1.79052189e-01 4.47596461e-01
5.23433030e-01 -4.24491942e-01 8.56898874e-02 8.41947831e-03
-3.90945941e-01 -3.72054785e-01 -6.36298716e-01 1.46593142e+00
-2.52266973e-01 8.73068035e-01 3.38967025e-01 -1.48974383e+00
9.03147340e-01 3.73108208e-01 6.03921533e-01 -4.19523656e-01
5.56975663e-01 5.87740898e-01 9.17155519e-02 -4.86145943e-01
-6.28758445e-02 -5.02223372e-01 3.45250785e-01 3.13902229e-01
2.40891844e-01 -4.24783170e-01 5.27870059e-01 -1.88927814e-01
9.01491106e-01 8.48662779e-02 4.51077893e-02 -4.64307815e-01
5.45367122e-01 3.55860323e-01 6.77642167e-01 5.90932131e-01
-1.79785773e-01 9.49268758e-01 5.45445919e-01 -8.18439305e-01
-1.29349732e+00 -4.73018676e-01 -3.44372988e-01 5.87388873e-01
4.09718566e-02 3.58425409e-01 -1.28288639e+00 -4.89903867e-01
-3.77038270e-01 4.50986214e-02 -6.52935445e-01 4.46279526e-01
-6.58786893e-01 -9.83964980e-01 4.95718598e-01 3.96031708e-01
3.59055370e-01 -1.26805794e+00 -9.06889081e-01 4.87477034e-01
-1.15370110e-01 -1.22558832e+00 8.03572163e-02 6.72710180e-01
-1.07649076e+00 -1.15208149e+00 -1.13864899e+00 -1.09182930e+00
1.22020268e+00 -1.06376179e-01 8.38921666e-01 3.90327662e-01
-6.00963593e-01 -9.24876891e-04 -1.37758464e-01 -5.80275536e-01
-4.17968005e-01 1.45078734e-01 -3.71191174e-01 -1.19184628e-02
4.04262483e-01 -6.22899652e-01 -6.66137397e-01 2.84841299e-01
-1.08793628e+00 7.15772361e-02 5.46541691e-01 1.08034730e+00
9.60859776e-01 1.38545915e-01 5.18459201e-01 -1.19077229e+00
2.90379584e-01 2.59374708e-01 -8.61988544e-01 -9.30670202e-02
-3.92928988e-01 -1.78806588e-01 9.10126507e-01 -4.11676377e-01
-7.31947184e-01 4.63056982e-01 -4.37662125e-01 -1.75824407e-02
-7.63240218e-01 5.14802516e-01 1.48216903e-01 -6.27439618e-01
7.02850163e-01 1.43532857e-01 3.90402943e-01 -2.62632132e-01
-2.40325063e-01 4.09858197e-01 5.11058927e-01 -3.37108403e-01
3.92282903e-01 8.91485691e-01 2.63022989e-01 -9.73029554e-01
-5.53848922e-01 -4.49257940e-01 -1.00714958e+00 -1.86586916e-01
1.03949606e+00 -3.11784953e-01 -5.81468761e-01 8.70519996e-01
-1.17681456e+00 -8.50982845e-01 -2.98718065e-01 4.46443439e-01
-6.85008466e-01 3.34620237e-01 -8.37672234e-01 -5.56114972e-01
-8.93171653e-02 -1.29364610e+00 8.32237661e-01 4.37211484e-01
-7.77501315e-02 -1.24185145e+00 8.28023553e-02 2.77889311e-01
4.33779478e-01 4.46747988e-01 9.55191195e-01 -6.04022026e-01
-5.52976847e-01 -3.24435055e-01 -3.00919443e-01 4.22228068e-01
2.22846523e-01 2.48150274e-01 -1.15966296e+00 -1.90598294e-01
-6.07642308e-02 -4.29716527e-01 7.82817245e-01 7.73816168e-01
1.33504069e+00 1.78111538e-01 -5.79553306e-01 3.85591954e-01
1.50087059e+00 4.96480137e-01 8.46632302e-01 2.87555307e-01
4.75039959e-01 7.25581467e-01 1.91767380e-01 -1.34086236e-01
-4.61428538e-02 4.08385009e-01 1.35387570e-01 -8.22324157e-01
1.13208234e-01 3.06009859e-01 -2.77563512e-01 5.64002216e-01
-3.80290180e-01 -2.67784029e-01 -1.02406418e+00 6.44255877e-01
-1.73792028e+00 -6.99736357e-01 -2.64260888e-01 2.07720828e+00
1.03937459e+00 1.38741329e-01 1.70529068e-01 4.53271329e-01
6.98587000e-01 -7.31620550e-01 -4.09738451e-01 -2.46399537e-01
-1.09880939e-01 4.50312883e-01 4.01894659e-01 4.44375545e-01
-1.10872805e+00 8.90430391e-01 7.43671083e+00 6.59580827e-01
-1.44261682e+00 -1.94555506e-01 9.56644833e-01 3.09118807e-01
8.27363431e-02 -1.91614348e-02 -4.83426839e-01 4.62127686e-01
6.94921315e-01 4.03231144e-01 2.34916925e-01 2.68542886e-01
4.35073912e-01 -6.34755254e-01 -1.03261554e+00 8.95960987e-01
-4.39389288e-01 -1.46239090e+00 -1.96322903e-01 3.31709951e-01
8.11130941e-01 -2.38038719e-01 -2.09200919e-01 -2.22653627e-01
1.44389659e-01 -1.28267097e+00 1.07357927e-01 5.76228321e-01
6.53070152e-01 -6.31685078e-01 1.28144455e+00 4.64135587e-01
-6.59570932e-01 2.68481076e-01 -5.25118887e-01 -1.32693723e-01
2.64402628e-01 8.56712282e-01 -8.89656186e-01 -8.17313697e-03
6.21984839e-01 2.21763358e-01 -4.13014561e-01 1.16571939e+00
2.74124384e-01 6.55681133e-01 -3.99065912e-01 -2.93910429e-02
3.23465854e-01 -3.05632621e-01 -3.16905901e-02 1.23789632e+00
2.59188026e-01 -1.06037609e-01 1.42726123e-01 8.87403786e-01
8.74751657e-02 4.35954705e-02 -3.86692643e-01 -3.47882688e-01
-5.81567697e-02 1.60863173e+00 -1.76051080e+00 -3.23265046e-01
-3.49588633e-01 7.41466522e-01 3.41685861e-01 4.97641712e-01
-3.68821234e-01 -6.00684226e-01 1.24399610e-01 3.74781758e-01
3.32206249e-01 -2.19691992e-01 -3.94695252e-01 -7.43827403e-01
-5.96046686e-01 -6.27450287e-01 1.30749037e-02 -6.31845653e-01
-1.02838659e+00 5.34239233e-01 -3.08167875e-01 -7.48859167e-01
-1.36160567e-01 -1.06279409e+00 -5.73326230e-01 7.36797392e-01
-1.54250741e+00 -9.36392009e-01 -3.15177262e-01 3.74010712e-01
2.73348421e-01 1.44013405e-01 8.95254672e-01 3.23297828e-01
-4.07273591e-01 -9.86300707e-02 7.89493695e-02 4.12316263e-01
4.43795055e-01 -1.52246475e+00 -1.77775379e-02 6.64106190e-01
8.82133842e-02 5.51357746e-01 7.62877464e-01 -1.79451078e-01
-6.93512082e-01 -8.17697167e-01 9.41527009e-01 -1.94469824e-01
3.11680228e-01 -3.22690427e-01 -9.59944665e-01 4.44541216e-01
3.68473858e-01 3.79634053e-01 8.96129489e-01 -3.37695748e-01
4.10785943e-01 8.71558189e-02 -1.21259665e+00 3.36938977e-01
4.09004807e-01 -3.67343158e-01 -1.40485510e-01 2.40878105e-01
7.10703209e-02 -2.88054168e-01 -7.80703664e-01 3.49265754e-01
4.32474017e-01 -9.62246656e-01 5.10488868e-01 -2.32602313e-01
2.11997569e-01 -5.19111931e-01 4.78795052e-01 -1.10920334e+00
-1.50878280e-02 -4.30260032e-01 4.02721286e-01 8.81473958e-01
5.88909447e-01 -6.74139857e-01 1.07072699e+00 5.53736687e-01
-1.08088218e-01 -8.75506997e-01 -7.85815001e-01 -5.29006064e-01
3.69145066e-01 1.37772560e-01 1.36444062e-01 8.36268902e-01
-2.24725492e-02 5.16580380e-02 1.09363720e-01 -3.16131562e-01
4.55519557e-01 2.00156018e-01 7.39208698e-01 -1.45934510e+00
9.48516466e-03 -5.13779879e-01 -6.29686236e-01 -9.79330063e-01
1.44373968e-01 -6.28833950e-01 3.19479048e-01 -1.60172689e+00
1.66927382e-01 -3.33810151e-01 -2.17356399e-01 5.33455610e-01
7.71306679e-02 8.05001915e-01 -9.58949104e-02 1.97026253e-01
-3.86450619e-01 -9.14892405e-02 1.55113506e+00 -8.37909989e-03
-4.96345162e-02 2.36705760e-03 -3.72790068e-01 7.97780633e-01
9.37213719e-01 -5.08946538e-01 -1.27699807e-01 -2.67209351e-01
9.36748385e-02 -3.15796077e-01 4.22193617e-01 -1.11096764e+00
3.20430368e-01 2.93943137e-02 7.31260598e-01 -2.51673222e-01
1.36510223e-01 -9.20340538e-01 1.64099097e-01 5.61531961e-01
-3.41490448e-01 -3.19936931e-01 1.31216079e-01 1.71778366e-01
-4.64463443e-01 -5.37013054e-01 1.21486533e+00 -6.77367449e-01
-4.12144065e-01 1.35411993e-01 -9.17515397e-01 -2.35449150e-01
1.04135692e+00 -6.56320274e-01 -2.89271306e-02 -9.53554064e-02
-9.60140586e-01 -5.15359528e-02 8.66989791e-01 -5.28179586e-01
3.63182545e-01 -9.67196405e-01 -2.40837753e-01 1.59589559e-01
-2.01335743e-01 3.96173954e-01 3.01027242e-02 1.12918448e+00
-1.17523265e+00 4.28359985e-01 -4.99999017e-01 -8.60883653e-01
-1.03817737e+00 5.17782271e-01 5.41198373e-01 -2.98115969e-01
-4.03324515e-01 7.77456343e-01 1.74809009e-01 -3.76355350e-01
-6.25870451e-02 -5.95254719e-01 -3.33245784e-01 -2.29895890e-01
4.17689502e-01 1.69618681e-01 -4.07197094e-03 -5.54819703e-01
-3.11377011e-02 6.98642194e-01 -1.15569130e-01 -6.24803603e-02
1.23792005e+00 -5.55386581e-02 -3.56669992e-01 5.12735546e-01
1.16121149e+00 -3.17520529e-01 -1.46030819e+00 1.92911744e-01
1.54206067e-01 -1.92964628e-01 1.99779436e-01 -5.74727178e-01
-1.15695179e+00 1.19612575e+00 4.35349822e-01 5.95458150e-01
1.12220669e+00 -1.77748352e-01 9.27887321e-01 4.11357343e-01
5.08857429e-01 -1.22800994e+00 -7.51994143e-04 3.37773085e-01
1.91486061e-01 -1.39110446e+00 -3.00491899e-01 -5.88164210e-01
-1.78445145e-01 1.32919252e+00 5.30515313e-01 -1.06039166e-01
4.62503910e-01 5.74959099e-01 5.83521843e-01 -2.35458702e-01
-3.43831629e-01 -3.40727836e-01 -2.98887014e-01 8.04680407e-01
6.31480277e-01 -3.76634389e-01 -4.01904345e-01 1.91773415e-01
3.45126688e-01 6.04698062e-01 5.70000231e-01 1.14710903e+00
-5.64578593e-01 -1.28896594e+00 -2.57002532e-01 3.84359121e-01
-9.45703983e-01 2.35126540e-01 -4.45273638e-01 6.86551452e-01
3.62846643e-01 6.81024671e-01 2.12930828e-01 1.38330564e-01
-2.68442240e-02 6.83456361e-02 7.17370689e-01 -6.82856560e-01
-4.44589853e-01 4.33239698e-01 -4.21037078e-01 -3.26866746e-01
-9.11843896e-01 -4.26277816e-01 -1.71334052e+00 -1.18957169e-01
-4.84003931e-01 3.02354515e-01 4.65532452e-01 1.00545239e+00
1.60165638e-01 3.50440472e-01 2.76962101e-01 -1.19277251e+00
2.71258324e-01 -8.58426690e-01 -8.00237715e-01 3.73589516e-01
3.13920379e-01 -3.59984070e-01 -3.28081310e-01 7.85855412e-01] | [14.52804183959961, -3.1511824131011963] |
a1c2f308-ad1b-4d3d-8d94-19d448b09125 | alive-caricature-from-2d-to-3d | 1803.06802 | null | http://arxiv.org/abs/1803.06802v3 | http://arxiv.org/pdf/1803.06802v3.pdf | Alive Caricature from 2D to 3D | Caricature is an art form that expresses subjects in abstract, simple and
exaggerated view. While many caricatures are 2D images, this paper presents an
algorithm for creating expressive 3D caricatures from 2D caricature images with
a minimum of user interaction. The key idea of our approach is to introduce an
intrinsic deformation representation that has a capacity of extrapolation
enabling us to create a deformation space from standard face dataset, which
maintains face constraints and meanwhile is sufficiently large for producing
exaggerated face models. Built upon the proposed deformation representation, an
optimization model is formulated to find the 3D caricature that captures the
style of the 2D caricature image automatically. The experiments show that our
approach has better capability in expressing caricatures than those fitting
approaches directly using classical parametric face models such as 3DMM and
FaceWareHouse. Moreover, our approach is based on standard face datasets and
avoids constructing complicated 3D caricature training set, which provides
great flexibility in real applications. | ['Yu-Kun Lai', 'Qianyi Wu', 'Juyong Zhang', 'Jianfei Cai', 'Jianmin Zheng'] | 2018-03-19 | alive-caricature-from-2d-to-3d-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Wu_Alive_Caricature_From_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Wu_Alive_Caricature_From_CVPR_2018_paper.pdf | cvpr-2018-6 | ['caricature'] | ['computer-vision'] | [ 3.35380696e-02 5.37448227e-01 1.70216829e-01 -4.39603895e-01
2.51163952e-02 -3.92557085e-01 5.87077200e-01 -9.45634425e-01
2.92621464e-01 3.72273862e-01 -3.09410840e-02 1.06153943e-01
-1.65985063e-01 -7.42485940e-01 -7.08234012e-01 -4.28134561e-01
2.77604729e-01 4.39508379e-01 -5.22390008e-01 -3.00066113e-01
1.51659682e-01 1.31742620e+00 -1.83541191e+00 -8.66922289e-02
7.91507125e-01 9.82193649e-01 6.42614216e-02 3.14126253e-01
-4.40925032e-01 1.78917721e-01 -5.90407073e-01 -8.13959062e-01
3.37026238e-01 -2.14194894e-01 -2.92457402e-01 7.01762080e-01
6.75988138e-01 -3.06709319e-01 -9.49953869e-02 8.80922198e-01
4.05314445e-01 -1.79705709e-01 8.77323985e-01 -1.38441145e+00
-1.09246981e+00 2.38352597e-01 -7.18615294e-01 -6.25063002e-01
4.34792817e-01 -2.54668504e-01 1.69469461e-01 -1.00992501e+00
7.17922926e-01 1.87863278e+00 7.81061947e-01 8.92666996e-01
-1.28622818e+00 -5.70496917e-01 -1.23184316e-01 -5.37239611e-01
-1.61027598e+00 -3.40722263e-01 1.39604497e+00 -3.89390588e-01
2.48604998e-01 5.43771505e-01 6.94512188e-01 9.51849818e-01
7.25156441e-02 5.53253353e-01 1.24732876e+00 -6.18490219e-01
-1.16535045e-01 4.78680253e-01 -7.17416853e-02 7.39447773e-01
8.93944688e-03 3.36380079e-02 4.65902500e-02 -3.43663365e-01
1.58988917e+00 1.01849608e-01 -2.48620242e-01 -4.13290828e-01
-7.05067158e-01 6.12503946e-01 -2.64324713e-02 4.13121939e-01
-1.99038729e-01 1.48430154e-01 -1.87210999e-02 8.55765790e-02
3.80459666e-01 2.36348361e-01 -1.25170514e-01 2.31242552e-01
-8.66888046e-01 5.89562833e-01 6.07791722e-01 1.24664164e+00
7.77701080e-01 4.10781026e-01 1.40529498e-01 9.46054995e-01
3.70541126e-01 7.22621500e-01 2.57601857e-01 -1.08813226e+00
-4.89267968e-02 9.13883448e-01 4.52681258e-02 -1.04625249e+00
2.76456331e-03 8.69389158e-03 -6.50431871e-01 8.43884468e-01
8.79351795e-02 1.85139831e-02 -6.91819131e-01 1.57943225e+00
4.98471797e-01 1.17514662e-01 1.12150311e-01 7.71189272e-01
9.21484292e-01 6.06554747e-01 -1.36708409e-01 -2.56787091e-01
1.49933767e+00 -4.52129871e-01 -1.18299830e+00 3.08326393e-01
3.37878436e-01 -1.03385842e+00 1.33623099e+00 4.42136854e-01
-1.25785470e+00 -7.75330245e-01 -8.78680587e-01 -3.08302134e-01
-1.01787038e-01 4.76590157e-01 5.60931146e-01 1.04705071e+00
-1.16843569e+00 5.03375590e-01 -2.09537461e-01 -7.62178600e-02
3.10137004e-01 5.17811537e-01 -6.82116270e-01 3.16121519e-01
-7.70826757e-01 1.03188407e+00 2.44088203e-01 2.47177571e-01
-5.47087491e-01 -8.35817099e-01 -8.84494126e-01 -2.20242605e-01
1.32500127e-01 -5.08439839e-01 9.07454967e-01 -1.22180367e+00
-1.92796338e+00 1.33164549e+00 -1.80760771e-01 1.50139049e-01
7.67062306e-01 -1.41092405e-01 -6.18577361e-01 -2.21065991e-02
-5.94713271e-01 7.65749633e-01 1.26848102e+00 -1.69386840e+00
1.31718293e-01 -7.07831085e-01 5.94630465e-02 2.04669535e-01
-2.15896681e-01 1.21403195e-01 -5.51862359e-01 -7.70637035e-01
2.01999731e-02 -8.54244411e-01 9.05151144e-02 5.89349985e-01
-2.43686318e-01 -2.26346955e-01 1.66919065e+00 -6.13938987e-01
9.57744420e-01 -2.11877060e+00 2.76868492e-01 4.00294930e-01
1.12140149e-01 3.68983090e-01 3.32300514e-02 2.15654939e-01
-2.67380565e-01 2.96114236e-01 -2.51681864e-01 -5.38936079e-01
4.00991030e-02 4.07699585e-01 -2.87452787e-01 2.78810948e-01
3.42639863e-01 8.28873038e-01 -2.77757436e-01 -8.33276808e-01
2.76571840e-01 1.19899786e+00 -6.80350721e-01 2.46100798e-01
-2.02979624e-01 4.95532095e-01 -5.08911669e-01 6.26301050e-01
1.31012416e+00 5.55832565e-01 3.71112605e-03 -3.02622050e-01
-1.61011770e-01 -8.38789940e-01 -1.40521181e+00 1.60293543e+00
-5.50721526e-01 1.96497798e-01 4.89427537e-01 -5.62675893e-01
1.67848635e+00 4.92754191e-01 3.17085683e-01 8.12495574e-02
4.44161505e-01 2.51124650e-01 -5.59566796e-01 -6.66551888e-01
2.12008655e-01 -4.66959000e-01 7.97927678e-02 8.22312608e-02
-7.87722319e-02 -9.27450001e-01 -3.54462415e-01 -2.83743113e-01
-4.77878675e-02 7.35071480e-01 4.71830405e-02 -4.75189358e-01
1.11294115e+00 -4.98199433e-01 3.87896538e-01 -1.83937028e-01
1.65190801e-01 5.97263515e-01 5.38241684e-01 -7.43784487e-01
-1.09157383e+00 -8.76494288e-01 -4.15018141e-01 1.76391974e-01
-3.99707370e-02 3.87249328e-02 -1.41613352e+00 -3.94340217e-01
1.95247710e-01 6.52057052e-01 -8.77990425e-01 8.39112550e-02
-7.57332385e-01 -1.34190440e-01 6.55905604e-01 2.75101900e-01
4.63719815e-01 -8.99810016e-01 -2.59804040e-01 -2.62377173e-01
1.96438789e-01 -8.68013024e-01 -7.68262148e-01 -7.44404316e-01
-1.11505079e+00 -8.06000948e-01 -7.61270165e-01 -9.66875494e-01
1.03211188e+00 -3.58818173e-02 9.27475572e-01 2.73223042e-01
-3.68551105e-01 3.84719849e-01 6.82700053e-02 -5.60989857e-01
-6.45111561e-01 -6.27932668e-01 1.69052839e-01 3.77293110e-01
2.96021789e-01 -8.50594223e-01 -2.94124365e-01 2.06166625e-01
-1.07446897e+00 -3.63029987e-02 4.42080140e-01 3.99342537e-01
7.71894932e-01 -1.40367612e-01 5.86686075e-01 -9.61999178e-01
6.32347286e-01 7.35636353e-02 -7.54194140e-01 1.77813530e-01
-3.05525601e-01 1.70554176e-01 6.88964546e-01 -5.20559549e-01
-1.59595239e+00 2.45126277e-01 -3.01650643e-01 -9.19787765e-01
-3.39531183e-01 -2.28081614e-01 -7.12578237e-01 -4.53829676e-01
4.96379912e-01 4.31298502e-02 8.29723954e-01 -8.63290727e-01
5.96783638e-01 8.11347008e-01 7.77368188e-01 -9.32111740e-01
1.02030051e+00 6.13432169e-01 3.13511670e-01 -9.30750787e-01
-1.16679206e-01 3.28593284e-01 -9.81525242e-01 -4.21497166e-01
7.29707897e-01 -6.05337858e-01 -1.09070528e+00 3.76107961e-01
-1.31519151e+00 4.02595401e-01 -3.23830932e-01 1.82019711e-01
-8.93233895e-01 5.75098336e-01 -3.44465882e-01 -1.04476881e+00
-3.99516284e-01 -1.19056022e+00 1.39725411e+00 1.28867581e-01
-1.72804534e-01 -1.01871896e+00 -6.22828752e-02 4.28318948e-01
2.39092574e-01 9.03525651e-01 1.04844236e+00 1.13657616e-01
-2.93421775e-01 -4.24367398e-01 2.55448133e-01 4.33905780e-01
2.28479952e-01 7.05697775e-01 -1.17007172e+00 1.06811821e-01
1.58485070e-01 -1.25303522e-01 9.62639228e-02 2.04780206e-01
1.39605725e+00 -3.00951004e-01 -1.89348489e-01 7.70971715e-01
1.48380435e+00 2.99166530e-01 1.01016068e+00 -4.30165708e-01
4.62510109e-01 8.61763656e-01 4.46844846e-01 3.39188039e-01
5.15960343e-02 9.70928073e-01 5.44638336e-01 -1.85034454e-01
-1.79337174e-01 -5.47041893e-01 2.26934880e-01 4.42951381e-01
-5.83851576e-01 3.93578678e-01 -3.24660420e-01 6.35248423e-02
-1.23609591e+00 -1.00970793e+00 -1.11177884e-01 2.18215466e+00
7.43010819e-01 -4.01571453e-01 1.10668458e-01 2.82535285e-01
8.60124111e-01 -2.58035272e-01 -2.32731715e-01 -8.94674599e-01
-1.82423934e-01 3.19660157e-01 -1.08510189e-01 6.27484500e-01
-6.59706175e-01 7.96307504e-01 6.57159710e+00 8.15835714e-01
-1.18141258e+00 -1.30007550e-01 5.23423433e-01 1.24313995e-01
-6.73100650e-01 -2.11802840e-01 -7.29685128e-01 2.88560688e-01
3.82777065e-01 -5.46072684e-02 3.55729043e-01 1.06710005e+00
4.98308301e-01 4.63625610e-01 -1.01607156e+00 1.23711812e+00
1.53420582e-01 -1.14528859e+00 5.16288817e-01 1.56082630e-01
7.07566261e-01 -1.31233549e+00 3.69712114e-01 4.85783555e-02
-2.88878858e-01 -1.29165423e+00 7.75513530e-01 7.72153854e-01
1.26911366e+00 -9.14971888e-01 3.84101540e-01 4.93981719e-01
-8.89212906e-01 1.98952779e-01 -4.91129369e-01 1.13404535e-01
2.20173448e-02 2.86971599e-01 -5.73725760e-01 6.36852384e-01
2.96734810e-01 8.28325227e-02 -1.64944753e-01 4.55822080e-01
3.27911265e-02 -1.77637413e-01 -2.51264125e-01 1.48465484e-01
-5.06361835e-02 -8.07829857e-01 4.16490614e-01 8.52936864e-01
7.13760734e-01 2.33720154e-01 -2.47693598e-01 1.15453577e+00
-1.23278104e-01 5.31799853e-01 -1.10190654e+00 3.76771949e-02
6.58339083e-01 1.28442729e+00 -1.08794577e-01 -1.48470476e-01
-2.66218454e-01 8.19890082e-01 1.39431074e-01 1.14371613e-01
-8.53646338e-01 -1.90767959e-01 6.16569042e-01 4.75693882e-01
9.25233238e-04 3.40191484e-03 -1.77172199e-01 -9.64312375e-01
-7.43925422e-02 -8.37063015e-01 -1.46896720e-01 -1.05576754e+00
-8.48791420e-01 8.85280192e-01 1.62187606e-01 -1.21895528e+00
-3.57127011e-01 -8.97688687e-01 -7.89715230e-01 1.07030857e+00
-1.16894650e+00 -1.86313033e+00 -4.80476499e-01 8.09722006e-01
3.18304569e-01 -1.44080728e-01 1.09064007e+00 1.17392011e-01
-4.49596673e-01 6.51982069e-01 -2.71664858e-01 -1.83846444e-01
4.58640218e-01 -8.54677558e-01 1.66515000e-02 3.82718623e-01
-1.74664348e-01 6.85795426e-01 7.33077765e-01 -3.97976190e-01
-1.73679113e+00 -8.99885833e-01 7.25092828e-01 -6.85240805e-01
-1.05894871e-01 -3.13936293e-01 -8.14335167e-01 8.50142002e-01
6.09186329e-02 4.80918474e-02 5.72158277e-01 -4.34372991e-01
-2.84352154e-01 -2.73023129e-01 -1.76744723e+00 6.81416035e-01
1.24798858e+00 -1.79920256e-01 -7.95182824e-01 1.67415336e-01
4.02216643e-01 -4.11159337e-01 -1.27272701e+00 3.48812163e-01
6.12323642e-01 -9.14964795e-01 1.05325329e+00 -3.72873545e-01
1.77713126e-01 -3.29533249e-01 1.59690037e-01 -1.04965067e+00
-1.30113721e-01 -1.10266674e+00 -1.72906622e-01 1.35773969e+00
-4.19627726e-02 -5.05134046e-01 8.22055221e-01 7.67676294e-01
-1.67148724e-01 -6.54944062e-01 -7.42659986e-01 -8.24633956e-01
2.40427032e-01 -1.59543697e-02 1.12072158e+00 8.74623120e-01
-3.44769031e-01 -1.07578069e-01 -5.86070001e-01 2.77903024e-02
7.63985634e-01 2.01653436e-01 1.07525814e+00 -1.60727549e+00
1.34109423e-01 -2.79215097e-01 -4.69137698e-01 -7.49253571e-01
5.55275619e-01 -7.94823706e-01 -5.53079367e-01 -9.22955155e-01
-1.67830825e-01 -4.79430020e-01 5.37897825e-01 1.52026787e-01
1.56690940e-01 2.25448534e-01 1.17965147e-01 2.21318960e-01
6.42609179e-01 6.37072504e-01 1.89371347e+00 2.54542679e-01
-2.75571167e-01 -7.00921416e-02 -8.08457315e-01 9.95298803e-01
5.47736824e-01 4.21800725e-02 -7.40330219e-01 -3.64193946e-01
-4.53857243e-01 1.21344507e-01 2.90589422e-01 -6.18219316e-01
-4.76038367e-01 -2.49780998e-01 4.65586245e-01 -5.46575963e-01
7.88506031e-01 -1.29607916e+00 9.76453364e-01 2.24742383e-01
-3.63467447e-02 5.86716682e-02 1.37578800e-01 1.24094635e-01
-1.49271935e-01 -4.96491909e-01 1.00337529e+00 -2.12898970e-01
-4.19941276e-01 5.03334999e-01 1.84780434e-01 -4.35719311e-01
1.33978283e+00 -6.92206800e-01 1.81917131e-01 -2.31202260e-01
-7.92461991e-01 -3.79338950e-01 8.56673062e-01 3.72426152e-01
7.31976449e-01 -1.85268593e+00 -7.40650475e-01 6.69227362e-01
-8.91785473e-02 -5.98952696e-02 3.21550965e-01 2.73202479e-01
-9.64246929e-01 3.68769854e-01 -6.02954865e-01 -2.80919760e-01
-1.43850350e+00 8.22257638e-01 5.16976953e-01 3.47112864e-01
-6.64082468e-01 4.79163587e-01 3.51384163e-01 -6.41902387e-01
5.05912490e-02 -2.48993337e-01 -4.78598595e-01 -2.94955015e-01
6.05258644e-01 2.27270573e-01 -2.27656767e-01 -1.17518473e+00
-1.82395294e-01 1.23018539e+00 3.15308452e-01 1.00935530e-02
1.10220361e+00 6.86752647e-02 -3.22918832e-01 -2.44759526e-02
1.21399605e+00 3.95209372e-01 -1.30105722e+00 9.01634246e-03
-4.83494133e-01 -8.11511159e-01 -1.41001999e-01 -3.08490515e-01
-1.09470284e+00 8.35183978e-01 3.99201572e-01 -1.03650436e-01
1.17966771e+00 -1.45331874e-01 5.29033065e-01 8.96624476e-02
4.18191969e-01 -9.00255382e-01 -1.19442090e-01 -1.47595465e-01
1.71473610e+00 -6.73474371e-01 -2.33091876e-01 -1.05178654e+00
-6.17597699e-01 1.45738006e+00 6.23679936e-01 -2.75023758e-01
7.13672101e-01 6.04396820e-01 1.26474723e-01 -2.49835461e-01
-3.27115178e-01 3.63487273e-01 3.98402274e-01 8.16042244e-01
3.83031696e-01 2.76901834e-02 -5.09132624e-01 7.04694867e-01
-4.23102021e-01 2.32922971e-01 4.32416052e-01 6.55634701e-01
-4.01248664e-01 -1.23905098e+00 -7.89763629e-01 -2.40705371e-01
-4.62316722e-01 4.29022282e-01 -4.95205879e-01 1.14149821e+00
4.36040610e-01 5.40913641e-01 1.85425892e-01 -1.34907857e-01
7.83423424e-01 3.26463431e-01 9.28917289e-01 -2.43442610e-01
-3.13494623e-01 1.82927847e-01 -1.37439311e-01 -2.06748694e-01
-4.82791454e-01 -3.28275472e-01 -1.11117148e+00 -5.97267926e-01
-2.31767192e-01 1.36871904e-01 7.27844059e-01 4.87905562e-01
4.51409250e-01 -1.28158972e-01 8.39429557e-01 -1.06416333e+00
-6.16719127e-01 -8.64726424e-01 -7.68575430e-01 6.98332310e-01
-1.09841950e-01 -8.96174610e-01 -2.23167002e-01 5.44339240e-01] | [12.734289169311523, -0.247842475771904] |
bcc859fb-86cd-4010-bef2-6d98fa3d9041 | a-computer-vision-based-optical-method-for | 2303.14233 | null | https://arxiv.org/abs/2303.14233v1 | https://arxiv.org/pdf/2303.14233v1.pdf | A computer vision based optical method for measuring fluid level in cell culture plates | For a transparent well with a known volume capacity, changes in fluid level result in predictable changes in magnification of an overhead light source. For a given well size and fluid, the relationship between volume and magnification can be calculated if the fluid's index of refraction is known or in a naive fashion with a calibration procedure. Light source magnification can be measured through a camera and processed using computer vision contour analysis with OpenCV. This principle was applied in the design of a 3D printable sensing device using a raspberry pi zero and a camera | ['Mircea Teodorescu', 'Pierre V. Baudin'] | 2023-03-24 | null | null | null | null | ['culture'] | ['speech'] | [ 1.63381666e-01 -1.78784579e-01 3.60289037e-01 2.76911929e-02
3.83053005e-01 -7.02661037e-01 1.02980144e-01 1.91599935e-01
-5.71190953e-01 6.94478810e-01 -1.79796681e-01 -2.26345927e-01
2.80320913e-01 -9.85012412e-01 -4.68946636e-01 -2.72582740e-01
3.76651466e-01 4.17212248e-01 6.63521588e-01 1.12325042e-01
5.77567279e-01 6.71911299e-01 -1.18200076e+00 -3.57995272e-01
3.28634411e-01 9.75806832e-01 2.78729171e-01 1.02252281e+00
-2.83838421e-01 3.93713057e-01 -4.28777069e-01 -2.37980098e-01
6.28718734e-01 -2.31931806e-01 -2.57977366e-01 1.50591388e-01
9.61255953e-02 -6.35901630e-01 2.11224571e-01 7.13754714e-01
1.05116099e-01 -1.18775122e-01 8.64713252e-01 -2.96487927e-01
-7.38995135e-01 -3.03153209e-02 -7.55490303e-01 -1.25427898e-02
7.01420128e-01 2.01480657e-01 2.14701608e-01 -7.89703250e-01
4.12642479e-01 1.04759729e+00 7.13902235e-01 3.72446835e-01
-1.41371071e+00 -1.65597200e-01 -4.35525149e-01 -6.01561189e-01
-1.11199975e+00 7.08628222e-02 6.84188426e-01 -7.36722946e-01
7.37461329e-01 3.91071975e-01 1.33986509e+00 -2.59443879e-01
1.12590230e+00 -4.15896744e-01 1.61620533e+00 -6.04375243e-01
4.96403098e-01 4.31062400e-01 -1.60668060e-01 7.41445065e-01
5.65137506e-01 1.20446786e-01 -2.25766599e-01 -6.89648017e-02
1.72110045e+00 1.08848080e-01 -2.22399399e-01 -3.16722274e-01
-9.49021578e-01 3.48291218e-01 2.32703164e-01 4.65241164e-01
5.60727417e-02 4.92410846e-02 -3.64042133e-01 2.51549035e-01
3.53719354e-01 8.63225937e-01 -3.91232520e-01 -1.77253380e-01
-4.32741344e-01 -6.18670620e-02 1.42185378e+00 5.03377914e-01
5.59534013e-01 -1.08474612e-01 2.99729649e-02 3.20211619e-01
7.60943592e-01 1.05394363e+00 2.91083544e-01 -9.14500952e-01
-2.95095742e-02 8.99425209e-01 6.57532513e-01 -6.50781989e-01
-4.19222236e-01 6.27425969e-01 2.63999822e-03 1.39520788e+00
8.27017009e-01 -2.44901061e-01 -8.18434715e-01 6.49426818e-01
4.18446571e-01 -3.29789162e-01 2.31330656e-02 7.75017142e-01
7.56948948e-01 3.46844226e-01 -4.58733708e-01 -5.35888612e-01
1.15517223e+00 -3.95600855e-01 -1.02648079e+00 9.44154616e-03
-2.02237934e-01 -9.18508053e-01 1.46867526e+00 1.52923152e-01
-1.14834559e+00 -3.98757100e-01 -1.20309699e+00 -1.61037639e-01
-2.75405675e-01 -2.43117750e-01 2.91336328e-01 7.90976882e-01
-6.64437354e-01 7.40522921e-01 -9.48182166e-01 -1.40192524e-01
-1.47760026e-02 3.68677199e-01 9.77515653e-02 6.27775729e-01
-4.12906736e-01 1.28565741e+00 -1.94309682e-01 8.26421604e-02
8.78408253e-02 -8.74938905e-01 -4.89261180e-01 -3.49196285e-01
-3.55556905e-01 -5.19912899e-01 1.07568157e+00 -5.00320196e-01
-2.51538205e+00 1.12805998e+00 7.53857046e-02 -1.62476167e-01
7.45242119e-01 -2.44103760e-01 -6.94882795e-02 3.85838658e-01
-3.16059977e-01 -1.53162092e-01 9.53671932e-01 -1.15213096e+00
-1.82928219e-01 -5.52934587e-01 2.98848245e-02 2.64009535e-01
1.35674641e-01 9.15400404e-03 -8.47266565e-05 -1.27020791e-01
2.35628143e-01 -8.14120591e-01 -1.77411214e-01 8.61307800e-01
7.07827657e-02 2.65242696e-01 7.41167426e-01 -5.09947002e-01
9.02605474e-01 -1.75184095e+00 -6.04155064e-01 5.35758913e-01
2.97214240e-01 8.05388168e-02 6.75268471e-01 1.76308498e-01
4.06356186e-01 -2.72960991e-01 -1.89503655e-01 5.93443990e-01
-3.65832895e-01 -3.22544962e-01 1.00433484e-01 4.14847583e-01
-3.94324362e-01 6.43623173e-01 -6.34705067e-01 -4.68203425e-01
7.11435497e-01 5.84643126e-01 -1.81222424e-01 2.89414734e-01
-4.16180380e-02 3.80416274e-01 -2.14228466e-01 7.12368727e-01
6.83154285e-01 -2.37440452e-01 -1.93275183e-01 6.29113838e-02
-6.04238451e-01 -4.50651109e-01 -1.26646543e+00 9.39169586e-01
-2.06588447e-01 7.23581553e-01 5.96382469e-02 3.44430774e-01
1.61971056e+00 2.98980922e-01 4.11520422e-01 -6.51357412e-01
2.84713477e-01 4.17592227e-01 -4.02619600e-01 -7.37926841e-01
3.73813152e-01 -8.49535763e-01 6.35921419e-01 6.29422307e-01
-5.34932673e-01 -9.02023435e-01 7.92085286e-03 -5.19096255e-01
6.10828817e-01 2.62691647e-01 5.80725610e-01 -4.19083297e-01
2.45379746e-01 -1.00813188e-01 -1.88577399e-01 2.95759737e-01
-3.56662087e-02 5.58267891e-01 9.08835009e-02 -7.24160612e-01
-1.10052824e+00 -1.04501915e+00 -7.32129157e-01 3.61262560e-01
5.69976866e-01 4.55920905e-01 -6.64942920e-01 5.73569834e-01
4.29714710e-01 1.56968981e-01 -4.48822886e-01 3.90502214e-01
-2.82571375e-01 -1.90098703e-01 -2.35784635e-01 2.98110217e-01
3.67591321e-01 -9.39502120e-01 -1.47369659e+00 2.85710543e-01
7.00027287e-01 -6.87623382e-01 -4.04770523e-02 -1.44702435e-01
-1.30537677e+00 -1.27066565e+00 -9.73044157e-01 -6.37488365e-01
7.26790309e-01 -2.49918066e-02 1.01018155e+00 -5.88514784e-04
-3.77269834e-01 6.50773168e-01 1.66912004e-01 -1.02368081e+00
-3.67383033e-01 -7.20865130e-01 -2.77382061e-02 -3.07924420e-01
5.38880706e-01 -3.68589729e-01 -1.07053995e+00 1.64024562e-01
-5.88677943e-01 -4.05081660e-02 -4.73779440e-02 -1.20272234e-01
8.06048036e-01 -1.81102484e-01 -2.11821511e-01 -6.55269086e-01
8.33867371e-01 4.47240531e-01 -1.09965885e+00 4.25408408e-02
-7.48903871e-01 -2.68465877e-01 2.64887512e-01 -5.22834837e-01
-1.05637240e+00 -4.83659562e-03 4.57978606e-01 -1.65906668e-01
1.14873283e-01 1.18712761e-01 4.44111526e-01 -7.26141989e-01
8.77083242e-01 -2.40300253e-01 4.81849253e-01 -1.89665541e-01
-2.41907820e-01 6.36638045e-01 5.12581348e-01 9.54120234e-03
8.11009586e-01 8.13794672e-01 3.98216873e-01 -7.03028440e-01
-1.41131982e-01 -5.96046448e-03 -7.41231024e-01 -7.16242790e-01
7.76785374e-01 -5.55902958e-01 -1.35091293e+00 7.20962822e-01
-6.70495510e-01 -5.17332971e-01 -7.30894923e-01 7.54360139e-01
-4.83589649e-01 -2.18618304e-01 -6.57883406e-01 -8.32148969e-01
-3.43780369e-01 -7.34126449e-01 5.19330084e-01 8.58927727e-01
-3.49302202e-01 -1.43027830e+00 4.00934637e-01 1.51878834e-01
5.95306933e-01 4.95289564e-01 4.54130083e-01 2.74549872e-01
-6.18904471e-01 -3.96003455e-01 1.30150616e-01 1.99160516e-01
6.26426816e-01 6.56611085e-01 -1.04499555e+00 2.32312400e-02
6.02328300e-01 1.93709508e-01 1.20535120e-01 9.13131177e-01
6.61988080e-01 -2.05542013e-01 -2.29539901e-01 5.14035761e-01
2.09811473e+00 5.77892840e-01 5.20100236e-01 3.39653939e-01
4.43507701e-01 9.42833349e-02 1.82311252e-01 4.60599750e-01
-6.18108660e-02 2.14012966e-01 -2.23578457e-02 -2.92982221e-01
-3.54737416e-03 -1.75170034e-01 -2.77243674e-01 5.06119609e-01
-8.15623820e-01 3.04922014e-01 -7.99876869e-01 -1.83844760e-01
-7.65600145e-01 -6.88369036e-01 -2.71323532e-01 2.32877588e+00
1.07717955e+00 1.73021287e-01 -4.65631066e-03 1.96976960e-01
4.39883977e-01 -4.29349184e-01 -5.54478467e-01 -8.02960873e-01
2.16252580e-01 2.42968604e-01 8.71769786e-01 9.49669719e-01
-4.40182924e-01 2.37872526e-01 8.32712269e+00 -1.72977999e-01
-1.39279675e+00 -5.44212341e-01 1.94681764e-01 -1.15839608e-01
-6.31426156e-01 -7.53045157e-02 -6.41024709e-01 7.04664767e-01
3.82904798e-01 -6.25530779e-01 4.84279722e-01 2.75634050e-01
2.17696622e-01 -9.79887664e-01 -1.14695978e+00 8.01248848e-01
-8.01266134e-02 -1.19081235e+00 -1.50482655e-01 -4.15178686e-02
4.85286176e-01 -6.45675302e-01 -2.79296804e-02 -5.44294596e-01
-2.56988317e-01 -5.94230533e-01 5.73794067e-01 1.03540480e+00
8.54638159e-01 -8.72744322e-02 3.48352253e-01 1.15233688e-02
-9.44871962e-01 2.66348600e-01 -3.72090101e-01 -7.98493803e-01
1.15804382e-01 4.90566283e-01 -9.29742336e-01 -5.76368153e-01
5.94379842e-01 1.71523932e-02 -1.39119327e-01 1.17014253e+00
7.26549253e-02 3.01708907e-01 -6.72685325e-01 -5.08522570e-01
-4.62909997e-01 -1.00398529e+00 5.99978089e-01 6.72761917e-01
4.63968553e-02 4.70098704e-01 -5.55196166e-01 1.15757954e+00
3.11805964e-01 2.02708304e-01 -4.81876552e-01 1.92171693e-01
2.82352060e-01 7.67835438e-01 -1.13546133e+00 6.89426810e-02
-7.17277586e-01 5.29953778e-01 -5.34889162e-01 5.53844392e-01
-3.63583118e-01 -4.52367038e-01 1.31984010e-01 6.26943827e-01
9.69727039e-02 -7.20620081e-02 -9.16900754e-01 -7.15416968e-01
1.59804061e-01 2.53818125e-01 -1.88506380e-01 -1.20637834e+00
-8.98158014e-01 4.21051942e-02 -1.71285957e-01 -1.38022268e+00
4.89623815e-01 -9.93622899e-01 -4.51420248e-01 1.02599955e+00
-1.38568962e+00 -5.94693542e-01 -3.81753117e-01 6.10011756e-01
2.06968322e-01 -8.43792334e-02 9.36433136e-01 -3.74433994e-01
1.64661527e-01 8.46887007e-02 2.43501246e-01 -1.80235147e-01
5.38420141e-01 -1.39845908e+00 -1.70944914e-01 3.61383945e-01
-6.98761523e-01 3.88924032e-01 8.50112498e-01 -7.79184401e-01
-1.55091178e+00 -3.85896146e-01 6.64781153e-01 -5.93057632e-01
6.80036664e-01 8.70361645e-03 -9.87441480e-01 6.80687666e-01
1.01087689e-01 1.90908462e-01 7.63237476e-01 -4.84342009e-01
2.74071336e-01 -2.65947431e-01 -1.61460555e+00 4.70033735e-01
2.94384181e-01 -5.14649868e-01 -7.60186136e-01 2.70052791e-01
4.18569416e-01 -1.18313050e+00 -1.18408751e+00 -2.20966116e-02
1.33215904e+00 -9.37038243e-01 7.00660646e-01 2.40915075e-01
3.61133754e-01 -4.02797610e-01 3.03224057e-01 -7.88967788e-01
-1.74592644e-01 -7.63614833e-01 9.76894498e-02 6.82151914e-01
3.74079734e-01 -1.05551600e+00 6.92172289e-01 1.57917964e+00
3.63003761e-01 -6.46978557e-01 -6.79269254e-01 -5.63359976e-01
-1.59408554e-01 2.86829084e-01 3.45708221e-01 1.53046548e-01
4.89391357e-01 1.38023496e-01 1.72258735e-01 -3.00176814e-02
6.51194930e-01 3.62192512e-01 5.46791255e-01 -1.65908396e+00
-2.17498410e-02 -1.73329040e-02 -4.75366116e-01 -5.90266347e-01
-7.83147156e-01 -1.81470752e-01 -2.78627753e-01 -1.68047905e+00
1.00765899e-01 -2.57561684e-01 1.74202040e-01 -4.80851009e-02
1.88139100e-02 1.64343938e-01 9.92409810e-02 5.12014329e-01
2.42415354e-01 -1.97421312e-01 1.98763800e+00 2.54711360e-01
-8.92815888e-01 2.83041537e-01 -2.70722538e-01 9.93468761e-01
4.49593127e-01 -2.66755849e-01 -2.63511688e-01 -2.67339766e-01
6.89906716e-01 3.72508973e-01 2.46228084e-01 -1.14266622e+00
1.10045038e-01 -3.87487352e-01 8.27215672e-01 -3.73991489e-01
9.55330580e-02 -1.65290058e+00 5.40522456e-01 6.82807028e-01
-1.89911410e-01 -1.79221138e-01 3.36863026e-02 4.14257079e-01
3.01156431e-01 -3.08523089e-01 8.98701668e-01 -3.81732553e-01
-2.77907848e-01 8.91907886e-03 -4.96005207e-01 -3.20390791e-01
1.35600352e+00 -1.23752439e+00 -3.02102208e-01 1.91812068e-02
-5.94730020e-01 -2.57959872e-01 1.19697607e+00 -2.08539635e-01
7.27860272e-01 -1.12144709e+00 -2.07953662e-01 6.88550174e-01
-2.43422896e-01 1.26074301e-02 -1.37784392e-01 4.77776468e-01
-1.57500148e+00 1.23989256e-03 -5.62619805e-01 -5.81244111e-01
-1.17786372e+00 3.36307526e-01 7.55581498e-01 2.29666799e-01
-7.37817049e-01 5.62867701e-01 -2.71731824e-01 1.84152201e-01
-1.30819470e-01 -8.38000000e-01 -3.65256369e-01 -3.79390746e-01
7.80034363e-01 7.27774978e-01 -2.26151064e-01 -1.84149995e-01
-1.36335537e-01 1.39151692e+00 7.68212199e-01 -1.77187532e-01
1.13530898e+00 -3.38629782e-01 -1.26111284e-01 1.18250310e+00
6.87149167e-01 3.62098962e-01 -1.71064794e+00 1.33389890e-01
-8.22396755e-01 -8.36272120e-01 -2.63715059e-01 -6.78714514e-01
-9.27566648e-01 4.11476791e-01 8.51945162e-01 5.03965437e-01
8.93230677e-01 -1.12129986e-01 4.21380699e-01 1.75673410e-01
3.10560942e-01 -1.26028299e+00 -3.05626839e-02 1.53524891e-01
9.61413026e-01 -1.08372176e+00 4.76400435e-01 -4.14418757e-01
-4.41684663e-01 1.55217075e+00 4.14711684e-01 -4.23410714e-01
1.11677372e+00 1.14143896e+00 5.58165491e-01 -3.20023566e-01
-6.79766536e-02 3.11814427e-01 1.21974029e-01 6.37819886e-01
4.73709643e-01 2.68165231e-01 -6.95232332e-01 1.69477835e-02
-1.52817145e-01 9.52933967e-01 7.83298850e-01 1.16929197e+00
-1.05999923e+00 -2.62398630e-01 -6.35099173e-01 6.37883186e-01
-4.92132038e-01 4.38116819e-01 -1.18013367e-01 4.66031194e-01
-7.54052252e-02 7.13964224e-01 6.34492636e-01 2.16382906e-01
6.79957092e-01 -1.72510386e-01 7.67436922e-01 -3.27169508e-01
-3.02661002e-01 3.23052198e-01 -4.12300199e-01 -4.00999874e-01
-9.01706338e-01 -4.46212083e-01 -1.64034975e+00 -2.26659864e-01
-1.71522453e-01 -2.45893359e-01 7.92425752e-01 6.58448577e-01
-3.59553307e-01 1.24979811e-02 4.01213676e-01 -7.48247743e-01
-7.82149360e-02 -6.89374864e-01 -1.28478456e+00 1.77876085e-01
4.08952564e-01 -3.97656739e-01 -6.55653238e-01 5.00445068e-01] | [13.090763092041016, -2.9535398483276367] |
7a362d84-c68f-4473-a68f-fb61ae417f56 | semantic-sentence-similarity-size-does-not | 2106.08648 | null | https://arxiv.org/abs/2106.08648v1 | https://arxiv.org/pdf/2106.08648v1.pdf | Semantic sentence similarity: size does not always matter | This study addresses the question whether visually grounded speech recognition (VGS) models learn to capture sentence semantics without access to any prior linguistic knowledge. We produce synthetic and natural spoken versions of a well known semantic textual similarity database and show that our VGS model produces embeddings that correlate well with human semantic similarity judgements. Our results show that a model trained on a small image-caption database outperforms two models trained on much larger databases, indicating that database size is not all that matters. We also investigate the importance of having multiple captions per image and find that this is indeed helpful even if the total number of images is lower, suggesting that paraphrasing is a valuable learning signal. While the general trend in the field is to create ever larger datasets to train models on, our findings indicate other characteristics of the database can just as important important. | ['Mirjam Ernestus', 'Stefan L. Frank', 'Danny Merkx'] | 2021-06-16 | null | null | null | null | ['learning-semantic-representations', 'grounded-language-learning'] | ['methodology', 'natural-language-processing'] | [ 3.76120657e-01 3.50576073e-01 -4.25843149e-03 -5.35373926e-01
-1.01919758e+00 -5.48676431e-01 8.79995167e-01 3.13550442e-01
-7.07189381e-01 5.80031574e-01 8.34359229e-01 -3.77680331e-01
2.45270625e-01 -5.87049246e-01 -7.93311059e-01 -3.99906754e-01
3.68165582e-01 5.40678740e-01 4.57964003e-01 -5.20469844e-01
5.44432044e-01 1.31995022e-01 -1.56604707e+00 4.14041579e-01
5.40628731e-01 7.54122913e-01 5.27140737e-01 5.81288636e-01
-2.30890930e-01 1.06421781e+00 -7.79760540e-01 -5.57659030e-01
1.51205584e-01 -6.81909740e-01 -9.72586036e-01 1.66398928e-01
8.41874599e-01 -1.42456934e-01 -4.28019345e-01 1.00884688e+00
5.33052862e-01 1.83162555e-01 6.81244493e-01 -1.12935758e+00
-1.24945593e+00 2.53505021e-01 1.97662599e-02 4.44964230e-01
5.49138904e-01 4.30146307e-01 1.11054361e+00 -7.69264758e-01
8.54064226e-01 1.29773939e+00 7.49648511e-01 4.16668206e-01
-1.11093998e+00 -2.73647457e-01 3.08806114e-02 2.67423660e-01
-1.31341779e+00 -7.08971381e-01 5.43758512e-01 -3.94636661e-01
1.15516531e+00 1.56551719e-01 6.04541183e-01 1.21676457e+00
-7.36678764e-02 5.28997779e-01 1.15394056e+00 -5.46352386e-01
3.72030258e-01 5.74152589e-01 -1.01767592e-02 6.14630818e-01
5.43284178e-01 -1.65180176e-01 -7.43071437e-01 -1.45076498e-01
6.69636190e-01 -3.49778295e-01 -4.20985103e-01 -5.83312690e-01
-1.42381358e+00 1.06653678e+00 6.04733765e-01 5.28971851e-01
-3.80011111e-01 1.66999698e-01 5.31658053e-01 4.96239424e-01
2.81489611e-01 7.75528491e-01 -5.80114918e-03 -2.87600070e-01
-1.03807545e+00 6.77192351e-03 8.12334836e-01 7.71403491e-01
6.56427801e-01 2.16148943e-01 6.20508604e-02 8.49093437e-01
1.90708831e-01 4.38260823e-01 7.05471098e-01 -1.04337835e+00
4.99411643e-01 6.16597950e-01 1.76246688e-01 -1.26771355e+00
2.05482952e-02 6.17356598e-03 -1.39193401e-01 8.97017717e-02
5.95507264e-01 2.62708545e-01 -7.50869930e-01 1.78828883e+00
-3.55001032e-01 -5.94231449e-02 3.03015172e-01 1.25147069e+00
9.64141548e-01 4.36783016e-01 3.82424444e-01 6.22787140e-02
1.38034225e+00 -6.84176505e-01 -4.41557080e-01 -8.64571810e-01
5.64254820e-01 -6.39648676e-01 1.63140559e+00 -3.36726978e-02
-9.23874736e-01 -5.03655016e-01 -1.27935684e+00 -1.43183783e-01
-5.65105736e-01 -2.89276451e-01 6.14538550e-01 7.33091354e-01
-1.35308087e+00 2.89170921e-01 -3.47512305e-01 -1.07990718e+00
3.30975562e-01 -1.26130775e-01 -5.83008647e-01 -2.17884198e-01
-1.08423781e+00 1.48930538e+00 2.03866512e-01 -2.63285816e-01
-7.94028342e-01 -8.84498581e-02 -1.06155014e+00 -1.34498462e-01
1.29366949e-01 -6.59009635e-01 1.21615779e+00 -1.42523611e+00
-8.47261608e-01 1.38434303e+00 -3.00202906e-01 -5.04527390e-01
2.31719077e-01 3.44657868e-01 -3.51558417e-01 5.33240259e-01
2.82278866e-01 8.34400773e-01 6.65731788e-01 -1.41551769e+00
-1.17368594e-01 -2.95709401e-01 7.13708252e-02 3.82866263e-01
-4.42651004e-01 1.02297202e-01 -5.44677349e-03 -6.09623194e-01
-1.53824762e-01 -6.38282120e-01 4.10731323e-03 1.81649372e-01
1.87224776e-01 -8.19493607e-02 3.73326659e-01 -7.16161549e-01
7.28753865e-01 -2.14772844e+00 -1.58155560e-01 -1.72262266e-01
-8.08086153e-03 1.27944320e-01 -5.46184838e-01 7.38685131e-01
2.98387744e-02 2.29277849e-01 -2.30727568e-01 -1.53759226e-01
6.89055175e-02 4.25438195e-01 -2.95148075e-01 4.37537283e-01
2.62525678e-01 1.21757233e+00 -1.05259335e+00 -7.66298354e-01
2.38252774e-01 2.52469599e-01 -3.08180779e-01 -1.58813506e-01
-2.64575444e-02 2.16243211e-02 -1.76869825e-01 2.46831402e-01
3.05234134e-01 -4.45966929e-01 9.81577560e-02 2.27249373e-04
9.71654132e-02 4.85854566e-01 -9.05657709e-01 1.74242520e+00
-3.41101617e-01 1.05682349e+00 -1.12282865e-01 -1.25048697e+00
9.03700948e-01 4.31469738e-01 2.90953945e-02 -1.17549622e+00
5.53062037e-02 2.33189121e-01 2.63726804e-02 -6.31254256e-01
6.55137718e-01 -6.63567126e-01 5.94903119e-02 5.39307654e-01
2.61989404e-02 -4.93773729e-01 -2.23101899e-02 4.38200504e-01
1.08766592e+00 -1.85870603e-01 2.52501845e-01 -2.68338740e-01
2.76979476e-01 5.99182844e-01 1.07781194e-01 8.65322351e-01
-4.68437016e-01 8.61264408e-01 3.01038802e-01 -1.77113995e-01
-1.22823238e+00 -1.17866337e+00 1.23660900e-02 9.98844266e-01
2.65581459e-01 -3.13724577e-01 -6.93542659e-01 -5.12233555e-01
-1.04224518e-01 1.14141631e+00 -6.57957792e-01 -3.28667730e-01
-7.16664121e-02 -3.82377625e-01 4.75007176e-01 7.21322477e-01
2.34836534e-01 -1.35547364e+00 -8.31984103e-01 -2.04473417e-02
-2.06830099e-01 -1.17624462e+00 -3.74359161e-01 -7.00397938e-02
-6.61013722e-01 -9.00547326e-01 -8.27180564e-01 -1.12707090e+00
6.20189190e-01 6.01249576e-01 1.39196849e+00 3.14433128e-02
-2.11799935e-01 8.65542591e-01 -5.91002345e-01 -5.10364413e-01
-6.80596948e-01 -2.72465646e-01 -1.77178204e-01 -3.48623574e-01
7.09613621e-01 -2.91774482e-01 -5.07059216e-01 2.60855496e-01
-8.06490541e-01 -7.68441334e-02 4.51145023e-01 8.43528628e-01
1.29248321e-01 -4.04757321e-01 8.28187227e-01 -5.84301651e-01
8.62814844e-01 -2.90316403e-01 -8.79193246e-02 3.74349713e-01
-3.43606293e-01 1.64280370e-01 2.02731326e-01 -3.83463472e-01
-7.59556413e-01 -9.26129594e-02 1.29800081e-01 -2.84570932e-01
-2.90213019e-01 3.75007182e-01 8.96534696e-02 -7.67134354e-02
9.98839080e-01 3.92750412e-01 3.85065496e-01 -1.03907898e-01
3.71799082e-01 8.11393261e-01 4.86629993e-01 -3.17187577e-01
7.08247602e-01 5.22332549e-01 -4.27090019e-01 -1.13894832e+00
-6.30627811e-01 -4.92046505e-01 -4.09042895e-01 -9.91157070e-02
9.81531858e-01 -1.04349232e+00 -2.83485204e-01 -1.27020761e-01
-1.15344238e+00 -3.09484780e-01 -5.37533283e-01 5.98329604e-01
-8.61196518e-01 3.77028346e-01 -3.96699846e-01 -7.02010095e-01
4.38606814e-02 -1.04685569e+00 9.58473682e-01 -4.83223870e-02
-7.01738060e-01 -1.04566026e+00 2.48702578e-02 7.07376719e-01
5.20559847e-01 1.14409335e-01 7.93365836e-01 -9.92115319e-01
-4.88112867e-01 -2.22334743e-01 -4.14292067e-01 3.55735093e-01
2.59925157e-01 -5.55504501e-01 -9.91917610e-01 -1.98871091e-01
1.77651867e-01 -7.81157553e-01 8.08295310e-01 4.66327444e-02
5.52574456e-01 -2.44735599e-01 6.52426779e-02 -4.40036058e-02
1.59568298e+00 1.90282747e-01 7.30970502e-01 5.48645556e-01
4.54803735e-01 8.10556412e-01 3.90919268e-01 -1.71252973e-02
6.46512568e-01 5.86911440e-01 2.08476976e-01 -9.76428539e-02
-3.49208891e-01 -5.32420158e-01 4.65162575e-01 6.54027879e-01
2.64624000e-01 -3.19768101e-01 -1.12407589e+00 9.75322247e-01
-1.63926339e+00 -1.25824153e+00 1.65275052e-01 2.22300053e+00
7.19663024e-01 1.60243675e-01 1.96890488e-01 1.07485443e-01
6.03408933e-01 3.80392134e-01 -3.33248585e-01 -6.29477143e-01
-3.58660132e-01 1.72470570e-01 4.17762488e-01 5.08005917e-01
-5.91099441e-01 9.96898711e-01 7.31502390e+00 4.03135240e-01
-1.03476763e+00 8.03535804e-02 3.82839203e-01 2.14684429e-03
-6.64957821e-01 -2.78362948e-02 -1.29706904e-01 4.48720187e-01
1.22304738e+00 -2.36849621e-01 2.79711723e-01 6.87175810e-01
2.50118643e-01 -3.98822337e-01 -1.13655102e+00 1.07746351e+00
6.39282584e-01 -1.39491332e+00 3.95507127e-01 -5.65089397e-02
3.37906957e-01 9.28146169e-02 -6.77629411e-02 -5.62131545e-03
2.16593668e-01 -1.38121951e+00 1.00986195e+00 5.09132206e-01
6.30338609e-01 -4.76857364e-01 6.47357821e-01 3.03745925e-01
-7.65181124e-01 1.88963205e-01 -4.94786829e-01 -1.10284433e-01
1.82846919e-01 -1.28683165e-01 -1.02130032e+00 1.33090660e-01
5.20633280e-01 5.15755594e-01 -1.24554682e+00 1.00472891e+00
1.41436644e-02 5.37691474e-01 -2.29736418e-02 -3.93464088e-01
3.45106155e-01 6.27910718e-02 3.37526292e-01 1.22308886e+00
6.14616685e-02 3.40215750e-02 -3.33005637e-01 6.12621546e-01
2.32473716e-01 1.63720995e-01 -1.14151967e+00 -5.18374205e-01
4.51318145e-01 7.88123071e-01 -9.16653156e-01 -3.82556379e-01
-9.93396819e-01 1.25207210e+00 2.40585357e-01 3.73220831e-01
-4.52725798e-01 -3.72378409e-01 3.90779793e-01 3.12797308e-01
4.11373377e-01 -3.37455064e-01 -4.26416993e-01 -1.07085943e+00
4.56331074e-02 -1.04246974e+00 2.05692559e-01 -1.44354463e+00
-1.37170589e+00 3.89169365e-01 -1.86023161e-01 -9.80004728e-01
-3.57703358e-01 -7.19375968e-01 -4.32200342e-01 7.90797949e-01
-1.28053188e+00 -9.90351081e-01 -3.98049206e-01 5.66493988e-01
5.73886275e-01 -1.79913238e-01 8.39821517e-01 -1.42481789e-01
1.49544151e-02 5.57326376e-01 -8.93103406e-02 3.08649808e-01
7.57284999e-01 -9.86640632e-01 5.48365533e-01 5.68278968e-01
7.74295211e-01 8.27327430e-01 9.32105660e-01 -4.20303613e-01
-1.21166050e+00 -4.53205526e-01 1.14687288e+00 -9.27672327e-01
6.16544485e-01 -4.46373343e-01 -9.76863921e-01 6.73421264e-01
6.38224661e-01 -2.73112297e-01 5.73457956e-01 -1.10762045e-01
-5.24272084e-01 2.60552704e-01 -1.11213446e+00 6.25150800e-01
1.27732229e+00 -1.08671474e+00 -1.31141245e+00 2.55233794e-01
6.44539177e-01 5.93233779e-02 -5.65020144e-01 -4.14477475e-02
1.82686985e-01 -9.29952919e-01 1.14064515e+00 -6.74134016e-01
5.16551912e-01 -2.04312533e-01 -5.84833205e-01 -1.31460953e+00
-1.38955131e-01 -1.03270654e-02 5.85800171e-01 1.16316116e+00
4.90635991e-01 -6.97974682e-01 8.71881008e-01 7.04779029e-01
-2.45371852e-02 -1.85544044e-01 -1.04525137e+00 -9.72921550e-01
1.11281022e-01 -4.45921868e-01 3.62771243e-01 1.17206442e+00
1.37487531e-01 6.68397129e-01 7.93146268e-02 -9.24732164e-02
4.15228635e-01 -1.18389525e-01 6.77056968e-01 -1.10016739e+00
1.18939318e-02 -3.38414818e-01 -8.38780999e-01 -6.26788080e-01
4.07712251e-01 -9.86995578e-01 1.98481511e-02 -1.98546433e+00
3.18396807e-01 -4.79073599e-02 -1.23520941e-01 3.73187244e-01
-6.35506399e-03 4.16639805e-01 2.33892187e-01 1.72106460e-01
-6.37197793e-01 4.40759450e-01 1.14169073e+00 -2.01081619e-01
1.79968461e-01 -5.33224106e-01 -1.20026124e+00 5.38019121e-01
6.11300707e-01 -2.84094453e-01 -8.05879593e-01 -5.83234787e-01
6.19625300e-02 -1.54176742e-01 8.84099066e-01 -9.75295663e-01
2.23301291e-01 -1.93871379e-01 3.83436233e-01 -1.66264921e-02
5.08252382e-01 -6.51986599e-01 -1.12354830e-01 1.96433142e-01
-5.81371963e-01 2.11436197e-01 4.31868613e-01 6.63428605e-01
-3.43394160e-01 -4.98536229e-01 7.70487189e-01 -4.31073725e-01
-1.16028297e+00 -3.00716817e-01 -4.82481211e-01 4.98988330e-01
7.75190115e-01 -8.15584183e-01 -3.89132231e-01 -8.72027040e-01
-4.42916453e-01 -9.69890952e-02 1.07302833e+00 7.62853444e-01
6.84130967e-01 -1.32388997e+00 -5.16260862e-01 -9.08254758e-02
7.19374955e-01 -5.56873202e-01 -2.48440012e-01 4.70619202e-01
-4.18523103e-01 6.17266119e-01 -1.63297057e-01 -3.77025872e-01
-1.06472361e+00 5.47796249e-01 2.06956208e-01 6.88055217e-01
-6.71696424e-01 8.42628360e-01 1.20649841e-02 -1.97696239e-01
1.21077985e-01 -1.63615108e-01 -9.44907665e-02 2.37298205e-01
4.51959014e-01 -6.17010072e-02 -7.43104517e-02 -1.04435802e+00
-5.65219581e-01 5.39533556e-01 1.57708451e-01 -5.16271234e-01
1.44894159e+00 -2.28924647e-01 3.13721865e-01 6.56499505e-01
1.20302951e+00 -1.58185497e-01 -1.04396582e+00 -2.27183849e-01
2.29813874e-01 -6.38982773e-01 -5.38149998e-02 -6.21857166e-01
-6.20703638e-01 9.35853839e-01 4.98149872e-01 5.80623522e-02
7.18969345e-01 4.79087472e-01 5.64649045e-01 4.13073093e-01
4.89807487e-01 -1.23130929e+00 4.69773322e-01 3.08466464e-01
1.09177279e+00 -1.41138864e+00 -1.43740755e-02 4.94090356e-02
-1.29617345e+00 6.43969119e-01 3.53096634e-01 -1.50781542e-01
1.58916309e-01 -2.16251850e-01 1.48395613e-01 -3.73423904e-01
-5.72023749e-01 -4.18103486e-01 5.57249272e-03 9.47668076e-01
4.85538632e-01 -3.77085477e-01 -1.48795024e-01 2.66699791e-01
-5.58559954e-01 -8.05529281e-02 5.86047828e-01 8.99728239e-01
-6.00184739e-01 -8.82689536e-01 -3.86006027e-01 4.06146824e-01
5.98280802e-02 -2.87775725e-01 -8.75264764e-01 8.49113286e-01
-2.18862534e-01 1.11226273e+00 3.19523752e-01 -1.90040320e-01
4.30870771e-01 3.80185425e-01 6.06033623e-01 -8.61739278e-01
-4.08472896e-01 -6.52994871e-01 3.23885173e-01 -4.25826788e-01
-4.76505339e-01 -6.73546851e-01 -1.06051123e+00 -2.48436674e-01
-6.77723289e-02 2.16874853e-01 5.30203938e-01 8.42044711e-01
1.90388665e-01 -6.33308515e-02 1.64689764e-01 -5.56685567e-01
-4.77622032e-01 -8.33020747e-01 -4.71977860e-01 7.96866477e-01
3.82423073e-01 -5.15563905e-01 -5.88679612e-01 1.58126250e-01] | [10.901494979858398, 1.716861605644226] |
11235a48-75b8-464a-a84c-99b589217f03 | effective-data-augmentation-with-multi-domain | 1912.11597 | null | https://arxiv.org/abs/1912.11597v1 | https://arxiv.org/pdf/1912.11597v1.pdf | Effective Data Augmentation with Multi-Domain Learning GANs | For deep learning applications, the massive data development (e.g., collecting, labeling), which is an essential process in building practical applications, still incurs seriously high costs. In this work, we propose an effective data augmentation method based on generative adversarial networks (GANs), called Domain Fusion. Our key idea is to import the knowledge contained in an outer dataset to a target model by using a multi-domain learning GAN. The multi-domain learning GAN simultaneously learns the outer and target dataset and generates new samples for the target tasks. The simultaneous learning process makes GANs generate the target samples with high fidelity and variety. As a result, we can obtain accurate models for the target tasks by using these generated samples even if we only have an extremely low volume target dataset. We experimentally evaluate the advantages of Domain Fusion in image classification tasks on 3 target datasets: CIFAR-100, FGVC-Aircraft, and Indoor Scene Recognition. When trained on each target dataset reduced the samples to 5,000 images, Domain Fusion achieves better classification accuracy than the data augmentation using fine-tuned GANs. Furthermore, we show that Domain Fusion improves the quality of generated samples, and the improvements can contribute to higher accuracy. | ["Shin'ya Yamaguchi", 'Sekitoshi Kanai', 'Takeharu Eda'] | 2019-12-25 | null | null | null | null | ['scene-recognition'] | ['computer-vision'] | [ 4.05517548e-01 -6.72534630e-02 -7.88516700e-02 -3.89353901e-01
-9.69702780e-01 -5.55228770e-01 5.48251629e-01 -2.99299419e-01
-2.68571705e-01 1.07664979e+00 -2.09144130e-02 -5.65335061e-03
4.01159376e-01 -1.20610833e+00 -9.29567873e-01 -8.96525741e-01
6.11920595e-01 5.49124658e-01 -1.08935580e-01 -5.53425774e-02
-3.50767702e-01 4.18445289e-01 -1.39277017e+00 4.24229294e-01
1.27801096e+00 1.25761366e+00 1.40440717e-01 5.31407297e-01
-2.61860758e-01 8.39501560e-01 -1.10145080e+00 -4.22461540e-01
3.95499796e-01 -5.60013473e-01 -4.82239574e-01 2.23250121e-01
2.42142886e-01 -5.97011864e-01 -8.62215757e-02 1.01530850e+00
2.99508870e-01 -1.60637237e-02 6.92999244e-01 -1.52475190e+00
-8.61017406e-01 3.63198787e-01 -5.02877891e-01 -2.24792778e-01
-1.84379190e-01 2.15222448e-01 5.06292641e-01 -5.63788474e-01
2.82071024e-01 1.04413199e+00 4.98890013e-01 8.13450396e-01
-1.11210895e+00 -1.11219645e+00 -9.14450139e-02 -1.28466323e-01
-1.16520190e+00 -3.10998350e-01 9.92250502e-01 -4.91258502e-01
8.86750817e-02 1.05841339e-01 5.20029008e-01 1.36341345e+00
-6.17095083e-02 8.60644281e-01 1.22085416e+00 -1.43467113e-01
2.25105867e-01 2.08857730e-01 -2.28449434e-01 3.74818593e-01
2.59987444e-01 1.84754223e-01 -2.08943084e-01 -1.10813389e-02
1.08860731e+00 4.04773861e-01 -2.56755382e-01 -6.12957142e-02
-1.11855996e+00 9.71382916e-01 5.64199626e-01 1.60184413e-01
-4.48484272e-01 6.34120479e-02 3.67672116e-01 3.56851876e-01
6.98624730e-01 3.34654421e-01 -3.58877450e-01 6.72856867e-02
-6.61323249e-01 7.76318237e-02 3.84087235e-01 1.10164034e+00
9.55676854e-01 4.70406502e-01 -2.86377013e-01 9.37937737e-01
1.99903235e-01 7.56248951e-01 6.68161750e-01 -7.56303847e-01
5.56617856e-01 7.41785347e-01 -1.59449726e-02 -6.25026584e-01
1.40637338e-01 -7.49249041e-01 -1.39087164e+00 1.27943203e-01
3.07578444e-01 -4.32649314e-01 -1.14684916e+00 1.87312496e+00
3.61677229e-01 3.16170722e-01 4.32302237e-01 5.57257295e-01
9.11742747e-01 8.97252679e-01 8.72683078e-02 -6.27969131e-02
1.02679825e+00 -1.14279842e+00 -8.27468336e-01 -3.85790050e-01
5.01871109e-01 -5.57110608e-01 1.14727557e+00 3.54911208e-01
-7.02633142e-01 -1.06254482e+00 -1.11994493e+00 6.76724836e-02
-2.17263877e-01 4.04625088e-01 5.91976881e-01 5.67557275e-01
-6.21805847e-01 2.74556011e-01 -4.50228661e-01 2.11087734e-01
9.92492616e-01 2.71862894e-01 -4.78056341e-01 -5.08565247e-01
-1.08256924e+00 4.31757718e-01 4.74385411e-01 -2.40281448e-01
-1.41120422e+00 -8.12454879e-01 -8.46973479e-01 -1.47411510e-01
1.76522240e-01 -6.42233610e-01 1.23745596e+00 -1.40640545e+00
-1.36674368e+00 6.72950268e-01 1.38208359e-01 -4.70605671e-01
5.10444582e-01 -1.89554691e-01 -4.76027280e-01 -1.01973854e-01
2.02833250e-01 6.97105110e-01 1.13496816e+00 -1.43176937e+00
-5.20306408e-01 -2.68719465e-01 4.11212025e-03 5.31843603e-02
-5.48344314e-01 -3.97944719e-01 -1.03237763e-01 -9.25924957e-01
-3.28313977e-01 -7.60632038e-01 -1.68879002e-01 -1.63242176e-01
-2.03274682e-01 2.57013962e-02 1.13993657e+00 -7.46393144e-01
6.72420681e-01 -2.45948005e+00 2.49025635e-02 -7.58649409e-02
3.84153128e-01 4.60769981e-01 -2.70075738e-01 -8.80801380e-02
-1.89431719e-02 1.62163734e-01 -3.88186008e-01 -3.97159934e-01
-3.71503979e-01 4.99471098e-01 -4.50857252e-01 1.31816920e-02
3.37026924e-01 1.07003379e+00 -7.64131367e-01 -2.44915515e-01
1.78962231e-01 3.46121132e-01 -4.27135319e-01 6.68971658e-01
-3.61282319e-01 8.74242485e-01 -5.41403770e-01 5.08300126e-01
7.71309733e-01 -2.98992693e-01 -2.20945045e-01 -1.66102767e-01
4.98025864e-01 -2.26417810e-01 -8.04722846e-01 1.73889232e+00
-6.02252901e-01 3.75071406e-01 -1.18963927e-01 -1.06613040e+00
1.32811522e+00 2.84052044e-01 2.22421944e-01 -7.92779922e-01
8.30128044e-02 1.16899028e-01 -7.43080378e-02 1.36520080e-02
1.27707839e-01 -3.60851586e-01 -2.96841443e-01 3.39048266e-01
3.32922012e-01 -2.76153862e-01 -7.58678168e-02 7.63762593e-02
9.52998102e-01 -6.12752438e-02 2.23870680e-01 1.43668860e-01
3.78108025e-01 1.51759032e-02 8.94324899e-01 3.57826769e-01
4.95656952e-02 5.97120464e-01 3.47613931e-01 -4.13758188e-01
-1.41239130e+00 -1.01265597e+00 2.00997636e-01 7.00617909e-01
-1.10628784e-01 4.67880331e-02 -8.82325947e-01 -1.01467347e+00
-9.39640403e-02 6.02774918e-01 -7.27080226e-01 -5.57013810e-01
-3.90972435e-01 -7.66768634e-01 6.33981884e-01 7.86079526e-01
1.25370860e+00 -9.99588370e-01 -9.43922922e-02 9.21206027e-02
-2.71435738e-01 -1.24677563e+00 -3.23256135e-01 -1.52385356e-02
-9.28919911e-01 -8.96142066e-01 -8.91047537e-01 -7.11424291e-01
8.89719784e-01 6.63751643e-03 1.01937222e+00 -1.99845523e-01
6.12415336e-02 -6.75510801e-03 -5.90876043e-01 -6.20474517e-01
-7.34067798e-01 -1.67526919e-02 -5.63246086e-02 2.52751797e-01
1.03930287e-01 -5.26244104e-01 -2.26904899e-01 3.37206274e-01
-9.87325788e-01 5.40367365e-01 8.29069614e-01 1.19327438e+00
7.27673233e-01 2.43289992e-01 9.74859715e-01 -1.19362247e+00
4.58238363e-01 -5.29386699e-01 -6.19807899e-01 1.44800127e-01
-2.50810981e-01 -1.98164284e-02 1.06233919e+00 -5.61821580e-01
-1.36749899e+00 5.50664328e-02 -3.53775024e-01 -8.04090202e-01
-3.75230253e-01 2.22324669e-01 -6.09438777e-01 -6.97903931e-02
8.03748548e-01 3.61654937e-01 6.50468469e-02 -4.82316554e-01
3.81807864e-01 6.80455983e-01 5.80134690e-01 -6.09110713e-01
1.03928471e+00 2.43052095e-01 -1.59939677e-01 -5.08770704e-01
-9.32099402e-01 5.91831505e-02 -4.14531738e-01 -4.76334617e-02
9.24053729e-01 -1.20748639e+00 -2.21621633e-01 9.76969838e-01
-1.19541478e+00 -6.27968252e-01 -6.48630381e-01 3.23665589e-01
-3.68321329e-01 -1.20633364e-01 -4.82277453e-01 -5.65776885e-01
-3.72107297e-01 -1.04487085e+00 1.00081289e+00 3.84072810e-01
2.84442991e-01 -9.28939164e-01 -1.76558539e-01 4.75946546e-01
4.11311328e-01 6.89126670e-01 8.22531223e-01 -7.82000363e-01
-6.12775445e-01 -1.20686673e-01 -1.95847586e-01 1.04226828e+00
5.26697755e-01 -2.65618443e-01 -1.17733610e+00 -2.28326157e-01
2.66178399e-01 -6.18123591e-01 6.60025001e-01 1.64884061e-01
1.70498526e+00 -3.85418236e-01 -1.21331744e-01 8.45036924e-01
1.41578114e+00 6.95664704e-01 6.94862545e-01 -9.21100527e-02
9.58804727e-01 3.14150631e-01 8.54118168e-01 3.24129045e-01
1.70688182e-01 3.44788164e-01 4.77885127e-01 -5.25686145e-01
-1.91379786e-01 -3.93238693e-01 1.36360973e-01 7.88089871e-01
-4.80320267e-02 -2.62010038e-01 -8.86420906e-01 4.36522514e-01
-1.57086253e+00 -7.44063318e-01 1.43218592e-01 2.05077124e+00
1.07295084e+00 1.18438387e-02 -7.54377246e-02 1.98229522e-01
7.45808542e-01 -3.00015546e-02 -8.90297592e-01 8.86953101e-02
-1.41272232e-01 4.01894867e-01 3.82666290e-01 8.98140296e-02
-9.81654823e-01 7.81835616e-01 5.91929579e+00 1.03232646e+00
-1.12489307e+00 3.50957125e-01 1.10882223e+00 7.87862688e-02
-3.03631157e-01 -4.21326488e-01 -7.15398610e-01 7.97504902e-01
8.59283090e-01 -1.39212966e-01 3.85287613e-01 1.13220549e+00
-1.74507082e-01 2.15997726e-01 -1.13062561e+00 1.15480196e+00
6.16929196e-02 -1.49643373e+00 1.93829551e-01 2.44353279e-01
1.01858079e+00 -2.43366793e-01 8.32058415e-02 5.24211705e-01
6.98161244e-01 -1.06990874e+00 4.39752996e-01 4.50396687e-01
1.19181049e+00 -9.21695530e-01 1.00277925e+00 6.07347608e-01
-9.39512551e-01 -1.08368486e-01 -3.47146958e-01 1.36161566e-01
-9.11673624e-03 1.00479019e+00 -7.85674810e-01 7.82604277e-01
5.17655194e-01 7.24930584e-01 -4.99046028e-01 6.82808578e-01
-3.34740043e-01 6.80302799e-01 -4.23218384e-02 3.35953146e-01
-8.71056616e-02 -2.81313479e-01 3.38983648e-02 6.07397020e-01
3.31857592e-01 1.09361604e-01 1.69072658e-01 1.00073183e+00
-6.70570791e-01 -3.27071577e-01 -8.75126660e-01 -2.62254775e-01
5.16415834e-01 1.23190451e+00 -2.53445059e-01 -6.85216486e-01
-3.32154185e-01 1.11999321e+00 2.45216802e-01 3.19271773e-01
-1.06057465e+00 -3.25394124e-01 7.28848338e-01 -1.83296457e-01
-6.60072491e-02 -1.39313683e-01 -5.14958501e-01 -1.21004772e+00
7.99481943e-02 -1.12994134e+00 1.32189050e-01 -7.55246401e-01
-1.37742019e+00 7.04981625e-01 -3.16670716e-01 -1.38487685e+00
-3.38267803e-01 -4.03498530e-01 -5.48584521e-01 1.07540238e+00
-1.34719694e+00 -1.39672124e+00 -9.30743396e-01 7.64153004e-01
5.93453228e-01 -5.07758439e-01 9.22465503e-01 6.06352270e-01
-5.76826870e-01 7.35675454e-01 2.01564163e-01 4.98055637e-01
6.45244777e-01 -1.05491614e+00 4.67190593e-01 8.15752268e-01
2.58466303e-02 1.66401520e-01 9.67650488e-02 -6.16571844e-01
-9.69729662e-01 -1.66129017e+00 6.57212362e-02 -2.22949296e-01
1.82411045e-01 -5.41179478e-01 -9.53728318e-01 8.28035653e-01
2.14408070e-01 1.57819539e-01 9.27500665e-01 -3.28057110e-01
-1.93730846e-01 -4.09441590e-01 -1.50658786e+00 1.31177574e-01
8.86305690e-01 -4.67509538e-01 -4.78371501e-01 2.86901742e-01
1.08791578e+00 -5.09467959e-01 -1.01825655e+00 6.01900935e-01
7.56486952e-02 -6.84283793e-01 8.55963230e-01 -5.82084894e-01
7.48750448e-01 -3.66791874e-01 -1.82698324e-01 -1.77152538e+00
-1.50200635e-01 -3.32125984e-02 -8.38548169e-02 1.68318403e+00
2.08378717e-01 -8.70844483e-01 7.47302413e-01 3.93474251e-01
-2.49063358e-01 -5.51301658e-01 -5.70604205e-01 -8.20102870e-01
1.35816649e-01 -1.74807191e-01 1.17309964e+00 1.09464109e+00
-9.81992543e-01 3.61698478e-01 -5.46333730e-01 -8.25929567e-02
7.15505898e-01 6.01558797e-02 1.19806695e+00 -1.29233205e+00
-3.36325884e-01 1.11089081e-01 -2.30142042e-01 -9.40503299e-01
1.86640099e-01 -6.49901390e-01 -1.66363433e-01 -1.49196219e+00
9.76834595e-02 -9.30754185e-01 -1.36132792e-01 7.38880098e-01
-1.66949406e-01 3.07017654e-01 1.23713657e-01 2.51885951e-01
-1.83737442e-01 8.80481899e-01 1.64279509e+00 -2.37604111e-01
1.21176660e-01 3.03151142e-02 -1.00447416e+00 5.54500580e-01
9.57292318e-01 -3.25160712e-01 -5.95006287e-01 -5.59430301e-01
-3.38135362e-01 2.73893997e-02 3.75223339e-01 -1.24759519e+00
-1.20806739e-01 -2.20942169e-01 8.68884563e-01 -4.02338624e-01
3.98924768e-01 -9.90121305e-01 3.79637897e-01 3.96402955e-01
-1.06454551e-01 -2.72437394e-01 3.83595437e-01 4.52161402e-01
-6.15514934e-01 3.21708359e-02 1.07466006e+00 -1.11107968e-01
-7.57217109e-01 5.92788875e-01 2.50523567e-01 1.11416899e-01
1.41162503e+00 1.66122336e-02 -4.87694621e-01 -2.80407071e-01
-5.43585002e-01 8.24204385e-02 4.86905783e-01 3.84289324e-01
6.71896756e-01 -1.87209392e+00 -7.15299189e-01 7.12114573e-01
4.92395461e-02 7.79057205e-01 3.38325471e-01 2.35787719e-01
-2.49172628e-01 -3.03325374e-02 -5.90531349e-01 -5.50131738e-01
-1.13824320e+00 5.33107519e-01 2.82023162e-01 -3.83636326e-01
-1.44079268e-01 8.65243793e-01 6.74063146e-01 -4.78000909e-01
-2.02170178e-01 -2.00807974e-01 -2.32957695e-02 -1.29112244e-01
5.19376397e-01 1.40147418e-01 -1.23233935e-02 -2.62286872e-01
6.46754876e-02 2.89777488e-01 -1.03313230e-01 9.98789072e-02
1.31764376e+00 4.03145194e-01 1.64999627e-02 2.35869065e-01
1.17889953e+00 -5.20244576e-02 -1.44517803e+00 -3.37010831e-01
-6.90595031e-01 -6.64174557e-01 -1.48545951e-01 -9.04913247e-01
-1.47139895e+00 1.02901578e+00 5.18623829e-01 6.02398925e-02
1.57796073e+00 -8.34008902e-02 8.52352381e-01 5.86658455e-02
5.39107025e-01 -7.81107843e-01 3.89757663e-01 2.15499133e-01
9.40322518e-01 -1.38556743e+00 -4.20635521e-01 -4.32101220e-01
-8.75776827e-01 6.24440849e-01 1.14451373e+00 -1.11847594e-01
2.29330450e-01 3.36719960e-01 -3.84078473e-02 2.06972480e-01
-5.01875877e-01 1.09547205e-01 1.05301283e-01 9.08032775e-01
1.88793205e-02 2.56786048e-01 2.50858486e-01 7.87529469e-01
-1.83212176e-01 9.43311229e-02 2.17659250e-01 6.20975733e-01
-1.83853805e-01 -1.48775959e+00 -4.19109821e-01 9.04836655e-01
-3.71534467e-01 -6.29638880e-02 -3.84820819e-01 7.00194716e-01
4.42193389e-01 7.06426978e-01 1.14044592e-01 -6.36224926e-01
2.20751956e-01 -5.26165813e-02 3.61215085e-01 -7.42990792e-01
-3.61128807e-01 -2.06636980e-01 -8.34934562e-02 -2.17124999e-01
-2.40217999e-01 -3.67082506e-01 -1.10700333e+00 -2.99384922e-01
-1.75830022e-01 2.08999723e-01 5.99537253e-01 8.75217319e-01
4.42164630e-01 1.00198281e+00 8.48085582e-01 -4.48914945e-01
-3.82363260e-01 -1.05629289e+00 -5.46213269e-01 6.59386158e-01
2.68089116e-01 -7.15197444e-01 -2.43742853e-01 2.51901925e-01] | [11.63126277923584, -0.26114383339881897] |
4cdb928f-1a63-4621-a1ee-dcd2974867ab | entire-space-counterfactual-learning-tuning | 2210.11039 | null | https://arxiv.org/abs/2210.11039v2 | https://arxiv.org/pdf/2210.11039v2.pdf | Entire Space Counterfactual Learning: Tuning, Analytical Properties and Industrial Applications | As a basic research problem for building effective recommender systems, post-click conversion rate (CVR) estimation has long been plagued by sample selection bias and data sparsity issues. To address the data sparsity issue, prevalent methods based on entire space multi-task model leverage the sequential pattern of user actions, i.e. exposure $\rightarrow$ click $\rightarrow$ conversion to construct auxiliary learning tasks. However, they still fall short of guaranteeing the unbiasedness of CVR estimates. This paper theoretically demonstrates two defects of these entire space multi-task models: (1) inherent estimation bias (IEB) for CVR estimation, where the CVR estimate is inherently higher than the ground truth; (2) potential independence priority (PIP) for CTCVR estimation, where the causality from click to conversion might be overlooked. This paper further proposes a principled method named entire space counterfactual multi-task model (ESCM$^2$), which employs a counterfactual risk minimizer to handle both IEB and PIP issues at once. To demonstrate the effectiveness of the proposed method, this paper explores its parameter tuning in practice, derives its analytic properties, and showcases its effectiveness in industrial CVR estimation, where ESCM$^2$ can effectively alleviate the intrinsic IEB and PIP issues and outperform baseline models. | ['Xinggao Liu', 'Weiming Liu', 'Yuxin Huang', 'Jiajun Fan', 'Zhichao Chen', 'Hao Wang'] | 2022-10-20 | null | null | null | null | ['auxiliary-learning', 'selection-bias'] | ['methodology', 'natural-language-processing'] | [ 2.47980461e-01 -2.89143860e-01 -5.42479634e-01 -1.17444299e-01
-1.12397683e+00 -3.25485677e-01 4.51617211e-01 -3.80313963e-01
-2.48989522e-01 1.09348249e+00 1.25482604e-01 -7.06632555e-01
-5.22575319e-01 -5.17887950e-01 -8.42025220e-01 -5.56498528e-01
9.64416414e-02 1.36263268e-02 -3.47330719e-01 -4.93680462e-02
3.00264746e-01 -2.60762632e-01 -1.37251532e+00 7.05636367e-02
1.24246991e+00 1.10662878e+00 1.85677916e-01 1.95124313e-01
-1.64847299e-01 4.79020238e-01 -5.35286665e-01 -7.24892735e-01
5.28460622e-01 -8.61811563e-02 -3.00137103e-01 -2.42351726e-01
3.32788169e-01 -2.76486367e-01 -8.24363381e-02 1.08090413e+00
3.64938289e-01 2.78215230e-01 7.92622089e-01 -1.59450805e+00
-5.96683562e-01 4.91919130e-01 -1.04984653e+00 1.21093333e-01
-7.65327960e-02 -3.02344654e-02 1.40941691e+00 -9.53756273e-01
1.40961483e-01 1.13131630e+00 8.34291577e-01 6.64022788e-02
-1.33356941e+00 -1.07535422e+00 4.86878783e-01 -2.52509296e-01
-1.49112070e+00 -1.80281743e-01 6.96429849e-01 -4.89784062e-01
4.48527455e-01 5.07261872e-01 3.14206213e-01 1.48766363e+00
4.47795928e-01 1.04712152e+00 1.41997039e+00 -8.68921056e-02
2.28511110e-01 5.48597336e-01 3.57540578e-01 1.69045776e-01
5.26517510e-01 4.51072335e-01 -6.08632445e-01 -4.23741311e-01
9.24308836e-01 2.88968444e-01 -5.93000054e-02 -2.92718470e-01
-9.38468516e-01 1.04908264e+00 6.88472241e-02 -1.42625272e-01
-2.22541377e-01 1.63074717e-01 2.64832884e-01 4.71172571e-01
6.21922016e-01 2.35929340e-01 -6.36952817e-01 5.10699116e-03
-1.18898523e+00 4.81671631e-01 5.25272548e-01 1.32233047e+00
4.39679623e-01 2.67479450e-01 -6.18479371e-01 8.46953094e-01
4.88856196e-01 7.22960114e-01 3.77083719e-01 -7.14056671e-01
9.71794844e-01 1.88992664e-01 5.70472062e-01 -9.44913566e-01
-6.10942356e-02 -8.93371701e-01 -1.00387681e+00 -1.24077819e-01
3.80530983e-01 -3.39383602e-01 -4.94690895e-01 1.89574301e+00
1.08438231e-01 1.35428016e-03 -3.51959795e-01 7.59599566e-01
1.94364890e-01 4.15645748e-01 2.29193464e-01 -7.05575287e-01
1.32324302e+00 -5.86267173e-01 -1.02578211e+00 -3.40900481e-01
5.49208045e-01 -6.42888427e-01 1.52619123e+00 5.30128181e-01
-7.97132194e-01 -4.98711199e-01 -8.99259806e-01 2.25620523e-01
-1.14726104e-01 4.05183911e-01 9.04640555e-01 9.95277107e-01
-3.10707808e-01 4.24095094e-01 -2.71571636e-01 3.54488552e-01
5.17147601e-01 1.41192764e-01 2.41173014e-01 -7.25759286e-03
-1.47345841e+00 5.28931618e-01 -1.39036164e-01 1.66039377e-01
-9.73776519e-01 -1.19658113e+00 -5.35956562e-01 8.58949944e-02
9.14566934e-01 -4.97433901e-01 1.21889031e+00 -5.91849685e-01
-1.20292306e+00 1.60060525e-01 -2.20808074e-01 -4.38276589e-01
9.72469926e-01 -5.65876007e-01 -4.40009862e-01 -7.79419601e-01
3.66307497e-01 -2.87130147e-01 1.09108090e+00 -1.23045599e+00
-7.80815244e-01 -4.18393314e-01 -3.21568489e-01 1.02659583e-01
-2.02273548e-01 -2.21712351e-01 -1.61454484e-01 -1.10881341e+00
-2.15157494e-01 -7.72117436e-01 -3.68300050e-01 -3.33003849e-01
-7.27653205e-01 -3.55175763e-01 4.96205002e-01 -6.30770504e-01
1.72674584e+00 -2.07576823e+00 -4.65227127e-01 2.25058287e-01
2.72084594e-01 1.89868987e-01 -9.90856662e-02 2.43911296e-01
-6.01627529e-02 4.67764974e-01 7.57901669e-02 -3.28846335e-01
9.67224613e-02 -2.73475409e-01 -6.70773387e-01 3.85136604e-01
-1.18445709e-01 7.52025723e-01 -6.57533407e-01 -3.85119021e-01
7.03963265e-02 1.65615439e-01 -4.52030152e-01 5.78619577e-02
-1.49710640e-01 9.73635837e-02 -7.99007535e-01 6.09114468e-01
8.45458567e-01 -3.30156028e-01 2.11853623e-01 -3.56659442e-01
-4.06154096e-01 1.99375987e-01 -1.50980747e+00 1.36164141e+00
-8.37178767e-01 1.87654570e-01 1.43602222e-01 -1.11906350e+00
7.16036320e-01 2.42621765e-01 5.49882472e-01 -7.07167387e-01
1.00756153e-01 2.44696453e-01 -2.59389490e-01 -1.91406474e-01
3.50699633e-01 -3.45776737e-01 -3.33020329e-01 4.56826538e-01
-1.31548107e-01 2.25766420e-01 -2.23531753e-01 1.54071972e-01
8.34554970e-01 1.93152860e-01 4.67467815e-01 -5.57994425e-01
1.54887646e-01 -2.70041406e-01 8.66950989e-01 1.24433064e+00
-2.49887735e-01 2.49704123e-01 3.51729900e-01 -2.57952988e-01
-8.90482128e-01 -1.24760568e+00 -3.17870200e-01 1.12883675e+00
7.86450878e-02 -1.85073391e-01 -1.74328476e-01 -8.17747295e-01
3.45557094e-01 1.18241453e+00 -7.14388072e-01 -7.78182819e-02
-2.42692202e-01 -1.02034140e+00 3.20435971e-01 3.13115448e-01
3.89358252e-01 -2.93769985e-01 -2.02047050e-01 1.73894793e-01
-4.34362143e-01 -7.07410395e-01 -8.02098095e-01 2.06238441e-02
-7.28729665e-01 -9.86632645e-01 -7.70882487e-01 1.02194794e-01
2.03558475e-01 7.00619519e-01 9.62744832e-01 -4.88247544e-01
-2.37049416e-01 1.41582608e-01 -1.12360112e-01 -7.76458859e-01
1.53339237e-01 -2.16427460e-01 2.25822389e-01 2.85681605e-01
3.70443135e-01 -5.22696614e-01 -7.69077063e-01 6.66155279e-01
-5.96439004e-01 -1.57234445e-01 7.35357463e-01 9.57689941e-01
7.07283795e-01 1.11388229e-01 1.08403015e+00 -1.07272542e+00
8.61691475e-01 -7.33585000e-01 -7.35056281e-01 6.03591084e-01
-1.11795712e+00 1.07587889e-01 4.85410661e-01 -6.84628308e-01
-1.38512981e+00 -2.48309061e-01 1.99454099e-01 -6.05119348e-01
3.35691869e-01 4.29957718e-01 -1.64847255e-01 3.48919839e-01
5.12986660e-01 2.83041418e-01 -2.62933492e-04 -6.83193684e-01
4.78750795e-01 7.86216199e-01 -9.66381952e-02 -5.59431732e-01
7.67405629e-01 3.65168840e-01 -9.68558490e-02 -6.50159955e-01
-1.27700830e+00 -5.00189364e-01 -9.75693688e-02 -1.08727828e-01
4.92753983e-01 -1.25193834e+00 -9.77532804e-01 -1.09151170e-01
-7.13072836e-01 -9.85923260e-02 -2.20448837e-01 6.84042275e-01
-6.39538646e-01 2.29381129e-01 -1.12630747e-01 -1.50376189e+00
-3.25438201e-01 -9.69922364e-01 8.75886023e-01 -1.96532130e-01
7.32925609e-02 -9.06378806e-01 -9.27344114e-02 5.21881878e-01
4.10371333e-01 6.90712556e-02 8.10399711e-01 -6.63710833e-01
-5.58950067e-01 -3.86836439e-01 -3.50366384e-01 2.18293145e-01
5.28171659e-02 -3.14518571e-01 -7.25741863e-01 -4.00951564e-01
3.03513259e-01 -1.93921119e-01 7.47131348e-01 9.38369989e-01
1.40453660e+00 -6.51040435e-01 -4.82081503e-01 3.82851154e-01
1.61586773e+00 1.56378537e-01 5.78762352e-01 1.09778188e-01
4.65036631e-01 3.77942175e-01 1.04604435e+00 8.36402714e-01
3.41312736e-01 7.74702668e-01 2.77420282e-01 8.02983493e-02
1.77173063e-01 -5.74752271e-01 2.93445349e-01 5.06069005e-01
-9.13674980e-02 -2.78906703e-01 -2.78782964e-01 3.70652199e-01
-1.78517509e+00 -7.83184767e-01 -3.43303770e-01 2.72090578e+00
6.82297885e-01 -8.22570454e-03 2.36781627e-01 -1.31700858e-01
8.53474438e-01 1.57257870e-01 -6.69424713e-01 -1.61867529e-01
1.00054890e-01 -5.83992153e-02 7.46051848e-01 2.04241231e-01
-1.00745344e+00 4.83680218e-01 6.18624306e+00 1.31499577e+00
-4.71547812e-01 4.75627810e-01 5.85150957e-01 -4.12155271e-01
-4.95592445e-01 8.67238268e-02 -1.20971096e+00 8.13982546e-01
7.28442013e-01 -3.42488348e-01 3.51920873e-01 8.67987394e-01
5.54918528e-01 -4.04801629e-02 -9.31399465e-01 1.07540536e+00
-2.39170656e-01 -1.09607291e+00 6.66807815e-02 4.70764041e-01
6.05174243e-01 -1.48775339e-01 4.53997821e-01 8.22159767e-01
5.34778178e-01 -8.38396370e-01 7.14540601e-01 3.22678357e-01
1.13762176e+00 -5.39049447e-01 5.38591266e-01 5.22979498e-01
-1.14612019e+00 -3.70039046e-01 -3.92414063e-01 -3.44353616e-02
3.75327408e-01 9.73507285e-01 -2.62684315e-01 7.82324791e-01
5.03949225e-01 3.74848008e-01 -1.44563705e-01 7.44438112e-01
4.44947071e-02 5.80715656e-01 -2.61223048e-01 -1.84214592e-01
3.02021243e-02 -4.09660429e-01 7.30840087e-01 9.43460166e-01
6.26543522e-01 -2.86926422e-02 -5.18189035e-02 1.18350935e+00
-1.94396675e-02 1.19380146e-01 -6.24503672e-01 8.03865418e-02
6.62270069e-01 9.99508798e-01 -1.47159323e-01 -9.20199323e-03
-6.90580130e-01 5.58640659e-01 7.72689059e-02 6.05241001e-01
-1.21199274e+00 -3.50879252e-01 6.97482944e-01 1.87122822e-01
3.00474346e-01 8.49535037e-03 -6.08820200e-01 -1.46992481e+00
1.70540512e-01 -8.61046553e-01 4.07654136e-01 -1.98582530e-01
-1.46957529e+00 3.95432785e-02 -3.41669358e-02 -1.49070644e+00
2.95954913e-01 -3.40843618e-01 -6.25259757e-01 9.52062666e-01
-1.51364028e+00 -9.99402106e-01 1.53815225e-01 5.36999345e-01
8.65029633e-01 -2.24743128e-01 4.59846109e-01 3.97372574e-01
-8.90222490e-01 1.02765620e+00 5.00536859e-01 -3.02265376e-01
6.55930519e-01 -1.09515715e+00 5.75151034e-02 6.47169411e-01
7.25641325e-02 1.01176441e+00 7.69728601e-01 -8.66876781e-01
-1.40445566e+00 -1.24410009e+00 5.68700314e-01 -5.42007029e-01
7.04215825e-01 -5.02368033e-01 -4.52332169e-01 6.74931884e-01
-3.53380412e-01 -2.71229714e-01 9.43310499e-01 7.64985502e-01
-4.63666499e-01 -3.15622151e-01 -1.00975406e+00 6.74334705e-01
1.01751280e+00 -2.70382792e-01 -4.71284240e-01 4.88190979e-01
8.24372828e-01 1.16039090e-01 -9.78806078e-01 4.44385201e-01
8.01393092e-01 -7.92771459e-01 1.14693749e+00 -8.43892992e-01
3.35329890e-01 8.60685557e-02 -3.62347007e-01 -1.15566754e+00
-3.10909063e-01 -7.42943704e-01 -3.78933072e-01 1.36151636e+00
6.94851339e-01 -5.53303540e-01 5.97635210e-01 8.17349195e-01
2.27581888e-01 -7.12315679e-01 -1.13199747e+00 -1.15223527e+00
8.86768624e-02 -7.15208709e-01 4.14625585e-01 9.30363476e-01
-2.89832592e-01 4.55562294e-01 -1.04936707e+00 -6.05067089e-02
1.13699424e+00 2.47919708e-01 6.75558746e-01 -1.24183059e+00
-4.17224914e-01 -2.79280901e-01 4.77147132e-01 -1.18716359e+00
-1.26225218e-01 -5.66428602e-01 -6.15913570e-02 -1.05189931e+00
2.80303270e-01 -7.95281231e-01 -5.67717195e-01 4.19861227e-02
-3.59938711e-01 -3.17808062e-01 1.28402174e-01 2.76164621e-01
-6.15095735e-01 7.34819174e-01 1.09932816e+00 9.31928828e-02
-3.75528902e-01 7.96479285e-01 -1.07844782e+00 5.45512199e-01
6.29730463e-01 -6.00638092e-01 -6.51167512e-01 -8.72982591e-02
3.76701921e-01 4.01765913e-01 2.68198133e-01 -4.49218959e-01
-1.39961630e-01 -4.01366800e-01 1.57436654e-01 -5.31709909e-01
2.49562532e-01 -7.98291147e-01 2.05330700e-01 2.69037515e-01
-5.66619813e-01 -2.15264305e-01 -1.00909650e-01 1.29992509e+00
6.91819042e-02 -5.98382130e-02 4.16971534e-01 -1.98105127e-01
-1.58774257e-01 4.74358112e-01 -2.08643422e-01 2.91684151e-01
8.60304534e-01 -2.28825980e-03 -2.40590870e-01 -2.83236384e-01
-4.32748288e-01 4.56625283e-01 -2.74903625e-01 5.36646187e-01
5.03891230e-01 -1.44822097e+00 -6.38528407e-01 2.38840487e-02
4.27902043e-02 -3.73991042e-01 6.73365235e-01 1.03699601e+00
5.54652691e-01 6.67868316e-01 4.24822241e-01 -2.20205843e-01
-8.66592765e-01 7.91088939e-01 4.25738730e-02 -5.96438646e-01
-3.22655886e-01 7.53992796e-01 4.77483660e-01 -2.24312335e-01
2.60792762e-01 -9.92199108e-02 -1.06151894e-01 9.66835096e-02
4.37334538e-01 7.13025451e-01 -1.74693525e-01 2.39883512e-02
-9.40350294e-02 2.27232248e-01 -2.52338856e-01 -1.21400207e-01
1.17577612e+00 -5.43221951e-01 3.56607348e-01 6.19374394e-01
1.08047020e+00 3.05588990e-02 -1.39756536e+00 -4.22376156e-01
-1.91329960e-02 -9.18474972e-01 2.33005062e-01 -8.79313648e-01
-7.75068045e-01 7.86563516e-01 4.63327140e-01 2.59755671e-01
8.44386220e-01 -4.85857338e-01 6.34895861e-01 9.18192044e-02
6.20092154e-01 -1.43697989e+00 7.59468228e-02 -6.50214702e-02
7.78498292e-01 -1.41088021e+00 2.52790809e-01 -4.43622142e-01
-9.92634594e-01 4.18769181e-01 5.49360216e-01 -1.20815514e-02
9.61826980e-01 -5.93438093e-03 -4.97859746e-01 -5.40168025e-02
-7.29234457e-01 9.49515626e-02 3.32722723e-01 4.59948063e-01
3.88840646e-01 2.32397676e-01 -8.10258210e-01 1.33747160e+00
1.13798238e-01 5.82514852e-02 3.72463912e-01 5.43466210e-01
-1.08571194e-01 -9.08244431e-01 -1.53080344e-01 1.07401216e+00
-7.76376784e-01 -2.98965454e-01 2.18734533e-01 9.44815755e-01
-9.67111364e-02 1.00145566e+00 -2.79017448e-01 -3.72361302e-01
5.58221102e-01 -8.67639333e-02 4.21577729e-02 -3.55343908e-01
-3.11194897e-01 3.79844218e-01 1.28318280e-01 -6.03070199e-01
-1.21239573e-01 -6.98722661e-01 -3.94209027e-01 -1.54136568e-01
-8.46978605e-01 3.53640020e-01 6.37115061e-01 9.71688032e-01
3.22626531e-01 6.10191345e-01 9.77028191e-01 -3.97048771e-01
-1.54994214e+00 -1.03475142e+00 -1.02471876e+00 2.66727000e-01
2.87161499e-01 -1.11370087e+00 -6.69289291e-01 -3.98269951e-01] | [9.675785064697266, 5.379745960235596] |
658b057b-2925-4a74-9f11-5a6b458a8ad3 | achieving-connectivity-between-wide-areas | 1807.04505 | null | http://arxiv.org/abs/1807.04505v1 | http://arxiv.org/pdf/1807.04505v1.pdf | Achieving Connectivity Between Wide Areas Through Self-Organising Robot Swarm Using Embodied Evolution | Abruptions to the communication infrastructure happens occasionally, where
manual dedicated personnel will go out to fix the interruptions, restoring
communication abilities. However, sometimes this can be dangerous to the
personnel carrying out the task, which can be the case in war situations,
environmental disasters like earthquakes or toxic spills or in the occurrence
of fire. Therefore, human casualties can be minimised if autonomous robots are
deployed that can achieve the same outcome: to establish a communication link
between two previously distant but connected sites. In this paper we
investigate the deployment of mobile ad hoc robots which relay traffic between
them. In order to get the robots to locate themselves appropriately, we take
inspiration from self-organisation and emergence in artificial life, where a
common overall goal may be achieved if the correct local rules on the agents in
system are invoked. We integrate the aspect of connectivity between two sites
into the multirobot simulation platform known as JBotEvolver. The robot swarm
is composed of Thymio II robots. In addition, we compare three heuristics, of
which one uses neuroevolution (evolution of neural networks) to show how
self-organisation and embodied evolution can be used within the integration.
Our use of embodiment in robotic controllers shows promising results and
provide solid knowledge and guidelines for further investigations. | ['Asieh Abolpour Mofrad', 'Hårek Haugerud', 'Anis Yazidi', 'Stefano Nichele', 'Erik Aaron Hansen', 'Alex Alcocer'] | 2018-07-12 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 9.37197506e-02 3.45850587e-01 5.29057860e-01 2.48047695e-01
7.02589571e-01 -5.28985858e-01 5.65389693e-01 2.52586573e-01
-6.03976846e-01 1.29420650e+00 -4.68959838e-01 -1.04110673e-01
-8.64320457e-01 -8.07183325e-01 -4.80118454e-01 -9.77046967e-01
-6.05199337e-01 6.56588674e-01 4.19603199e-01 -1.03077042e+00
3.17992389e-01 7.54270792e-01 -1.81577861e+00 -5.26845336e-01
7.78689444e-01 4.83664781e-01 6.10561430e-01 5.40344954e-01
1.25700384e-01 4.57613230e-01 -9.77519155e-01 7.61968866e-02
3.56549293e-01 -3.31060141e-01 -7.32818127e-01 3.22998166e-02
-9.75284755e-01 2.31154025e-01 2.34717920e-01 7.03264058e-01
4.19510305e-01 3.03944558e-01 8.12688768e-01 -1.62014019e+00
-7.19940811e-02 6.18202150e-01 -3.79885972e-01 -5.84204197e-02
3.80997658e-01 1.43416911e-01 1.64739698e-01 -5.34810536e-02
6.80082560e-01 1.11388814e+00 5.42255461e-01 2.97591358e-01
-7.56989121e-01 -9.39222500e-02 -1.55170172e-01 1.31308764e-01
-1.40703166e+00 -2.79822260e-01 2.67775953e-01 -1.52879022e-02
9.27868783e-01 2.93014020e-01 8.84558141e-01 9.16039348e-01
9.35850024e-01 -2.06919655e-01 7.15092182e-01 -6.14469588e-01
3.87192190e-01 1.31667435e-01 -5.17548442e-01 4.13451433e-01
7.83308387e-01 4.36348580e-02 -4.67413291e-02 -3.08338236e-02
6.97871685e-01 -2.16805950e-01 -2.95354784e-01 -4.00905758e-01
-1.35989368e+00 4.98852611e-01 5.72817802e-01 8.57120752e-01
-9.28380191e-01 1.84502751e-01 1.66060328e-01 6.09937310e-01
-2.72122324e-01 1.03519368e+00 -2.83668071e-01 -1.02111764e-01
-3.76896411e-02 -1.00063026e-01 1.08553779e+00 6.30559802e-01
5.46633303e-01 -1.70927960e-02 5.27820766e-01 4.00616884e-01
1.05653457e-01 6.46951720e-02 4.64532554e-01 -1.08256328e+00
-1.68856815e-01 6.82541072e-01 2.14726284e-01 -1.32054794e+00
-7.96122193e-01 -2.70933867e-01 -8.98601830e-01 6.12387121e-01
-4.74792942e-02 -4.44296598e-01 -3.57285470e-01 1.32492065e+00
4.62038219e-01 -2.35379145e-01 4.76518333e-01 8.29633176e-01
9.54412483e-03 6.84505403e-01 -2.02851683e-01 -3.87527347e-01
1.11760163e+00 -7.92149663e-01 -7.23745644e-01 -1.46959066e-01
5.13715804e-01 -4.49636221e-01 2.75844008e-01 4.61083144e-01
-8.28926682e-01 -2.04388157e-01 -1.16656852e+00 8.31448197e-01
-7.33731329e-01 -4.44883585e-01 5.22354007e-01 4.18176949e-01
-1.30847394e+00 7.77830601e-01 -7.39079237e-01 -1.03230047e+00
-3.97107363e-01 7.09086001e-01 -3.72530162e-01 3.26579362e-01
-1.18589008e+00 1.35084438e+00 5.23858249e-01 3.06468189e-01
-7.67017782e-01 1.06543027e-01 -3.30212235e-01 -2.95145307e-02
4.90448058e-01 -8.35881889e-01 7.53522813e-01 -1.12061954e+00
-1.63368058e+00 4.16413128e-01 6.56297624e-01 -3.95794928e-01
4.82434094e-01 4.81736422e-01 -7.03242719e-02 -1.54089350e-02
6.94690123e-02 7.78418660e-01 4.33669120e-01 -1.36960781e+00
-7.55025446e-01 1.28549598e-02 2.86854267e-01 5.51510215e-01
-3.82150322e-01 -7.60456175e-02 9.98180658e-02 -2.86406398e-01
-1.32812738e-01 -1.21346915e+00 -5.01452684e-01 -3.11904788e-01
-3.18728894e-01 -1.30228862e-01 6.96076989e-01 -4.19199765e-02
5.50007164e-01 -1.82814884e+00 7.58913517e-01 3.27128649e-01
-1.18849449e-01 3.20520662e-02 -6.80245236e-02 1.03906691e+00
3.40010285e-01 1.80550009e-01 -2.50819445e-01 8.27137679e-02
-1.28492534e-01 5.98720551e-01 3.51492524e-01 3.89099479e-01
-9.52053145e-02 1.79287985e-01 -7.32514501e-01 -2.19052717e-01
-1.76771641e-01 2.42976323e-01 -1.67497784e-01 -4.41274717e-02
5.33401556e-02 5.83022058e-01 -5.08161008e-01 3.15190136e-01
1.30233765e-01 3.20337921e-01 3.36762816e-01 5.62235892e-01
-5.92028201e-01 -2.82291681e-01 -1.08143330e+00 1.37322378e+00
-4.25893694e-01 4.97362465e-01 5.79519451e-01 -1.16885412e+00
9.12051380e-01 3.86994243e-01 6.03225052e-01 -4.78083879e-01
4.65738803e-01 3.33657354e-01 4.06999469e-01 -5.27131975e-01
5.41496754e-01 2.41335690e-01 -2.72547632e-01 5.56488514e-01
-3.51753414e-01 -5.05839944e-01 4.04011428e-01 -4.83590588e-02
1.33063316e+00 1.76881242e-03 4.27638263e-01 -4.00600791e-01
4.66287732e-01 4.22501504e-01 3.69870484e-01 6.63610935e-01
-9.86968875e-02 5.44142611e-02 4.04036611e-01 -3.34045500e-01
-9.60230470e-01 -6.91601217e-01 2.66810566e-01 8.12132835e-01
5.99573910e-01 1.87459335e-01 -6.57745063e-01 -1.39128774e-01
-1.77273527e-01 6.75010622e-01 -5.58840394e-01 -4.81924087e-01
-5.79601228e-01 -6.97398603e-01 4.86808032e-01 -2.98249722e-01
6.25822842e-01 -1.42447054e+00 -1.51256657e+00 4.84101892e-01
1.14408471e-01 -6.35042310e-01 1.89646468e-01 6.12266898e-01
-4.40314978e-01 -1.10673046e+00 -4.75267500e-01 -9.40972447e-01
7.87519336e-01 5.24818838e-01 5.17566264e-01 7.16474116e-01
-3.04840058e-01 4.76487219e-01 -6.21091604e-01 -4.81747329e-01
-7.57256687e-01 4.11106855e-01 5.63216388e-01 -3.67637008e-01
-4.43781167e-01 -8.58416378e-01 -3.26994777e-01 7.56010294e-01
-8.90798748e-01 -1.37924299e-01 5.88392138e-01 5.42485654e-01
5.52555844e-02 7.43186712e-01 7.46995211e-01 1.04502976e-01
1.05231118e+00 -7.51274705e-01 -3.65094215e-01 3.30070585e-01
-3.33030701e-01 -6.64710999e-02 6.50198996e-01 -5.53475916e-01
-7.35366464e-01 4.31982987e-02 3.98888171e-01 -6.77331388e-02
-1.87545776e-01 5.25845289e-01 1.07973047e-01 -2.75524527e-01
5.42454243e-01 7.22501799e-02 4.67680782e-01 8.92078578e-02
-1.43458089e-02 8.24807405e-01 2.46987969e-01 -3.82236570e-01
5.07040441e-01 3.25464576e-01 3.20145965e-01 -9.94400978e-01
3.78379077e-01 5.57768419e-02 -5.46305954e-01 -6.06084049e-01
6.48149550e-01 -2.00992122e-01 -1.01010680e+00 3.92865390e-01
-1.27925694e+00 -3.67002517e-01 -1.39992952e-01 7.54128993e-02
-5.56758881e-01 -1.77404322e-02 -2.47268733e-02 -8.80295813e-01
2.57092342e-02 -1.13860059e+00 4.94504541e-01 5.08487761e-01
-1.99829057e-01 -9.76364315e-01 2.29767352e-01 -2.62169629e-01
7.45609403e-01 5.28999805e-01 6.14238799e-01 -5.75711608e-01
-3.60115945e-01 5.89340143e-02 5.57466745e-01 -1.86167508e-01
3.87603164e-01 2.82288522e-01 -1.11512490e-01 -3.56830627e-01
-1.08558357e-01 5.88526353e-02 8.47099125e-02 1.13916159e-01
1.67793527e-01 -2.82333612e-01 -7.68670797e-01 -7.68832564e-02
1.01071990e+00 8.37524414e-01 7.51093149e-01 9.09857571e-01
-7.36332834e-02 1.38828206e+00 8.58395100e-01 4.33907479e-01
3.01013708e-01 7.70029545e-01 8.17257345e-01 4.83255498e-02
3.22823018e-01 4.41968679e-01 2.96753883e-01 8.02054703e-01
-4.53821480e-01 -6.09889150e-01 -9.57081199e-01 4.76016581e-01
-1.92227936e+00 -7.26515234e-01 -1.24472573e-01 1.92914867e+00
3.31893981e-01 9.73564535e-02 8.05231854e-02 2.29227111e-01
1.04735804e+00 -5.66120088e-01 -2.83710957e-01 -7.90735841e-01
-1.21645726e-01 -5.36022425e-01 4.80845571e-01 3.36452305e-01
-5.28552234e-01 6.32576585e-01 5.36314440e+00 3.89504910e-01
-1.03697705e+00 9.07708108e-02 1.44990832e-01 -1.15691029e-01
2.26288363e-01 1.19618751e-01 -2.20639467e-01 5.14269769e-01
9.95904267e-01 -1.99631765e-01 7.22276270e-01 3.12126130e-01
5.10513127e-01 -6.31349981e-01 -4.52912211e-01 4.25954252e-01
-1.54737160e-01 -8.70909691e-01 -3.85594994e-01 2.97657818e-01
3.92083913e-01 -2.24835515e-01 -3.80982101e-01 -4.64221872e-02
4.29524392e-01 -7.68575609e-01 8.19150448e-01 7.70144820e-01
2.07550511e-01 -1.06628060e+00 8.17173719e-01 8.09344828e-01
-8.72217178e-01 -4.42451447e-01 -1.81544825e-01 -3.86450589e-01
5.17956793e-01 1.37377679e-01 -1.17035973e+00 7.57291794e-01
7.48798907e-01 5.48521848e-03 -1.85936883e-01 1.25069571e+00
-1.93601027e-01 -3.78556371e-01 -6.01291060e-01 -7.18391120e-01
3.09668034e-01 -4.59993601e-01 1.02522123e+00 4.27175373e-01
6.70768380e-01 1.14212319e-01 -1.23335309e-01 5.05700946e-01
5.26529133e-01 -1.17338538e-01 -9.94521260e-01 2.39393547e-01
5.60680926e-01 1.29956269e+00 -1.33365703e+00 2.20130503e-01
4.42114890e-01 1.06374300e+00 1.46735787e-01 6.32056966e-02
-9.53236401e-01 -8.30883861e-01 5.68881035e-01 -8.75522643e-02
2.10761130e-02 -4.32892889e-01 1.47620574e-01 -2.79409409e-01
-3.65615606e-01 -7.15418100e-01 -2.48314559e-01 -9.51388597e-01
-6.85416102e-01 8.68411779e-01 1.15321606e-01 -1.12458158e+00
-3.97908539e-01 -2.17461661e-01 -6.52323306e-01 7.09593967e-02
-9.59525168e-01 -6.82435215e-01 -2.61098593e-01 2.64647633e-01
2.34227583e-01 -3.40338498e-01 6.23747885e-01 -3.58694345e-02
-7.20000982e-01 -1.51552692e-01 1.15343794e-01 -7.12975323e-01
4.92494851e-01 -7.91967869e-01 -1.34444505e-01 6.02868974e-01
-4.07303065e-01 5.32970011e-01 9.76964116e-01 -9.00912166e-01
-1.48184097e+00 -6.38416171e-01 5.46811104e-01 1.75069198e-01
5.77511072e-01 -1.33378372e-01 -4.10821110e-01 3.21212411e-01
6.54501081e-01 -6.43344879e-01 1.49739943e-02 -4.27035689e-01
7.60450959e-01 7.23784044e-02 -1.23079872e+00 9.26662326e-01
1.10368896e+00 4.82530534e-01 -3.96898597e-01 3.20519716e-01
7.11115956e-01 4.87780198e-02 -5.93637168e-01 2.19355479e-01
4.36403215e-01 -1.00061619e+00 7.01465070e-01 -2.83824742e-01
-6.94074705e-02 -4.72736835e-01 2.53239602e-01 -1.85584104e+00
-1.53643906e-01 -8.99204969e-01 7.20145166e-01 1.20248902e+00
5.13150275e-01 -1.17174244e+00 9.57361832e-02 1.53347701e-01
-3.13532025e-01 -2.94022888e-01 -1.24573803e+00 -9.08043683e-01
-1.99525699e-01 3.08909237e-01 6.27745271e-01 9.09525692e-01
3.97811294e-01 1.84511200e-01 -1.78821161e-01 2.13162020e-01
2.27721199e-01 -5.81932902e-01 7.82970250e-01 -1.38712943e+00
-1.60063237e-01 -5.00218987e-01 -4.80498493e-01 -9.64532569e-02
7.69527927e-02 -3.55526417e-01 3.99642408e-01 -1.53270411e+00
-4.16718245e-01 -9.12700117e-01 2.77090430e-01 5.97561419e-01
5.43396771e-01 -2.52865255e-02 2.39280313e-01 2.70318240e-01
-6.25669837e-01 4.42657769e-01 1.05893576e+00 1.36727303e-01
-5.29960454e-01 -7.05280676e-02 -4.43428189e-01 6.75870538e-01
1.08119583e+00 -7.63987422e-01 -2.67716080e-01 -1.26792222e-01
6.77060246e-01 3.54596198e-01 6.39184490e-02 -1.31936967e+00
8.36673856e-01 -2.14981303e-01 -3.29203308e-01 2.14210182e-01
3.08490664e-01 -1.43252492e+00 7.76606202e-01 1.20140779e+00
1.22779675e-01 5.30743062e-01 -4.76247119e-03 5.42859197e-01
2.95361094e-02 -5.88404417e-01 4.62990254e-01 -1.11379378e-01
-3.42540711e-01 -5.24247646e-01 -1.38014579e+00 -7.76984036e-01
1.83491063e+00 -4.55412269e-01 -4.91331756e-01 -4.43037987e-01
-4.75877285e-01 7.07741678e-01 7.32917249e-01 4.40448910e-01
3.43827784e-01 -7.68260658e-01 -4.74427879e-01 -2.14611620e-01
-3.34753722e-01 -1.61909714e-01 -1.68526232e-01 1.18926334e+00
-1.00021040e+00 2.15694502e-01 -8.42843592e-01 -2.95437932e-01
-1.07419753e+00 4.56226736e-01 6.10759199e-01 7.64984339e-02
-1.51364967e-01 2.92873681e-01 -3.58421028e-01 -3.51181835e-01
2.38972232e-02 -9.94218290e-02 -5.60314119e-01 1.80927023e-01
9.24131796e-02 6.80530369e-01 6.34784773e-02 -4.61231440e-01
-5.76341689e-01 3.97367001e-01 3.42978686e-01 -2.91700631e-01
1.52231812e+00 -5.32087624e-01 -7.03782260e-01 2.61858135e-01
3.72632265e-01 -2.12262973e-01 -8.12331200e-01 7.26723313e-01
3.58234271e-02 8.45338106e-02 -2.95993835e-01 -6.49365902e-01
-7.47764826e-01 1.43164426e-01 2.28081346e-01 9.99921262e-01
1.19897175e+00 -2.48507217e-01 3.25987488e-01 7.69331872e-01
1.11225796e+00 -1.04267514e+00 1.14800565e-01 5.47119141e-01
1.08632743e+00 -6.49890125e-01 -1.62940636e-01 -3.04945022e-01
-6.37419343e-01 1.19533014e+00 5.31706572e-01 -2.29454845e-01
3.34409654e-01 4.97300088e-01 -2.50724763e-01 -2.32857838e-01
-7.66769290e-01 -2.17510149e-01 -6.16555452e-01 9.68989611e-01
-2.87934601e-01 -1.86638653e-01 -6.48260832e-01 1.06428318e-01
-2.20504045e-01 -4.43137378e-01 1.08234286e+00 1.26828253e+00
-8.62838924e-01 -1.08480644e+00 -7.80494034e-01 -1.58959255e-01
-1.10690603e-02 5.14343500e-01 -7.54545569e-01 1.18078566e+00
3.79640013e-01 1.20198619e+00 2.53402293e-01 -2.92770267e-01
4.72517997e-01 -4.40621048e-01 2.08616152e-01 -2.87265301e-01
-8.73192370e-01 -3.27145219e-01 3.19708169e-01 -2.01981544e-01
-6.81291878e-01 -7.26613224e-01 -1.52743018e+00 -3.33622664e-01
-2.40764961e-01 5.33116817e-01 1.14947367e+00 7.51591921e-01
4.24303323e-01 8.50056708e-01 6.73288167e-01 -1.29534781e+00
-1.26799524e-01 -9.07701671e-01 -5.34494400e-01 -3.42235178e-01
4.56035323e-02 -9.73038495e-01 -5.50124109e-01 -2.35465050e-01] | [4.941695690155029, 1.7709872722625732] |
2b542738-8802-405f-adf4-289b97c52eed | cluzh-at-sigmorphon-2022-shared-tasks-on | null | null | https://aclanthology.org/2022.sigmorphon-1.21 | https://aclanthology.org/2022.sigmorphon-1.21.pdf | CLUZH at SIGMORPHON 2022 Shared Tasks on Morpheme Segmentation and Inflection Generation | This paper describes the submissions of the team of the Department of Computational Linguistics, University of Zurich, to the SIGMORPHON 2022 Shared Tasks on Morpheme Segmentation and Inflection Generation. Our submissions use a character-level neural transducer that operates over traditional edit actions. While this model has been found particularly wellsuited for low-resource settings, using it with large data quantities has been difficult. Existing implementations could not fully profit from GPU acceleration and did not efficiently implement mini-batch training, which could be tricky for a transition-based system. For this year’s submission, we have ported the neural transducer to PyTorch and implemented true mini-batch training. This has allowed us to successfully scale the approach to large data quantities and conduct extensive experimentation. We report competitive results for morpheme segmentation (including sharing first place in part 2 of the challenge). We also demonstrate that reducing sentence-level morpheme segmentation to a word-level problem is a simple yet effective strategy. Additionally, we report strong results in inflection generation (the overall best result for large training sets in part 1, the best results in low-resource learning trajectories in part 2). Our code is publicly available. | ['Peter Makarov', 'Simon Clematide', 'Silvan Wehrli'] | null | null | null | null | naacl-sigmorphon-2022-7 | ['morpheme-segmentaiton', 'morphological-inflection'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.19706386e-01 1.92665696e-01 1.32934526e-01 -3.40378940e-01
-1.32720947e+00 -7.15935826e-01 4.55914497e-01 3.10478181e-01
-8.22641313e-01 6.87663436e-01 2.49329254e-01 -7.63003707e-01
4.56049770e-01 -7.80375361e-01 -7.75092185e-01 -2.14601249e-01
-2.85172075e-01 7.57145107e-01 1.13006108e-01 -6.37117088e-01
5.39941907e-01 1.84107441e-02 -1.48577976e+00 4.91046101e-01
8.90440762e-01 2.04323694e-01 7.85852820e-02 9.79487836e-01
-2.69245684e-01 7.98665285e-02 -6.28356695e-01 -4.88020480e-01
4.67485815e-01 -4.52691823e-01 -1.21441150e+00 -2.80181825e-01
9.25247371e-01 3.44320163e-02 3.62694621e-01 8.84797037e-01
5.91785669e-01 4.73837465e-01 4.73625988e-01 -7.34707415e-01
-5.79531789e-01 1.20090568e+00 -3.78663450e-01 3.91528577e-01
2.43959144e-01 1.65495992e-01 1.27284408e+00 -8.96174610e-01
8.00189137e-01 1.19951200e+00 8.59529257e-01 5.44244826e-01
-1.21567500e+00 -4.02172238e-01 1.21797673e-01 2.47288644e-02
-1.12915313e+00 -6.47490561e-01 2.26626128e-01 -1.93360403e-01
1.78643155e+00 4.33995634e-01 7.61952937e-01 6.78034067e-01
6.43861890e-02 1.07181501e+00 1.05330074e+00 -5.78407824e-01
6.24213777e-02 4.40780856e-02 2.04836130e-01 6.76096916e-01
4.57078032e-02 -1.55600220e-01 -3.22523892e-01 4.55267578e-02
4.54475999e-01 -8.62723708e-01 -3.04316655e-02 4.04556036e-01
-1.24923170e+00 8.83954704e-01 2.99641460e-01 5.02403975e-01
-3.19237337e-02 1.07201248e-01 6.78899348e-01 4.73636389e-01
7.02944040e-01 7.84508526e-01 -5.93233168e-01 -7.38738775e-01
-1.36473989e+00 5.42461514e-01 1.09045970e+00 1.00699675e+00
6.62201524e-01 1.36418432e-01 -1.10939264e-01 9.37799275e-01
-5.73237501e-02 2.53985971e-02 7.97219753e-01 -9.42275882e-01
6.49280131e-01 1.92517012e-01 -1.20690979e-01 -5.78260660e-01
-5.48586726e-01 -7.93001242e-03 -5.04154921e-01 6.91518635e-02
7.11721718e-01 -5.63729465e-01 -1.13520670e+00 1.73535347e+00
2.26814359e-01 -1.32701442e-01 -2.80275419e-02 7.23579884e-01
5.58061838e-01 8.47627699e-01 1.22910470e-01 -1.12646818e-01
1.37458646e+00 -1.05920136e+00 -4.94676858e-01 -4.72792804e-01
1.23523605e+00 -9.97337520e-01 1.28555572e+00 5.64912140e-01
-1.63012564e+00 -3.62967223e-01 -1.06484151e+00 -5.44283032e-01
-6.25045240e-01 -2.34244913e-01 9.82779384e-01 8.23378980e-01
-1.44831789e+00 8.46242726e-01 -1.05350351e+00 -5.75939715e-01
2.48830579e-02 3.92775297e-01 -1.90334991e-01 2.01096818e-01
-1.32291675e+00 1.03943586e+00 6.01829171e-01 1.66964397e-01
-2.40255669e-01 -7.90338635e-01 -8.40324342e-01 -2.65417993e-01
1.76613107e-01 -5.10955513e-01 1.70730340e+00 -8.50065529e-01
-1.53841257e+00 1.01577973e+00 -2.02602923e-01 -6.50627792e-01
5.25499940e-01 -1.02306321e-01 -6.58345446e-02 -3.70260566e-01
8.21909085e-02 1.05203950e+00 3.15283954e-01 -6.64937317e-01
-7.83800006e-01 -1.84779018e-01 -4.55461293e-02 3.96410823e-01
-2.39140823e-01 1.20554537e-01 -2.15232342e-01 -6.47775471e-01
4.32031676e-02 -9.18246031e-01 -3.04249972e-01 -9.11419928e-01
-4.30161029e-01 -5.20871162e-01 2.29562074e-01 -8.52254331e-01
1.18274403e+00 -1.83579373e+00 5.99313378e-02 -1.23669999e-02
-2.49203965e-01 2.20069945e-01 -6.20602854e-02 4.58617240e-01
-4.56168912e-02 4.64679301e-01 -6.58804297e-01 -7.11234450e-01
2.13327289e-01 1.22548334e-01 -1.80785760e-01 1.06914319e-01
2.38078594e-01 1.05064905e+00 -8.96957397e-01 -4.69310373e-01
-1.55249819e-01 1.30826794e-02 -8.04045022e-01 -1.43511623e-01
-3.57540458e-01 -9.95107666e-02 1.76220134e-01 2.89681554e-01
6.63588285e-01 1.98865458e-01 -3.36942151e-02 9.27733630e-02
-6.12128437e-01 6.43650234e-01 -1.08737028e+00 2.00577474e+00
-5.98088086e-01 6.78333163e-01 1.55744329e-02 -6.04559600e-01
6.98013127e-01 1.51465401e-01 9.51628089e-02 -5.03653467e-01
-1.27651677e-01 4.92579967e-01 2.52095789e-01 -3.00874472e-01
1.42731416e+00 -4.06620085e-01 -4.02575374e-01 6.65833771e-01
1.23561816e-02 -7.63919890e-01 7.95262218e-01 2.50979871e-01
8.63011539e-01 2.78096795e-01 9.25056338e-02 -4.91528451e-01
-7.31361508e-02 4.28359360e-01 2.56536573e-01 6.75536335e-01
-1.50081769e-01 7.15908885e-01 3.44168156e-01 -3.13768566e-01
-1.18272817e+00 -8.66726041e-01 -2.17406198e-01 1.49833596e+00
-3.30840081e-01 -7.51331806e-01 -1.12222183e+00 -4.31004018e-01
-3.10780436e-01 1.05855608e+00 -5.24765968e-01 1.99386060e-01
-1.04534233e+00 -1.06496239e+00 1.13170087e+00 5.11769652e-01
3.51108223e-01 -1.31091833e+00 -4.88469154e-01 3.96523148e-01
-1.72803953e-01 -7.36523509e-01 -6.88054800e-01 3.08045536e-01
-8.40151489e-01 -6.11892700e-01 -5.52706957e-01 -1.08362949e+00
3.72640014e-01 -3.19146752e-01 1.21439886e+00 5.00761569e-02
-4.19708669e-01 -4.98143956e-02 -1.36453941e-01 -5.35887599e-01
-5.84836125e-01 7.78461695e-01 -1.72939256e-01 -6.44036651e-01
5.34522474e-01 -2.62420654e-01 -4.33126658e-01 -2.24411815e-01
-8.52751195e-01 1.31367654e-01 3.15034121e-01 6.43203676e-01
3.77158642e-01 -4.79429960e-01 5.49175680e-01 -1.45344150e+00
9.91179764e-01 -4.25066769e-01 -5.33537209e-01 -1.01663291e-01
-5.88764071e-01 5.10564186e-02 6.97683156e-01 6.67144433e-02
-1.03941584e+00 -1.43558877e-02 -7.12172866e-01 3.53239745e-01
-2.42023915e-01 6.83916390e-01 2.28638694e-01 3.45618725e-01
7.66102970e-01 -9.27906409e-02 -1.17688775e-01 -5.56595206e-01
6.24652326e-01 7.06183195e-01 5.83321691e-01 -7.13290989e-01
4.18588012e-01 5.70819639e-02 -3.30786645e-01 -8.81791353e-01
-6.22036457e-01 -3.20321947e-01 -5.90603888e-01 3.88696849e-01
9.12415683e-01 -7.88926065e-01 -2.52110422e-01 5.47813952e-01
-1.25259614e+00 -7.94993222e-01 -4.19118583e-01 1.14427462e-01
-6.59797370e-01 3.95793170e-01 -9.92777526e-01 -4.30488318e-01
-6.26789689e-01 -1.15676439e+00 9.23109472e-01 2.31965110e-01
-5.36903143e-01 -1.37057316e+00 3.86391193e-01 2.50176847e-01
6.70451164e-01 9.12198424e-02 1.02297938e+00 -8.37234080e-01
-3.60281289e-01 1.42186675e-02 2.15297509e-02 2.67295301e-01
-2.17489321e-02 2.94177175e-01 -8.88432205e-01 -2.85641432e-01
-2.61869401e-01 -5.06767690e-01 1.09659827e+00 3.13207746e-01
9.67893362e-01 -1.71354443e-01 -9.90233719e-02 5.89066446e-01
1.13355947e+00 -3.53463501e-01 5.50812542e-01 4.02933359e-01
7.06738770e-01 6.20258093e-01 4.85873729e-01 1.63949072e-01
8.13291848e-01 5.95304549e-01 -8.88632610e-02 4.07130178e-03
-2.50936538e-01 -1.37111589e-01 5.44085979e-01 1.44379604e+00
-5.30383177e-02 -3.12976748e-01 -1.08458877e+00 9.18294072e-01
-1.75923693e+00 -7.91195750e-01 -7.48823404e-01 2.18530440e+00
1.35008788e+00 2.98674464e-01 3.07243764e-01 -7.83447996e-02
5.19137859e-01 5.76902106e-02 -1.19996384e-01 -1.43367970e+00
-6.48852587e-02 6.08971059e-01 5.73965907e-01 9.36251998e-01
-1.01360846e+00 1.57131624e+00 7.08967781e+00 8.91212046e-01
-9.23564434e-01 2.90071279e-01 4.68560606e-01 -2.60184407e-01
-5.06630838e-01 6.44640997e-02 -1.07370651e+00 4.19650882e-01
1.59016490e+00 -4.10704613e-01 5.04575849e-01 4.60685819e-01
4.42235507e-02 -2.84900635e-01 -1.22551155e+00 5.29828787e-01
9.80779231e-02 -1.21370018e+00 -7.40932152e-02 -6.64236099e-02
7.98079133e-01 4.73019779e-01 -1.66821834e-02 6.92121863e-01
3.12142789e-01 -1.21750665e+00 5.52296042e-01 5.25835939e-02
7.40385115e-01 -1.01256812e+00 4.53857303e-01 2.48163447e-01
-1.08624506e+00 4.30084735e-01 -5.28260887e-01 -2.56352544e-01
4.24535632e-01 3.59019578e-01 -1.18250334e+00 4.00661856e-01
3.40658456e-01 4.77379084e-01 -7.12375581e-01 1.07189500e+00
-2.24865854e-01 8.39396715e-01 -5.67460775e-01 -6.30830675e-02
5.61967194e-01 -1.63634077e-01 5.06708503e-01 1.88382745e+00
1.18122049e-01 -5.75975440e-02 1.41481966e-01 7.01109469e-01
-2.54040986e-01 4.44558501e-01 -4.55559433e-01 -5.82038574e-02
2.79034287e-01 1.47104549e+00 -8.01191032e-01 -4.72255856e-01
-5.89991398e-02 1.06379747e+00 6.09233797e-01 -7.05771744e-02
-5.70409656e-01 -8.64407003e-01 5.62514603e-01 -2.32366189e-01
2.49563500e-01 -3.88636500e-01 -6.31043732e-01 -1.00922108e+00
-1.92173392e-01 -9.67759907e-01 4.88196880e-01 -2.87538052e-01
-1.09416783e+00 5.75013936e-01 1.11903273e-01 -5.04140139e-01
-7.49116063e-01 -6.04448259e-01 -6.57936454e-01 1.06889403e+00
-1.18125927e+00 -9.84784782e-01 2.09575012e-01 1.97078571e-01
8.95503402e-01 1.39488101e-01 1.01461470e+00 2.79699773e-01
-5.69135725e-01 1.00802541e+00 -5.38473651e-02 -2.69603673e-02
9.61795628e-01 -1.87901115e+00 1.33876300e+00 9.88164544e-01
4.97536987e-01 7.09647059e-01 7.94372141e-01 -6.98668957e-01
-1.23850071e+00 -9.36608136e-01 1.30771649e+00 -6.32449389e-01
7.73509145e-01 -6.50696695e-01 -9.14133728e-01 9.66567695e-01
6.61472499e-01 -6.45909905e-01 6.32669032e-01 4.44222957e-01
-2.54602190e-02 5.29774427e-01 -7.95322955e-01 8.13639283e-01
1.08527911e+00 -2.55107462e-01 -8.89175355e-01 6.27892792e-01
8.86711836e-01 -1.01529646e+00 -8.45441818e-01 -3.08771376e-02
2.77720839e-01 -7.43419945e-01 4.41610605e-01 -6.52448893e-01
5.42772293e-01 2.73344424e-02 1.35601282e-01 -1.73676395e+00
-7.12652728e-02 -8.08884859e-01 2.42511839e-01 1.28830278e+00
9.63135004e-01 -5.47582150e-01 7.43863225e-01 6.15514100e-01
-6.65667117e-01 -6.79185688e-01 -8.59508395e-01 -5.92798829e-01
7.99208462e-01 -7.36152887e-01 4.44472760e-01 7.83048868e-01
3.06150764e-01 3.96470994e-01 1.71560980e-02 -2.44989812e-01
3.25855613e-01 8.01956132e-02 5.74400365e-01 -7.72341669e-01
-4.89131421e-01 -7.29927897e-01 -7.15977177e-02 -9.55173790e-01
2.21742153e-01 -1.44903851e+00 4.69047576e-01 -1.63557386e+00
-1.10568307e-01 -4.64810312e-01 5.87331913e-02 7.30650663e-01
-4.11158860e-01 5.76768100e-01 3.05847466e-01 -2.65674163e-02
-2.93091208e-01 7.12604448e-02 1.01086032e+00 3.43173966e-02
-2.97131598e-01 -1.34607002e-01 -7.99602985e-01 4.90156323e-01
9.71201897e-01 -3.56491327e-01 -2.37487778e-01 -8.50960493e-01
3.68555307e-01 -4.47017223e-01 -1.02264933e-01 -7.88587332e-01
3.28079909e-01 1.19640812e-01 -5.34503870e-02 -4.76587176e-01
1.81056157e-01 6.95490837e-02 -2.82884389e-01 2.12523192e-01
-3.66685063e-01 4.26419824e-01 7.28181362e-01 -1.27336597e-02
-7.10972399e-02 -6.30954266e-01 5.78509152e-01 -3.90096426e-01
-6.47801161e-01 9.88199785e-02 -5.53135335e-01 6.39815807e-01
7.22333014e-01 -3.69627267e-01 -2.52470821e-01 -8.64669681e-02
-6.13720775e-01 2.71674812e-01 6.07344627e-01 4.32804376e-01
1.76921308e-01 -8.48813534e-01 -1.03936684e+00 2.25088000e-01
-1.78485289e-01 1.40945747e-01 3.16778868e-02 7.82941282e-01
-7.82630026e-01 4.12200868e-01 -1.64533347e-01 -3.30273151e-01
-1.09327567e+00 5.27473427e-02 2.78303534e-01 -3.39078069e-01
-4.95214731e-01 1.23850942e+00 -3.71865839e-01 -7.75828958e-01
-1.29408032e-01 -5.20160735e-01 6.24038167e-02 4.24037308e-01
4.80445832e-01 3.35999280e-01 6.83421373e-01 -4.14543271e-01
-2.59621382e-01 2.54238248e-01 -5.06566167e-01 -5.35508037e-01
1.24713957e+00 5.84764071e-02 -2.46247873e-01 6.74106300e-01
1.21401072e+00 1.69025525e-01 -8.22235227e-01 2.33096570e-01
2.88751453e-01 1.20878750e-02 6.53038248e-02 -7.51524866e-01
-6.41001821e-01 1.01786852e+00 1.88125908e-01 3.23077440e-01
7.50090301e-01 -3.86624962e-01 1.22163773e+00 5.84489107e-01
3.36058319e-01 -1.43552458e+00 -3.80322278e-01 1.16643274e+00
5.64548433e-01 -9.98584867e-01 -1.28858805e-01 -8.80411565e-02
-5.69457114e-01 1.11309040e+00 6.95300221e-01 -1.79574251e-01
2.14614943e-01 4.67684656e-01 9.37739164e-02 -5.49573032e-03
-8.68642628e-01 -1.18193418e-01 -9.47592594e-03 3.31583470e-01
1.05242395e+00 2.11886391e-01 -6.29186749e-01 3.20862412e-01
-1.02993131e+00 -3.13995868e-01 8.82146239e-01 9.13078189e-01
-4.86353636e-01 -1.48516023e+00 -6.85807243e-02 5.60725391e-01
-5.63925564e-01 -6.40151322e-01 -3.97818834e-01 9.00942683e-01
1.06052561e-02 7.58819818e-01 3.93020928e-01 -1.86670333e-01
3.88956428e-01 4.26961899e-01 7.52696872e-01 -1.08166599e+00
-1.28259838e+00 -2.06509694e-01 4.73763347e-01 -4.76618290e-01
1.00724809e-01 -9.58489001e-01 -1.58101964e+00 -5.80499351e-01
-8.58535767e-02 2.63127923e-01 9.16685522e-01 8.58438075e-01
2.14352757e-01 2.61439532e-01 4.98240255e-02 -1.02240658e+00
-6.82796955e-01 -1.16616213e+00 -3.61083478e-01 1.66352049e-01
1.81119218e-02 9.13346633e-02 -3.02838832e-01 1.18488118e-01] | [10.675230979919434, 9.868005752563477] |
604d6918-571b-417b-91e0-dc46bf3890b2 | banknote-net-open-dataset-for-assistive | 2204.03738 | null | https://arxiv.org/abs/2204.03738v1 | https://arxiv.org/pdf/2204.03738v1.pdf | BankNote-Net: Open dataset for assistive universal currency recognition | Millions of people around the world have low or no vision. Assistive software applications have been developed for a variety of day-to-day tasks, including optical character recognition, scene identification, person recognition, and currency recognition. This last task, the recognition of banknotes from different denominations, has been addressed by the use of computer vision models for image recognition. However, the datasets and models available for this task are limited, both in terms of dataset size and in variety of currencies covered. In this work, we collect a total of 24,826 images of banknotes in variety of assistive settings, spanning 17 currencies and 112 denominations. Using supervised contrastive learning, we develop a machine learning model for universal currency recognition. This model learns compliant embeddings of banknote images in a variety of contexts, which can be shared publicly (as a compressed vector representation), and can be used to train and test specialized downstream models for any currency, including those not covered by our dataset or for which only a few real images per denomination are available (few-shot learning). We deploy a variation of this model for public use in the last version of the Seeing AI app developed by Microsoft. We share our encoder model and the embeddings as an open dataset in our BankNote-Net repository. | ['Juan Lavista Ferres', 'Saqib Shaikh', 'Hemant Malhotra', 'Eugene Seleznev', 'Srinivas Vinnakota', 'Felipe Oviedo'] | 2022-04-07 | null | null | null | null | ['vision-based-navigation-with-language-based'] | ['robots'] | [-1.51387796e-01 -5.38800836e-01 -1.05496496e-01 -2.13488549e-01
-5.56809902e-01 -7.84489036e-01 6.49781287e-01 -1.78197473e-01
-5.57785451e-01 6.81406200e-01 4.17121440e-01 2.09044991e-03
5.48145652e-01 -8.62154007e-01 -4.62724715e-01 -3.86463225e-01
4.42285478e-01 3.66075248e-01 -2.91148484e-01 -2.28760973e-01
3.91299427e-01 5.08890986e-01 -1.19515407e+00 4.29687470e-01
4.29696411e-01 1.33148074e+00 -2.37402059e-02 9.77750540e-01
1.41103938e-01 9.63794470e-01 -8.59965801e-01 -1.06886101e+00
5.96834958e-01 8.83026272e-02 -5.69673061e-01 1.74790233e-01
8.45395982e-01 -7.12503374e-01 -9.16775107e-01 9.67795730e-01
8.39014471e-01 -1.56923249e-01 8.87289941e-01 -9.88721967e-01
-1.47773194e+00 7.16260448e-02 -1.32581875e-01 4.29016471e-01
4.13578957e-01 3.80209863e-01 9.43710804e-01 -1.15414095e+00
5.01986802e-01 1.18344331e+00 7.25176036e-01 5.51547289e-01
-6.79630220e-01 -5.35670698e-01 -2.40463540e-01 2.61146098e-01
-1.08362019e+00 -7.12896228e-01 4.68853265e-01 -5.50156713e-01
1.29558086e+00 -4.44265343e-02 9.25755382e-01 1.59201765e+00
5.00221737e-02 1.04520953e+00 1.08329034e+00 -4.61790264e-01
-3.85620706e-02 6.15261436e-01 3.34300727e-01 8.55359256e-01
1.50921375e-01 -8.84166211e-02 -7.93804109e-01 -1.78054377e-01
1.06355774e+00 2.02779770e-01 -4.48154509e-01 -2.11681761e-02
-1.20424175e+00 1.04793108e+00 7.68952370e-02 -3.23842503e-02
-3.53115648e-01 -5.97471371e-02 4.58573759e-01 5.48549950e-01
4.62827444e-01 4.25982475e-01 -4.97436345e-01 -3.30800325e-01
-6.85067832e-01 2.41075546e-01 1.21098518e+00 9.97201502e-01
4.51458067e-01 1.79119632e-01 -5.63872792e-03 1.16732371e+00
-1.59973800e-02 5.75528860e-01 7.42579937e-01 -5.48112750e-01
6.31079495e-01 6.11570179e-01 -3.34701091e-02 -6.14011824e-01
-1.05220238e-02 2.21532792e-01 -8.26020598e-01 1.84627235e-01
5.79792440e-01 -4.31321472e-01 -9.36508000e-01 1.08288324e+00
-9.25641432e-02 3.96507025e-01 3.99871558e-01 1.09248376e+00
9.71596718e-01 5.22173226e-01 -7.03738555e-02 3.16186249e-01
1.66788244e+00 -1.00909984e+00 -2.33652696e-01 -2.09830761e-01
2.00238392e-01 -9.38712835e-01 9.87767577e-01 5.06053448e-01
-6.88119948e-01 -3.84803683e-01 -1.08002925e+00 -3.31267953e-01
-8.60206425e-01 1.99045584e-01 9.26258326e-01 9.69389498e-01
-8.09903383e-01 3.33131492e-01 -1.92562267e-01 -6.56040788e-01
5.97820044e-01 1.43081784e-01 -5.45906663e-01 -1.88992947e-01
-1.14805734e+00 1.06443691e+00 1.05744652e-01 -1.47074044e-01
-8.63007843e-01 -5.88799775e-01 -9.98117507e-01 -1.41877875e-01
-9.10249650e-02 -5.82004726e-01 1.07450056e+00 -1.22391737e+00
-1.37651348e+00 1.20510626e+00 5.58515251e-01 -6.80586398e-01
5.26358843e-01 -3.58331621e-01 -5.47143877e-01 2.58692771e-01
-1.91229820e-01 6.08819127e-01 8.00991654e-01 -4.72917259e-01
-3.91653806e-01 -4.53174025e-01 3.02222162e-01 2.61901975e-01
-7.45254815e-01 2.76117027e-01 -6.72729731e-01 -1.06026101e+00
-6.67194903e-01 -7.89552569e-01 3.00486684e-01 1.92076847e-01
-3.01665276e-01 2.26467811e-02 7.83272147e-01 -1.20524704e+00
5.96383870e-01 -2.05493140e+00 -1.17902547e-01 -2.35151678e-01
-1.57717809e-01 4.53914791e-01 -1.87262997e-01 1.10192820e-01
8.45396519e-02 -2.23607302e-01 -9.66120511e-02 -4.22317237e-01
5.22204265e-02 -1.15216047e-01 -4.26458806e-01 5.54729819e-01
2.10281700e-01 9.65829670e-01 -5.01067400e-01 -3.49949569e-01
3.59123647e-01 5.16982436e-01 -2.64442980e-01 4.20077831e-01
9.18328539e-02 -1.38173476e-01 -2.21019417e-01 1.45444834e+00
7.27636993e-01 6.41474202e-02 -1.78264052e-01 -2.84465909e-01
1.55360669e-01 -1.84440583e-01 -1.09567726e+00 1.74027801e+00
-4.65721965e-01 7.71663725e-01 -1.96565211e-01 -1.09363759e+00
9.34498668e-01 3.58565331e-01 2.56827593e-01 -6.57136619e-01
-1.96412392e-02 7.55249858e-02 -5.47864437e-01 -5.97478747e-01
1.00368094e+00 -1.28856763e-01 -4.48451132e-01 5.51382959e-01
4.81406450e-01 2.08747722e-02 2.51416247e-02 7.87263215e-02
9.93647337e-01 -1.56697616e-01 2.48642460e-01 2.08510309e-01
1.01147763e-01 -2.15324108e-02 4.79650557e-01 7.12331355e-01
-5.18296123e-01 1.09068477e+00 1.36987016e-01 -7.67501116e-01
-1.35793352e+00 -1.12407780e+00 -3.12842727e-01 9.26941872e-01
2.19072178e-02 7.52089247e-02 -5.21314383e-01 -5.49012244e-01
3.61575514e-01 2.59105384e-01 -2.77978629e-01 -9.56365317e-02
-3.26855928e-01 -8.23408961e-01 9.81167197e-01 7.29553580e-01
1.05273736e+00 -1.42361021e+00 -7.01768160e-01 2.14755565e-01
2.13694223e-03 -1.30323219e+00 -7.33704686e-01 2.56486963e-02
-4.75806862e-01 -1.43318701e+00 -1.37994778e+00 -8.30996513e-01
1.85832590e-01 2.57850170e-01 1.22488821e+00 -5.96493296e-02
-8.00561190e-01 1.11979079e+00 -1.83920041e-01 -5.63709378e-01
9.02293026e-02 -1.50922045e-01 2.29806021e-01 1.49218217e-01
9.16654587e-01 -1.14900433e-03 -5.80908120e-01 -8.73628855e-02
-6.30625367e-01 -3.99037242e-01 4.51790631e-01 9.64349389e-01
4.06158678e-02 -6.03441656e-01 4.91336554e-01 -6.21994317e-01
8.13564599e-01 -5.10052264e-01 -6.18473947e-01 7.25927293e-01
-2.84194946e-01 -1.77548334e-01 3.08333218e-01 -6.02594137e-01
-8.18426013e-01 -5.20208627e-02 2.80045241e-01 -6.23585641e-01
-1.03693731e-01 2.76944131e-01 -1.46241665e-01 -2.55490607e-03
4.20607746e-01 2.92364448e-01 -1.64295718e-01 -6.07218385e-01
3.42549294e-01 1.34803998e+00 9.29190338e-01 -4.36726093e-01
2.94453591e-01 5.08810766e-02 -8.15245450e-01 -1.03126049e+00
-5.13510764e-01 -5.71318328e-01 -5.86055219e-01 -1.68718219e-01
8.41912627e-01 -1.34726143e+00 -5.29149413e-01 1.33204150e+00
-1.11677659e+00 -1.59142762e-01 -1.72058389e-01 5.57531118e-01
-3.93505573e-01 4.84079182e-01 -9.40897167e-01 -9.54375863e-01
-6.87484145e-01 -8.80010962e-01 9.61873114e-01 5.62029421e-01
-7.52609747e-04 -9.70804989e-01 1.18137732e-01 5.69736302e-01
2.28265226e-01 8.20282921e-02 6.86422646e-01 -5.57709455e-01
-5.30022919e-01 -3.26600999e-01 -4.82811987e-01 7.72424161e-01
3.54580246e-02 -1.87765718e-01 -1.14415038e+00 -2.79065937e-01
-3.61468703e-01 -9.68581319e-01 1.30317974e+00 2.78991222e-01
7.88664818e-01 -2.23413736e-01 2.31721029e-01 7.80426383e-01
1.45711446e+00 -9.03918520e-02 7.21414983e-01 6.58960104e-01
5.32383740e-01 3.35338742e-01 4.25123394e-01 7.53605962e-01
5.04374385e-01 5.69806933e-01 2.81957716e-01 5.10334335e-02
2.19484121e-02 3.47188972e-02 6.65145040e-01 2.26614371e-01
-2.15126753e-01 -6.95409775e-02 -8.07300150e-01 7.59250998e-01
-1.69273376e+00 -1.14621055e+00 5.06351292e-01 2.23813534e+00
7.93144107e-01 -1.78797707e-01 2.54718155e-01 -1.69513017e-01
7.31144667e-01 8.29370171e-02 -6.50742114e-01 -5.72083890e-01
-4.25863564e-01 2.55833209e-01 7.53666043e-01 3.71377282e-02
-1.44169176e+00 1.11531508e+00 6.74843073e+00 3.54959428e-01
-1.18994391e+00 -1.12541348e-01 9.56732750e-01 -1.99179724e-01
1.94702804e-01 -2.74945796e-01 -9.20300782e-01 5.79190016e-01
9.60033834e-01 1.19121457e-02 5.46838284e-01 1.10433745e+00
-2.75896221e-01 -2.46995613e-01 -1.09466827e+00 1.65535867e+00
6.53741419e-01 -1.50027871e+00 7.91532360e-03 7.53141241e-03
5.28025270e-01 2.09464431e-02 2.68974304e-01 5.70068240e-01
4.37665164e-01 -1.22350228e+00 7.75251865e-01 7.99956977e-01
1.18594551e+00 -7.35189259e-01 8.45551670e-01 1.57787073e-02
-9.35591102e-01 -2.34417826e-01 -8.23958516e-01 1.29384726e-01
5.50690293e-03 3.72098595e-01 -4.30079162e-01 1.37432620e-01
8.39625120e-01 9.31656778e-01 -5.13199747e-01 1.06495464e+00
4.62926365e-02 2.91323721e-01 -1.17176272e-01 4.12605926e-02
1.05738197e-03 -2.59106785e-01 1.21155016e-01 1.40796518e+00
4.20960963e-01 1.14089563e-01 1.44513741e-01 5.30872941e-01
-4.96360689e-01 -2.10162085e-02 -7.55687416e-01 -2.82745719e-01
5.78673899e-01 1.24200439e+00 -1.35523677e-01 -6.73639774e-01
-6.96327150e-01 1.56767428e+00 2.95064002e-01 2.93598443e-01
-7.43863702e-01 -6.29301012e-01 9.47133899e-01 -3.55498582e-01
3.00007612e-01 -1.62630558e-01 -1.34622961e-01 -1.80139077e+00
9.91201475e-02 -1.14926541e+00 7.07197309e-01 -8.24805379e-01
-1.64460480e+00 3.48277450e-01 -2.14594379e-01 -1.12306404e+00
-1.10832430e-01 -1.20774949e+00 -6.99599266e-01 1.00099158e+00
-1.53813159e+00 -1.47502971e+00 -4.22282428e-01 1.02846193e+00
8.52046549e-01 -1.25818348e+00 1.18201578e+00 3.46942216e-01
-6.35044515e-01 6.96447492e-01 1.83508620e-02 8.25932026e-01
1.04110038e+00 -1.29503298e+00 4.02522624e-01 6.09674037e-01
4.08516884e-01 3.97272676e-01 2.78025687e-01 -3.44024271e-01
-1.63121164e+00 -9.99448419e-01 8.64363730e-01 -4.45403010e-01
6.18943334e-01 -4.59235787e-01 -4.91296440e-01 7.37474978e-01
2.26826131e-01 2.04924390e-01 8.62068474e-01 -1.73320249e-02
-7.03162491e-01 -5.26275905e-03 -1.37993193e+00 3.22555661e-01
6.54662013e-01 -9.51307297e-01 -7.93873191e-01 4.13443059e-01
2.29233220e-01 -2.49927565e-01 -9.45649624e-01 -2.88507283e-01
7.99045742e-01 -8.06094170e-01 1.18086672e+00 -6.48902893e-01
3.88810396e-01 3.76553506e-01 -4.36006367e-01 -1.25488102e+00
-4.20223251e-02 -2.18411356e-01 -4.15246636e-01 1.28038061e+00
1.87636018e-01 -5.71061373e-01 8.99336457e-01 7.74428904e-01
1.75264508e-01 -5.03337681e-01 -9.16394114e-01 -5.62426209e-01
9.55018997e-02 -3.45945150e-01 5.09185910e-01 1.00634909e+00
4.32019942e-02 3.00574601e-02 -8.80434752e-01 -6.10390268e-02
4.60137099e-01 1.32455572e-01 8.07710111e-01 -1.01036739e+00
-5.96929789e-01 -1.36046559e-02 -9.26567614e-01 -1.08570707e+00
1.56118736e-01 -6.66778088e-01 -4.11961526e-01 -1.30221272e+00
4.65337515e-01 -2.95007229e-01 -3.49502325e-01 4.95108157e-01
1.07549287e-01 5.73044538e-01 4.24962133e-01 4.41942722e-01
-6.33291453e-02 4.49049652e-01 6.30579352e-01 -8.60729456e-01
8.94907638e-02 -1.32645518e-01 -7.75850713e-01 7.20479786e-01
6.94853425e-01 -1.78133976e-02 1.79022327e-01 -5.05544066e-01
-3.12528074e-01 -7.38717020e-02 6.50264680e-01 -1.09207439e+00
6.88544512e-02 -4.44677509e-02 8.43817353e-01 -3.99585575e-01
7.74365723e-01 -5.21237850e-01 -4.64148670e-01 2.40062639e-01
8.26929808e-02 2.48316452e-01 -6.92184791e-02 5.29939353e-01
-1.94539890e-01 -3.73590261e-01 6.95282578e-01 -5.56136370e-01
-1.28565145e+00 5.40741742e-01 -2.37574831e-01 6.46342710e-02
1.03878319e+00 -4.22166049e-01 -4.32597876e-01 -3.81866843e-01
-6.28716230e-01 -2.13615268e-01 4.40397352e-01 6.63066864e-01
8.68851900e-01 -1.46180844e+00 -9.09683108e-01 1.48171589e-01
3.49316001e-01 -5.34542918e-01 3.81866563e-03 4.62189138e-01
-6.99425280e-01 4.66524273e-01 -5.89709342e-01 -1.73360497e-01
-1.39595330e+00 4.27525938e-01 4.23576385e-01 -2.77065903e-01
-8.66854131e-01 7.07806408e-01 -3.00190806e-01 -3.94726872e-01
3.38236541e-01 -5.69640025e-02 -3.68938088e-01 3.39592218e-01
9.52825427e-01 1.56238124e-01 6.85336962e-02 -6.83214962e-01
-1.98038548e-01 5.26298821e-01 -3.11702460e-01 -1.82683989e-02
1.61754394e+00 2.11462408e-01 3.25594366e-01 3.22075665e-01
1.15033126e+00 -2.97421008e-01 -1.52582788e+00 -2.92079687e-01
-1.69434339e-01 -5.50784826e-01 -1.12013735e-01 -8.00760388e-01
-1.30483913e+00 8.54363143e-01 9.19189751e-01 -3.14870089e-01
9.20351923e-01 -2.75010675e-01 8.47138941e-01 7.39961624e-01
5.71396589e-01 -1.65840316e+00 3.10068250e-01 6.62676632e-01
7.51877367e-01 -1.55169475e+00 -3.26337926e-02 1.93509072e-01
-1.07948816e+00 1.42556500e+00 3.90759081e-01 -3.15603733e-01
3.42772514e-01 2.46089324e-01 3.20594728e-01 1.03180707e-02
-5.38485467e-01 2.87204999e-02 -9.25055817e-02 1.06392050e+00
4.70212609e-01 2.67451443e-02 2.04504207e-01 7.13000357e-01
-1.76813558e-01 4.61368859e-02 8.10293496e-01 9.23542380e-01
-3.49771708e-01 -9.45348501e-01 -5.38779914e-01 4.94350940e-01
-5.74020743e-01 -4.94024158e-01 -5.26895940e-01 4.95261431e-01
5.92736788e-02 9.09726262e-01 2.44438738e-01 -1.87530518e-01
1.90905318e-01 2.76127160e-01 4.30697143e-01 -5.86262107e-01
-6.65927529e-01 -6.33423746e-01 3.19394112e-01 -1.34881243e-01
-1.46285474e-01 -6.97319210e-01 -5.26701272e-01 -6.27210379e-01
-1.04593389e-01 -3.76892358e-01 6.06297612e-01 7.52418518e-01
9.52197686e-02 -4.73482423e-02 5.69846332e-01 -9.29694295e-01
-8.20507288e-01 -1.06870508e+00 -1.07378972e+00 4.96966928e-01
8.91119093e-02 -3.75534624e-01 -7.33889714e-02 3.44327480e-01] | [10.08338737487793, 2.0955958366394043] |
19648563-58db-4001-8d3c-39393b4892a5 | anticipation-and-next-action-forecasting-in | 1901.03728 | null | http://arxiv.org/abs/1901.03728v1 | http://arxiv.org/pdf/1901.03728v1.pdf | Anticipation and next action forecasting in video: an end-to-end model with memory | Action anticipation and forecasting in videos do not require a hat-trick, as
far as there are signs in the context to foresee how actions are going to be
deployed. Capturing these signs is hard because the context includes the past.
We propose an end-to-end network for action anticipation and forecasting with
memory, to both anticipate the current action and foresee the next one.
Experiments on action sequence datasets show excellent results indicating that
training on histories with a dynamic memory can significantly improve
forecasting performance. | ['Valsamis Ntouskos', 'Lorenzo Mauro', 'Elham Omrani', 'Fiora Pirri', 'Edoardo Alati', 'Mahdieh Izadpanahkakhk'] | 2019-01-11 | null | null | null | null | ['action-anticipation'] | ['computer-vision'] | [ 4.69805419e-01 -3.71051580e-02 -5.68949163e-01 -6.37518525e-01
-2.25418672e-01 -3.55087370e-01 7.09545016e-01 -3.28208774e-01
-3.31656426e-01 5.66765487e-01 8.92617345e-01 -1.01627395e-01
1.20179608e-01 -3.46861064e-01 -6.86190307e-01 -4.53512818e-01
-7.46713042e-01 5.73910438e-02 4.24019605e-01 -1.72428787e-01
1.27787843e-01 2.45278984e-01 -1.44631517e+00 7.83080578e-01
9.92151052e-02 1.16228509e+00 1.96138710e-01 1.22783160e+00
3.55234951e-01 1.75204396e+00 -5.02313375e-01 -1.55343592e-01
2.45445400e-01 -4.01557475e-01 -7.36917675e-01 2.86919624e-01
3.83290291e-01 -1.12362731e+00 -7.93947577e-01 4.69966173e-01
2.83893287e-01 3.57072502e-01 2.10420623e-01 -1.12999451e+00
-2.68778473e-01 7.67022908e-01 -7.72739574e-02 3.55720103e-01
8.08710098e-01 7.17315078e-01 7.66269505e-01 -3.07421118e-01
9.11864340e-01 1.22825730e+00 6.73490703e-01 7.83580720e-01
-3.40999424e-01 -3.01032126e-01 7.53705144e-01 6.85211837e-01
-6.97648585e-01 -6.12917066e-01 6.08146966e-01 -4.73410666e-01
1.21708059e+00 1.57685742e-01 8.13161254e-01 1.38345158e+00
4.24645275e-01 1.15749907e+00 6.51195765e-01 -1.84812218e-01
2.98604488e-01 -4.90925997e-01 -3.21251184e-01 5.46717644e-01
-4.36856776e-01 1.13877445e-01 -8.06031168e-01 1.29973084e-01
5.05969107e-01 4.03817922e-01 -1.48491800e-01 1.45270646e-01
-1.29248548e+00 4.44540143e-01 2.04021573e-01 -9.83761717e-03
-9.39998388e-01 6.85922444e-01 7.07341135e-01 2.66388804e-01
3.04614723e-01 2.41989285e-01 -4.00070459e-01 -8.55673552e-01
-8.04746091e-01 2.84574896e-01 6.64458454e-01 8.08173001e-01
3.29155862e-01 1.42705515e-01 -3.87292236e-01 1.14000887e-01
-8.53122771e-02 3.95995229e-01 4.31054860e-01 -1.59755301e+00
4.61842895e-01 4.22201127e-01 6.41963840e-01 -9.43743646e-01
-3.73070061e-01 2.28555530e-01 -3.42651218e-01 1.75349206e-01
3.99862319e-01 -5.74491799e-01 -9.89668906e-01 1.60817146e+00
2.55080640e-01 8.89071524e-01 1.45725757e-01 7.68766284e-01
3.85664463e-01 7.93999076e-01 3.62938315e-01 -3.13716441e-01
9.28702474e-01 -9.15394068e-01 -7.59635270e-01 -7.88806260e-01
8.46423447e-01 -4.61236864e-01 6.17302358e-01 3.56409132e-01
-9.65739787e-01 -5.60781300e-01 -7.21837997e-01 1.62963465e-01
7.46760219e-02 1.78289816e-01 1.01128376e+00 -4.86824885e-02
-9.39722240e-01 9.69707906e-01 -1.25107837e+00 -3.25871259e-01
4.23237085e-01 1.17314719e-01 -3.69831502e-01 -1.60394028e-01
-1.21462822e+00 9.13879395e-01 6.83784664e-01 4.43981200e-01
-1.19294417e+00 -1.06218010e-01 -8.12251627e-01 -2.97813360e-02
6.69552445e-01 -3.20625842e-01 1.57290196e+00 -1.15084791e+00
-1.60183847e+00 1.21790521e-01 -3.08396190e-01 -9.50482368e-01
6.53737009e-01 -6.64963543e-01 -5.64231515e-01 2.19243154e-01
-2.85627514e-01 7.83476710e-01 8.61972928e-01 -5.24646401e-01
-1.01909113e+00 6.90142661e-02 6.22662663e-01 2.29211748e-01
-3.36611152e-01 1.47080749e-01 -3.04119438e-01 -2.71104395e-01
-2.93864757e-01 -1.21294022e+00 -4.07489330e-01 6.57364875e-02
-1.70212269e-01 -2.11933196e-01 1.01832163e+00 -6.16945386e-01
1.44014621e+00 -1.82098806e+00 -2.13193819e-01 -2.70859271e-01
-1.68384403e-01 4.77259606e-01 -2.57087260e-01 8.90091121e-01
3.01848594e-02 -3.03256303e-01 4.10523623e-01 -1.11739509e-01
-2.75763813e-02 5.08307934e-01 -7.52506495e-01 2.96595603e-01
2.05036879e-01 6.93784118e-01 -1.25350690e+00 -1.90056041e-01
2.48925015e-01 5.38828790e-01 -4.75908786e-01 4.12689209e-01
-6.99408650e-01 5.13580382e-01 -5.56974292e-01 4.82593715e-01
1.31956011e-01 -3.91804188e-01 5.54440439e-01 1.65215179e-01
-1.86639994e-01 4.57481056e-01 -1.05946469e+00 1.62154877e+00
-3.91276062e-01 8.36749852e-01 -3.93655002e-01 -5.68873346e-01
4.75153714e-01 5.06168306e-01 5.89948058e-01 -4.54587996e-01
-1.81462511e-01 -1.33764207e-01 8.79770815e-02 -1.11396778e+00
5.51264822e-01 1.01569071e-01 -3.98486033e-02 4.18346286e-01
-3.28064442e-01 4.07003582e-01 5.20075858e-01 2.21945807e-01
1.55833840e+00 5.87711275e-01 1.48285136e-01 6.84069812e-01
2.64827490e-01 2.02422619e-01 1.02385318e+00 7.84590006e-01
-3.23628187e-01 1.61025360e-01 6.32123590e-01 -9.97520089e-01
-5.44338942e-01 -3.89346808e-01 6.85579598e-01 1.43282807e+00
-2.16129109e-01 -5.60990572e-01 -4.61862385e-01 -8.02727282e-01
-4.21980530e-01 9.36366260e-01 -5.91144621e-01 -2.22850561e-01
-9.32871282e-01 -1.64115541e-02 1.86565202e-02 1.12838852e+00
4.58451897e-01 -1.35394919e+00 -1.33748055e+00 4.11610812e-01
-2.31400639e-01 -1.30371511e+00 -6.66180968e-01 -2.12085918e-01
-8.22108626e-01 -1.04877222e+00 -4.40839320e-01 -2.80516356e-01
2.29053766e-01 1.22580037e-01 1.13046038e+00 1.12316944e-01
1.53147936e-01 6.26030564e-01 -4.12103027e-01 -3.00168872e-01
-4.35988128e-01 -5.52186131e-01 1.22896574e-01 -7.95201957e-03
3.18246067e-01 -6.14355206e-01 -9.50191975e-01 2.30148271e-01
-6.79552257e-01 1.38153315e-01 4.76799250e-01 6.10031128e-01
2.65594810e-01 1.26739711e-01 3.59647572e-01 -7.34345675e-01
2.18075767e-01 -3.51624280e-01 -3.25505167e-01 4.69186306e-01
-2.18635216e-01 -1.45411998e-01 6.02678418e-01 -7.81887174e-01
-1.17791069e+00 5.88793159e-01 -8.66875052e-02 -4.89044458e-01
-2.00312391e-01 4.37454671e-01 9.67150182e-02 3.88632745e-01
3.11184704e-01 1.68574512e-01 -3.65687639e-01 -3.26095968e-01
4.39008802e-01 2.74930328e-01 6.17579520e-01 -9.20029804e-02
1.93760291e-01 6.30022407e-01 4.69977409e-02 -3.75239521e-01
-1.06238508e+00 -3.75533402e-01 -6.63552284e-01 -7.30605423e-01
6.15620375e-01 -1.09976947e+00 -1.05601859e+00 4.79199857e-01
-1.30739510e+00 -5.97803414e-01 -2.31932491e-01 6.73337877e-01
-7.85778701e-01 2.74092734e-01 -6.60182714e-01 -1.04795134e+00
-4.57042828e-02 -1.05682135e+00 8.54085565e-01 1.88951433e-01
-5.12988389e-01 -9.69901800e-01 6.65352121e-02 -4.91782986e-02
3.62524420e-01 4.70526546e-01 1.87493011e-01 -4.98383403e-01
-8.33591700e-01 -4.90217745e-01 3.62010211e-01 3.06513011e-01
-7.17144459e-02 9.27396938e-02 -7.20584393e-01 -7.26141632e-02
-2.12211251e-01 -2.91435152e-01 9.16099906e-01 4.92169082e-01
9.91245687e-01 -8.36730361e-01 -2.88977474e-01 2.66238421e-01
1.01452982e+00 6.35320246e-01 8.66561294e-01 1.42867133e-01
5.61482906e-01 5.29253125e-01 1.04723203e+00 8.71519566e-01
5.22657752e-01 5.94605863e-01 5.32989800e-01 3.39227349e-01
-1.47600576e-01 -5.28684855e-01 9.39762235e-01 3.89751434e-01
-2.98591018e-01 -5.29670537e-01 -7.31272638e-01 4.34904188e-01
-2.18278408e+00 -1.59014893e+00 9.11270157e-02 1.91781616e+00
5.45826554e-01 4.10987109e-01 2.87146926e-01 -2.08298981e-01
4.59794968e-01 6.43522739e-01 -8.80006075e-01 -2.95950949e-01
2.51541466e-01 -3.60070825e-01 4.36058372e-01 3.71904731e-01
-1.28501499e+00 9.85850573e-01 7.39530945e+00 5.02514780e-01
-1.29477942e+00 -1.49121046e-01 6.67525709e-01 -3.88377547e-01
1.15600102e-01 2.98689216e-01 -7.43132472e-01 6.18074834e-01
1.15244734e+00 1.97930291e-01 2.24966273e-01 1.01903689e+00
5.41519403e-01 -4.16259080e-01 -1.28321934e+00 6.91357851e-01
-1.35243580e-01 -1.51011562e+00 -1.45116016e-01 -3.89669210e-01
5.63784063e-01 2.52084266e-02 -1.53163284e-01 2.03343764e-01
5.95191896e-01 -7.67154336e-01 6.98641539e-01 1.00665951e+00
4.67727214e-01 -4.74530429e-01 3.93208861e-01 5.38954198e-01
-1.07106054e+00 -5.52509367e-01 -3.73391032e-01 -5.67193806e-01
6.75495505e-01 2.37518728e-01 -1.09383011e+00 -2.37094592e-02
3.74650627e-01 1.25597286e+00 -3.41080546e-01 8.93907666e-01
-5.13315737e-01 9.14265990e-01 -9.21555385e-02 -1.20552041e-01
3.84906590e-01 1.94991082e-01 5.08674145e-01 1.20540249e+00
4.43572313e-01 2.79721856e-01 4.93804038e-01 4.06796625e-03
3.01732063e-01 -5.73317826e-01 -7.39544272e-01 -4.85229343e-01
3.24830115e-01 9.72237945e-01 -5.16908944e-01 -6.31411135e-01
-4.54263330e-01 9.84769940e-01 -9.95170325e-02 3.23909789e-01
-9.32391047e-01 8.10438097e-02 6.31700695e-01 3.83677185e-02
5.67982376e-01 -3.17456096e-01 1.83499396e-01 -9.75275636e-01
2.31752060e-02 -6.89583957e-01 5.96915722e-01 -1.11501813e+00
-7.31138229e-01 4.27821159e-01 -1.12031087e-01 -1.64350784e+00
-9.54215527e-01 -4.63749409e-01 -8.99107397e-01 1.79267406e-01
-1.27971888e+00 -1.09664333e+00 -1.02060154e-01 3.70054364e-01
7.24504232e-01 1.92898914e-01 5.57847857e-01 -3.21511552e-02
-3.25323075e-01 1.20533248e-02 -1.85353965e-01 1.35547474e-01
4.82838362e-01 -8.99925172e-01 5.56399524e-01 1.05835092e+00
1.23818047e-01 2.93925971e-01 1.09786999e+00 -8.35812271e-01
-1.77460968e+00 -8.68954301e-01 1.02849877e+00 -5.78722298e-01
8.19417238e-01 -1.64598995e-03 -6.86009943e-01 1.14973497e+00
2.85531819e-01 -7.02599529e-03 3.17675382e-01 -1.01508480e-02
-2.10504159e-01 -5.41636720e-02 -6.00097001e-01 6.04754210e-01
1.26855731e+00 -4.46461260e-01 -3.42956275e-01 6.97488546e-01
7.28446305e-01 -6.46684706e-01 -6.39501572e-01 2.21808091e-01
8.87129128e-01 -1.05890238e+00 9.26321507e-01 -9.58867133e-01
6.79281235e-01 -1.24161445e-01 -2.55915001e-02 -1.14943087e+00
-1.47764623e-01 -1.14693928e+00 -6.44783020e-01 9.41004813e-01
1.56841010e-01 -2.63909549e-01 8.42513204e-01 1.22660327e+00
-1.91989899e-01 -8.24323118e-01 -7.01653004e-01 -8.37957263e-01
-7.00007498e-01 -6.16772056e-01 7.22579598e-01 4.13193583e-01
1.09615043e-01 3.29002254e-02 -1.16398275e+00 -9.81547125e-03
-9.17169154e-02 3.14240783e-01 7.55528390e-01 -6.60075903e-01
-4.57025737e-01 3.35963853e-02 -5.42797804e-01 -1.45790029e+00
1.96840569e-01 -2.65175462e-01 3.68703157e-01 -1.53776193e+00
3.95453982e-02 7.11111203e-02 -2.87489563e-01 7.04434097e-01
-1.89442590e-01 -1.65089145e-01 4.09618199e-01 8.38865787e-02
-1.24286711e+00 3.29272062e-01 1.22190297e+00 -1.57421436e-02
-7.65772462e-02 3.21955174e-01 -2.92205006e-01 1.10277855e+00
6.21600628e-01 -4.85963196e-01 -5.67887306e-01 -4.44994718e-01
2.56086320e-01 7.05051661e-01 4.02219713e-01 -1.28414917e+00
4.20923740e-01 -6.45278454e-01 3.60589027e-01 -7.78162599e-01
8.34260821e-01 -7.31406868e-01 3.36609006e-01 5.43026328e-01
-8.01452041e-01 3.56347799e-01 -9.99435261e-02 9.65590954e-01
-4.27525900e-02 -1.05318408e-02 2.65992194e-01 -1.73175767e-01
-1.55913866e+00 4.46559876e-01 -5.61086059e-01 -1.46943808e-01
9.73043561e-01 -2.38343984e-01 -3.20481002e-01 -1.06417322e+00
-8.83913100e-01 4.07128781e-01 3.13874543e-01 4.01223361e-01
6.85814977e-01 -1.25474954e+00 -5.08885503e-01 -3.43562663e-01
-2.17007205e-01 -5.45507371e-01 5.58114886e-01 7.58634210e-01
-2.51784980e-01 2.59691149e-01 -3.44536513e-01 -3.94687176e-01
-1.37123489e+00 6.73531890e-01 1.10794537e-01 -4.12757725e-01
-7.92990208e-01 9.68757868e-01 -6.98879063e-02 3.72493178e-01
6.02042079e-01 -3.39478940e-01 -3.76964152e-01 -1.94530934e-02
1.11885250e+00 4.61054534e-01 -5.51069975e-01 -5.72667241e-01
-3.93197477e-01 1.57218590e-01 -8.92481953e-02 1.51806476e-03
1.61261034e+00 -2.35247806e-01 2.44816318e-01 5.69290221e-01
8.83114517e-01 -4.21255469e-01 -2.22041798e+00 -1.42440110e-01
2.28638947e-01 -7.59619236e-01 -9.50345024e-02 -9.05514657e-01
-1.00000668e+00 8.04719508e-01 5.33510208e-01 -2.20848303e-02
1.33991873e+00 -3.96043539e-01 1.22819555e+00 7.59296894e-01
3.83131385e-01 -1.42285538e+00 1.60635725e-01 7.53544450e-01
9.33883190e-01 -1.15436304e+00 2.13545963e-01 2.75861621e-02
-1.23009622e+00 1.33079541e+00 5.17140806e-01 -9.58414376e-02
5.47630310e-01 1.68075219e-01 8.70876312e-02 -2.35646769e-01
-1.49979329e+00 -2.59056002e-01 1.27822548e-01 4.63532865e-01
2.65717685e-01 7.47218430e-02 1.71364263e-01 -4.57172021e-02
2.94020891e-01 4.90085423e-01 6.54550552e-01 1.11932158e+00
-4.65672404e-01 -8.58850062e-01 -2.81289779e-02 3.46688986e-01
-4.26241219e-01 1.67378157e-01 -3.31165969e-01 4.86141175e-01
3.52937169e-02 1.12493515e+00 1.17157944e-01 -6.06634378e-01
1.89828739e-01 9.95388031e-02 3.70493144e-01 -3.07505935e-01
-4.36252028e-01 -1.48667082e-01 6.14522576e-01 -1.15666771e+00
-7.66029000e-01 -7.32022285e-01 -1.21641278e+00 -3.78128231e-01
-4.04288843e-02 -2.98486382e-01 4.43527460e-01 1.02160418e+00
6.21765316e-01 4.07196581e-01 6.29697442e-01 -1.05433369e+00
-5.72611392e-01 -8.75983357e-01 -1.41820118e-01 3.65971982e-01
4.31133837e-01 -5.64317882e-01 -4.73295525e-02 3.17778945e-01] | [8.057389259338379, 0.4959203898906708] |
84edc4be-65b0-4caf-8193-68d817ebc6c8 | leveraging-intra-and-inter-dataset-variations | null | null | http://openaccess.thecvf.com/content_cvpr_2017_workshops/w33/html/Wu_Leveraging_Intra_and_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017_workshops/w33/papers/Wu_Leveraging_Intra_and_CVPR_2017_paper.pdf | Leveraging intra and inter-dataset variations for robust face alignment | Face alignment is a critical topic in the computer vision community. Numerous efforts have been made and various benchmark datasets have been released in recent decades. However, two significant issues remain in recent datasets, e.g., Intra-Dataset Variation and Inter-Dataset Variation. Inter-Dataset Variation refers to bias on expression, head pose, etc. inside one certain dataset, while Intra-Dataset Variation refers to different bias across different datasets. To address the mentioned problems, we proposed a novel Deep Variation Leveraging Network (DVLN), which consists of two strong coupling sub-networks, e.g., Dataset-Across Network (DA-Net) and Candidate-Decision Network (CD-Net). Extensive evaluations show that our approach demonstrates real-time performance and dramatically outperforms state-of-the-art methods on the challenging 300-W dataset. | ['Wenyan Wu', 'Shuo Yang'] | 2017-07-21 | null | null | null | cvpr-2017-2017-7 | ['robust-face-alignment'] | ['computer-vision'] | [ 6.26622811e-02 -2.36933544e-01 -1.94637671e-01 -1.02559805e+00
-3.74769360e-01 -1.31551683e-01 5.26096702e-01 -5.61176360e-01
-8.94134566e-02 5.64748347e-01 1.84244052e-01 2.86332160e-01
-2.38257438e-01 -3.65071416e-01 -2.72071451e-01 -5.92019141e-01
1.27547055e-01 2.89281309e-01 1.56786680e-01 -2.62941360e-01
-6.49311617e-02 6.19390428e-01 -1.32648706e+00 -1.11354873e-01
3.53297323e-01 1.02190316e+00 -4.22867954e-01 -1.20691940e-01
-2.20097095e-01 3.70787919e-01 -6.24752402e-01 -6.87704563e-01
4.26887631e-01 -4.43555892e-01 -2.55945146e-01 2.31713772e-01
9.14800227e-01 -1.55057967e-01 -4.32433635e-01 1.30398858e+00
1.16674793e+00 -1.31726682e-01 1.83227271e-01 -1.92081785e+00
-5.78762531e-01 5.71635365e-01 -1.25309467e+00 2.80896455e-01
-2.28451993e-02 3.22947949e-01 7.21849501e-01 -1.03785443e+00
6.76273763e-01 1.63808060e+00 8.24752569e-01 8.39558601e-01
-1.11921680e+00 -1.35736346e+00 1.75835893e-01 2.36875236e-01
-1.57313657e+00 -7.45715737e-01 1.15613711e+00 -3.37468624e-01
4.43478674e-01 1.53575704e-01 4.46756631e-01 1.32074666e+00
1.92474321e-01 6.91409707e-01 1.02682304e+00 -1.22837909e-02
-5.11505716e-02 -9.01961029e-02 2.03566208e-01 5.20877182e-01
2.82445073e-01 9.29119159e-03 -7.64387369e-01 -2.10702688e-01
4.96709913e-01 -1.30494773e-01 -4.59140271e-01 -4.60674763e-01
-6.36962116e-01 7.58015037e-01 2.51528472e-01 2.45089710e-01
2.21165255e-01 1.11302333e-02 7.63099432e-01 1.41395032e-01
5.46517849e-01 -2.09989786e-01 -6.02849901e-01 -1.56595856e-02
-9.88913476e-01 3.60103697e-01 5.12070179e-01 1.07348967e+00
5.37658930e-01 2.55699784e-01 -3.77852768e-01 1.36334324e+00
3.56769681e-01 4.40079033e-01 5.73607802e-01 -5.33782303e-01
8.37242544e-01 5.11168540e-01 -5.41658342e-01 -1.42149603e+00
-7.62795746e-01 -3.76698762e-01 -1.31587124e+00 -1.50093600e-01
2.06623092e-01 -3.57529670e-01 -9.05682325e-01 2.28424311e+00
5.76447308e-01 4.32277054e-01 -3.92368436e-01 7.79808819e-01
1.36259365e+00 1.67831089e-02 -2.86947452e-02 -3.42740655e-01
1.03784871e+00 -9.49241161e-01 -9.61314976e-01 -2.80931294e-01
1.92962244e-01 -8.63977313e-01 6.62466645e-01 1.81519333e-02
-8.56086671e-01 -7.71808982e-01 -9.74581838e-01 1.07212313e-01
-1.94018319e-01 6.20573834e-02 4.68258888e-01 7.83539355e-01
-1.20376325e+00 2.61254698e-01 -4.60425615e-01 -4.35677767e-01
8.77851963e-01 7.80813575e-01 -5.39571643e-01 -2.18890145e-01
-9.93712842e-01 5.03236473e-01 4.56766738e-03 5.10506809e-01
-6.74058795e-01 -9.02462661e-01 -7.41468310e-01 -2.79996634e-01
4.94243294e-01 -3.91450047e-01 1.05005717e+00 -1.02629256e+00
-1.47139800e+00 9.74124253e-01 -1.30154103e-01 5.65601699e-02
6.99773252e-01 -1.24625854e-01 -9.41834748e-01 -4.92798209e-01
-1.55503616e-01 5.79584301e-01 7.60043085e-01 -1.10565495e+00
-2.68999487e-01 -9.58232343e-01 -3.88394654e-01 -1.54059246e-01
-2.53016531e-01 3.62267941e-01 -1.03447425e+00 -8.45952749e-01
2.59992462e-02 -1.17589223e+00 2.16094136e-01 4.60260808e-01
-7.29590237e-01 -4.32599455e-01 1.21648955e+00 -5.07130802e-01
1.31567597e+00 -2.10112047e+00 9.04042944e-02 3.88639987e-01
4.59993035e-01 6.97367132e-01 -5.10009944e-01 -1.39472894e-02
-4.89669442e-01 -8.94590914e-02 -1.67846382e-01 -5.02575755e-01
4.44203094e-02 2.87916780e-01 -2.57600714e-02 7.41010010e-01
7.97740519e-02 8.18294823e-01 -4.24863428e-01 -4.99603242e-01
-1.24561675e-01 5.67501605e-01 -4.50048983e-01 1.92994863e-01
1.11964621e-01 4.45919156e-01 -3.90901119e-01 8.97392154e-01
1.26548386e+00 5.83123490e-02 2.66787320e-01 -5.73138952e-01
2.04935506e-01 -3.72480929e-01 -1.21818483e+00 1.67933404e+00
-6.64722770e-02 8.06015611e-01 7.52628297e-02 -7.99577236e-01
1.06059813e+00 7.25833327e-02 4.63689297e-01 -7.41061330e-01
4.72761303e-01 4.52836864e-02 3.87927204e-01 -5.54142773e-01
3.37187737e-01 2.63795018e-01 3.59481037e-01 2.51511067e-01
1.70557678e-01 4.33050603e-01 4.58694771e-02 -1.72436997e-01
8.31072688e-01 2.26804297e-02 9.64003950e-02 -2.73953527e-01
6.14502370e-01 -7.68037140e-01 1.12537026e+00 1.35214582e-01
-7.89010108e-01 6.79977000e-01 5.68350375e-01 -4.70274508e-01
-6.17721438e-01 -8.00552785e-01 -3.56114686e-01 8.83149326e-01
1.16965979e-01 -2.45287299e-01 -6.95663810e-01 -6.93761110e-01
2.35228434e-01 1.97382867e-02 -8.70258868e-01 -1.76453546e-01
-6.68940723e-01 -9.98174787e-01 6.80683315e-01 6.58117950e-01
9.95694220e-01 -7.61580169e-01 1.45524964e-01 2.67083161e-02
1.22632891e-01 -1.31699491e+00 -7.38714159e-01 -2.00447574e-01
-5.26560187e-01 -1.09025788e+00 -4.97897804e-01 -7.24356830e-01
6.88840032e-01 2.14600310e-01 1.18110812e+00 4.62596864e-02
-6.66358352e-01 9.17902514e-02 4.90967073e-02 -4.59177285e-01
3.27637196e-01 -8.84754956e-02 1.67963102e-01 2.52321213e-01
7.17712581e-01 -4.60486293e-01 -6.52164996e-01 6.72240138e-01
-6.37336373e-01 -5.20325661e-01 3.01265001e-01 8.33330691e-01
5.89166284e-01 -2.67220467e-01 5.51953018e-01 -9.95288491e-01
5.81280053e-01 -5.01225233e-01 -6.70456350e-01 3.43675256e-01
-5.21246493e-01 -2.39011154e-01 1.99084505e-01 -5.36433220e-01
-8.06364357e-01 -1.08770944e-01 -1.00422680e-01 -4.71048564e-01
9.27907676e-02 8.13049823e-02 -7.63551116e-01 -3.73320311e-01
3.35956424e-01 -1.03618823e-01 1.58154979e-01 -1.98130265e-01
1.31025299e-01 6.43644810e-01 5.85865855e-01 -5.74902654e-01
9.83109295e-01 2.72501409e-01 2.06149280e-01 -6.15447462e-01
-8.05390179e-01 -3.03437352e-01 -6.44629240e-01 -2.76900709e-01
6.12556219e-01 -1.02032804e+00 -5.51254988e-01 1.00490856e+00
-1.08410001e+00 -1.50612831e-01 2.05019861e-01 1.89638302e-01
9.56566110e-02 1.50708914e-01 -2.56352276e-01 -3.85282844e-01
-5.52921951e-01 -1.35078716e+00 8.25506032e-01 5.23244858e-01
-1.61587924e-01 -1.00024378e+00 8.82716328e-02 3.11649740e-01
6.10963047e-01 4.55245584e-01 5.26081085e-01 -6.16343319e-01
-2.42645621e-01 4.14102897e-02 -4.86610532e-01 3.42388481e-01
4.76487130e-01 4.16158408e-01 -1.04896331e+00 -4.42581207e-01
-2.34207571e-01 -2.64832914e-01 4.33643639e-01 4.46640283e-01
1.60762143e+00 -3.56010795e-02 -4.53334183e-01 9.43089485e-01
1.31519413e+00 1.08541422e-01 5.94731212e-01 2.31215525e-02
8.82655442e-01 6.26453400e-01 4.13785726e-01 4.42849129e-01
3.94407243e-01 1.13516426e+00 2.83893883e-01 -1.71653673e-01
-2.60458171e-01 1.12429626e-01 8.46884549e-02 6.13261282e-01
2.17564881e-01 -2.28782162e-01 -8.09080422e-01 4.17899072e-01
-1.81660736e+00 -8.57736111e-01 -1.23428114e-01 1.83712220e+00
8.27044189e-01 -6.89500347e-02 2.29368523e-01 -1.69778988e-01
9.76566017e-01 5.06192088e-01 -9.64926302e-01 -2.04087228e-01
-2.85107851e-01 2.20983520e-01 2.90022880e-01 8.44558999e-02
-9.62764144e-01 8.18623245e-01 6.23525143e+00 9.52623487e-01
-1.33597291e+00 2.28394881e-01 8.84528816e-01 -3.31710428e-01
-1.29954562e-01 -6.36332512e-01 -1.23672783e+00 7.33778775e-01
3.06223571e-01 -1.08154625e-01 1.85154378e-01 8.50438833e-01
4.36109267e-02 1.47906244e-01 -1.09137690e+00 1.47086263e+00
7.09562957e-01 -8.54136586e-01 -8.75305906e-02 2.88199279e-02
8.49225163e-01 1.65820464e-01 2.38683045e-01 2.56358594e-01
1.87230572e-01 -9.73870277e-01 1.89023718e-01 1.49941862e-01
7.27070868e-01 -1.03196561e+00 8.00548136e-01 -2.71581322e-01
-1.33731914e+00 2.35516325e-01 -4.43826854e-01 5.78301430e-01
-1.84777647e-01 9.02772307e-01 -2.15429276e-01 3.99558187e-01
7.53729820e-01 7.59687483e-01 -6.98155820e-01 9.41003323e-01
-2.09748954e-01 3.01314414e-01 -2.12011084e-01 2.66806185e-01
-2.68027298e-02 -2.93138921e-01 3.63324225e-01 9.80668545e-01
1.91194043e-01 -2.67137557e-01 1.61622856e-02 5.98652959e-01
-5.70782244e-01 -1.74641106e-02 -2.00872034e-01 3.59768242e-01
6.25030756e-01 1.56322610e+00 -3.66611153e-01 3.17041911e-02
-5.36105037e-01 7.86636591e-01 3.33916247e-01 2.48994991e-01
-1.07043540e+00 -2.63756901e-01 1.27323282e+00 -1.23970956e-01
3.01647156e-01 1.00170923e-02 -8.65473449e-02 -9.37865496e-01
2.49632642e-01 -1.04991257e+00 2.69975394e-01 -4.29513037e-01
-1.49027693e+00 7.70761490e-01 -1.39114842e-01 -1.14741910e+00
6.44084364e-02 -5.21741331e-01 -5.73007286e-01 6.79357529e-01
-1.71175599e+00 -1.17831123e+00 -7.24067926e-01 8.27174127e-01
4.13140267e-01 -6.91301584e-01 4.12202686e-01 7.05400884e-01
-1.28909850e+00 1.37245905e+00 5.67453029e-03 4.55533683e-01
1.20915294e+00 -6.76601052e-01 3.30322385e-01 8.70334148e-01
4.84094359e-02 7.93693602e-01 4.45991814e-01 -4.38308328e-01
-1.43090308e+00 -1.49733281e+00 7.48059750e-01 -1.46251470e-01
5.52643538e-01 -3.87789249e-01 -8.45349073e-01 8.05193782e-01
1.47379011e-01 6.96261168e-01 9.37174380e-01 2.76514381e-01
-7.22645164e-01 -7.04299331e-01 -1.35384572e+00 6.02361858e-01
1.44003916e+00 -4.65081394e-01 -1.72286496e-01 2.47128531e-01
3.60568196e-01 -8.70521963e-01 -6.58037663e-01 5.79106927e-01
6.96199775e-01 -9.51763630e-01 7.79352784e-01 -5.17204285e-01
4.02541906e-01 -3.76438618e-01 -2.08188817e-01 -1.43820548e+00
-2.18456581e-01 -6.85839593e-01 -8.30123648e-02 1.91674209e+00
1.69142529e-01 -5.82198024e-01 7.89637685e-01 8.89332592e-01
1.30096421e-01 -1.00604427e+00 -1.04283011e+00 -7.87425458e-01
-1.13257542e-01 -4.49494213e-01 1.09467268e+00 1.19750977e+00
-6.53625548e-01 2.54805028e-01 -6.08561277e-01 1.19246878e-01
6.77182078e-01 6.82597831e-02 1.02768135e+00 -1.24295640e+00
2.60935407e-02 -5.42420149e-01 -7.09681809e-01 -7.12097168e-01
4.38003898e-01 -7.96185434e-01 -1.05887868e-01 -9.90616798e-01
5.90658665e-01 -4.41047460e-01 -2.81296402e-01 6.10708892e-01
-2.93996960e-01 5.12728810e-01 1.25811785e-01 -5.53418435e-02
-4.71969604e-01 6.44303322e-01 1.13006783e+00 -2.08914042e-01
-9.77439433e-02 -2.09269613e-01 -7.78552890e-01 7.04099536e-01
8.14709961e-01 -3.46227288e-01 -3.90701771e-01 -6.06060207e-01
-1.45556599e-01 -4.64942813e-01 9.15697590e-02 -9.01858807e-01
3.42896134e-01 -7.18884319e-02 6.03741765e-01 -6.58858478e-01
2.03793347e-01 -8.38756680e-01 1.72316089e-01 1.18335456e-01
-1.28499642e-01 2.45320618e-01 4.77640241e-01 3.39724779e-01
-4.15503442e-01 3.51814657e-01 1.01952910e+00 4.26776260e-01
-7.63532281e-01 9.48504031e-01 4.52931851e-01 4.25157040e-01
1.07807732e+00 -5.71699105e-02 -5.64446926e-01 -1.30597681e-01
-3.15729231e-01 4.12253320e-01 -2.85699237e-02 8.60458434e-01
4.32662964e-01 -1.76651621e+00 -7.40688503e-01 2.50705570e-01
4.11566854e-01 -8.13254565e-02 3.58891815e-01 7.76191473e-01
7.34467153e-03 1.90854713e-01 -4.33996707e-01 -7.77224243e-01
-1.80314386e+00 1.84820499e-03 4.30501938e-01 -1.92536235e-01
-1.05013199e-01 1.25172138e+00 1.87549040e-01 -6.79705322e-01
4.37626302e-01 -3.92938592e-02 -2.54400551e-01 3.72297198e-01
6.62983537e-01 4.03594136e-01 -4.04410437e-02 -1.07272863e+00
-6.57527924e-01 1.21937084e+00 -2.69255012e-01 5.47412872e-01
1.30747831e+00 -6.42991439e-02 -2.83370018e-01 9.37511921e-02
1.56493306e+00 -2.32908681e-01 -1.05446506e+00 -4.89099503e-01
-1.55662075e-01 -6.70519114e-01 -4.21874151e-02 -7.49121368e-01
-2.00628281e+00 6.46974564e-01 1.00307155e+00 -4.70252067e-01
1.20944262e+00 -3.49370360e-01 7.92511523e-01 6.12763427e-02
4.09295678e-01 -1.35081840e+00 1.50840431e-01 3.75917286e-01
1.15174270e+00 -1.56890023e+00 3.42282504e-01 -5.78588128e-01
-5.05564809e-01 1.06148696e+00 1.10956180e+00 1.90249473e-01
9.96964276e-01 3.93857539e-01 3.09063464e-01 -3.13794583e-01
-4.70808536e-01 4.83659934e-03 4.43337768e-01 6.18482530e-01
5.08732736e-01 -8.35812762e-02 -1.75259292e-01 4.84268546e-01
-2.94586480e-01 2.03371480e-01 -2.71405187e-02 6.22047961e-01
1.88907668e-01 -1.35697889e+00 -1.85847595e-01 3.85901034e-01
-5.07853270e-01 2.18113720e-01 -5.85080862e-01 1.07445085e+00
4.77806211e-01 9.73452866e-01 1.11516580e-01 -5.39670229e-01
4.66364026e-01 -1.79808021e-01 5.14047503e-01 -3.70918095e-01
-7.54924536e-01 -1.42814994e-01 -4.54327837e-02 -9.87562060e-01
-7.32412338e-01 -6.18403018e-01 -8.89748633e-01 -4.89068270e-01
-2.92001039e-01 -3.31691921e-01 5.73602378e-01 9.11270738e-01
4.85378563e-01 6.09089136e-01 8.82003725e-01 -6.00776196e-01
-3.07108670e-01 -1.10685515e+00 -6.30085528e-01 3.51713955e-01
2.85979420e-01 -8.01500618e-01 -4.55813020e-01 -3.02735031e-01] | [13.362777709960938, 0.6002155542373657] |
9485c4e2-47e1-4c4e-926f-cf3e0df6616c | dense-siamese-network | 2203.11075 | null | https://arxiv.org/abs/2203.11075v2 | https://arxiv.org/pdf/2203.11075v2.pdf | Dense Siamese Network for Dense Unsupervised Learning | This paper presents Dense Siamese Network (DenseSiam), a simple unsupervised learning framework for dense prediction tasks. It learns visual representations by maximizing the similarity between two views of one image with two types of consistency, i.e., pixel consistency and region consistency. Concretely, DenseSiam first maximizes the pixel level spatial consistency according to the exact location correspondence in the overlapped area. It also extracts a batch of region embeddings that correspond to some sub-regions in the overlapped area to be contrasted for region consistency. In contrast to previous methods that require negative pixel pairs, momentum encoders or heuristic masks, DenseSiam benefits from the simple Siamese network and optimizes the consistency of different granularities. It also proves that the simple location correspondence and interacted region embeddings are effective enough to learn the similarity. We apply DenseSiam on ImageNet and obtain competitive improvements on various downstream tasks. We also show that only with some extra task-specific losses, the simple framework can directly conduct dense prediction tasks. On an existing unsupervised semantic segmentation benchmark, it surpasses state-of-the-art segmentation methods by 2.1 mIoU with 28% training costs. Code and models are released at https://github.com/ZwwWayne/DenseSiam. | ['Chen Change Loy', 'Kai Chen', 'Jiangmiao Pang', 'Wenwei Zhang'] | 2022-03-21 | null | null | null | null | ['unsupervised-semantic-segmentation'] | ['computer-vision'] | [ 2.19199006e-02 4.81765151e-01 -3.55654299e-01 -5.48564315e-01
-6.46011710e-01 -2.99808204e-01 4.68206167e-01 -6.15604185e-02
-4.72462147e-01 4.20250863e-01 5.19363806e-02 1.42216921e-01
1.29341036e-01 -6.81580007e-01 -1.01888335e+00 -5.95483005e-01
2.13746112e-02 4.26128834e-01 6.11864150e-01 -7.19270781e-02
1.37173653e-01 2.57526368e-01 -1.27136981e+00 3.61167103e-01
9.50473368e-01 1.23585451e+00 4.98779893e-01 2.74402052e-01
-9.12361667e-02 6.84775054e-01 -1.02489837e-01 -2.77063370e-01
5.91062963e-01 -4.36858028e-01 -9.98514771e-01 2.44500637e-01
7.12842286e-01 -4.05365914e-01 -4.44028437e-01 1.20546579e+00
1.52141973e-01 1.88706592e-01 6.36764765e-01 -1.18696439e+00
-8.68257403e-01 6.22954667e-01 -9.08309460e-01 2.02402860e-01
-2.32102081e-01 1.49014816e-01 1.43628144e+00 -1.05525303e+00
7.49154270e-01 1.01483548e+00 6.84980154e-01 2.68020064e-01
-1.31203902e+00 -3.88084948e-01 4.18807089e-01 2.57629722e-01
-1.39399338e+00 -1.58465788e-01 6.60554349e-01 -2.76935726e-01
1.03614044e+00 -3.18997651e-02 7.27285981e-01 7.05779254e-01
1.05842486e-01 1.12096584e+00 1.08358932e+00 -1.79765806e-01
2.38264069e-01 1.36049222e-02 1.02052920e-01 8.72106910e-01
6.37443885e-02 1.52733058e-01 -4.06731814e-01 4.48725641e-01
1.04988277e+00 2.25305200e-01 -3.22660774e-01 -7.65891790e-01
-1.29905260e+00 9.11522090e-01 1.10524678e+00 3.51279616e-01
-2.31411129e-01 2.09414586e-01 1.90988705e-01 -1.33074773e-02
4.84282851e-01 5.48379242e-01 -3.79460037e-01 8.78308788e-02
-1.22675216e+00 3.00105419e-02 5.04546344e-01 9.86565292e-01
1.16681349e+00 -7.07598999e-02 -1.70496881e-01 8.83648217e-01
2.45518982e-01 1.90344065e-01 3.12770307e-01 -1.16759932e+00
4.24844831e-01 7.21800327e-01 -2.12767512e-01 -9.07433152e-01
-3.16124350e-01 -4.95721817e-01 -9.76394594e-01 4.19263542e-01
5.31387985e-01 6.74197227e-02 -1.31348062e+00 1.65849638e+00
2.43936345e-01 3.78970176e-01 -1.48548216e-01 1.10498834e+00
8.61508012e-01 6.71577215e-01 -1.47320786e-02 3.89099002e-01
1.15877640e+00 -1.54545486e+00 -3.75768125e-01 -5.56224048e-01
4.39876378e-01 -5.61396420e-01 1.23365939e+00 6.05088733e-02
-1.34259820e+00 -4.96336788e-01 -1.09554887e+00 -4.31555301e-01
-3.51392955e-01 1.58136189e-01 7.64562428e-01 8.79405588e-02
-1.24981141e+00 7.22451806e-01 -1.08332741e+00 -2.29787111e-01
8.07042718e-01 2.38876164e-01 -3.12342167e-01 -5.23505732e-02
-9.10054922e-01 7.92535126e-01 3.05630326e-01 8.88150334e-02
-7.12263405e-01 -9.63178933e-01 -1.22637820e+00 2.67185539e-01
2.65712827e-01 -4.94551212e-01 1.08699799e+00 -1.12277591e+00
-1.21063292e+00 1.22818851e+00 -1.26418829e-01 -6.84856534e-01
4.96907473e-01 -1.73964843e-01 8.00844282e-02 3.67346436e-01
3.92662764e-01 1.33198678e+00 6.38204098e-01 -1.20807600e+00
-5.11597276e-01 -1.92141652e-01 -2.61280444e-02 2.21151575e-01
-8.01298693e-02 -3.52497011e-01 -9.59992528e-01 -8.00995708e-01
4.12518322e-01 -7.56754160e-01 -4.27326173e-01 3.85119408e-01
-6.39390230e-01 -8.28925818e-02 7.11467326e-01 -6.79457188e-01
8.73965740e-01 -2.13885021e+00 3.48268360e-01 4.44839537e-01
5.03079414e-01 6.23822398e-02 -3.55201811e-01 -9.84435063e-03
-1.16565116e-02 1.04893513e-01 -7.70328760e-01 -5.18505514e-01
-4.70686257e-02 2.46231511e-01 -1.30285993e-01 4.34147328e-01
3.85821760e-01 1.40291178e+00 -7.49859631e-01 -5.87256491e-01
4.29813445e-01 3.46771419e-01 -7.13680327e-01 2.11872712e-01
-1.30836755e-01 2.12931842e-01 -1.56509340e-01 6.83876455e-01
7.56976604e-01 -7.19542742e-01 4.21744697e-02 -4.20537651e-01
3.04507278e-02 8.93221125e-02 -9.81099725e-01 2.05849814e+00
-4.07469571e-01 7.38844156e-01 4.57170419e-02 -1.27195489e+00
7.26626396e-01 -2.40653202e-01 4.85927492e-01 -8.89602482e-01
3.32047343e-02 2.11481109e-01 -1.51316270e-01 -1.30513459e-01
3.42948794e-01 1.66395903e-01 -3.15918922e-02 2.73095936e-01
4.38946337e-01 -2.88894534e-01 1.54122621e-01 4.48815107e-01
8.14629197e-01 2.16139019e-01 2.46979490e-01 -3.92130017e-01
1.14368849e-01 -2.54162606e-02 6.43204808e-01 7.02374756e-01
-2.31152594e-01 1.16911495e+00 6.55101418e-01 -3.21251780e-01
-1.04764557e+00 -1.48290479e+00 -1.53223172e-01 7.91454017e-01
6.92338943e-01 -2.79334486e-01 -7.58331001e-01 -7.64830828e-01
4.22644522e-03 3.42228591e-01 -7.74439573e-01 8.84328634e-02
-5.09562016e-01 -4.18927908e-01 2.04391748e-01 8.83774817e-01
8.37606311e-01 -1.16704047e+00 -4.31814194e-01 -8.42700228e-02
3.69219594e-02 -1.17634976e+00 -7.73560882e-01 3.47660154e-01
-8.24936032e-01 -1.12540138e+00 -8.51844370e-01 -1.20003498e+00
9.53229010e-01 3.37738156e-01 1.41465831e+00 1.56859890e-01
-3.74351799e-01 2.71433368e-02 -3.08811456e-01 1.33757427e-01
1.23957977e-01 2.65222192e-01 -5.23951769e-01 -2.85653144e-01
6.94884360e-02 -7.23998904e-01 -8.81454408e-01 3.57056886e-01
-7.20145881e-01 4.19572294e-01 6.79914534e-01 1.07084692e+00
1.11484039e+00 -1.88533083e-01 2.15540677e-01 -9.45067525e-01
1.16982346e-03 -4.46377218e-01 -6.09900057e-01 3.08448315e-01
-6.25564575e-01 1.78977758e-01 4.17489320e-01 -1.10447370e-01
-7.95348048e-01 2.10092708e-01 -2.72462130e-01 -5.65048933e-01
-1.77767873e-01 1.95916682e-01 -1.68141797e-01 -2.04667412e-02
3.12167734e-01 1.83469296e-01 2.04987615e-01 -3.50224733e-01
6.51833594e-01 1.11052975e-01 7.04417944e-01 -3.74302238e-01
7.70963013e-01 6.05918705e-01 -4.05539364e-01 -4.70529854e-01
-1.06401825e+00 -6.38121605e-01 -6.78855181e-01 5.82264960e-02
1.07406712e+00 -1.02965260e+00 -3.44336569e-01 4.25896347e-01
-8.76153946e-01 -7.99345732e-01 -6.24363720e-01 2.69904345e-01
-6.62951529e-01 3.96859378e-01 -8.88976216e-01 -1.85727075e-01
-2.39481419e-01 -1.26436090e+00 1.23234916e+00 2.11908296e-01
-8.42188746e-02 -9.48439300e-01 -1.32781312e-01 2.69752741e-01
2.93185145e-01 7.36721829e-02 7.01791883e-01 -3.34513247e-01
-8.14029574e-01 1.93619624e-01 -5.99582314e-01 4.90742177e-01
2.65750848e-02 -7.05432519e-02 -7.92249858e-01 -1.01258457e-01
-3.05682749e-01 -4.30087417e-01 1.43750215e+00 7.63486028e-01
1.53773832e+00 -2.88156569e-01 -1.62848279e-01 1.07794082e+00
1.53409290e+00 -7.11815059e-02 8.10067058e-01 2.48784125e-01
9.77683008e-01 4.57627386e-01 4.68413830e-01 1.55311897e-01
4.25007433e-01 6.17176771e-01 6.16624832e-01 -6.34674847e-01
-3.81044984e-01 -4.52657104e-01 1.63931593e-01 5.10646522e-01
1.30613327e-01 8.75614490e-03 -6.17042840e-01 7.17929482e-01
-1.91883433e+00 -9.64060307e-01 1.84490472e-01 1.92183661e+00
9.01444376e-01 1.18557513e-01 9.60550383e-02 -1.94829211e-01
6.93483651e-01 5.14961004e-01 -8.24767530e-01 -3.20486188e-01
-2.93896198e-01 5.11438370e-01 6.72079980e-01 6.37626171e-01
-1.24981189e+00 1.22161996e+00 5.78104019e+00 1.13466406e+00
-9.13246095e-01 2.15208352e-01 1.09599280e+00 -3.08578074e-01
-5.66240847e-01 -1.01513751e-01 -7.72221386e-01 4.97061223e-01
2.56880611e-01 3.80170226e-01 3.77905399e-01 9.03077900e-01
8.37886333e-02 -3.46875161e-01 -1.02254009e+00 9.67015743e-01
-1.39461299e-02 -1.82216144e+00 -6.13512434e-02 -6.91823140e-02
1.17377341e+00 3.77731800e-01 2.16813236e-01 1.71809018e-01
3.73345017e-01 -1.18238723e+00 9.21135247e-01 2.43143618e-01
6.89095378e-01 -5.61417580e-01 7.30341792e-01 7.45766163e-02
-1.20543802e+00 8.77260417e-02 -3.89319628e-01 2.07385197e-01
3.50763321e-01 6.95484757e-01 -3.05859238e-01 1.83196321e-01
9.07831073e-01 1.19527447e+00 -6.50081754e-01 1.04542720e+00
-4.52731907e-01 4.75697458e-01 -3.93134177e-01 3.43259811e-01
5.96987188e-01 -4.48040217e-01 1.59857377e-01 1.10426879e+00
1.76995546e-01 -2.85801500e-01 1.34279177e-01 1.33240068e+00
-2.16904640e-01 -6.98670670e-02 -4.16293889e-01 2.24911317e-01
4.72365737e-01 1.20000172e+00 -8.67102444e-01 -4.08156931e-01
-4.06640857e-01 1.31596863e+00 6.05987430e-01 5.15473127e-01
-8.71555746e-01 -2.26185575e-01 6.47297859e-01 2.50549734e-01
7.07637906e-01 -1.19883530e-01 -7.52687693e-01 -1.22633481e+00
4.96252775e-02 -5.61249256e-01 3.26721042e-01 -7.10341692e-01
-1.31355536e+00 5.63750684e-01 -1.98974773e-01 -1.14405382e+00
-2.85691787e-02 -4.95091110e-01 -9.26186383e-01 6.92753792e-01
-1.62821293e+00 -1.16615546e+00 -2.95982659e-01 5.17332494e-01
6.88540041e-01 -2.29130760e-02 4.90370125e-01 9.54805315e-02
-7.37840712e-01 6.95474684e-01 -1.09093878e-02 2.32635275e-01
4.14540052e-01 -1.48988724e+00 5.92830777e-01 8.97812009e-01
4.55570489e-01 3.77049118e-01 2.60188162e-01 -4.36016262e-01
-6.85781837e-01 -1.19840229e+00 7.65196800e-01 -9.69207361e-02
6.09342873e-01 -5.12569308e-01 -9.86845553e-01 6.49979591e-01
4.25082147e-01 3.81647408e-01 1.96216598e-01 2.93470845e-02
-5.31278551e-01 -3.54464129e-02 -1.03566992e+00 6.10603034e-01
1.12967730e+00 -4.74855691e-01 -5.31099379e-01 3.17523569e-01
8.27553749e-01 -5.04056811e-01 -7.66098022e-01 2.74846792e-01
3.15789729e-01 -1.29957390e+00 1.07412362e+00 -3.61420959e-01
8.60897183e-01 -3.57508510e-01 -1.16547853e-01 -1.18655908e+00
-3.01551431e-01 -1.73232824e-01 -3.44538502e-02 1.02652693e+00
6.33675277e-01 -4.48818982e-01 9.53620374e-01 5.30738890e-01
-2.91292876e-01 -1.26713884e+00 -7.79674411e-01 -7.11787224e-01
1.92732766e-01 -3.93247932e-01 4.42745417e-01 7.14451253e-01
-2.47512668e-01 8.58824626e-02 -1.84394434e-01 1.20064013e-01
5.93315184e-01 4.36934441e-01 3.72188300e-01 -7.81059206e-01
-3.93743455e-01 -7.34806716e-01 -4.41130161e-01 -1.51406336e+00
2.30484486e-01 -1.09986413e+00 1.77389324e-01 -1.81385684e+00
3.34620059e-01 -4.50487196e-01 -4.25264567e-01 6.59551799e-01
-1.07240766e-01 6.52392447e-01 2.31442541e-01 2.77473480e-01
-8.44724953e-01 6.61959350e-01 1.33622921e+00 -3.44336540e-01
-2.91986853e-01 -2.22862631e-01 -6.82868659e-01 7.20510244e-01
9.28319335e-01 -1.66531444e-01 -3.74883562e-01 -4.67521131e-01
-8.30583647e-02 -4.00049686e-01 7.08761394e-01 -9.52594876e-01
9.25469846e-02 -3.83644775e-02 5.08111477e-01 -6.12504125e-01
2.91555911e-01 -5.93583643e-01 -1.67986602e-01 3.56609970e-01
-5.50742030e-01 -1.02511473e-01 1.10549942e-01 4.34656233e-01
-4.74627376e-01 -9.54878107e-02 9.75629807e-01 -1.76893428e-01
-1.33075309e+00 5.61433673e-01 8.95549543e-03 3.19810361e-01
1.08297122e+00 -4.18313950e-01 -2.31971249e-01 -1.95410281e-01
-7.49266028e-01 3.70730072e-01 7.00688004e-01 2.46920317e-01
7.54056871e-01 -1.31332612e+00 -4.11790967e-01 3.37307811e-01
-2.87765227e-02 3.92540663e-01 4.34618443e-01 1.02919996e+00
-6.05378151e-01 1.93229303e-01 -3.55934948e-01 -7.69807100e-01
-9.21348870e-01 2.95025200e-01 4.48977649e-01 -2.97447890e-01
-1.02105784e+00 1.19618332e+00 7.90646970e-01 -2.41561562e-01
3.32316458e-01 -5.23388088e-01 1.39182881e-01 -2.41993159e-01
3.10480624e-01 -5.32222763e-02 -1.58968762e-01 -5.54960847e-01
-3.95801932e-01 7.33511567e-01 -1.15266263e-01 5.67012886e-03
1.29712009e+00 -1.61414281e-01 -1.19837448e-02 1.38234094e-01
1.51873767e+00 -2.64937997e-01 -1.93920791e+00 -2.78917909e-01
-1.77852288e-01 -4.14844513e-01 1.81795076e-01 -6.76436961e-01
-1.83601904e+00 1.00979292e+00 4.58474547e-01 -1.66614205e-01
1.10966158e+00 3.95028234e-01 9.37082887e-01 5.59412539e-02
9.80328396e-02 -1.18111157e+00 1.63187504e-01 3.84228468e-01
8.92186642e-01 -1.55690217e+00 -4.36775684e-02 -4.90849972e-01
-9.85181391e-01 6.51858866e-01 8.80269468e-01 -4.92695868e-01
7.35767066e-01 3.26311320e-01 -1.21550523e-01 -2.86561072e-01
-4.50575680e-01 -5.29269874e-01 5.60591459e-01 5.99158347e-01
3.32725823e-01 6.65929243e-02 -1.40973657e-01 5.12145519e-01
-7.18809217e-02 -4.38964427e-01 7.84204006e-02 3.89188528e-01
-4.59939212e-01 -7.72826791e-01 6.94018230e-02 7.41182983e-01
-1.35966033e-01 -2.84308076e-01 -1.58825099e-01 8.98800850e-01
2.47074857e-01 5.68818748e-01 4.58280146e-01 -2.60082513e-01
1.30831540e-01 -3.89696926e-01 3.52016747e-01 -6.27817750e-01
-3.79704297e-01 1.29129484e-01 -2.40985826e-01 -1.00525904e+00
-3.77116591e-01 -4.78307426e-01 -1.59708667e+00 -1.67929858e-01
-3.34286410e-03 -9.26031619e-02 1.98623165e-01 8.09633493e-01
3.88353467e-01 3.47554266e-01 4.34024721e-01 -9.93592262e-01
-3.70618969e-01 -7.65414238e-01 -7.86469281e-01 5.05272090e-01
1.38244569e-01 -5.99287689e-01 -3.07902694e-01 4.39140759e-02] | [9.665946960449219, 0.4939509332180023] |
dd7003e6-fd4e-421d-8fc1-fbf72b530a25 | monopair-monocular-3d-object-detection-using | 2003.00504 | null | https://arxiv.org/abs/2003.00504v1 | https://arxiv.org/pdf/2003.00504v1.pdf | MonoPair: Monocular 3D Object Detection Using Pairwise Spatial Relationships | Monocular 3D object detection is an essential component in autonomous driving while challenging to solve, especially for those occluded samples which are only partially visible. Most detectors consider each 3D object as an independent training target, inevitably resulting in a lack of useful information for occluded samples. To this end, we propose a novel method to improve the monocular 3D object detection by considering the relationship of paired samples. This allows us to encode spatial constraints for partially-occluded objects from their adjacent neighbors. Specifically, the proposed detector computes uncertainty-aware predictions for object locations and 3D distances for the adjacent object pairs, which are subsequently jointly optimized by nonlinear least squares. Finally, the one-stage uncertainty-aware prediction structure and the post-optimization module are dedicatedly integrated for ensuring the run-time efficiency. Experiments demonstrate that our method yields the best performance on KITTI 3D detection benchmark, by outperforming state-of-the-art competitors by wide margins, especially for the hard samples. | ['Mingyang Li', 'Lei Tai', 'Kai Sun', 'Yongjian Chen'] | 2020-03-01 | monopair-monocular-3d-object-detection-using-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Chen_MonoPair_Monocular_3D_Object_Detection_Using_Pairwise_Spatial_Relationships_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Chen_MonoPair_Monocular_3D_Object_Detection_Using_Pairwise_Spatial_Relationships_CVPR_2020_paper.pdf | cvpr-2020-6 | ['vehicle-pose-estimation'] | ['computer-vision'] | [-2.28144914e-01 1.60218123e-03 -2.44556472e-01 -5.30593514e-01
-7.39779353e-01 -4.25171673e-01 3.58047038e-01 -8.75792801e-02
-6.09926224e-01 3.74050170e-01 -1.91829488e-01 -2.39975438e-01
1.24779955e-01 -2.89395899e-01 -9.23409939e-01 -8.23537409e-01
1.32882714e-01 3.86661291e-01 7.14581370e-01 2.31922746e-01
1.78375468e-01 6.16436958e-01 -1.74608886e+00 -1.93811059e-01
9.28073525e-01 1.35046017e+00 6.30048692e-01 2.12688699e-01
2.04267591e-01 4.92707223e-01 -2.68199533e-01 -5.69845662e-02
5.04358172e-01 2.17432946e-01 2.24297374e-01 1.88310042e-01
6.43470824e-01 -5.74539363e-01 -5.35784304e-01 1.15632427e+00
4.48935032e-01 1.53543681e-01 7.47538209e-01 -1.28948009e+00
-2.16831490e-01 2.34064922e-01 -9.88844454e-01 1.73978657e-01
1.24971231e-03 4.26360101e-01 7.99382091e-01 -1.25469959e+00
3.02623481e-01 1.38919055e+00 3.18050086e-01 2.74188131e-01
-8.14172089e-01 -7.80220389e-01 7.37713099e-01 5.58637619e-01
-1.57208097e+00 -3.26348156e-01 7.65801966e-01 -4.97411996e-01
8.09882462e-01 7.45765790e-02 5.59299409e-01 6.52734041e-01
1.73898086e-01 1.13698518e+00 8.32756877e-01 4.72362805e-03
1.20291032e-01 3.55429262e-01 2.21293062e-01 6.06594443e-01
5.74375510e-01 4.06803578e-01 -4.98526484e-01 1.18457451e-01
3.08630556e-01 4.40127216e-02 -3.18158008e-02 -9.07981932e-01
-1.04706025e+00 5.46837866e-01 7.33306587e-01 -2.70732731e-01
-2.20227078e-01 -2.75682826e-02 1.78987160e-02 -1.57146364e-01
4.50507581e-01 -1.06356211e-01 -3.74519765e-01 2.40504548e-01
-7.10529745e-01 4.57975239e-01 1.24670185e-01 1.32438409e+00
8.45506966e-01 -2.21345156e-01 -4.54869419e-01 5.73988080e-01
8.10949445e-01 7.69263506e-01 -2.61539463e-02 -5.96265733e-01
8.06494534e-01 8.51119459e-01 4.40129071e-01 -8.45548570e-01
-4.95053232e-01 -7.22187757e-01 -5.04154682e-01 5.72932661e-01
3.55953544e-01 8.21659267e-02 -9.95767891e-01 1.43381953e+00
8.34389865e-01 2.18617380e-01 -7.73184672e-02 1.41052520e+00
8.76311183e-01 4.94427085e-01 -1.84510425e-01 -1.44412592e-01
1.28162599e+00 -1.05719352e+00 -3.19148242e-01 -8.90688360e-01
3.38180840e-01 -7.74890244e-01 6.54428661e-01 9.96909589e-02
-7.72460639e-01 -6.91555619e-01 -1.25969160e+00 -3.12688023e-01
-4.10210006e-02 6.77549303e-01 3.11412096e-01 4.48782772e-01
-3.44800800e-01 9.68016312e-02 -9.27888274e-01 1.37088433e-01
7.52411067e-01 1.94348812e-01 -2.23099768e-01 -2.82003134e-01
-6.18307352e-01 1.03418255e+00 4.09940332e-01 3.49818885e-01
-1.00467134e+00 -6.73898280e-01 -9.98331428e-01 -1.38177499e-01
7.68458545e-01 -4.14231300e-01 1.09127414e+00 -2.95066591e-02
-1.07860970e+00 8.73826504e-01 -4.78337556e-01 -5.57976186e-01
8.96443665e-01 -3.75223756e-01 2.28227684e-04 -3.30828339e-01
2.04269141e-01 9.23036516e-01 1.01099741e+00 -1.36301291e+00
-1.18405998e+00 -7.92770445e-01 -2.30124354e-01 6.15421236e-01
7.17782080e-02 -4.21735287e-01 -8.99403274e-01 -1.91639870e-01
7.27311492e-01 -9.39242661e-01 -3.28403890e-01 3.44174027e-01
-5.55490077e-01 -4.15624082e-01 9.41629171e-01 -3.61182719e-01
6.57405674e-01 -2.26668382e+00 1.97075278e-01 -2.32101846e-02
2.91247308e-01 1.19758449e-01 7.42017925e-02 -2.51798064e-01
4.12986577e-01 -4.66418624e-01 -6.19181544e-02 -8.08521390e-01
2.44098529e-01 3.40211391e-02 -2.29813829e-01 6.77906275e-01
4.76294786e-01 8.74976814e-01 -7.69594908e-01 -4.89562154e-01
5.69414556e-01 4.24858898e-01 -3.62093508e-01 1.81594625e-01
-3.81883085e-01 4.76202339e-01 -6.73921049e-01 8.18335652e-01
1.18206370e+00 1.51761010e-01 -4.61202204e-01 -1.44357726e-01
-2.52940893e-01 1.55718312e-01 -1.35031450e+00 1.61542642e+00
-4.52991724e-02 4.72321659e-01 3.77682336e-02 -5.60863197e-01
1.14322603e+00 -3.04404527e-01 6.82469606e-02 -6.00601912e-01
1.88871145e-01 3.34713310e-01 -1.43299156e-04 -2.46095479e-01
5.87328434e-01 2.91551560e-01 -5.09274080e-02 -1.96005311e-02
-3.90228450e-01 -2.88282067e-01 -1.13443345e-01 2.11328478e-03
7.76128590e-01 2.95356333e-01 2.50461102e-01 -1.59027129e-02
5.31265616e-01 3.71271074e-02 8.92166495e-01 7.63856113e-01
-5.64175069e-01 6.00009859e-01 3.35031033e-01 -1.67942360e-01
-8.53086412e-01 -1.03609133e+00 -2.91610211e-01 3.47565413e-01
9.88458514e-01 1.77728727e-01 -3.58827382e-01 -8.49559605e-01
5.02212286e-01 6.69186413e-01 -5.87240934e-01 -2.25399628e-01
-3.47324312e-01 -4.67283428e-01 -1.69690445e-01 5.25998771e-01
2.69663841e-01 -6.68604851e-01 -9.73037481e-01 -5.42677641e-02
1.68816954e-01 -1.43814886e+00 -4.36353207e-01 4.42566007e-01
-7.63501048e-01 -9.59932387e-01 -6.84876263e-01 -7.46173799e-01
7.31057882e-01 9.26830947e-01 6.44497693e-01 -1.82963386e-01
-3.87853265e-01 -3.46891284e-01 -1.78406879e-01 -6.21047676e-01
7.66167566e-02 -6.46491498e-02 1.15479149e-01 4.06577252e-02
6.65943027e-01 -1.71753138e-01 -7.23745346e-01 5.99153280e-01
-2.40752861e-01 1.37640849e-01 1.02654815e+00 7.32247055e-01
1.07133698e+00 -2.72989888e-02 1.49847418e-01 -4.00512338e-01
-1.36640102e-01 -4.13323104e-01 -1.11874485e+00 2.44258977e-02
-3.06448698e-01 2.98847631e-02 2.33480096e-01 -3.64682734e-01
-9.96442854e-01 6.51137292e-01 2.43670925e-01 -8.07149291e-01
-1.52783155e-01 6.36063367e-02 -4.85406697e-01 -2.66811579e-01
6.11546516e-01 1.92688316e-01 -1.16234839e-01 -5.74282169e-01
2.00766638e-01 7.24416316e-01 5.17474949e-01 -1.66874215e-01
1.07672083e+00 6.12249315e-01 8.15531462e-02 -5.32923162e-01
-1.05366504e+00 -8.29995334e-01 -6.52191162e-01 -2.45687127e-01
7.34280765e-01 -1.38763809e+00 -5.81600428e-01 4.69144881e-01
-1.09476697e+00 4.33706157e-02 -1.01089880e-01 7.22939968e-01
-3.22187275e-01 2.74558455e-01 -1.93917863e-02 -1.16009235e+00
5.94311319e-02 -1.40439153e+00 1.34112763e+00 4.28542793e-01
4.82886493e-01 -3.95957008e-02 -3.18423182e-01 5.12635410e-01
-1.14705123e-01 -8.06734115e-02 4.51347619e-01 -3.30993563e-01
-1.12303889e+00 -4.93532717e-01 -6.66424751e-01 2.00055629e-01
-1.14618294e-01 -2.48097479e-01 -1.08848774e+00 -1.96588486e-01
3.16975601e-02 -1.61098421e-01 1.25824320e+00 5.02599716e-01
9.89288688e-01 4.04206842e-01 -7.78240502e-01 5.31804919e-01
1.03873146e+00 1.24096453e-01 2.07015231e-01 3.63176130e-02
8.30532908e-01 6.70932889e-01 1.37673140e+00 5.10738373e-01
4.78138447e-01 8.74965668e-01 8.33963692e-01 1.80514976e-01
3.43246683e-02 -3.27580601e-01 1.71907201e-01 2.43432701e-01
4.52902317e-01 -1.33971572e-01 -5.85881650e-01 4.72562850e-01
-2.19018412e+00 -6.42812192e-01 -2.73758471e-01 2.34637690e+00
3.67868483e-01 5.84516764e-01 -2.28072312e-02 -9.56404433e-02
7.26465106e-01 1.44794807e-01 -1.28800952e+00 4.54531521e-01
-1.70917690e-01 -3.74950379e-01 8.76484752e-01 4.89123732e-01
-1.35480750e+00 9.85024631e-01 5.17063951e+00 9.23833311e-01
-6.34871602e-01 1.94484089e-02 5.70764899e-01 -4.77953702e-01
-6.98004365e-02 3.86157893e-02 -1.58537889e+00 4.34657335e-01
3.22759673e-02 1.32923067e-01 9.85490605e-02 1.05307043e+00
4.08402592e-01 -4.54970092e-01 -9.94621336e-01 1.16442943e+00
6.71587884e-02 -8.38693500e-01 -4.58996415e-01 1.20962538e-01
7.90340960e-01 3.79244000e-01 1.79445356e-01 1.31916091e-01
-4.47034948e-02 -5.80292642e-01 1.27480364e+00 3.68864387e-01
2.63774425e-01 -7.77711213e-01 6.29212439e-01 7.63210118e-01
-1.17359304e+00 -2.67392755e-01 -8.74626219e-01 -2.38844231e-02
2.07032204e-01 8.66428018e-01 -7.44639158e-01 4.02066946e-01
8.04321647e-01 8.10454845e-01 -5.38167059e-01 1.57514215e+00
-5.60728312e-01 1.37340985e-02 -5.13355553e-01 -2.61260748e-01
2.29328975e-01 -9.84661058e-02 9.13796663e-01 6.95090532e-01
3.65711629e-01 5.07805943e-02 3.33071440e-01 1.08181727e+00
1.67879462e-01 -4.24786992e-02 -4.47877139e-01 4.62276548e-01
7.68508852e-01 1.33408773e+00 -4.69215006e-01 -1.07486494e-01
-5.17406523e-01 5.94730556e-01 4.57124859e-01 1.45911455e-01
-9.36664224e-01 -7.60517195e-02 8.54149163e-01 -1.85315255e-02
7.03149438e-01 -4.07754570e-01 -4.30163652e-01 -1.04912329e+00
6.93153441e-01 -4.95590299e-01 8.20030458e-03 -7.60541916e-01
-1.16406822e+00 4.43505138e-01 -6.79708347e-02 -1.59029555e+00
1.05544642e-01 -7.50256240e-01 -2.99764305e-01 9.37656403e-01
-1.99495244e+00 -9.96879280e-01 -6.74427569e-01 1.57288715e-01
5.29175103e-01 -4.32087518e-02 2.57035736e-02 3.17129165e-01
-7.52517760e-01 4.93149877e-01 -9.35654193e-02 -4.30160820e-01
7.09224701e-01 -1.00029397e+00 4.48556244e-01 1.07705879e+00
1.28619865e-01 2.27882396e-02 6.27246082e-01 -7.45229721e-01
-1.44460428e+00 -1.18466759e+00 6.01506114e-01 -3.67194861e-01
2.51812667e-01 -5.86827338e-01 -7.92195737e-01 3.89913052e-01
-4.79847819e-01 3.69273901e-01 -2.30755899e-02 -1.85530052e-01
-2.88007259e-01 -2.22515911e-01 -9.71609592e-01 6.00460947e-01
1.27108800e+00 -2.39569545e-01 -4.78249490e-01 4.32823300e-01
7.63729453e-01 -9.00630414e-01 -2.37385318e-01 6.05396032e-01
5.37082851e-01 -9.42530215e-01 1.14862025e+00 -3.45573900e-03
-1.05966004e-02 -9.87075448e-01 -2.13845417e-01 -8.73632073e-01
-1.60365134e-01 -2.51679689e-01 -4.08576310e-01 9.27895606e-01
4.43234861e-01 -4.85352337e-01 1.09999263e+00 5.01073539e-01
-5.00576019e-01 -5.69652140e-01 -1.21590734e+00 -9.12457824e-01
-4.35365617e-01 -6.65288866e-01 4.19268042e-01 2.31729656e-01
-4.34014946e-01 2.31649980e-01 -5.00057280e-01 7.22985208e-01
8.80204260e-01 3.53159457e-01 1.06599736e+00 -1.26442230e+00
-3.40419471e-01 -4.33236331e-01 -6.92055881e-01 -1.81352174e+00
1.11268960e-01 -6.79630518e-01 6.98565841e-01 -1.26865101e+00
2.89461106e-01 -7.90643573e-01 -1.07431322e-01 2.03370899e-01
-5.45360982e-01 2.49427676e-01 1.11179344e-01 2.39417493e-01
-6.32618666e-01 9.35590565e-01 1.31918776e+00 -3.34050626e-01
-3.23464334e-01 2.44921178e-01 -4.81860369e-01 5.78735471e-01
4.17087525e-01 -5.05989015e-01 -3.68805319e-01 -5.17879248e-01
-1.99757770e-01 -2.35488370e-01 5.94195664e-01 -1.29557192e+00
4.02104169e-01 -1.77075982e-01 5.46606958e-01 -1.57330501e+00
7.20599532e-01 -9.90652621e-01 -1.05138756e-01 3.80536795e-01
1.44687369e-02 -5.07654727e-01 2.10027710e-01 8.86591315e-01
-8.11152756e-02 -3.34897310e-01 8.04088593e-01 1.96950421e-01
-9.94006753e-01 5.96850038e-01 1.51461929e-01 -1.27979487e-01
1.44833922e+00 -2.31467530e-01 -2.53988504e-01 7.61614218e-02
-3.66544396e-01 8.11005890e-01 4.66314048e-01 6.73893631e-01
6.58082128e-01 -1.28316772e+00 -6.95578635e-01 3.34899545e-01
4.98874038e-01 6.25216663e-01 4.19627309e-01 8.88354599e-01
-1.12715773e-01 5.04824281e-01 3.50798368e-02 -1.15831399e+00
-1.30612898e+00 7.13464022e-01 2.83914775e-01 1.76888868e-01
-6.17683113e-01 1.04698479e+00 4.60414261e-01 -2.89012998e-01
5.87714791e-01 -5.98611832e-01 -1.02674037e-01 -4.74871360e-02
5.27007699e-01 1.72763482e-01 4.58469754e-03 -6.77074969e-01
-5.45087397e-01 8.20559680e-01 -2.31117442e-01 2.05217138e-01
1.01986802e+00 -3.48360747e-01 4.22861606e-01 4.41372901e-01
9.17521119e-01 -1.86785296e-01 -2.07532525e+00 -5.18162191e-01
-2.91142523e-01 -8.53883624e-01 3.29905748e-01 -4.89628553e-01
-9.08661962e-01 1.06042254e+00 7.51536310e-01 -3.10969204e-01
5.91978788e-01 1.32974982e-01 5.34166753e-01 5.08662224e-01
4.85919863e-01 -8.81700933e-01 -1.44152194e-01 4.71074730e-01
6.85141087e-01 -1.75269020e+00 2.30565801e-01 -7.75047243e-01
-4.16866153e-01 6.97676241e-01 8.82895172e-01 -9.91378129e-02
5.50837934e-01 3.30473036e-02 -4.42683920e-02 2.67790649e-02
-6.16282046e-01 -3.64086002e-01 7.48399913e-01 5.35377145e-01
-2.78592795e-01 1.13145150e-01 -8.80133584e-02 6.05635047e-01
1.63077503e-01 -4.26801980e-01 -1.96022335e-02 7.80169129e-01
-7.97995865e-01 -8.08995247e-01 -4.38989848e-01 4.43054318e-01
1.77154869e-01 1.32736266e-01 -3.14746052e-01 7.09311604e-01
2.68838078e-01 8.56227100e-01 8.38276297e-02 -4.02841091e-01
5.05164146e-01 -3.21937025e-01 4.73488897e-01 -6.40354335e-01
5.35140687e-04 1.84413284e-01 -2.20453009e-01 -5.28484464e-01
-1.05980113e-01 -1.01890862e+00 -1.13244140e+00 1.59833148e-01
-7.76619792e-01 -2.90890723e-01 7.79822707e-01 9.80410099e-01
4.58648264e-01 4.17823970e-01 7.47004688e-01 -1.32065785e+00
-7.41656363e-01 -7.80246913e-01 -4.24865097e-01 -1.02977127e-01
3.73937547e-01 -1.15960109e+00 -4.26139861e-01 -5.46932518e-01] | [7.93573522567749, -2.3988301753997803] |
80360a08-6bb9-4d31-ba78-9e4e6c1f9ad7 | sam-helps-shadow-when-segment-anything-model | 2306.06113 | null | https://arxiv.org/abs/2306.06113v1 | https://arxiv.org/pdf/2306.06113v1.pdf | SAM-helps-Shadow:When Segment Anything Model meet shadow removal | The challenges surrounding the application of image shadow removal to real-world images and not just constrained datasets like ISTD/SRD have highlighted an urgent need for zero-shot learning in this field. In this study, we innovatively adapted the SAM (Segment anything model) for shadow removal by introducing SAM-helps-Shadow, effectively integrating shadow detection and removal into a single stage. Our approach utilized the model's detection results as a potent prior for facilitating shadow detection, followed by shadow removal using a second-order deep unfolding network. The source code of SAM-helps-Shadow can be obtained from https://github.com/zhangbaijin/SAM-helps-Shadow. | ['Shanying Zhu', 'Chaochen Gu', 'Xiaofeng Zhang'] | 2023-06-01 | null | null | null | null | ['shadow-removal', 'shadow-detection-and-removal', 'shadow-detection', 'image-shadow-removal'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 4.95604843e-01 2.13050425e-01 3.27370346e-01 -3.66746217e-01
-3.32448632e-01 -2.43576109e-01 5.62883496e-01 -4.75046605e-01
-2.70510852e-01 5.73338568e-01 1.96169838e-01 -6.87842071e-01
4.10583884e-01 -5.39677262e-01 -5.26022136e-01 -7.49949515e-01
8.92680138e-02 1.06878906e-01 7.49161363e-01 -1.16465271e-01
1.11375950e-01 4.82475966e-01 -1.20369089e+00 5.38786463e-02
1.00004649e+00 6.58705711e-01 9.27363336e-01 9.15745854e-01
1.18249826e-01 7.13723540e-01 -7.01938033e-01 -8.78544301e-02
4.32100624e-01 -3.78177464e-01 -3.35174412e-01 -7.24844486e-02
3.73905420e-01 -1.10283053e+00 -7.23680615e-01 4.78268564e-01
7.49549508e-01 4.21997279e-01 3.49231362e-01 -1.30001223e+00
-5.94310343e-01 4.99033295e-02 -6.05328023e-01 2.63246804e-01
2.99601883e-01 4.76552248e-01 6.61378622e-01 -1.16719747e+00
5.63177764e-01 1.21423304e+00 7.13775992e-01 4.48553860e-01
-1.03449011e+00 -5.77372730e-01 8.46535265e-02 3.49637985e-01
-1.15981901e+00 -4.83940601e-01 8.16638887e-01 -1.71016499e-01
9.20446217e-01 3.21321070e-01 7.67758846e-01 1.39991665e+00
4.31967884e-01 1.17770231e+00 1.20711601e+00 -3.32917273e-01
1.36854351e-01 -3.32149416e-02 1.51342303e-01 7.02026129e-01
2.51105517e-01 2.83243030e-01 -5.41787863e-01 2.59273108e-02
7.01066256e-01 1.61421910e-01 -3.55529517e-01 -2.90107250e-01
-8.50625634e-01 6.02862597e-01 5.49704134e-01 -1.22435220e-01
-1.98564008e-01 2.86857128e-01 3.92583460e-02 -2.50107944e-01
6.00861490e-01 1.05252035e-01 -3.59596461e-01 1.75941393e-01
-1.18465662e+00 1.03762925e-01 7.02093720e-01 7.68050849e-01
7.79079616e-01 2.78507590e-01 -2.97832996e-01 4.74844128e-01
1.77881956e-01 8.33303332e-01 -2.50450283e-01 -1.11328065e+00
-2.19203115e-01 1.08718097e-01 3.32393169e-01 -8.65076900e-01
-5.32124639e-01 -1.03770263e-01 -6.35594070e-01 5.02749026e-01
4.99994382e-02 -4.03285146e-01 -1.53459036e+00 1.31619537e+00
5.80902994e-01 7.24973083e-01 -5.21820225e-02 1.18505096e+00
1.19480371e+00 6.12520695e-01 -2.47283697e-01 5.21708988e-02
1.17622089e+00 -1.14845991e+00 -6.75180018e-01 -6.12304986e-01
8.34117755e-02 -7.99934447e-01 1.23820865e+00 1.01497047e-01
-4.67182934e-01 -2.48635560e-01 -1.05872965e+00 -3.93974096e-01
-5.28013408e-01 -2.24301796e-02 7.92341948e-01 1.93060219e-01
-9.57590580e-01 3.61892611e-01 -7.32978344e-01 -4.17110413e-01
7.24791646e-01 -7.51651451e-02 7.29528591e-02 -1.77871883e-01
-1.14901435e+00 9.91339386e-01 8.90407637e-02 3.95608038e-01
-1.46025860e+00 -7.62444556e-01 -7.85873711e-01 -5.01524583e-02
9.62388694e-01 -5.38065135e-01 1.34179330e+00 -7.51367211e-01
-1.46909249e+00 5.13696074e-01 -4.28265691e-01 -4.64329749e-01
7.64030039e-01 -7.89454520e-01 -5.97561002e-02 1.33623004e-01
-2.77145077e-02 4.26544666e-01 1.01397741e+00 -1.87911606e+00
-1.18367575e-01 4.67314683e-02 1.67997614e-01 3.24473441e-01
1.94243938e-01 -1.23997428e-01 -5.78442812e-01 -5.51024079e-01
-2.01000690e-01 -9.65774953e-01 -2.88252056e-01 1.26112372e-01
-6.52238846e-01 2.10382119e-01 1.52186215e+00 -1.17931724e+00
1.10217667e+00 -1.88438320e+00 -2.24271953e-01 -1.30952429e-02
5.25814533e-01 5.83972633e-01 -1.76022634e-01 4.84746665e-01
1.58194542e-01 -2.52816379e-01 -8.00577641e-01 -7.25835681e-01
3.02406051e-03 3.03828806e-01 -4.56488401e-01 5.59419632e-01
6.29636347e-02 1.12592781e+00 -9.68217432e-01 -4.36401963e-01
7.26427615e-01 8.78647804e-01 8.60540196e-02 4.96445090e-01
-4.04951602e-01 2.98357248e-01 -1.48208052e-01 9.78177369e-01
1.05023098e+00 -1.17251240e-01 3.81737798e-02 -3.65090527e-04
-1.90528169e-01 3.54159512e-02 -1.06434989e+00 1.36880326e+00
-3.73308688e-01 1.10703933e+00 3.64136994e-01 -2.85389334e-01
5.36992192e-01 -3.96346897e-02 6.34497046e-01 -9.16804254e-01
2.12572828e-01 -1.99538544e-01 -2.76049912e-01 -3.72375369e-01
5.39398313e-01 1.32560164e-01 1.74630150e-01 5.49443245e-01
-3.11892718e-01 -3.01736355e-01 -2.18473613e-01 5.81045628e-01
1.13203526e+00 6.35185838e-01 -1.97239388e-02 -1.43267646e-01
9.34852660e-02 3.45611647e-02 7.56230354e-01 8.55467141e-01
-4.83472705e-01 9.49911594e-01 2.35716283e-01 -2.33072162e-01
-8.17590714e-01 -1.21713471e+00 -7.43374452e-02 1.16084814e+00
5.14386714e-01 -1.61055833e-01 -7.19523907e-01 -6.44958913e-01
3.02916914e-01 1.08010173e+00 -8.12061310e-01 -1.10775031e-01
-4.77463573e-01 -6.34786248e-01 4.87178683e-01 3.75503361e-01
5.62864661e-01 -1.33427382e+00 -1.35901797e+00 -7.31606334e-02
-1.83259606e-01 -9.13524687e-01 -4.76003647e-01 5.01319289e-01
-1.79488689e-01 -1.17078245e+00 -9.61907029e-01 -1.26305953e-01
4.09595788e-01 1.00072765e+00 1.13357794e+00 4.89753723e-01
-7.35120296e-01 4.38340425e-01 -3.67126346e-01 -8.45541060e-01
2.62356680e-02 -2.98261940e-01 -1.62089363e-01 -9.74285007e-02
-2.26971321e-03 -5.11291981e-01 -9.71963286e-01 1.82928115e-01
-1.00653458e+00 3.78553718e-01 6.71445608e-01 6.05419219e-01
1.63026243e-01 -3.61753047e-01 6.78476598e-03 -9.41230416e-01
4.91222739e-01 -3.42285305e-01 -6.59235537e-01 1.16006009e-01
-6.28540814e-01 -5.00155747e-01 1.52935341e-01 -1.13807559e-01
-1.69873524e+00 1.00041837e-01 8.15388002e-03 -6.49981856e-01
-1.13913380e-01 -5.05930409e-02 -2.02688903e-01 -1.13915384e-01
4.87407565e-01 9.35573131e-02 -2.26402014e-01 -2.95520753e-01
4.82772082e-01 4.62261915e-01 4.87596512e-01 -1.93022877e-01
1.06978703e+00 8.40868056e-01 -1.00501589e-01 -1.17991483e+00
-1.11916757e+00 -5.59709549e-01 -8.42394412e-01 -5.16985714e-01
8.56065631e-01 -7.41167426e-01 -3.76920491e-01 9.84764278e-01
-1.04909205e+00 -1.35468292e+00 -9.49823931e-02 -1.63244367e-01
-2.52173424e-01 6.60746932e-01 -3.90144408e-01 -1.19361866e+00
-5.32489777e-01 -6.32286906e-01 1.17176604e+00 5.51409543e-01
-6.12576827e-02 -9.12852049e-01 1.89736307e-01 3.82371515e-01
5.53750217e-01 3.41335684e-01 2.16089070e-01 3.29845883e-02
-9.88499165e-01 1.13915235e-01 -5.35283864e-01 3.71826321e-01
5.81727065e-02 1.77009016e-01 -1.33343530e+00 -1.66802526e-01
-2.24589020e-01 -1.85036883e-01 1.37382805e+00 7.28227973e-01
8.91430795e-01 -4.71552424e-02 -2.43442565e-01 8.05933475e-01
1.43198884e+00 5.42743132e-02 9.17874217e-01 3.90582711e-01
1.14850092e+00 2.54064232e-01 8.84075940e-01 6.45180225e-01
3.74445289e-01 4.06116784e-01 5.45146167e-01 -7.72142291e-01
-5.70344090e-01 -9.65411291e-02 2.63121396e-01 1.27165109e-01
-3.96364518e-02 -4.77422655e-01 -1.09326530e+00 4.44701761e-01
-2.00957704e+00 -9.03185666e-01 -3.25489759e-01 1.77236998e+00
4.39693689e-01 1.00936539e-01 -1.83130220e-01 -4.27181900e-01
3.01952600e-01 8.18177879e-01 -8.12941849e-01 -9.42591950e-02
-3.33050221e-01 2.21622307e-02 7.14837790e-01 9.81274664e-01
-1.18823946e+00 1.49467778e+00 5.65436411e+00 4.76373374e-01
-1.00767672e+00 1.73290193e-01 3.71404797e-01 -1.15886703e-01
-3.96897644e-01 4.15249228e-01 -4.18557316e-01 4.27831739e-01
5.80145240e-01 1.44742548e-01 4.13163126e-01 6.56350136e-01
5.30730546e-01 -9.68688309e-01 -4.28141892e-01 4.95469481e-01
1.44711509e-01 -1.08121777e+00 -4.86380368e-01 6.29433468e-02
7.06294239e-01 2.87264884e-01 -4.32015583e-02 2.16538683e-01
4.88424033e-01 -7.77773082e-01 5.43837726e-01 8.12443316e-01
6.89533412e-01 -3.02743793e-01 5.83687603e-01 1.47342071e-01
-1.24551904e+00 -1.54513412e-03 -3.16313952e-01 -1.37888446e-01
3.93692017e-01 9.20059085e-01 -1.28552163e+00 5.32523990e-01
8.81381392e-01 5.55998445e-01 -5.79942763e-01 9.28562820e-01
-5.19863665e-01 5.80019534e-01 -3.36366773e-01 3.33602577e-01
4.09560539e-02 -3.88418585e-01 6.88559294e-01 1.55018842e+00
-6.08486235e-02 3.04106981e-01 2.06517726e-01 6.30886018e-01
2.35481575e-01 -5.28152645e-01 -6.75594628e-01 1.90943033e-01
3.87118399e-01 1.39126229e+00 -1.04326832e+00 -5.16555727e-01
-2.47796357e-01 1.49566710e+00 -5.86885773e-02 7.29162157e-01
-1.22341537e+00 -2.29589418e-01 6.12897098e-01 9.75878537e-02
2.55047619e-01 -4.65715110e-01 -4.51516539e-01 -9.75755274e-01
-1.37109950e-01 -5.51272452e-01 -4.15417366e-03 -1.15863550e+00
-7.05665350e-01 3.05717856e-01 9.06493738e-02 -7.54241586e-01
4.83177781e-01 -2.35803217e-01 -1.02416229e+00 7.83397436e-01
-1.64597631e+00 -1.49201155e+00 -9.78242576e-01 5.88095307e-01
6.66874826e-01 3.10492575e-01 6.08570874e-01 7.62780458e-02
-7.62130976e-01 1.72328517e-01 9.26658288e-02 -2.16091089e-02
9.41582739e-01 -1.37505066e+00 8.10673773e-01 1.38802433e+00
-1.96204692e-01 3.45319092e-01 1.11072314e+00 -1.24371505e+00
-1.27428877e+00 -8.22493494e-01 3.59377384e-01 -6.30977511e-01
5.92908084e-01 -6.51241720e-01 -1.10995221e+00 6.24489844e-01
5.31298637e-01 -1.33841515e-01 2.54879504e-01 -1.34328142e-01
-1.19036987e-01 -7.35015497e-02 -9.43097293e-01 8.48537564e-01
1.25367486e+00 -4.65729296e-01 -3.23191136e-01 4.55394626e-01
8.56663525e-01 -5.29099762e-01 -1.59687236e-01 3.67805481e-01
6.07579827e-01 -1.32990062e+00 1.01818359e+00 1.44634724e-01
2.28488520e-01 -3.85531574e-01 -1.20592080e-01 -1.04304087e+00
-3.14936966e-01 -7.14116991e-01 -7.26206839e-01 8.61005783e-01
-3.24042588e-02 -6.27135158e-01 8.72914016e-01 6.24026716e-01
-4.75705445e-01 -9.29466844e-01 -7.37334311e-01 -4.15824264e-01
-4.28908229e-01 -3.24720114e-01 1.20035902e-01 5.67853034e-01
-7.08646536e-01 1.31244212e-01 -8.10290515e-01 4.35290754e-01
9.48527098e-01 1.87742665e-01 1.16880786e+00 -1.05123496e+00
-3.50282788e-01 -9.48089734e-02 3.37088883e-01 -9.68567133e-01
7.76933506e-02 -2.31367767e-01 5.87976098e-01 -2.18560529e+00
3.07432353e-01 -2.80827463e-01 -1.12297729e-01 6.95222557e-01
-4.16387320e-01 1.76418051e-01 4.45199221e-01 1.42225325e-01
-5.51479757e-01 8.56956661e-01 1.37719774e+00 1.21032529e-01
-3.27642947e-01 1.20451830e-01 -4.95040148e-01 6.90241218e-01
1.00643766e+00 -3.65074158e-01 -4.24860448e-01 -2.64149934e-01
-3.28641564e-01 -1.59418285e-01 7.49970794e-01 -8.84268224e-01
4.38098144e-03 -4.14031982e-01 4.93553787e-01 -8.88871491e-01
8.68467629e-01 -7.77026653e-01 -1.95131376e-02 4.94105101e-01
2.50209689e-01 -3.90522569e-01 2.94659972e-01 6.40213072e-01
5.13704360e-01 8.23111981e-02 6.65073872e-01 -1.83629960e-01
-1.16023457e+00 1.28993958e-01 -4.33283240e-01 -2.41773292e-01
1.17066050e+00 -3.36870074e-01 -6.59774423e-01 -4.54896927e-01
-4.22475368e-01 4.22717929e-01 7.07576513e-01 4.02996719e-01
9.94394898e-01 -6.56706452e-01 -5.18983126e-01 1.50965080e-01
-3.09179455e-01 2.48665601e-01 5.73955297e-01 9.16074395e-01
-5.20829856e-01 1.65178820e-01 -5.87201230e-02 -4.05713290e-01
-1.47380650e+00 3.72180849e-01 4.49625790e-01 6.08212650e-02
-1.09010279e+00 1.00901425e+00 5.91624439e-01 -4.41802621e-01
3.16235691e-01 9.55845881e-03 4.21314955e-01 -7.92438909e-02
2.92524129e-01 6.49196625e-01 -1.51025712e-01 -4.84938413e-01
-5.69807947e-01 1.34811714e-01 1.82233363e-01 -1.64631829e-01
1.43173516e+00 -3.22928190e-01 2.98179239e-02 5.03099263e-01
7.81421185e-01 -4.89986278e-02 -2.06546068e+00 -1.41644314e-01
-1.88369885e-01 -6.33499920e-01 2.00330794e-01 -1.12029862e+00
-9.00377154e-01 7.01861441e-01 6.83035791e-01 -6.75558001e-02
1.14833295e+00 4.35090577e-03 9.92147923e-01 3.37977856e-01
3.03944886e-01 -1.25414622e+00 2.70670801e-01 7.77191222e-01
1.01818204e+00 -1.47311592e+00 3.74970675e-01 -3.60673010e-01
-8.48144770e-01 7.80909479e-01 7.30839133e-01 -1.79978698e-01
8.19435418e-01 5.75362444e-01 4.49421555e-01 -2.55713850e-01
-5.20957947e-01 -5.14818192e-01 1.43025905e-01 7.78817475e-01
-3.20141613e-02 1.83465213e-01 1.18284531e-01 2.17802867e-01
1.76342010e-01 -5.13254926e-02 7.73870409e-01 1.33989966e+00
-6.37731791e-01 -7.20162153e-01 -5.42704046e-01 4.64254111e-01
2.22066164e-01 -4.33442324e-01 -7.98963666e-01 4.93479371e-01
1.18388705e-01 8.84924412e-01 -3.01083744e-01 -4.45540875e-01
1.44185543e-01 -1.27434626e-01 1.45644844e-01 -5.55269003e-01
-5.53103685e-02 2.76725650e-01 1.79361627e-01 -8.83657694e-01
-1.01306371e-01 -7.56727874e-01 -1.41700518e+00 -3.47982556e-01
-2.05803588e-01 -4.42548096e-01 5.34901142e-01 7.99613476e-01
4.48190212e-01 7.73748219e-01 4.38684642e-01 -1.50173354e+00
5.45708686e-02 -8.66063118e-01 -4.76645023e-01 -8.16517696e-02
7.21039593e-01 -8.60996664e-01 -5.12664139e-01 -1.13316730e-01] | [10.847152709960938, -4.1112775802612305] |
80ae1741-dcc9-410a-a5eb-bb45d13263be | fast-localization-and-single-pixel-imaging-of | 2208.07371 | null | https://arxiv.org/abs/2208.07371v1 | https://arxiv.org/pdf/2208.07371v1.pdf | Fast localization and single-pixel imaging of the moving object using time-division multiplexing | When imaging moving objects, single-pixel imaging produces motion blur. This paper proposes a new single-pixel imaging method, which can achieve anti-motion blur imaging of a fast-moving object. The geometric moment patterns and Hadamard patterns are used to alternately encode the position information and the image information of the object with time-division multiplexing. In the reconstruction process, the object position information is extracted independently and combining motion-compensation reconstruction algorithm to decouple the object motion from image information. As a result, the anti-motion blur image and the high frame rate object positions are obtained. Experimental results show that for a moving object with an angular velocity of up to 0.5rad/s relative to the imaging system, the proposed method achieves a localization frequency of 5.55kHz, and gradually reconstructs a clear image of the fast-moving object with a pseudo resolution of 512x512. The method has application prospects in single-pixel imaging of the fast-moving object. | ['Yingjian Wang', 'Yafeng Chen', 'Jian Huang', 'Wei Yang', 'Linbin Zha', 'Dongfeng Shi', 'Wenwen Meng', 'Zijun Guo'] | 2022-08-15 | null | null | null | null | ['motion-compensation'] | ['computer-vision'] | [ 3.73227775e-01 -5.84161341e-01 1.81779668e-01 7.15562031e-02
-1.53057575e-01 -4.07382071e-01 3.44873160e-01 -8.88429582e-01
-7.00261533e-01 8.51085603e-01 4.87342961e-02 -1.63929731e-01
-2.47770667e-01 -2.56510496e-01 -2.26142168e-01 -1.22523928e+00
2.08328664e-01 -6.36500865e-02 5.66202283e-01 7.15602458e-01
5.18394351e-01 3.97147208e-01 -1.39281416e+00 2.00753242e-01
7.48102367e-01 9.64929461e-01 1.00210035e+00 9.03980315e-01
5.66311963e-02 1.00002503e+00 -8.31461132e-01 3.85194212e-01
1.83936507e-01 -5.80650091e-01 -5.30566633e-01 2.72254914e-01
1.67195976e-01 -9.68368828e-01 -7.84187973e-01 1.27997828e+00
2.50005901e-01 2.80862185e-03 4.08413023e-01 -5.77634394e-01
-9.57224727e-01 2.15169042e-01 -1.12104177e+00 5.89536846e-01
2.59092953e-02 4.00047153e-01 -9.73712876e-02 -7.13834405e-01
4.89953429e-01 1.01673472e+00 3.26220512e-01 5.25065243e-01
-1.04001129e+00 -4.80552673e-01 -7.35109746e-01 5.51433563e-01
-1.31484950e+00 -4.90062118e-01 8.20032239e-01 -3.61237019e-01
2.23881319e-01 4.77330446e-01 3.08207542e-01 2.58507580e-01
6.27500355e-01 2.03889579e-01 1.38431680e+00 -4.41056222e-01
-1.16671741e-01 6.03980664e-03 2.96423107e-01 4.03593183e-01
6.82082593e-01 2.39168420e-01 -2.55423427e-01 1.52528226e-01
1.22954154e+00 2.68645316e-01 -1.12440503e+00 1.52053401e-01
-1.57415783e+00 -3.45138200e-02 7.70309567e-02 7.68027544e-01
-2.70481557e-01 2.45571569e-01 -1.96554512e-01 -8.42639506e-02
1.52870929e-02 3.49602908e-01 3.44600305e-02 -3.79759461e-01
-1.01527107e+00 -1.25660583e-01 4.52391654e-01 7.93186605e-01
3.88743907e-01 1.82612687e-01 -4.11113530e-01 4.51790690e-01
3.39150965e-01 9.14654195e-01 8.01957786e-01 -1.15578294e+00
1.92500487e-01 -1.12047553e-01 8.94891322e-01 -6.18542194e-01
-3.07867136e-02 -6.81614503e-02 -1.01409078e+00 5.22244334e-01
7.77324796e-01 -2.09831789e-01 -8.56547058e-01 1.03575909e+00
3.13157469e-01 4.42505121e-01 1.11592516e-01 1.51198578e+00
6.26530588e-01 9.93515015e-01 -6.22401118e-01 -9.17275965e-01
1.64906967e+00 -9.85355198e-01 -1.51043367e+00 -8.28963518e-02
-3.14010739e-01 -1.30320287e+00 5.01479983e-01 3.42136502e-01
-1.28564215e+00 -9.47997987e-01 -1.06183541e+00 1.27673093e-02
5.48744202e-01 5.23034334e-01 4.39819284e-02 3.74951035e-01
-7.23387539e-01 4.64087486e-01 -8.07272077e-01 3.15153390e-01
9.82546657e-02 -2.15458777e-02 6.22662716e-03 -2.96374768e-01
-8.22339416e-01 9.20075893e-01 4.35888350e-01 2.14274466e-01
-3.71941060e-01 -8.06021631e-01 -2.78041661e-01 -4.71813530e-02
2.41909809e-02 -8.29229116e-01 1.38262200e+00 -7.15213478e-01
-1.68707049e+00 4.38208699e-01 -5.15123785e-01 -2.17650190e-01
5.58043301e-01 -2.93917447e-01 -7.62031615e-01 8.14287603e-01
6.16807044e-02 -2.26762686e-02 1.48851776e+00 -1.05162096e+00
-7.04621136e-01 -1.14595272e-01 -6.78078949e-01 2.22025514e-01
1.25089005e-01 9.69702378e-02 -3.35155159e-01 -4.17202175e-01
1.88958302e-01 -5.83650827e-01 2.07062121e-02 1.37198702e-01
2.44186055e-02 5.00900865e-01 1.70117342e+00 -6.10830069e-01
1.43033135e+00 -2.51702809e+00 -2.36025915e-01 -7.14557111e-01
2.01991871e-01 5.72912693e-01 1.84710711e-01 8.51767883e-02
1.06446192e-01 -6.34810627e-01 -2.74317473e-01 1.86199442e-01
-8.72778118e-01 -1.39496014e-01 -3.13998848e-01 6.85768604e-01
-1.27039477e-01 8.84101391e-01 -8.92144978e-01 -3.51584047e-01
4.15365547e-01 5.56197047e-01 8.94140918e-03 2.20192447e-01
3.65646958e-01 9.31718826e-01 -3.75294238e-01 3.91722053e-01
1.46495247e+00 -3.56379271e-01 2.90087704e-02 -6.81686163e-01
-7.67352521e-01 -3.28814834e-01 -1.08478200e+00 1.21142912e+00
-1.52828515e-01 9.06811178e-01 4.64337766e-01 -1.76390469e-01
6.94948435e-01 3.07198524e-01 6.10682070e-01 -6.28313661e-01
6.20671511e-02 2.94445843e-01 1.41173810e-01 -1.02757454e+00
6.15597904e-01 -3.70048106e-01 4.82320219e-01 4.46270376e-01
-3.79574150e-01 -1.20987780e-01 9.39959064e-02 -3.83996218e-01
8.69680822e-01 -2.61785924e-01 1.56758413e-01 -3.72260779e-01
5.76443017e-01 -3.78503948e-02 3.31788093e-01 6.82397783e-01
-2.41929695e-01 7.90947437e-01 -4.21850860e-01 -3.18196446e-01
-1.26842511e+00 -1.07783377e+00 -4.28276867e-01 -3.48919332e-02
1.03776515e+00 3.59432757e-01 -5.96013486e-01 9.42653269e-02
-5.73235601e-02 5.53962350e-01 1.04694180e-02 -8.04774463e-03
-8.96851480e-01 -8.07073653e-01 -3.17574386e-03 7.44044259e-02
1.13261139e+00 -7.07432210e-01 -1.17940795e+00 2.48108923e-01
-3.71907622e-01 -1.22588706e+00 -7.40743756e-01 -5.47515094e-01
-8.09834659e-01 -1.16095209e+00 -1.11662447e+00 -1.06067872e+00
8.37519228e-01 1.02632546e+00 3.40144068e-01 -3.35238606e-01
-4.87369001e-01 1.62540078e-01 1.61704302e-01 1.36581257e-01
-3.14530611e-01 -9.45050955e-01 -5.19107245e-02 4.28593725e-01
7.59455860e-02 -2.33061343e-01 -1.07577205e+00 5.20238817e-01
-1.00494123e+00 4.03371245e-01 7.77858734e-01 7.76560068e-01
6.50604293e-02 5.92412114e-01 8.53953362e-02 -2.68161952e-01
3.49397004e-01 1.89426124e-01 -1.12410426e+00 -8.53007287e-02
-3.03620100e-01 -1.47673011e-01 4.98042017e-01 -1.02041531e+00
-1.74086320e+00 -2.24250361e-01 5.83491087e-01 -6.21831894e-01
-1.15158543e-01 1.62152246e-01 5.80625013e-02 -2.68415272e-01
5.06920934e-01 6.30066693e-01 4.88022715e-01 -5.73718011e-01
1.62548959e-01 9.80479360e-01 1.07972360e+00 1.41506478e-01
8.99927020e-01 8.73060465e-01 1.40559137e-01 -1.16680396e+00
-1.58221319e-01 -6.28513336e-01 -2.47536793e-01 -4.34259564e-01
1.02240443e+00 -1.04238439e+00 -9.12878156e-01 1.26956308e+00
-1.40370715e+00 -5.63340075e-02 -3.63843068e-02 1.16023374e+00
-1.91431835e-01 6.93779409e-01 -1.22912979e+00 -6.36399388e-01
-2.51039386e-01 -9.49302793e-01 6.89925015e-01 8.03067923e-01
1.80558488e-01 -6.94636226e-01 -4.71593291e-01 3.79177958e-01
7.89863706e-01 -2.24740133e-01 4.31802630e-01 5.86272717e-01
-1.46150255e+00 -2.37084758e-02 -5.87335944e-01 1.11577921e-01
6.03165865e-01 -2.65852869e-01 -9.59263682e-01 -2.20252484e-01
1.04885089e+00 5.64950526e-01 6.72944725e-01 8.77400875e-01
8.69828939e-01 -5.94887376e-01 -4.01523083e-01 5.58650315e-01
1.75855672e+00 8.11711848e-01 9.23007011e-01 1.28851816e-01
6.69897199e-01 4.16701287e-02 8.82893324e-01 3.41078609e-01
-2.98818737e-01 7.57484674e-01 -7.79723050e-03 -1.03450299e-03
-5.34413040e-01 2.69364327e-01 2.66005509e-02 8.56075764e-01
-9.42902416e-02 -2.68884040e-02 -3.92752737e-01 4.34310108e-01
-1.43471885e+00 -1.29382050e+00 -8.74145985e-01 2.25311518e+00
8.10124815e-01 -2.91501731e-01 -5.86654544e-01 -8.98015127e-02
1.04730272e+00 3.18846911e-01 -4.07048404e-01 2.74008125e-01
-8.29915851e-02 -1.65961385e-01 8.30679834e-01 9.04271543e-01
-8.77742946e-01 4.03453082e-01 5.88945913e+00 8.24518442e-01
-1.55027175e+00 1.55385733e-01 1.88885853e-01 -1.61811933e-02
4.25538979e-02 -1.35554597e-01 -7.68778622e-01 1.19226182e+00
5.66444993e-01 -4.77794290e-01 5.51987588e-01 2.98040390e-01
5.31638920e-01 -7.63046801e-01 -4.64906543e-01 1.41546297e+00
-1.40712783e-01 -1.10214114e+00 -2.28440389e-01 9.73042548e-02
8.27881038e-01 -5.72396815e-01 3.00582528e-01 -6.31269455e-01
-5.08720815e-01 -5.14919281e-01 6.72809660e-01 8.90730083e-01
1.12948120e+00 -2.95785189e-01 5.20686746e-01 5.65071225e-01
-9.08949077e-01 -2.24921152e-01 -4.22034889e-01 -2.95100242e-01
6.68323398e-01 1.07466042e+00 -7.61257887e-01 4.71834242e-01
4.39988703e-01 5.05510151e-01 -3.40497606e-02 1.25606465e+00
3.39162424e-02 3.63967299e-01 -8.82513151e-02 1.37556538e-01
-2.33883914e-02 -6.39486074e-01 8.19229782e-01 7.92448342e-01
9.36365843e-01 6.33928597e-01 -3.77803892e-01 9.74060297e-01
3.44004244e-01 -8.35932553e-01 -3.10889661e-01 1.01408988e-01
5.57278812e-01 1.18216896e+00 -5.19057393e-01 -5.34446776e-01
-3.49955797e-01 1.29935145e+00 -6.18339419e-01 6.30880654e-01
-9.66485023e-01 -7.71504343e-01 5.76043248e-01 1.10117473e-01
6.45292997e-01 -4.64499086e-01 -1.61492154e-01 -1.13062668e+00
1.11565985e-01 -2.35008031e-01 -2.79366344e-01 -1.37869191e+00
-7.05509007e-01 1.88306004e-01 -1.10035360e-01 -1.56460440e+00
8.11761022e-02 -5.63537240e-01 -5.75829446e-01 1.22822714e+00
-1.36438751e+00 -6.32707059e-01 -4.32047307e-01 2.61344761e-01
6.19595766e-01 2.64688671e-01 2.53909439e-01 3.49341929e-01
-2.22352296e-01 -3.28109592e-01 6.93219841e-01 -1.43400729e-01
7.62862325e-01 -7.97221899e-01 -9.78858545e-02 1.16860998e+00
-5.30575097e-01 4.68980968e-01 7.74160862e-01 -6.49474919e-01
-1.60677969e+00 -9.10356224e-01 7.46711493e-01 -2.08192438e-01
4.62906033e-01 1.54584914e-01 -1.17475426e+00 2.42499597e-02
4.98134792e-01 2.36947849e-01 -1.53732440e-02 -1.21716535e+00
4.27386016e-01 -3.28487724e-01 -1.25845635e+00 4.23910975e-01
7.58191824e-01 -4.25396889e-01 -6.92561686e-01 -1.35274872e-01
7.51821041e-01 -7.96783388e-01 -5.70890963e-01 4.32831049e-01
6.05139732e-01 -9.45043325e-01 9.67571497e-01 4.03556019e-01
3.28272700e-01 -1.23822904e+00 4.23743874e-01 -1.08192635e+00
-7.64702916e-01 -8.82649004e-01 -2.29337737e-01 1.04249203e+00
-3.07610810e-01 -8.29738319e-01 4.33849961e-01 3.06179851e-01
-7.31986910e-02 -1.00441501e-01 -9.83085811e-01 -6.89020574e-01
-9.04146731e-01 4.03468043e-01 9.39759985e-02 7.62913287e-01
-1.47994846e-01 2.48138919e-01 -6.47961497e-01 6.41267538e-01
9.92132664e-01 4.59884495e-01 3.00644010e-01 -6.68174505e-01
-2.52690077e-01 -2.31342942e-01 -4.11774009e-01 -1.71913493e+00
-5.25668323e-01 -1.10851116e-01 3.82631496e-02 -1.30909145e+00
1.96018696e-01 1.58782914e-01 -1.33900508e-01 -4.59080607e-01
-3.38703483e-01 3.02736282e-01 1.11674257e-01 7.05140829e-01
-4.79114763e-02 9.12183151e-02 1.92014623e+00 -1.25136718e-01
-8.75257999e-02 1.17257230e-01 -1.62490472e-01 3.75365973e-01
5.17177045e-01 -2.66094971e-02 -4.40249085e-01 -6.70516312e-01
-7.95327723e-01 4.39982235e-01 5.64072847e-01 -1.11987078e+00
5.01020908e-01 -1.29036516e-01 7.33891189e-01 -6.36446774e-01
3.19432616e-01 -1.10230649e+00 9.04126704e-01 8.97064865e-01
1.87562093e-01 -3.49780262e-01 9.09045935e-02 5.57204902e-01
-3.86238188e-01 -5.32683313e-01 1.07760715e+00 -1.06936194e-01
-8.33840311e-01 -1.44753039e-01 -5.26732624e-01 -3.43134701e-01
1.23760164e+00 -4.40404415e-01 -9.20274973e-01 -1.67522296e-01
-3.12423944e-01 -2.96017021e-01 6.50926888e-01 -3.25188600e-02
7.92231321e-01 -1.32229280e+00 -3.06602180e-01 5.27578771e-01
-4.86756653e-01 -8.23301449e-02 7.90196657e-01 1.20155573e+00
-1.04052079e+00 4.87425894e-01 -2.37517625e-01 -8.32034528e-01
-1.51733625e+00 8.10562670e-01 3.51556808e-01 2.61056870e-01
-7.14336038e-01 6.37031972e-01 1.99051857e-01 9.19516504e-01
-3.65161091e-01 -3.82150084e-01 4.98323031e-02 -3.12248349e-01
1.19215882e+00 6.80767894e-01 -3.79970193e-01 -4.39023823e-01
-1.80362463e-02 9.17077303e-01 7.67437965e-02 -2.50330478e-01
8.47889900e-01 -5.87293684e-01 -2.42467165e-01 4.95878071e-01
1.10275936e+00 4.32489157e-01 -1.61435664e+00 -9.75108370e-02
-2.37300098e-01 -1.20716786e+00 -8.18544477e-02 -6.57138526e-01
-8.30265641e-01 7.48676360e-01 6.39259756e-01 3.09530765e-01
1.35074461e+00 -7.59286061e-02 7.60516107e-01 -3.10267031e-01
3.94596756e-01 -7.30371594e-01 1.00241147e-01 2.35776260e-01
5.22840977e-01 -7.11399496e-01 9.06324908e-02 -5.13729572e-01
-3.30534905e-01 1.28670859e+00 5.76597691e-01 1.85019687e-01
2.32729062e-01 3.31566870e-01 1.88957900e-01 3.04451257e-01
-5.05490482e-01 5.48242107e-02 2.52321124e-01 4.61373866e-01
1.15601011e-01 -7.17754737e-02 -5.99057019e-01 2.60391440e-02
4.31278259e-01 3.18250000e-01 8.92429173e-01 8.01249862e-01
-7.45781422e-01 -5.44616044e-01 -9.16180491e-01 2.90920250e-02
-5.36413372e-01 2.36620814e-01 4.00962412e-01 4.35543776e-01
3.55139002e-02 8.52661133e-01 4.31248575e-01 6.71548396e-02
8.92073140e-02 -2.36898854e-01 7.47059405e-01 -3.34042823e-04
3.26566339e-01 4.62035060e-01 -2.15760037e-01 -3.63908917e-01
-5.62885046e-01 -3.97197008e-01 -1.23449492e+00 -1.22134700e-01
-2.24430412e-01 2.49945208e-01 5.31575203e-01 3.47384036e-01
3.27535957e-01 4.48481619e-01 6.83503330e-01 -7.59229600e-01
-4.10454601e-01 -9.17646110e-01 -8.59659970e-01 2.91721553e-01
9.13401127e-01 -1.42189369e-01 -6.78421319e-01 3.41611862e-01] | [11.454476356506348, -2.650191068649292] |
0928108e-5c3d-4ed4-addf-dc567cf2c627 | a-contrario-horizon-first-vanishing-point | null | null | http://openaccess.thecvf.com/content_ECCV_2018/html/Gilles_Simon_A_Contrario_Horizon-First_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Gilles_Simon_A_Contrario_Horizon-First_ECCV_2018_paper.pdf | A-Contrario Horizon-First Vanishing Point Detection Using Second-Order Grouping Laws | We show that, in images of man-made environments, the horizon line can usually be hypothesized based on an a contrario detection of second-order grouping events. This allows constraining the extraction of the horizontal vanishing points on that line, thus reducing false detections. Experiments made on three datasets show that our method, not only achieves state-of-the-art performance w.r.t. horizon line detection on two datasets, but also yields much less spurious vanishing points than the previous top-ranked methods. | ['Marie-Odile Berger', 'Antoine Fond', 'Gilles Simon'] | 2018-09-01 | null | null | null | eccv-2018-9 | ['line-detection', 'horizon-line-estimation'] | ['computer-vision', 'computer-vision'] | [ 1.79055870e-01 1.51778519e-01 1.62976414e-01 -2.42531046e-01
-3.79462183e-01 -5.82795858e-01 8.85766745e-01 3.77922207e-01
-2.75502354e-01 3.23993176e-01 4.34900969e-02 -1.26238421e-01
-4.74972241e-02 -6.28028929e-01 -7.74945259e-01 -6.30408823e-01
-8.34329247e-01 3.09287161e-01 1.13565528e+00 -2.43057355e-01
5.97677350e-01 4.65977550e-01 -1.72026396e+00 3.28286320e-01
2.04560295e-01 9.56756353e-01 4.07287327e-04 7.46145427e-01
4.25734311e-01 5.21677852e-01 -6.38104558e-01 -3.48746777e-01
4.86314744e-01 -3.00679445e-01 -5.30674815e-01 4.27854806e-01
9.43794906e-01 -2.71851420e-01 -3.18509966e-01 1.03684127e+00
2.03764454e-01 2.79063229e-02 6.80527925e-01 -9.12379980e-01
5.63227348e-02 1.56533197e-01 -9.65972006e-01 4.74514455e-01
5.53686023e-01 -1.38280302e-01 1.23631024e+00 -1.08489966e+00
1.00400555e+00 1.16858256e+00 7.81286776e-01 -2.06504226e-01
-1.20108342e+00 -1.64903477e-01 1.44703716e-01 1.59390733e-01
-1.70823622e+00 -2.12813094e-01 5.41164160e-01 -5.02856135e-01
8.77434731e-01 5.16933262e-01 6.42250597e-01 5.94945550e-01
4.20230985e-01 6.72270536e-01 9.43740487e-01 -5.78388751e-01
1.31256491e-01 -5.81977218e-02 -4.94675972e-02 8.92068028e-01
4.35842067e-01 1.54569224e-01 -7.09395766e-01 -2.01302335e-01
8.23952615e-01 -3.75639200e-01 -1.36464000e-01 -6.08124435e-01
-1.35730398e+00 6.75337315e-01 5.44231713e-01 1.36073917e-01
-1.78567067e-01 -2.07292870e-01 3.92901421e-01 -4.10130471e-02
4.21574473e-01 6.57852530e-01 -1.34661183e-01 2.06643239e-01
-9.86480713e-01 3.86860251e-01 8.12045515e-01 9.66933191e-01
7.25600183e-01 -2.61995018e-01 1.77655876e-01 3.87169987e-01
1.04831336e-02 3.91152650e-01 -4.30562124e-02 -7.22056746e-01
2.18025714e-01 4.99843299e-01 3.45435560e-01 -1.20743549e+00
-9.14408028e-01 -5.81158400e-01 -5.84975958e-01 2.33815506e-01
4.88025486e-01 -1.56694558e-02 -9.07758892e-01 1.25671780e+00
4.32093561e-01 1.36741623e-01 -1.52300268e-01 8.03089678e-01
4.64520723e-01 5.04809797e-01 -6.21312737e-01 -2.79599875e-01
1.41434562e+00 -7.13854373e-01 -4.60137725e-01 -4.15456861e-01
4.02372301e-01 -1.00372970e+00 6.21091068e-01 8.10724735e-01
-4.15669441e-01 -5.87507248e-01 -1.21736407e+00 1.57993287e-01
-3.81883979e-01 1.14372887e-01 8.18869770e-01 4.09424424e-01
-7.30976045e-01 5.92734516e-01 -7.51560271e-01 -6.63001060e-01
-1.09725587e-01 -1.14985583e-02 -3.30216497e-01 4.40594018e-01
-7.30282307e-01 6.84197783e-01 4.41241384e-01 1.01470932e-01
-5.99469662e-01 -1.21195205e-01 -8.33837748e-01 -2.28030980e-01
6.73595846e-01 -2.14360669e-01 1.17755437e+00 -4.09149319e-01
-9.97766614e-01 1.16522729e+00 -2.33655766e-01 -9.48251247e-01
9.26055312e-01 -8.08050513e-01 -5.75483441e-01 4.35213447e-01
3.54944140e-01 6.49678707e-01 9.07830358e-01 -1.40215242e+00
-1.21364915e+00 -1.36057496e-01 -5.76298088e-02 1.19079188e-01
5.93445972e-02 1.66844984e-03 -3.79502237e-01 -3.13311487e-01
5.27050078e-01 -1.04063427e+00 -3.80369633e-01 -1.06095120e-01
-8.27023923e-01 -2.31412753e-01 9.07197237e-01 -5.07455647e-01
1.20344722e+00 -2.30033255e+00 -1.66664228e-01 4.80322659e-01
2.37652957e-01 -6.17033653e-02 4.53500509e-01 6.34672523e-01
1.04415566e-01 7.27747232e-02 2.56721489e-02 -2.86547899e-01
-1.64543509e-01 1.82885557e-01 -6.50867701e-01 1.02887106e+00
5.99741153e-02 2.44140327e-01 -8.29791486e-01 -3.62552106e-01
6.60308838e-01 -2.02231124e-01 -2.81511217e-01 -1.39784813e-03
2.28090018e-01 7.87255168e-02 4.97642308e-02 3.53176802e-01
6.04994655e-01 -1.58929661e-01 1.24941148e-01 -1.70096934e-01
-6.98014617e-01 2.14556262e-01 -1.36351061e+00 9.66589093e-01
1.90800324e-01 1.21397030e+00 -4.59577352e-01 -6.21934533e-01
1.12030888e+00 -5.25035616e-03 2.51521766e-01 -3.77227247e-01
7.88545609e-02 1.84729010e-01 1.70104533e-01 -3.31268251e-01
8.66124153e-01 3.48746747e-01 -3.41670275e-01 -5.25109887e-01
-9.86098871e-02 -3.57724518e-01 3.58935803e-01 1.03150070e-01
8.94935668e-01 -9.49249789e-02 8.03530574e-01 -5.09211659e-01
3.68195802e-01 -2.68624220e-02 6.13246679e-01 1.21419621e+00
-1.59135595e-01 8.05905879e-01 5.95885992e-01 -6.54347181e-01
-1.22834444e+00 -1.15858567e+00 -2.37343773e-01 7.26748049e-01
6.89834893e-01 -7.57478774e-01 -4.64894891e-01 -3.74724567e-01
7.06666261e-02 6.70383275e-01 -8.61169577e-01 7.87104294e-02
-6.80439174e-01 -4.74124402e-01 3.47224891e-01 2.88744032e-01
4.29871559e-01 -5.45120358e-01 -1.22685230e+00 6.17320649e-02
3.79613340e-02 -1.53438115e+00 -8.82439464e-02 3.47131938e-01
-6.10330641e-01 -1.23526418e+00 -2.75878549e-01 -4.97454911e-01
6.60322785e-01 7.00517297e-01 1.08438754e+00 -4.18861955e-03
-4.67850536e-01 6.49020821e-02 -4.70947981e-01 -5.81332445e-01
5.91424890e-02 -2.61625558e-01 1.16797224e-01 1.12652272e-01
2.66625851e-01 -2.54357517e-01 -4.89073426e-01 7.28442729e-01
-2.50668973e-01 1.03933536e-01 3.70485723e-01 5.50857306e-01
6.49188221e-01 3.64960223e-01 2.92301625e-02 -6.17191732e-01
-1.40936285e-01 -1.83108628e-01 -1.10022926e+00 8.20594579e-02
-1.52115181e-01 -3.03195775e-01 5.73379099e-01 -2.21646547e-01
-6.53078616e-01 2.04733253e-01 3.77605140e-01 -3.55407745e-01
-4.56645131e-01 1.18081227e-01 4.21543241e-01 -2.86246032e-01
9.71485853e-01 8.27690288e-02 -7.43387699e-01 -2.93095917e-01
3.11848402e-01 2.20748290e-01 1.03619576e+00 -2.07467396e-02
1.01332760e+00 1.07176054e+00 3.57668191e-01 -1.55212867e+00
-8.38561893e-01 -1.05254400e+00 -9.85180497e-01 -4.83980060e-01
7.64987886e-01 -6.81921005e-01 -3.13724488e-01 3.09646070e-01
-1.15448880e+00 8.09560791e-02 -1.22074701e-01 4.37166631e-01
-5.41833043e-01 4.19615209e-01 -1.81912631e-01 -1.16149151e+00
2.19334662e-01 -5.73023200e-01 1.22341239e+00 2.37358928e-01
-3.29451650e-01 -6.03519261e-01 1.08054914e-02 -1.10668965e-01
-4.86293465e-01 7.24201679e-01 2.54493058e-01 -6.10925734e-01
-5.62308967e-01 -5.13824403e-01 -9.39992964e-02 -1.57310978e-01
-8.59183595e-02 4.07513380e-01 -9.62983072e-01 -1.67979762e-01
-1.09592311e-01 1.91974893e-01 1.05647647e+00 4.97609347e-01
7.17175305e-01 -1.24197714e-01 -5.98093152e-01 6.94307148e-01
1.19645298e+00 1.77423909e-01 5.77737749e-01 6.14014387e-01
4.31876957e-01 6.75806344e-01 1.02207899e+00 6.53795242e-01
6.63368106e-02 8.18796039e-01 5.26138008e-01 -3.38340491e-01
1.58789247e-01 -1.66788951e-01 2.63639949e-02 2.06404805e-01
-1.46292165e-01 -2.47935399e-01 -1.09502208e+00 8.51466298e-01
-2.22101212e+00 -9.88766015e-01 -6.94136798e-01 2.29838872e+00
-7.74208922e-03 7.36962616e-01 3.94226581e-01 1.94151416e-01
8.20171535e-01 2.74082333e-01 -3.23719919e-01 -1.00698888e-01
-3.86897653e-01 -3.48357946e-01 1.04724574e+00 2.86456734e-01
-1.69659221e+00 9.01842654e-01 7.70365047e+00 4.45703089e-01
-9.62911725e-01 -2.65044272e-01 2.01697826e-01 7.38276467e-02
4.23337191e-01 1.29146457e-01 -1.18140054e+00 3.77651751e-02
4.04605240e-01 -2.02362284e-01 -6.47423267e-02 9.77916181e-01
1.96813107e-01 -5.45238018e-01 -1.02297974e+00 9.89797533e-01
4.28550206e-02 -1.26801288e+00 -2.47574344e-01 1.84329331e-01
7.44759798e-01 1.45488992e-01 -3.59090656e-01 -2.48280093e-01
1.56658366e-01 -5.36717951e-01 1.16719377e+00 1.71382457e-01
2.19080731e-01 -8.81046236e-01 8.10580313e-01 3.57966900e-01
-1.37247753e+00 1.23897605e-01 -4.63940293e-01 -4.09290642e-01
2.92298883e-01 8.27269077e-01 -1.22343135e+00 7.91717112e-01
9.80839610e-01 6.02549314e-01 -9.16232288e-01 1.67404318e+00
-6.67585671e-01 4.87521619e-01 -7.76558995e-01 -2.35887468e-02
5.22204578e-01 1.30252153e-01 8.75589252e-01 1.60167122e+00
2.21681356e-01 3.65406871e-02 6.51935279e-01 2.94248760e-01
4.98566478e-01 1.38740037e-02 -1.01583397e+00 5.07296085e-01
3.62361521e-01 1.14447105e+00 -1.36632681e+00 -3.21429521e-01
-2.98623621e-01 9.08257544e-01 2.07244158e-01 -5.58023667e-03
-6.79853141e-01 -4.88337785e-01 2.08950445e-01 2.34376639e-01
5.96222281e-01 -6.96658194e-01 -2.38847896e-01 -1.04374492e+00
7.62973204e-02 -3.81520659e-01 4.30663526e-01 -5.66058636e-01
-9.35904741e-01 4.37918305e-01 1.71568885e-01 -1.47573411e+00
-3.36715907e-01 -7.26668596e-01 -7.39146590e-01 2.15485051e-01
-8.64077389e-01 -9.53709483e-01 -3.19097131e-01 2.38297924e-01
5.79814613e-01 1.42205030e-01 6.23404801e-01 -1.74245998e-01
-3.85671496e-01 1.54567704e-01 1.10625774e-01 5.46985529e-02
6.40859246e-01 -1.56671870e+00 6.87624276e-01 1.47958195e+00
4.78614032e-01 3.76347750e-01 1.45641422e+00 -7.37774968e-01
-7.89819479e-01 -7.73861885e-01 5.60480654e-01 -2.84048706e-01
8.05964470e-01 -6.94244087e-01 -8.65167081e-01 7.73619533e-01
1.75132096e-01 -9.73426998e-02 3.03575844e-01 3.36040944e-01
-3.36020708e-01 1.25638902e-01 -7.03715146e-01 7.28732824e-01
1.07748604e+00 -3.96614373e-01 -7.09461510e-01 7.79150963e-01
3.60659868e-01 -5.16230762e-01 -5.72215557e-01 5.91921508e-01
4.56996590e-01 -1.36486375e+00 1.08177793e+00 -2.13165760e-01
5.43953851e-02 -7.07807541e-01 -1.86242566e-01 -1.03626359e+00
-4.34764832e-01 -7.78739870e-01 1.39474161e-02 8.32191527e-01
2.19951227e-01 -4.32720065e-01 7.61959851e-01 -1.25701874e-01
-2.66778529e-01 -2.59160399e-01 -1.10413647e+00 -1.16645491e+00
-7.94767261e-01 -2.76670814e-01 -5.73285781e-02 7.43364334e-01
-4.47719311e-03 4.05926377e-01 -6.67777658e-01 7.25969434e-01
8.33565116e-01 1.58034474e-01 1.23021758e+00 -1.26162112e+00
-3.83423805e-01 -3.47626895e-01 -1.10616910e+00 -1.07950306e+00
-3.14000338e-01 -9.63577107e-02 3.60456914e-01 -1.16645408e+00
-2.50012040e-01 1.50081247e-01 -6.63953498e-02 2.05719471e-01
-4.69839722e-02 4.94697273e-01 1.19727245e-02 3.43727827e-01
-7.91078448e-01 1.71237469e-01 6.43056035e-01 3.16370219e-01
-8.09688196e-02 -4.38495493e-03 -7.84174427e-02 1.48943043e+00
7.22879708e-01 -2.90218890e-01 -1.46941394e-01 7.83283710e-02
1.99901029e-01 -1.48763016e-01 5.94873786e-01 -1.43683529e+00
2.31930718e-01 -2.24068522e-01 5.33998311e-01 -1.27732861e+00
3.73331934e-01 -4.54955995e-01 -1.10793225e-01 6.24607205e-01
1.42075911e-01 1.39451819e-02 3.40923607e-01 6.20924890e-01
-2.53834784e-01 -3.36685121e-01 8.60753357e-01 -1.16075553e-01
-1.26535511e+00 -4.10285801e-01 -4.40951616e-01 -3.28654766e-01
1.09902585e+00 -3.16570967e-01 -4.62511212e-01 -3.69466245e-01
-5.35370111e-01 1.84270456e-01 5.37237346e-01 5.28334498e-01
4.68813002e-01 -9.36937034e-01 -5.46969414e-01 1.17778048e-01
3.15383315e-01 -5.33463433e-02 -1.00927725e-01 7.48060107e-01
-8.57021272e-01 3.07187825e-01 6.33114995e-03 -1.03082073e+00
-1.59883130e+00 5.15211642e-01 2.78617918e-01 -1.24693461e-01
-9.80601013e-01 8.78854215e-01 5.66182017e-01 9.29624140e-02
9.53098387e-02 -1.49860635e-01 -2.49721870e-01 4.91294488e-02
7.00593591e-01 4.45540607e-01 -4.77024913e-02 -7.31192172e-01
-6.11378133e-01 6.09559655e-01 -2.01712295e-01 -2.93144077e-01
1.00155056e+00 -1.57356858e-01 1.01335198e-01 7.44339883e-01
7.56895721e-01 2.76960045e-01 -1.13719296e+00 -5.30696474e-02
3.00028950e-01 -8.10827374e-01 -1.10001102e-01 -3.15472305e-01
-2.05026239e-01 5.99054277e-01 5.18937349e-01 7.74566054e-01
7.31338084e-01 8.71350393e-02 1.69408202e-01 6.19582534e-01
6.06182277e-01 -9.84007001e-01 -2.69959457e-02 5.35303831e-01
9.19771194e-01 -1.10773480e+00 2.84256816e-01 -1.00810003e+00
-2.86950380e-01 1.36181974e+00 5.60160220e-01 -6.46695375e-01
3.77812237e-01 9.06882510e-02 -7.89987892e-02 -3.60757351e-01
-5.32230318e-01 -3.44638288e-01 4.80130732e-01 4.96334612e-01
1.52259439e-01 2.24686801e-01 -2.90724874e-01 -2.77128577e-01
-6.23119831e-01 -4.80056107e-01 7.67333448e-01 9.20843959e-01
-1.00117421e+00 -3.94898325e-01 -7.74397194e-01 2.92585522e-01
-3.00783515e-01 7.41556287e-02 -7.84769714e-01 1.22553360e+00
3.20641249e-01 9.85541224e-01 3.56497854e-01 -5.09434819e-01
5.75283468e-01 -3.38071436e-01 4.30360764e-01 -5.16193628e-01
-2.54736006e-01 4.55472469e-01 4.83297765e-01 -6.70233071e-01
-1.77477315e-01 -1.05408955e+00 -1.04959595e+00 -3.98506783e-02
-6.32043004e-01 1.18943781e-01 3.16501081e-01 7.84107149e-01
3.24607939e-02 4.76826906e-01 6.17338896e-01 -1.15139151e+00
-1.81763574e-01 -8.10508072e-01 -7.09831119e-01 4.10839438e-01
4.82775152e-01 -9.07764375e-01 -5.98038375e-01 7.27988183e-02] | [8.130681037902832, -1.8470226526260376] |
c8025f0b-a292-4318-be8e-f10ad07f254b | removing-radio-frequency-interference-from | 2210.12931 | null | https://arxiv.org/abs/2210.12931v3 | https://arxiv.org/pdf/2210.12931v3.pdf | Removing Radio Frequency Interference from Auroral Kilometric Radiation with Stacked Autoencoders | Radio frequency data in astronomy enable scientists to analyze astrophysical phenomena. However, these data can be corrupted by radio frequency interference (RFI) that limits the observation of underlying natural processes. In this study, we extend recent developments in deep learning algorithms to astronomy data. We remove RFI from time-frequency spectrograms containing auroral kilometric radiation (AKR), a coherent radio emission originating from the Earth's auroral zones that is used to study astrophysical plasmas. We propose a Denoising Autoencoder for Auroral Radio Emissions (DAARE) trained with synthetic spectrograms to denoise AKR signals collected at the South Pole Station. DAARE achieves 42.2 peak signal-to-noise ratio (PSNR) and 0.981 structural similarity (SSIM) on synthesized AKR observations, improving PSNR by 3.9 and SSIM by 0.064 compared to state-of-the-art filtering and denoising networks. Qualitative comparisons demonstrate DAARE's capability to effectively remove RFI from real AKR observations, despite being trained completely on a dataset of simulated AKR. The framework for simulating AKR, training DAARE, and employing DAARE can be accessed at github.com/Cylumn/daare. | ['Philip J. Erickson', 'Ryan Volz', 'John Swoboda', 'James LaBelle', 'Mary Knapp', 'Allen Chang'] | 2022-10-24 | null | null | null | null | ['astronomy'] | ['miscellaneous'] | [ 3.25381905e-02 -4.44017857e-01 6.12493336e-01 5.54884858e-02
-6.57073438e-01 -5.06141543e-01 9.37998891e-01 -6.44403696e-01
-1.74618602e-01 7.39640832e-01 4.53830808e-01 -2.82198131e-01
-3.16612124e-01 -8.77497077e-01 -6.90941811e-01 -1.13540196e+00
-5.67195043e-02 -5.07415980e-02 -4.43900466e-01 -3.17036152e-01
2.72646360e-02 1.09451604e+00 -1.52055299e+00 3.54225636e-01
4.85358715e-01 1.11599112e+00 2.90589064e-01 9.70638216e-01
2.70136178e-01 9.11982775e-01 -1.09148479e+00 2.86074549e-01
3.58080864e-01 -4.48346317e-01 -4.67046440e-01 -3.10694128e-01
3.39183927e-01 -2.56730974e-01 -1.05606031e+00 9.76507068e-01
7.24646568e-01 5.58441877e-01 8.85358691e-01 -5.53010106e-01
-9.42335486e-01 1.02319568e-01 -5.69033623e-01 9.69950616e-01
-1.15214571e-01 1.68823645e-01 4.76841003e-01 -9.69388843e-01
3.07027817e-01 1.02937579e+00 6.67716026e-01 3.56856823e-01
-1.21903872e+00 -5.39561987e-01 -7.27130294e-01 2.88989693e-01
-1.18664920e+00 -6.03340447e-01 6.32673740e-01 -4.04769957e-01
1.26894176e+00 6.15060568e-01 4.97543961e-01 1.23274922e+00
5.14904618e-01 9.88261774e-02 1.08726215e+00 -4.41706479e-01
1.62705079e-01 -3.79194379e-01 -1.86368376e-01 1.62082434e-01
-1.07528634e-01 8.38517845e-01 -5.57735801e-01 -1.54590234e-01
8.54233921e-01 -3.94007564e-01 -8.55025530e-01 6.76449001e-01
-1.13112831e+00 8.12262356e-01 5.22353768e-01 6.49333715e-01
-7.00443685e-01 3.48983705e-01 -2.19525266e-02 9.54462051e-01
6.72756195e-01 7.25610018e-01 -3.24804813e-01 4.14740518e-02
-6.40022814e-01 3.43514949e-01 6.21739626e-01 2.82474577e-01
3.02813262e-01 9.00108576e-01 1.88485999e-02 1.07653999e+00
1.17978491e-01 1.12861371e+00 7.77555346e-01 -1.12453020e+00
-3.81369919e-01 -5.57930470e-01 2.45157614e-01 -9.61700439e-01
-3.63013744e-01 -8.71355355e-01 -1.23600101e+00 3.28889161e-01
1.44298196e-01 -2.59957165e-01 -9.66974974e-01 1.44169140e+00
1.24268673e-01 4.27673995e-01 4.86545384e-01 1.19582534e+00
1.19963729e+00 1.30556715e+00 -5.16667664e-01 -3.11717182e-01
1.34495544e+00 -6.29141688e-01 -7.67720103e-01 -2.84227878e-01
1.97806314e-01 -1.04666018e+00 6.34483159e-01 7.89691150e-01
-9.47659910e-01 -7.08703756e-01 -9.08784330e-01 5.89801744e-02
-7.84228817e-02 -6.45465963e-03 5.11690438e-01 4.84859645e-01
-9.67539847e-01 7.52178550e-01 -5.68806291e-01 2.10117489e-01
-1.93035584e-02 -1.50064439e-01 -1.98466778e-01 1.84630305e-01
-1.48275912e+00 8.08504403e-01 2.65203640e-02 1.88934565e-01
-1.34779978e+00 -1.09213209e+00 -2.87862062e-01 2.83198833e-01
-5.15809003e-03 -6.96776867e-01 1.67737257e+00 -9.28671300e-01
-1.57120323e+00 3.31438899e-01 1.62655562e-01 -8.94535184e-01
1.18085973e-01 -4.38494861e-01 -1.30248249e+00 4.96321917e-01
-3.07690173e-01 1.49058670e-01 1.24214745e+00 -1.02056396e+00
-3.95441532e-01 2.08531208e-02 -5.54879308e-01 -5.18693663e-02
3.33205760e-02 5.17590977e-02 3.15665990e-01 -1.11727405e+00
2.17184201e-01 -7.57213235e-01 7.38395676e-02 -3.62081945e-01
1.33171797e-01 5.30422106e-02 1.02962434e+00 -9.31237578e-01
6.77091360e-01 -2.32042098e+00 -1.23478016e-02 3.05512585e-02
1.87761471e-01 3.33096921e-01 -1.98379874e-01 2.93826967e-01
-6.06709182e-01 -3.30591559e-01 -2.44983852e-01 7.91003257e-02
-3.24827135e-01 -1.30006433e-01 -9.10922289e-01 7.84941554e-01
3.49953622e-02 4.11356390e-01 -6.42770588e-01 6.14295483e-01
3.13460886e-01 9.22659397e-01 -1.37082309e-01 1.39097154e-01
-5.28210998e-02 7.29745984e-01 -1.89533144e-01 3.64120424e-01
9.48371232e-01 1.63646966e-01 -4.84763026e-01 -3.94707859e-01
-3.76351804e-01 2.89566815e-01 -8.29036236e-01 1.20241594e+00
-7.99868762e-01 1.19445825e+00 4.34845805e-01 -9.97562885e-01
1.17895651e+00 4.67668176e-01 3.21327835e-01 -1.21115613e+00
2.09146678e-01 2.79174685e-01 1.02234572e-01 -3.92779827e-01
3.66664320e-01 -5.12858152e-01 3.42107713e-01 4.75388736e-01
1.32166028e-01 -5.14337659e-01 -1.58040777e-01 1.36165112e-01
1.20580006e+00 -4.92136449e-01 -1.65041178e-01 -5.07190347e-01
6.10939264e-01 -1.53064579e-01 2.34924793e-01 1.16491055e+00
1.96356118e-01 8.10267448e-01 -1.97566390e-01 -7.69855380e-01
-1.33999836e+00 -1.13505125e+00 -5.98476768e-01 7.12337375e-01
-3.64132255e-01 2.08473176e-01 -4.33324099e-01 3.69348861e-02
-9.49859023e-02 1.12292838e+00 -4.60482806e-01 -4.43913430e-01
-4.56071079e-01 -1.18733978e+00 6.91741467e-01 2.27932632e-01
6.77656412e-01 -1.17455697e+00 -3.83534759e-01 2.40404248e-01
-4.07016397e-01 -8.91993999e-01 2.15934236e-02 1.64587557e-01
-5.33270478e-01 -4.92587328e-01 -6.35669291e-01 -1.87388241e-01
5.95975518e-02 5.76188266e-01 1.30474687e+00 -8.23210180e-02
-7.13289142e-01 4.45094824e-01 -3.04211706e-01 -6.32117033e-01
-8.14123154e-01 -6.27434313e-01 4.57484841e-01 -9.70465541e-02
2.37458616e-01 -8.98596764e-01 -5.50227284e-01 2.74269521e-01
-1.08095837e+00 -2.50430077e-01 5.05649388e-01 8.78072143e-01
5.71669221e-01 4.14218664e-01 8.44271421e-01 -1.95204705e-01
6.43668771e-01 -5.17607093e-01 -8.89973700e-01 -3.52722019e-01
-1.47054046e-01 -1.78949967e-01 8.82538319e-01 -3.43442321e-01
-1.42420626e+00 -5.48408628e-01 -2.96653271e-01 -5.78367829e-01
-3.32212836e-01 4.33467984e-01 3.56228113e-01 -2.15290874e-01
1.27390182e+00 4.62170273e-01 -1.37481257e-01 -4.64849830e-01
2.05852523e-01 9.86468673e-01 1.28073812e+00 -2.67995209e-01
8.24803770e-01 7.89565861e-01 3.56078446e-02 -1.56108081e+00
-9.24038768e-01 -4.62817967e-01 3.15219134e-01 -3.51915896e-01
6.41509891e-01 -1.18900967e+00 -5.91832399e-01 6.02388144e-01
-9.18767571e-01 -1.65616587e-01 -3.07387203e-01 8.85321498e-01
-4.68356133e-01 2.04195157e-01 -6.96866810e-01 -6.85911059e-01
-6.86339617e-01 -5.63354850e-01 7.53956378e-01 2.68188685e-01
4.30645794e-02 -7.28334129e-01 3.53957891e-01 2.99882770e-01
8.01626921e-01 1.70193195e-01 4.37364399e-01 -2.04526305e-01
-3.22809070e-01 -5.03558144e-02 -3.85005847e-02 7.67075837e-01
1.10859439e-01 -2.01368883e-01 -1.41321921e+00 -4.81183708e-01
1.06399500e+00 -2.82341659e-01 1.21050692e+00 6.77936375e-01
1.31404710e+00 -1.32729143e-01 1.37973040e-01 9.13915157e-01
1.26504910e+00 3.21333915e-01 9.73761559e-01 4.08208042e-01
3.03704381e-01 2.27917388e-01 2.62100160e-01 5.71103692e-01
-9.46613848e-01 2.50141054e-01 3.62852693e-01 -1.18680008e-01
-3.14797580e-01 5.68698406e-01 3.56410086e-01 6.00421906e-01
-1.80183172e-01 -5.32793760e-01 -7.30432987e-01 6.00488544e-01
-1.10668564e+00 -1.16241741e+00 -6.03029847e-01 1.90469825e+00
5.44929624e-01 -1.79417104e-01 -5.09568751e-01 2.68107116e-01
4.98116046e-01 3.58008921e-01 -3.49560648e-01 -5.58656037e-01
-3.61525446e-01 5.40848970e-01 5.86503983e-01 5.35114288e-01
-9.97154534e-01 3.14715922e-01 5.53319216e+00 7.55693316e-01
-1.41852379e+00 8.36306885e-02 4.64983106e-01 -2.25789279e-01
-2.83611774e-01 -6.46254361e-01 -1.39650136e-01 3.28157455e-01
1.45922792e+00 -1.53510720e-01 9.45224762e-01 3.97539139e-01
5.06360650e-01 -5.36482185e-02 -4.23622578e-01 1.03060591e+00
-1.54011235e-01 -1.37840009e+00 -1.55644119e-01 -8.05157423e-02
7.82194734e-01 5.57270169e-01 3.44654977e-01 6.03638701e-02
2.01834291e-01 -1.26868379e+00 3.06404233e-01 1.06069422e+00
6.68743193e-01 -1.11321568e+00 6.52399182e-01 2.48020977e-01
-4.53092963e-01 -9.42329690e-02 -7.62996197e-01 -7.54502863e-02
-7.01876031e-03 1.44358397e+00 -7.94030964e-01 7.92435169e-01
1.25392342e+00 3.00557613e-01 2.51243226e-02 6.64834499e-01
-6.94749877e-02 8.77663851e-01 -5.64619839e-01 4.64957714e-01
3.86326879e-01 -4.17834014e-01 1.00178027e+00 1.09604931e+00
8.04729700e-01 3.98828387e-01 -6.48630977e-01 8.10605407e-01
-3.34988445e-01 -4.31079924e-01 -5.85368454e-01 -1.85210019e-01
3.34437609e-01 1.39037538e+00 -4.90236044e-01 -4.05483693e-01
-3.82314175e-01 6.76530898e-01 -5.20786285e-01 4.00026739e-01
-8.46820712e-01 -5.28497219e-01 8.23596835e-01 9.32300743e-03
3.58134836e-01 -1.61977440e-01 -1.00265309e-01 -8.63069832e-01
-3.59541923e-01 -1.05745852e+00 6.61077648e-02 -1.31947827e+00
-1.42127299e+00 7.05951095e-01 -2.81475991e-01 -1.20840883e+00
-8.57560933e-02 -5.24979591e-01 -6.26514375e-01 1.11436415e+00
-1.35660160e+00 -3.87364656e-01 -6.99821830e-01 4.43502903e-01
3.56763244e-01 -5.22845864e-01 8.84765565e-01 2.51400948e-01
-1.91363692e-02 -2.39596739e-01 1.03749800e+00 -3.07681441e-01
6.77710414e-01 -1.25774610e+00 8.09544921e-01 8.76844168e-01
3.31593335e-01 2.48625368e-01 1.25268102e+00 -4.23194349e-01
-1.28951693e+00 -1.23153412e+00 3.88981819e-01 -1.47863358e-01
7.09722102e-01 -7.59632513e-02 -1.33627701e+00 4.22539741e-01
4.75263655e-01 2.30070978e-01 4.54850286e-01 -4.87815857e-01
-1.49427235e-01 -1.51618004e-01 -1.25115085e+00 4.06348944e-01
6.13944650e-01 -4.88188803e-01 -8.31935823e-01 6.70849323e-01
5.67078948e-01 -3.47768158e-01 -9.36396062e-01 4.72022623e-01
1.09096356e-01 -1.05760145e+00 1.25331831e+00 -2.74375796e-01
5.62508941e-01 -6.44818604e-01 -5.79946563e-02 -1.71533346e+00
-5.61674595e-01 -6.34191513e-01 -2.66912371e-01 6.56170964e-01
-5.16088270e-02 -6.86528206e-01 2.30156690e-01 -4.71548021e-01
-5.82951188e-01 2.38318928e-02 -9.69910264e-01 -8.37966561e-01
5.56310266e-02 -4.02469307e-01 2.68981457e-01 1.00980544e+00
-9.01564717e-01 1.77462980e-01 -5.45779049e-01 7.56599903e-01
6.74559355e-01 1.13589108e-01 5.16544938e-01 -1.21687865e+00
-6.54435396e-01 -1.04719281e-01 4.05673571e-02 -7.25153446e-01
9.78512131e-03 -7.13644922e-01 2.24650592e-01 -1.26721680e+00
-6.84270859e-01 1.38327688e-01 -2.92843938e-01 1.03626922e-01
4.11532879e-01 6.30579054e-01 -2.06168905e-01 1.56415343e-01
4.44240391e-01 7.44877458e-01 1.10457277e+00 -1.50538713e-01
-1.12265218e-02 -1.81640059e-01 -3.54718864e-01 7.43861794e-01
1.02980292e+00 -5.07546306e-01 -2.19884217e-01 -5.56529343e-01
1.02194495e-01 1.97968632e-01 6.35938883e-01 -1.45117402e+00
3.81388217e-02 8.74431282e-02 8.05865288e-01 -3.62795115e-01
3.09205234e-01 -4.41390932e-01 6.15046740e-01 3.43656838e-01
-5.94725497e-02 -3.49882901e-01 5.12094259e-01 6.37744784e-01
-4.31496829e-01 -2.13880062e-01 1.20577264e+00 -2.00952739e-01
-3.97155166e-01 -1.98378429e-01 -9.88356709e-01 -1.88942030e-01
5.09543180e-01 4.32510018e-01 -6.53477192e-01 -6.33009791e-01
-6.17845118e-01 -3.59598935e-01 -4.23932299e-02 2.34230608e-01
5.22030234e-01 -9.55776989e-01 -1.01550961e+00 4.13386852e-01
-4.32859749e-01 -1.94123626e-01 6.76115572e-01 6.77490592e-01
-7.86087334e-01 3.11390758e-01 -1.05791800e-01 -3.61352682e-01
-1.01845956e+00 4.10270989e-01 8.95612001e-01 3.58497351e-01
-1.13120461e+00 9.07066345e-01 3.70860696e-01 -2.06660673e-01
-2.15364099e-01 -1.51387841e-01 -1.08272806e-01 -1.83559060e-01
9.60840583e-01 4.18182790e-01 7.90758371e-01 -3.64534050e-01
1.97898038e-02 1.30301058e-01 -1.69932507e-02 -9.14829224e-02
1.61746049e+00 1.27225876e-01 -1.79602236e-01 8.31173956e-02
1.12992406e+00 2.09980458e-01 -1.01811433e+00 -3.98235470e-02
-5.27424455e-01 -4.58649278e-01 1.00338984e+00 -1.03873849e+00
-1.23316157e+00 6.17965102e-01 8.84508967e-01 6.97580159e-01
1.55227327e+00 -9.82041210e-02 8.41897130e-01 8.31182003e-01
-2.59623706e-01 -8.90324950e-01 -5.03013022e-02 6.33808136e-01
1.42899442e+00 -5.45491934e-01 -6.19022399e-02 8.43841583e-02
-1.60382748e-01 1.28424704e+00 -2.13756692e-02 -4.83595312e-01
6.55052304e-01 4.96897757e-01 2.72257984e-01 -2.12191194e-01
-7.11483359e-01 2.77740777e-01 5.18244952e-02 3.25038373e-01
2.47306094e-01 -2.08748817e-01 -1.50714651e-01 1.44341379e-01
-5.74805379e-01 -2.30776802e-01 1.05188787e+00 6.71924591e-01
-6.75382972e-01 -3.87429208e-01 -1.08507740e+00 5.20429730e-01
-8.87378514e-01 -1.44275308e-01 -7.09226448e-03 2.73085028e-01
-2.64417797e-01 9.43162799e-01 3.33793193e-01 1.63995460e-01
2.57323593e-01 -1.27737492e-01 2.00355321e-01 1.10627338e-02
-4.35863912e-01 3.73957336e-01 1.62585199e-01 -2.74388045e-01
-4.31586117e-01 -4.07778710e-01 -1.02046251e+00 -4.51163590e-01
-1.97532639e-01 4.26722109e-01 8.70998502e-01 6.40169740e-01
2.52392173e-01 1.14957130e+00 8.99881780e-01 -8.18972111e-01
-3.28950882e-01 -1.21532047e+00 -8.83637965e-01 9.18854326e-02
9.53089297e-01 -4.90611374e-01 -1.25090945e+00 -3.72022539e-02] | [11.23192310333252, -2.540212869644165] |
968edd1a-d022-455c-9ac4-bb09b9185a0c | detecting-and-accommodating-novel-types-and | 2211.04555 | null | https://arxiv.org/abs/2211.04555v1 | https://arxiv.org/pdf/2211.04555v1.pdf | Detecting and Accommodating Novel Types and Concepts in an Embodied Simulation Environment | In this paper, we present methods for two types of metacognitive tasks in an AI system: rapidly expanding a neural classification model to accommodate a new category of object, and recognizing when a novel object type is observed instead of misclassifying the observation as a known class. Our methods take numerical data drawn from an embodied simulation environment, which describes the motion and properties of objects when interacted with, and we demonstrate that this type of representation is important for the success of novel type detection. We present a suite of experiments in rapidly accommodating the introduction of new categories and concepts and in novel type detection, and an architecture to integrate the two in an interactive system. | ['Nikhil Krishnaswamy', 'Sadaf Ghaffari'] | 2022-11-08 | null | null | null | null | ['type'] | ['speech'] | [ 2.73812771e-01 4.40841280e-02 2.49148920e-01 -2.52671063e-01
1.38469249e-01 -8.73202980e-01 7.16542542e-01 6.64887905e-01
-6.04110658e-01 5.64435065e-01 -1.34756729e-01 -3.34429950e-01
-1.54222578e-01 -8.00295651e-01 -6.05773449e-01 -2.55048394e-01
-5.56570351e-01 4.98961896e-01 3.29367042e-01 -1.85085356e-01
9.08462405e-01 6.96796417e-01 -2.05465269e+00 5.39680004e-01
8.17751706e-01 6.95914984e-01 4.40924615e-01 8.31700325e-01
1.60263777e-01 9.24451709e-01 -7.95554280e-01 2.92409152e-01
3.89347002e-02 -5.65988719e-01 -8.79793227e-01 -6.16064370e-02
3.66999179e-01 -3.32133919e-01 -2.17096861e-02 9.10356283e-01
1.12964153e-01 7.79187262e-01 1.09109700e+00 -1.44458508e+00
-8.69750500e-01 3.01103473e-01 3.46601188e-01 6.39930189e-01
1.00951600e+00 2.45390400e-01 2.28710696e-01 -8.26716304e-01
4.27194536e-01 1.28174996e+00 6.23661518e-01 7.09688067e-01
-1.25150585e+00 -5.68851829e-01 3.56344670e-01 3.81891161e-01
-1.31610131e+00 -6.22440815e-01 3.78704071e-01 -6.74763680e-01
1.21178579e+00 3.57721627e-01 1.37571633e+00 9.70965743e-01
3.85886222e-01 6.54280663e-01 1.26285827e+00 -4.76842850e-01
9.68227446e-01 4.26692277e-01 4.38439667e-01 4.59839344e-01
3.72534841e-01 6.80078566e-01 -6.00693822e-01 -1.73555598e-01
1.08731496e+00 -2.85045624e-01 -1.29629940e-01 -4.16570693e-01
-1.10955346e+00 1.38800144e-01 2.51718491e-01 4.95653570e-01
-4.14197445e-01 3.01432401e-01 -5.44804670e-02 4.15329039e-01
9.36647654e-02 1.06383121e+00 -1.47097558e-01 -3.97139788e-01
-4.84919578e-01 4.33035254e-01 8.82007360e-01 9.01943803e-01
2.05868542e-01 2.10043356e-01 -3.29652280e-02 3.86155546e-01
1.56239241e-01 2.23728150e-01 9.33013260e-01 -1.24393487e+00
-3.56599689e-01 6.32588685e-01 3.85210842e-01 -8.07476997e-01
-6.53705537e-01 -3.46519083e-01 6.40326831e-03 6.61173344e-01
4.44272041e-01 1.45647541e-01 -9.83628154e-01 1.81808364e+00
7.47615173e-02 2.87380606e-01 3.80815327e-01 6.63212776e-01
5.91275394e-01 5.34722984e-01 3.60039502e-01 -2.50040114e-01
1.19811857e+00 -5.70888638e-01 -6.75759554e-01 -5.34198403e-01
7.10387290e-01 3.34536433e-02 1.08618438e+00 7.01090574e-01
-1.39737606e+00 -6.59115672e-01 -1.24902630e+00 -9.25968960e-03
-8.87766838e-01 -4.69550520e-01 9.02120709e-01 6.42051339e-01
-1.24655497e+00 8.07422101e-01 -8.67548347e-01 -6.01335168e-01
3.05072039e-01 2.46788487e-01 -8.54885206e-02 2.97813505e-01
-8.99648130e-01 1.46613610e+00 5.82756519e-01 -9.57582071e-02
-1.09805322e+00 -4.02614176e-01 -7.60727525e-01 3.69062781e-01
6.64304420e-02 -7.39979625e-01 1.67561388e+00 -1.17031205e+00
-1.35736895e+00 8.02381575e-01 -9.49585512e-02 -1.05880432e-01
1.09931510e-02 4.52609621e-02 -5.63934684e-01 1.20774500e-01
-1.29835919e-01 6.57516420e-01 5.39783359e-01 -1.42619443e+00
-6.81586981e-01 -3.41083169e-01 2.50832111e-01 5.44661164e-01
-1.99365616e-01 -1.13454886e-01 4.91363883e-01 -6.76284492e-01
4.09996539e-01 -6.84697688e-01 2.76126545e-02 1.43986687e-01
4.06709760e-01 -3.12452048e-01 3.82407457e-01 -1.41357034e-01
9.42779064e-01 -2.09409809e+00 2.02057704e-01 1.54846281e-01
2.57952362e-01 -6.63173422e-02 -1.79610997e-02 3.29882652e-01
-2.64666259e-01 1.63231105e-01 -3.51836085e-02 1.05416477e-01
6.91180676e-02 1.80901531e-02 -1.33122966e-01 1.75673999e-02
6.03709416e-03 7.07329273e-01 -1.30823374e+00 -1.23715676e-01
-1.34204835e-01 -1.05872385e-01 -5.79393923e-01 2.38593131e-01
-2.06933483e-01 1.36476293e-01 -1.92838013e-01 5.83088040e-01
2.59357274e-01 -1.99941527e-02 7.72743598e-02 3.28715056e-01
-1.87347725e-01 4.40440238e-01 -1.33173943e+00 1.34290481e+00
-7.63388053e-02 5.69621921e-01 -2.00218067e-01 -8.27156126e-01
6.36809289e-01 3.66565704e-01 -2.65719831e-01 -6.76253378e-01
2.69748747e-01 5.69031090e-02 5.39140224e-01 -6.44558907e-01
7.80295551e-01 -1.30390659e-01 4.90346476e-02 8.96452069e-01
1.17020674e-01 -6.29345119e-01 1.93506241e-01 3.26275736e-01
9.92540658e-01 9.51831117e-02 5.56199014e-01 -4.99441385e-01
-2.62829870e-01 2.80777633e-01 3.22476000e-01 1.43295932e+00
-3.63356739e-01 1.51264712e-01 2.06511810e-01 -3.39074463e-01
-7.39584029e-01 -1.36795092e+00 -2.75018662e-01 1.17809510e+00
2.24488750e-01 -9.49272960e-02 -6.18124723e-01 -1.82617903e-01
2.63080769e-03 1.24682128e+00 -9.31588054e-01 -8.06847930e-01
-1.33540154e-01 -6.41916037e-01 3.18506807e-01 8.91357422e-01
3.00381929e-01 -1.44933617e+00 -1.38080847e+00 2.31751904e-01
2.12830767e-01 -4.95224565e-01 4.32596058e-02 5.40969253e-01
-1.14410567e+00 -1.15114450e+00 1.06461588e-02 -1.14975357e+00
9.22481179e-01 2.11622387e-01 1.01863289e+00 5.34000278e-01
-1.01451963e-01 8.88480186e-01 -3.00656080e-01 -5.29903412e-01
-5.27503967e-01 -6.05144024e-01 4.78321493e-01 -5.95314264e-01
2.69014239e-01 -5.65292954e-01 -3.60947728e-01 8.05184692e-02
-8.55666101e-01 3.63266408e-01 2.94582367e-01 6.10843182e-01
-8.48569870e-02 2.14917660e-02 6.76628351e-01 -5.87358594e-01
1.04304218e+00 -8.18353236e-01 -9.34391171e-02 3.30114841e-01
-5.98915040e-01 -2.32496604e-01 3.94652456e-01 -1.01802206e+00
-9.74269152e-01 -1.50713667e-01 3.77557933e-01 2.21693337e-01
-4.57902730e-01 7.82084703e-01 -4.88786697e-02 -1.33544654e-01
9.22695100e-01 4.36023712e-01 -3.13769169e-02 -9.16683674e-02
2.48622224e-01 4.88800973e-01 7.67765224e-01 -8.09582829e-01
1.90542504e-01 3.88382934e-02 -2.72524983e-01 -5.46039760e-01
-2.23537639e-01 -4.78873681e-03 -6.45080447e-01 -4.95999008e-01
3.43752772e-01 -6.17715955e-01 -1.00233507e+00 7.84658790e-01
-8.71073902e-01 -8.49511921e-01 -4.54525173e-01 5.96987724e-01
-7.99246073e-01 -3.23942490e-02 -3.87241036e-01 -9.31843877e-01
2.76704907e-01 -9.64286566e-01 3.23640764e-01 5.08363843e-01
-5.41098595e-01 -8.74744654e-01 -4.77896370e-02 -4.29542184e-01
5.29528141e-01 -2.25179225e-01 1.19366050e+00 -1.06427050e+00
-1.93762496e-01 -1.94239795e-01 1.00873843e-01 -2.68698990e-01
-3.66499871e-02 -2.92281747e-01 -8.66350412e-01 -2.41047591e-01
3.06889921e-01 -3.94448757e-01 5.03154039e-01 -6.37004972e-02
1.21995568e+00 -2.17313230e-01 -5.85883975e-01 2.80254155e-01
8.74645948e-01 1.20502782e+00 5.73206365e-01 5.15912712e-01
8.40611290e-03 3.50046128e-01 5.19876659e-01 3.45403254e-01
4.74208087e-01 5.48955500e-01 1.26464292e-01 3.89170438e-01
3.44504595e-01 -3.04643344e-02 1.81935281e-01 3.79596710e-01
-7.50512183e-02 -3.30281883e-01 -9.79638517e-01 5.19170165e-01
-1.73103011e+00 -1.22087729e+00 3.99105132e-01 2.18893480e+00
1.01407778e+00 2.64999896e-01 -3.74046229e-02 8.02273452e-02
5.27450383e-01 -5.69743276e-01 -9.41181839e-01 -6.61375582e-01
6.80748895e-02 2.83051759e-01 -2.51161516e-01 4.70350325e-01
-6.56803012e-01 9.95298207e-01 8.24345779e+00 2.37779737e-01
-6.35206223e-01 -1.65268123e-01 2.96187669e-01 -3.95139828e-02
-5.89464977e-02 1.52578857e-02 -3.71669978e-01 2.47905582e-01
9.90789473e-01 -7.89446354e-01 7.04990745e-01 6.64036632e-01
-4.15991135e-02 -8.91771495e-01 -1.81652653e+00 6.53323114e-01
3.71725500e-01 -1.05547428e+00 2.12244973e-01 -3.55472565e-01
1.79937109e-01 -7.76888669e-01 3.12894523e-01 6.29189014e-01
5.19291043e-01 -1.16578436e+00 1.10602570e+00 9.20599043e-01
2.67353207e-01 -2.54325867e-01 -9.22006369e-03 5.21700084e-01
-6.38725638e-01 -3.24465603e-01 -9.91320163e-02 -7.64519870e-01
-4.02040780e-01 -4.04753894e-01 -7.51428187e-01 -3.55892003e-01
7.58337677e-01 5.33400536e-01 -8.63547146e-01 1.51012349e+00
-4.03881967e-02 3.64400685e-01 -3.77973020e-01 -4.44245547e-01
-2.80172050e-01 2.52718270e-01 6.95513070e-01 9.08800662e-01
3.36811543e-01 9.73292708e-01 2.21294742e-02 1.12066960e+00
3.60356599e-01 -4.05837655e-01 -6.05459332e-01 1.18704597e-02
1.02103817e+00 8.43856812e-01 -1.04599833e+00 -9.46153402e-01
2.05845401e-01 8.27198505e-01 3.83731961e-01 4.16000009e-01
-3.92844468e-01 -2.45964110e-01 3.27273399e-01 3.22706476e-02
-1.01520061e-01 -4.20144260e-01 -4.61837023e-01 -1.09977782e+00
-3.46916705e-01 -9.30866361e-01 2.63332993e-01 -1.52985847e+00
-9.27338481e-01 3.65140945e-01 4.15431499e-01 -9.02579546e-01
-2.58196682e-01 -6.78543091e-01 -6.89296663e-01 5.78632176e-01
-5.02950549e-01 -5.04808843e-01 -6.04206920e-01 3.38588983e-01
3.62606406e-01 -6.97083324e-02 1.23990464e+00 -4.13425535e-01
-3.05167079e-01 4.69073743e-01 -1.34629281e-02 -4.01352912e-01
4.88488734e-01 -1.44492829e+00 3.92359942e-01 5.43457985e-01
-1.84362754e-01 1.13555014e+00 7.96194136e-01 -9.47417438e-01
-1.03868139e+00 -3.82235825e-01 2.97608882e-01 -7.58475542e-01
3.98512214e-01 -5.67665517e-01 -1.27047038e+00 9.58213210e-01
6.57176077e-02 -4.62799251e-01 6.98759377e-01 -6.29551485e-02
-1.48192868e-01 5.25584340e-01 -1.32008314e+00 9.48879480e-01
1.33981121e+00 -4.88089412e-01 -1.39425099e+00 3.01410049e-01
3.57549250e-01 -5.39315403e-01 -5.55097401e-01 1.86218977e-01
9.84423637e-01 -7.51741529e-01 8.35841060e-01 -1.06437457e+00
2.05659002e-01 -3.02769274e-01 4.83188145e-02 -1.87458026e+00
-8.13812971e-01 -1.55080050e-01 -3.48727494e-01 5.77389061e-01
1.74494937e-01 -7.38045216e-01 3.75695497e-01 1.01078820e+00
-2.90567577e-01 -3.89617741e-01 -7.54929304e-01 -7.25029349e-01
3.30852151e-01 -3.61150563e-01 5.49604952e-01 1.00680089e+00
8.75601590e-01 -8.70850682e-02 4.06966686e-01 2.37967998e-01
3.45932752e-01 -2.00431734e-01 2.57233888e-01 -1.40294170e+00
-2.10909799e-01 -7.48322010e-01 -6.95681989e-01 -9.88241017e-01
-1.02456935e-01 -8.25592458e-01 1.72696754e-01 -1.22800291e+00
3.84978890e-01 -5.55655479e-01 -3.87956202e-01 5.44441402e-01
-1.05859600e-01 -1.58917182e-03 2.03726918e-01 1.51795685e-01
-5.20229697e-01 3.42442930e-01 7.25476861e-01 7.39476597e-03
-4.62756485e-01 -3.18429708e-01 -9.44278836e-01 9.77608979e-01
6.93482637e-01 -4.12681639e-01 -2.89415151e-01 -3.04431826e-01
4.49052334e-01 1.23372048e-01 9.32112634e-01 -1.50459361e+00
6.22523665e-01 -3.70941162e-01 9.49451208e-01 -3.63597959e-01
4.63203669e-01 -7.78524458e-01 1.72721803e-01 6.42374992e-01
-6.92404628e-01 2.44392052e-01 8.70457411e-01 5.06229818e-01
4.91792381e-01 -5.37240326e-01 5.83766937e-01 -3.10049176e-01
-1.23356140e+00 -4.68804270e-01 -1.32197893e+00 -1.36968821e-01
1.13302255e+00 -9.22896504e-01 -6.69402897e-01 -3.15719366e-01
-1.42375386e+00 1.45477012e-01 6.71536624e-01 5.52985787e-01
8.51573110e-01 -1.10731232e+00 -1.09519929e-01 4.62787390e-01
1.61524251e-01 -4.48798090e-01 -1.49960950e-01 3.25711340e-01
-3.03214788e-01 -4.16922569e-02 -6.78335726e-01 -2.19641104e-01
-9.45213735e-01 5.15324771e-01 7.63680995e-01 7.11646199e-01
-3.52625936e-01 7.61600494e-01 3.47713709e-01 -1.99564621e-01
2.96613932e-01 -4.61053401e-01 -5.62340617e-01 -1.62076458e-01
6.07400656e-01 1.38176203e-01 -6.40098080e-02 -1.25802249e-01
-2.58178711e-01 9.81582180e-02 -1.39067665e-01 -3.53195280e-01
1.07582772e+00 -1.57353744e-01 -3.50436792e-02 1.25077045e+00
2.22992986e-01 -5.19771695e-01 -1.01379347e+00 9.53978151e-02
-1.30865231e-01 -3.96289527e-01 -1.03222728e-01 -1.48655581e+00
-1.11087829e-01 4.63545829e-01 8.20156336e-01 1.33821890e-01
9.13167655e-01 -1.48198515e-01 2.69256439e-02 9.37459469e-01
6.71742439e-01 -1.18349254e+00 4.21104550e-01 5.62678814e-01
1.07601094e+00 -7.41238117e-01 1.39645010e-01 -9.22880620e-02
-3.91014785e-01 1.20198727e+00 1.06654346e+00 -1.27477378e-01
5.68503320e-01 1.67454839e-01 -2.40259215e-01 -4.73093957e-01
-1.11284745e+00 1.20733939e-01 3.84240821e-02 8.58902395e-01
3.27253520e-01 -3.61841475e-03 -1.25061125e-01 8.33395779e-01
-3.92421544e-01 2.04315245e-01 1.16896796e+00 1.57821584e+00
-7.40394473e-01 -2.75691450e-01 -4.07882810e-01 7.21062243e-01
4.53700423e-02 -2.28312984e-01 -4.94326860e-01 7.48353541e-01
1.04629762e-01 9.73772466e-01 5.52482843e-01 -3.56760621e-01
3.31116587e-01 3.29741061e-01 8.90232861e-01 -1.06718373e+00
-6.59628868e-01 -6.48722589e-01 -1.61180928e-01 -3.70665550e-01
-1.75666779e-01 -9.52602267e-01 -1.38431692e+00 -2.27089915e-02
-3.31288785e-01 3.39152306e-01 5.02005517e-01 9.15679693e-01
1.68231845e-01 5.48102319e-01 1.51707143e-01 -9.03561532e-01
-1.97180748e-01 -1.04358315e+00 -6.18202984e-01 3.45465899e-01
2.04541251e-01 -1.25329423e+00 -6.76186562e-01 3.01035047e-01] | [4.378599643707275, 1.272653579711914] |
174f6ebd-bd95-4b75-bd09-386f77cb496c | bottlenecks-club-unifying-information | 2207.04895 | null | https://arxiv.org/abs/2207.04895v1 | https://arxiv.org/pdf/2207.04895v1.pdf | Bottlenecks CLUB: Unifying Information-Theoretic Trade-offs Among Complexity, Leakage, and Utility | Bottleneck problems are an important class of optimization problems that have recently gained increasing attention in the domain of machine learning and information theory. They are widely used in generative models, fair machine learning algorithms, design of privacy-assuring mechanisms, and appear as information-theoretic performance bounds in various multi-user communication problems. In this work, we propose a general family of optimization problems, termed as complexity-leakage-utility bottleneck (CLUB) model, which (i) provides a unified theoretical framework that generalizes most of the state-of-the-art literature for the information-theoretic privacy models, (ii) establishes a new interpretation of the popular generative and discriminative models, (iii) constructs new insights to the generative compression models, and (iv) can be used in the fair generative models. We first formulate the CLUB model as a complexity-constrained privacy-utility optimization problem. We then connect it with the closely related bottleneck problems, namely information bottleneck (IB), privacy funnel (PF), deterministic IB (DIB), conditional entropy bottleneck (CEB), and conditional PF (CPF). We show that the CLUB model generalizes all these problems as well as most other information-theoretic privacy models. Then, we construct the deep variational CLUB (DVCLUB) models by employing neural networks to parameterize variational approximations of the associated information quantities. Building upon these information quantities, we present unified objectives of the supervised and unsupervised DVCLUB models. Leveraging the DVCLUB model in an unsupervised setup, we then connect it with state-of-the-art generative models, such as variational auto-encoders (VAEs), generative adversarial networks (GANs), as well as the Wasserstein GAN (WGAN), Wasserstein auto-encoder (WAE), and adversarial auto-encoder (AAE) models through the optimal transport (OT) problem. We then show that the DVCLUB model can also be used in fair representation learning problems, where the goal is to mitigate the undesired bias during the training phase of a machine learning model. We conduct extensive quantitative experiments on colored-MNIST and CelebA datasets, with a public implementation available, to evaluate and analyze the CLUB model. | ['Slava Voloshynovskiy', 'Deniz Gunduz', 'Flavio P. Calmon', 'Behrooz Razeghi'] | 2022-07-11 | null | null | null | null | ['privacy-preserving-deep-learning', 'privacy-preserving-deep-learning'] | ['methodology', 'natural-language-processing'] | [ 3.25040787e-01 2.61083663e-01 -2.12005779e-01 -1.89990133e-01
-1.06793618e+00 -5.70649266e-01 6.77078605e-01 -6.68248385e-02
-2.58415371e-01 8.89486313e-01 2.64232963e-01 -5.32399714e-01
-5.18298745e-01 -8.91376138e-01 -8.23439837e-01 -1.04369521e+00
-2.28142310e-02 3.50598454e-01 -3.05103779e-01 -1.29509969e-02
-1.13424733e-01 3.36215883e-01 -9.08947766e-01 -1.89852968e-01
7.94367433e-01 1.16927707e+00 -3.50383103e-01 3.86862099e-01
2.40642458e-01 8.11369121e-01 -3.29188466e-01 -9.85276461e-01
5.37173748e-01 -6.76916063e-01 -1.05416965e+00 -1.49623185e-01
-1.22438103e-01 -4.54491526e-01 -6.49273455e-01 1.13797712e+00
5.39395750e-01 1.62822276e-01 7.99965918e-01 -1.56034732e+00
-8.78573000e-01 6.83148623e-01 -4.00003314e-01 -7.55966306e-02
-1.02506310e-01 -1.58136532e-01 1.19602072e+00 -3.83367658e-01
6.18483603e-01 1.20196545e+00 5.13297677e-01 8.73721063e-01
-1.39955711e+00 -6.41303539e-01 -1.66140795e-01 4.49615307e-02
-1.41046965e+00 -4.22581732e-01 5.76681912e-01 -4.28544044e-01
2.05989659e-01 5.54111719e-01 4.92671877e-01 1.28148091e+00
6.41763583e-02 1.15301752e+00 1.16372907e+00 -2.99474597e-01
3.99585009e-01 3.91045153e-01 -8.70436837e-04 6.62965357e-01
1.47822797e-01 3.84243995e-01 -3.64639223e-01 -5.02799392e-01
5.93234539e-01 1.27373293e-01 -6.07723355e-01 -5.64238369e-01
-6.65584624e-01 1.21727586e+00 3.74466151e-01 4.11878675e-02
-1.24708623e-01 3.12755167e-01 2.03114986e-01 4.12202746e-01
5.76861322e-01 1.44469976e-01 1.97200291e-02 1.50260836e-01
-8.40850353e-01 3.64814818e-01 1.12698662e+00 1.16423810e+00
8.98896158e-01 -3.35822767e-03 -7.53287673e-01 7.36124873e-01
4.55750465e-01 3.64188135e-01 3.10461409e-02 -9.69461143e-01
4.55055445e-01 -8.49151835e-02 -3.15888375e-02 -6.08618081e-01
4.77479510e-02 -5.56595504e-01 -1.29867804e+00 -2.20387727e-01
2.28320464e-01 -1.63685054e-01 -3.95419270e-01 2.12538695e+00
1.63629696e-01 -5.28822504e-02 7.53119662e-02 7.03540802e-01
3.28864604e-01 7.45578468e-01 -1.60107195e-01 -5.21882534e-01
1.09304667e+00 -6.60817623e-01 -6.70382202e-01 3.24288845e-01
3.71121705e-01 -3.63410443e-01 8.26498032e-01 1.10992253e-01
-1.17029428e+00 -7.13035837e-02 -9.91840005e-01 -2.76685297e-01
-2.34554574e-01 -4.33452308e-01 7.23183095e-01 1.11438596e+00
-1.35502279e+00 7.35141456e-01 -7.42119312e-01 -2.74393469e-01
8.49456668e-01 2.07545921e-01 -1.50509194e-01 -8.32361206e-02
-1.25976253e+00 4.62394863e-01 1.44272342e-01 -2.20707189e-02
-1.44829714e+00 -8.34287941e-01 -8.03933144e-01 3.25549752e-01
5.72236776e-01 -1.17842269e+00 1.04537332e+00 -8.84205997e-01
-1.64964211e+00 8.12629998e-01 1.97034869e-02 -6.76628828e-01
9.71187651e-01 2.91248714e-03 1.84085686e-02 -1.77111417e-01
-1.73335522e-01 3.67392391e-01 7.99714267e-01 -1.19570410e+00
-2.58580595e-01 -4.53692675e-01 4.70834970e-02 -1.00872681e-01
-4.37845141e-01 -1.69908211e-01 -3.85780185e-01 -7.47892678e-01
-3.90608341e-01 -8.48763168e-01 -2.68055648e-01 2.12851107e-01
-9.22314703e-01 3.03271227e-02 7.91882038e-01 -5.65961957e-01
1.23654509e+00 -2.14829946e+00 4.94885236e-01 4.34105545e-01
3.41551602e-01 2.68277451e-02 -5.78759518e-03 5.17561078e-01
2.33102232e-01 5.18527865e-01 -6.04351699e-01 -7.86704183e-01
4.72686559e-01 4.90724683e-01 -3.67858559e-01 7.09171712e-01
-2.35107958e-01 1.14195430e+00 -5.99582851e-01 -2.39855096e-01
1.58831626e-02 5.09263813e-01 -8.78013790e-01 4.64006007e-01
-1.83458582e-01 5.71329951e-01 -4.86476809e-01 5.25558770e-01
8.99430990e-01 -1.66834489e-01 1.01666506e-02 9.00277272e-02
1.42415017e-01 -5.64763322e-02 -8.19568098e-01 1.57099736e+00
-4.64804560e-01 4.60826129e-01 2.45477915e-01 -1.07639241e+00
5.02383351e-01 2.53511667e-01 3.80872965e-01 -3.08989197e-01
2.64651000e-01 8.65940228e-02 -4.39243704e-01 -2.46851087e-01
3.93186003e-01 -3.61513525e-01 -2.80236870e-01 5.13361514e-01
2.52279609e-01 1.18614726e-01 -1.33297071e-01 3.15485656e-01
9.23579574e-01 -4.08673584e-01 3.98678243e-01 -4.50217366e-01
4.92779255e-01 -8.88576031e-01 5.06604075e-01 1.04834700e+00
-2.03437090e-01 5.82611501e-01 8.37981641e-01 -1.01690046e-01
-1.05269504e+00 -1.25697827e+00 -2.38209188e-01 9.21269000e-01
1.52226791e-01 -5.15864015e-01 -9.98697042e-01 -6.70878947e-01
-1.64712165e-02 7.81220675e-01 -7.76664853e-01 -2.16268331e-01
4.83937301e-02 -1.04033864e+00 7.67094493e-01 1.71313614e-01
7.61721432e-01 -4.25359100e-01 -3.10741719e-02 -1.69546381e-01
-2.66499639e-01 -7.65676677e-01 -6.21309996e-01 1.61143020e-01
-5.11712730e-01 -7.62807488e-01 -8.21668923e-01 -5.48582412e-02
3.80859971e-01 -1.68858603e-01 8.51923943e-01 -1.92459583e-01
-1.82838012e-02 6.57624364e-01 -1.35331631e-01 -3.40995967e-01
-7.04195261e-01 2.12757930e-01 -1.61415339e-01 4.59696710e-01
-5.44771105e-02 -8.10713291e-01 -6.02129519e-01 3.04739177e-01
-1.30707455e+00 -3.87438643e-03 4.40647811e-01 1.11907840e+00
8.89926672e-01 -2.99861163e-01 2.57968217e-01 -1.29065406e+00
8.36614192e-01 -7.76928723e-01 -8.15433919e-01 4.27814156e-01
-9.55022216e-01 5.12916386e-01 7.43115425e-01 7.83298612e-02
-9.26980734e-01 -4.17766809e-01 -4.89098042e-01 -6.22382879e-01
2.97711611e-01 3.25510174e-01 -6.37943864e-01 -2.02773631e-01
3.45559388e-01 3.99896771e-01 -1.12855002e-01 -4.13880676e-01
5.57077408e-01 8.22197974e-01 3.96135032e-01 -8.34436893e-01
8.68663371e-01 5.42753756e-01 2.49877438e-01 -6.62631750e-01
-9.52938080e-01 -1.55360997e-01 -2.70622164e-01 1.35122776e-01
7.82585502e-01 -7.44667947e-01 -8.69837403e-01 3.62859994e-01
-1.06377721e+00 -1.76280826e-01 -6.61561668e-01 8.80191848e-02
-9.02964473e-01 4.44337875e-01 -5.96434414e-01 -1.24054170e+00
-5.22821128e-01 -1.25284123e+00 9.34631646e-01 8.44443366e-02
2.92327553e-01 -1.22860599e+00 -3.38757294e-03 3.42838377e-01
5.61576068e-01 5.54966033e-01 1.08474922e+00 -6.42141283e-01
-9.34699059e-01 -3.94836329e-02 -1.30047679e-01 7.02420890e-01
-3.44514698e-01 -4.47313517e-01 -1.20982206e+00 -5.14082134e-01
2.08602771e-01 -3.12173218e-01 1.18654442e+00 4.16261315e-01
1.72384536e+00 -8.98423135e-01 -1.57705188e-01 1.41813815e+00
1.62721789e+00 -5.47385179e-02 7.50387967e-01 -2.46391073e-01
5.21516323e-01 3.00456107e-01 -1.88964635e-01 7.69648373e-01
4.06110257e-01 6.44726038e-01 5.86200476e-01 5.51452823e-02
3.81025136e-01 -6.56516552e-01 2.93929607e-01 8.05790961e-01
-1.63455024e-01 -5.07887900e-01 -1.66304797e-01 2.82635123e-01
-1.92659783e+00 -9.64297652e-01 3.26538652e-01 2.52301455e+00
7.79995978e-01 -3.70139569e-01 2.32121378e-01 -1.18248172e-01
5.89261651e-01 4.36659992e-01 -5.77523470e-01 -5.38657188e-01
-1.83996543e-01 2.80787915e-01 7.77487218e-01 5.03182888e-01
-1.20862484e+00 5.23997605e-01 5.71419191e+00 1.13673961e+00
-5.47940016e-01 7.27087319e-01 1.17260814e+00 6.70670941e-02
-8.50614905e-01 5.87565117e-02 -6.63378596e-01 5.73790669e-01
9.64483559e-01 -4.62892890e-01 1.09013069e+00 9.68269110e-01
-1.98240593e-01 3.47009838e-01 -1.46394360e+00 1.07107115e+00
-1.25866771e-01 -1.35598445e+00 9.84855890e-02 5.70941985e-01
8.80908310e-01 -2.95616575e-02 4.94889855e-01 2.52876490e-01
5.55679500e-01 -1.09319937e+00 6.52325332e-01 5.30944586e-01
1.07731974e+00 -8.42995167e-01 5.93045473e-01 3.57540786e-01
-8.02708268e-01 -2.28196219e-01 -6.34333968e-01 4.23675746e-01
1.16815090e-01 6.18456841e-01 2.38836352e-02 9.63146865e-01
5.17499387e-01 5.85082591e-01 6.44603604e-03 8.17340076e-01
-2.34885246e-01 5.29372931e-01 -3.24743390e-01 -7.45727792e-02
1.56349614e-01 -3.76105309e-01 7.96980679e-01 9.77430582e-01
3.65765333e-01 -1.71461165e-01 -3.33751798e-01 1.45585084e+00
-7.25697815e-01 2.12251730e-02 -4.59720701e-01 -3.94767039e-02
4.56639171e-01 1.00367665e+00 -6.27660602e-02 1.06053777e-01
-1.67574838e-01 1.18565047e+00 3.46395463e-01 5.32854915e-01
-7.37768352e-01 -2.14104593e-01 8.88095140e-01 -5.26832417e-02
1.06315792e-01 3.39518487e-01 1.44629143e-02 -1.48686182e+00
4.59953696e-02 -6.04914188e-01 6.44461334e-01 -1.55951962e-01
-1.56363165e+00 3.86822730e-01 5.88109717e-02 -1.03852379e+00
-1.60627320e-01 -3.33397895e-01 -5.39131939e-01 9.35351074e-01
-1.75049424e+00 -9.91747141e-01 -1.55714706e-01 9.82312202e-01
-1.10183589e-01 -1.32060692e-01 1.00697649e+00 4.05434370e-01
-6.92204535e-01 1.16649830e+00 8.76861036e-01 5.99549077e-02
2.01553240e-01 -1.32095325e+00 5.76569475e-02 9.45367873e-01
1.26846045e-01 6.34872496e-01 3.40961516e-01 -6.09460063e-02
-1.49320078e+00 -1.39693689e+00 5.28079510e-01 -2.33622208e-01
4.28842127e-01 -7.40156651e-01 -5.53237438e-01 8.49467695e-01
9.18954760e-02 2.20906511e-01 9.33509707e-01 -5.18186465e-02
-3.31260920e-01 -3.03863794e-01 -1.39874327e+00 3.62012655e-01
1.14545882e+00 -6.02615833e-01 1.88480422e-01 5.20553470e-01
8.82432580e-01 -3.68184268e-01 -7.83740044e-01 -8.78392085e-02
5.45748055e-01 -1.20226276e+00 8.77416909e-01 -6.93039536e-01
4.81683612e-01 2.68188745e-01 -4.53324646e-01 -1.29799092e+00
-1.29231676e-01 -1.43927014e+00 -3.89274776e-01 1.43208706e+00
2.83479929e-01 -8.22622359e-01 5.83917141e-01 6.47631347e-01
1.25868842e-01 -9.93671358e-01 -1.31642663e+00 -7.63174176e-01
4.15113777e-01 -4.30318147e-01 8.15508842e-01 7.98898220e-01
-1.97056785e-01 -4.81481804e-03 -8.91803741e-01 -9.48868319e-02
8.66247296e-01 -9.50865149e-02 7.22369194e-01 -1.03264046e+00
-7.34313130e-01 -5.10987997e-01 -3.48845482e-01 -1.30756629e+00
1.92804247e-01 -1.34129536e+00 -2.21665069e-01 -1.07456791e+00
4.75833833e-01 -4.23709244e-01 -4.42005247e-01 2.76130438e-01
4.85294536e-02 -2.72065639e-01 3.87410939e-01 2.68473715e-01
-5.04763961e-01 9.58089411e-01 9.70473826e-01 -2.02290013e-01
7.11889192e-02 5.83132863e-01 -1.08687782e+00 2.94263929e-01
3.86108547e-01 -4.73644972e-01 -5.16826153e-01 -2.51204550e-01
2.12513730e-01 2.55982399e-01 6.27260208e-01 -6.34263754e-01
1.49927720e-01 -1.94439236e-02 4.66708913e-02 -1.07956804e-01
2.26660684e-01 -8.95511627e-01 2.31847122e-01 4.49930400e-01
-5.67246199e-01 -4.75434154e-01 -2.96824038e-01 8.03524494e-01
-2.61285696e-02 -1.76544055e-01 1.01333034e+00 -3.78997102e-02
3.01583335e-02 9.99902427e-01 -9.75672528e-02 4.18426722e-01
8.81139994e-01 2.71807134e-01 -3.88091832e-01 -7.65039802e-01
-5.66014290e-01 2.29238331e-01 1.97680458e-01 -2.62478571e-02
4.92809415e-01 -1.42113101e+00 -7.81093061e-01 2.95559406e-01
1.00668140e-01 -1.21311732e-02 2.86426038e-01 8.61658156e-01
-1.98024601e-01 3.42382014e-01 1.13426127e-01 -3.66906524e-01
-6.65222883e-01 5.87693810e-01 4.46293861e-01 -6.41705573e-01
-2.62596786e-01 1.00683320e+00 5.18251002e-01 -2.62583345e-01
4.19273704e-01 -1.03169857e-02 4.25603390e-01 -3.56357455e-01
4.28124964e-01 4.89326447e-01 -2.87164032e-01 -4.79222417e-01
-8.79758373e-02 -3.15440185e-02 -3.61639559e-02 -1.79987505e-01
1.22168899e+00 -2.34903678e-01 -1.83553457e-01 1.81857571e-01
1.67436934e+00 -1.60443425e-01 -1.22793782e+00 -4.94007915e-01
-2.79226273e-01 -4.74417746e-01 -2.42561605e-02 -5.06559372e-01
-1.30031025e+00 1.15778792e+00 5.17434120e-01 3.92013460e-01
1.29564595e+00 1.07967421e-01 8.50010872e-01 3.24148685e-02
5.34749448e-01 -7.99864233e-01 -3.99395645e-01 1.89648166e-01
7.61119604e-01 -8.37092817e-01 -2.64453828e-01 -3.05069983e-01
-4.57562000e-01 8.45963001e-01 -2.93859616e-02 2.01930955e-01
9.69508350e-01 2.99295872e-01 -6.16612256e-01 2.24991649e-01
-6.15023851e-01 7.27630500e-03 1.58785239e-01 6.06126785e-01
-4.14768755e-02 2.98238575e-01 -3.16425741e-01 1.10442865e+00
-2.35641107e-01 -7.61969388e-02 3.13379556e-01 5.20858109e-01
1.44321486e-01 -1.20668626e+00 7.12295026e-02 4.85629112e-01
-7.79263258e-01 -1.96184129e-01 -3.84180039e-01 5.45325577e-01
7.00050965e-02 6.81089163e-01 -1.88717008e-01 -4.88666564e-01
-3.19104269e-03 -1.36815682e-01 2.71576852e-01 -1.43448487e-01
-4.42938447e-01 3.85385603e-02 -3.19883823e-01 -5.32230258e-01
-3.02569151e-01 -5.88295400e-01 -3.33485663e-01 -7.89741397e-01
-4.07010794e-01 3.01261187e-01 4.73495007e-01 8.99234831e-01
3.06383789e-01 3.65210116e-01 9.66020703e-01 -3.73814970e-01
-1.02222848e+00 -7.60424793e-01 -9.00777519e-01 2.61754215e-01
3.42613459e-01 -2.30782330e-01 -5.12500405e-01 -2.97577202e-01] | [7.234932899475098, 3.997911214828491] |
0b4abf6a-fe7e-49e4-a991-e80d9edfcced | a-light-transformer-for-speech-to-intent | null | null | https://www.researchgate.net/publication/350395426_A_Light_Transformer_For_Speech-To-Intent_Applications | https://ieeexplore.ieee.org/document/9383559 | A Light Transformer For Speech-To-Intent Applications | Spoken language understanding (SLU) systems can make life more agreeable, safer (e.g. in a car) or can increase the independence of physically challenged users. However, due to the many sources of variation in speech, a well-trained system is hard to transfer to other conditions like a different language or to speech impaired users. A remedy is to design a user-taught SLU system that can learn fully from scratch from users’ demonstrations, which in turn requires that the system’s model quickly converges after only a few training samples. In this paper, we propose a light transformer structure by using a simplified relative position encoding with the goal to reduce the model size and improve efficiency. The light transformer works as an alternative speech encoder for an existing user-taught multitask SLU system. Experimental results on three datasets with challenging speech conditions prove our approach outperforms the existed system and other state-of-art models with half of the original model size and training time. | ['Hugo', 'Pu; Van hamme', 'Wang'] | 2021-01-19 | null | null | null | ieee-spoken-language-technology-workshop-slt | ['spoken-language-understanding', 'spoken-language-understanding'] | ['natural-language-processing', 'speech'] | [ 1.23223348e-03 3.54556203e-01 -3.51812430e-02 -5.06242275e-01
-7.49770403e-01 -3.31748664e-01 4.01359111e-01 -2.72725731e-01
-4.06923294e-01 8.32804263e-01 6.82576820e-02 -4.42199469e-01
2.56892323e-01 -5.70396125e-01 -7.99402833e-01 -3.41029406e-01
4.06039596e-01 4.41395044e-01 5.81877351e-01 -4.11985159e-01
-7.20736757e-02 1.83651894e-01 -1.76422930e+00 1.61402717e-01
1.28187835e+00 6.44221246e-01 8.83331358e-01 7.25442410e-01
-2.35650584e-01 7.67937124e-01 -5.84886849e-01 1.68800309e-01
-1.79434195e-01 -5.65246165e-01 -8.41855466e-01 -1.74656361e-02
3.84584904e-01 -7.92403042e-01 -4.83605623e-01 6.73135042e-01
7.25144148e-01 4.22440648e-01 3.91950667e-01 -1.13804567e+00
-6.78428173e-01 3.83021832e-01 -6.34827316e-02 -1.09326079e-01
5.54729521e-01 4.81642112e-02 4.39470887e-01 -5.34653485e-01
2.00742602e-01 1.41795897e+00 3.23461711e-01 1.09920764e+00
-8.81926715e-01 -6.75183535e-01 2.70505011e-01 3.54297936e-01
-1.25134885e+00 -7.15098023e-01 4.49563563e-01 -1.20966703e-01
9.47138011e-01 1.33302480e-01 3.07053268e-01 1.36730468e+00
-2.57123590e-01 1.03095555e+00 1.01750422e+00 -3.96750063e-01
2.31599167e-01 3.17475796e-01 2.18471274e-01 7.85406828e-01
-2.92532563e-01 -1.24666318e-01 -5.66157520e-01 2.23749772e-01
5.80345154e-01 -3.95396389e-02 -5.03379107e-01 -3.91441464e-01
-1.01117194e+00 6.81792915e-01 1.73830584e-01 3.56896400e-01
3.93949859e-02 -6.70955479e-02 3.85393351e-01 4.25844550e-01
2.80867904e-01 -7.75508285e-02 -5.64534485e-01 -6.66677058e-01
-6.87623620e-01 1.08950250e-02 8.61248910e-01 8.80701065e-01
5.36423564e-01 5.61640551e-03 5.19135855e-02 1.28230095e+00
2.88708359e-01 6.19067609e-01 7.50317693e-01 -6.27364516e-01
4.36159521e-01 3.56912434e-01 -5.24697360e-03 -2.09075779e-01
-2.86810845e-01 -2.13043228e-01 -4.20641184e-01 3.16981971e-01
3.48240137e-01 -3.33171427e-01 -9.37935591e-01 1.88238537e+00
3.03787649e-01 4.59442049e-01 1.55605063e-01 8.02725732e-01
7.91572630e-01 9.10681963e-01 -3.69722433e-02 -4.62050848e-02
1.15541601e+00 -1.20760024e+00 -8.75887930e-01 -5.01667857e-01
8.65208209e-01 -4.44682598e-01 1.58640242e+00 5.67221880e-01
-8.13524723e-01 -7.73033738e-01 -1.03776646e+00 -2.27373898e-01
-3.94582480e-01 9.57382172e-02 3.76565367e-01 9.14084792e-01
-1.00262511e+00 5.56255460e-01 -7.81302273e-01 -4.52658713e-01
1.51838884e-01 3.05364281e-01 -3.62123311e-01 -1.20496817e-01
-1.35973597e+00 1.21274769e+00 4.06181216e-01 -5.78429438e-02
-9.11308944e-01 -6.96048975e-01 -1.05079818e+00 2.53673315e-01
6.13335371e-01 -4.62433159e-01 1.58086956e+00 -6.46570385e-01
-2.19586110e+00 3.04261327e-01 -2.38179192e-01 -2.79421359e-01
5.09985089e-01 -5.15222430e-01 -2.75007367e-01 2.26332881e-02
-2.32361823e-01 5.76740921e-01 6.35365188e-01 -1.03912747e+00
-6.10849440e-01 -1.57381341e-01 2.96511561e-01 3.50174248e-01
-4.96490717e-01 -2.24680185e-01 -2.55882621e-01 -2.01018155e-01
-2.56207854e-01 -8.98578405e-01 2.93810405e-02 7.53167346e-02
8.36277381e-02 -5.53470254e-01 1.27560246e+00 -8.44265044e-01
1.18758416e+00 -2.13733983e+00 2.81338453e-01 -1.00642562e-01
-1.51062340e-01 6.98458493e-01 -2.12401420e-01 4.10353422e-01
5.63002750e-02 -1.35227621e-01 -1.46540478e-01 -4.93344754e-01
-5.78884035e-02 6.10674977e-01 6.12507649e-02 1.36895701e-01
-1.22819923e-01 5.19149184e-01 -1.07530308e+00 -3.15385014e-01
5.68182945e-01 5.23197591e-01 -6.71711922e-01 4.73557740e-01
-8.24137926e-02 6.51609242e-01 -3.49978805e-01 1.06261201e-01
4.33202773e-01 5.07791303e-02 -3.27320583e-02 8.41286406e-02
-6.00773357e-02 5.51239371e-01 -1.13142335e+00 2.05305719e+00
-1.16129625e+00 5.75315058e-01 7.92925507e-02 -1.16655993e+00
8.19143474e-01 5.78350306e-01 8.20952654e-02 -8.08786511e-01
9.46360081e-02 2.37206504e-01 1.43959269e-01 -8.26659262e-01
4.38560694e-02 -9.07463208e-02 7.43065774e-02 4.53842014e-01
1.85838789e-01 -1.77748337e-01 -3.05587817e-02 1.45008564e-01
9.35777962e-01 1.54530242e-01 2.08859786e-01 -1.48483053e-01
6.96613431e-01 -7.09482491e-01 1.82893038e-01 6.41902506e-01
-3.21445882e-01 3.39986444e-01 7.41360635e-02 -1.37619883e-01
-7.84091294e-01 -1.09296072e+00 7.62142763e-02 1.18030834e+00
-4.30213548e-02 -3.07361007e-01 -1.12568343e+00 -6.55111432e-01
-4.02371347e-01 1.13313115e+00 -2.14241698e-01 -4.00325418e-01
-4.85531598e-01 5.55731058e-02 4.26763266e-01 5.71998298e-01
6.42457128e-01 -1.05428612e+00 -5.24564922e-01 1.23691387e-01
-5.53782344e-01 -1.21963263e+00 -5.51961005e-01 -7.27917999e-02
-5.29219985e-01 -7.49015927e-01 -7.60868788e-01 -8.79449427e-01
6.23290420e-01 3.41537327e-01 5.24438202e-01 8.88975337e-02
-3.35355587e-02 4.71508920e-01 -5.64729214e-01 -4.91246521e-01
-6.87973857e-01 1.01679958e-01 4.90859985e-01 -8.20146576e-02
4.55118977e-02 -5.57901680e-01 -4.37471241e-01 2.45425433e-01
-7.95510769e-01 2.85732269e-01 4.42958981e-01 9.89486992e-01
7.50460699e-02 -1.53971374e-01 9.30214882e-01 -6.11329734e-01
7.50888407e-01 -1.14220493e-01 -2.11952671e-01 4.69382018e-01
-2.66048014e-01 1.67329162e-01 7.45250046e-01 -7.52951920e-01
-1.23989761e+00 3.64772715e-02 -4.08187449e-01 -3.05472583e-01
-2.83007890e-01 3.31215829e-01 -3.57214391e-01 3.26904394e-02
3.84514272e-01 4.40086544e-01 2.98277915e-01 -5.32009244e-01
4.58291590e-01 1.21941459e+00 4.42695916e-01 -5.15368640e-01
5.00134468e-01 -2.32259259e-02 -4.93028373e-01 -1.46957517e+00
-7.67958462e-01 -3.76026332e-01 -6.42054796e-01 -3.24522316e-01
7.38634109e-01 -7.94529140e-01 -7.59958804e-01 6.29425108e-01
-1.20827746e+00 -7.60241807e-01 1.92401148e-02 5.74757695e-01
-5.35750091e-01 5.65989792e-01 -3.07389200e-01 -1.01829338e+00
-1.57435611e-01 -1.22595000e+00 9.72319961e-01 3.93659860e-01
-3.49384397e-01 -8.68125319e-01 -1.37912463e-02 6.49345994e-01
5.21759927e-01 -2.82961577e-01 6.96045637e-01 -6.06516838e-01
-2.31737450e-01 -5.48522137e-02 8.81062821e-02 8.31479490e-01
4.38950956e-01 -4.71935362e-01 -1.12669146e+00 -4.24548775e-01
1.52754247e-01 -7.70943940e-01 4.30700183e-01 1.08173899e-01
1.28887880e+00 -2.74429709e-01 -2.69334733e-01 1.24912299e-01
6.14054978e-01 4.20804888e-01 7.43592620e-01 -9.95377228e-02
6.19886458e-01 7.02647865e-01 4.98256743e-01 7.90699646e-02
4.74431038e-01 7.98228443e-01 4.68115397e-02 7.75953531e-02
-2.06581607e-01 -2.39861295e-01 6.35852218e-01 1.12894547e+00
7.36685097e-02 -2.01780528e-01 -7.13565826e-01 5.69992542e-01
-1.82010794e+00 -8.44919086e-01 2.01063529e-01 2.28401661e+00
7.02069938e-01 -2.14708801e-02 -1.30368233e-01 1.68977037e-01
3.58560413e-01 -9.10543874e-02 -3.22858185e-01 -4.76587653e-01
4.03339893e-01 2.91363329e-01 1.01371266e-01 8.39491189e-01
-6.89593434e-01 9.04843569e-01 6.23546457e+00 8.79349232e-01
-1.14480460e+00 4.37575847e-01 4.05000955e-01 -1.02500513e-01
5.77981882e-02 -2.87023515e-01 -6.61942661e-01 5.41263461e-01
1.32755017e+00 -1.43688738e-01 4.66790825e-01 8.27563286e-01
5.37980258e-01 -2.54003793e-01 -1.27193713e+00 1.16651821e+00
2.68412530e-01 -7.80087352e-01 -1.18361495e-01 -2.49462202e-01
2.83175915e-01 -8.47998410e-02 -1.18662328e-01 6.40339017e-01
1.04402276e-02 -9.60228145e-01 5.20294785e-01 2.32432738e-01
7.66848922e-01 -7.36422002e-01 4.53308344e-01 9.69433129e-01
-9.95440602e-01 -2.17872113e-02 -2.37444505e-01 -2.15660214e-01
3.65310639e-01 -7.20279515e-02 -1.03103948e+00 3.57770145e-01
6.53033435e-01 1.36546820e-01 -1.83158666e-01 6.85723364e-01
-3.11064750e-01 6.58755720e-01 -5.41494370e-01 -4.78876382e-01
3.02557647e-01 -1.55737698e-01 3.44577581e-01 9.23980236e-01
4.61789846e-01 3.86686921e-01 2.48056978e-01 3.84481996e-01
1.66202411e-01 1.32834017e-01 -8.28235090e-01 1.24670304e-01
4.01781291e-01 8.35672557e-01 -2.95775920e-01 -4.01579887e-01
-6.97385609e-01 1.41255474e+00 4.62484509e-01 2.04117820e-01
-9.15736377e-01 -3.59418184e-01 5.44063330e-01 1.46895200e-01
8.57169181e-02 -4.28765923e-01 1.67225882e-01 -1.09189343e+00
1.06490225e-01 -9.21131551e-01 -1.07332908e-01 -9.71703291e-01
-6.71410263e-01 6.53903425e-01 1.90514866e-02 -1.23787558e+00
-4.52650905e-01 -7.38609254e-01 -5.46298027e-01 8.31262946e-01
-1.56745827e+00 -1.12043786e+00 -2.26873755e-01 5.88209927e-01
1.13635778e+00 -1.85009524e-01 9.80599284e-01 4.98197019e-01
-5.56541026e-01 6.56239748e-01 -8.18524733e-02 -9.06157270e-02
7.19032884e-01 -1.02560258e+00 1.91545591e-01 6.70475304e-01
-4.68368307e-02 5.94136775e-01 7.61823356e-01 -3.86407286e-01
-1.19716513e+00 -6.96202576e-01 8.62879992e-01 -5.15595257e-01
5.24130940e-01 -5.94112635e-01 -1.12170255e+00 5.51299214e-01
3.37179452e-01 -2.42042020e-01 6.42790496e-01 1.62728429e-01
-3.66768055e-02 -1.94969669e-01 -1.03775060e+00 6.77685440e-01
1.07650864e+00 -6.49185479e-01 -6.77494228e-01 4.23753589e-01
7.76663184e-01 -5.78427732e-01 -5.86451828e-01 -1.92745216e-02
6.28518820e-01 -6.57526374e-01 8.10579836e-01 -6.27143919e-01
1.65919006e-01 -1.31591484e-01 4.24639843e-02 -1.74672222e+00
1.12033617e-02 -6.75285220e-01 -1.72105074e-01 1.09274173e+00
3.98137093e-01 -7.31549263e-01 4.33136135e-01 6.57638788e-01
-4.77457464e-01 -7.55551219e-01 -1.15941381e+00 -9.46248889e-01
2.04296689e-03 -5.17134905e-01 2.92995811e-01 6.11029863e-01
3.25887829e-01 7.55486727e-01 -6.43324196e-01 3.36026937e-01
3.04673761e-01 -4.07393426e-01 9.52623069e-01 -1.16378880e+00
-4.20555472e-01 -6.57295510e-02 4.48175408e-02 -1.50328994e+00
4.00921345e-01 -5.74496269e-01 2.97865838e-01 -1.68261778e+00
-1.67139713e-02 -4.25410271e-01 2.50019711e-02 6.30196273e-01
-2.42469653e-01 -3.67560089e-01 1.66889444e-01 -3.62856269e-01
-3.56833249e-01 8.27128053e-01 1.41939068e+00 -1.43992156e-01
-2.83585638e-01 8.55156332e-02 -3.52711439e-01 6.87341452e-01
7.60244608e-01 -1.87464952e-01 -1.13910174e+00 -4.53486353e-01
-3.36621106e-01 8.76053497e-02 1.11859538e-01 -1.10596514e+00
3.13246638e-01 -8.23013783e-02 -1.17877260e-01 -4.28946376e-01
5.71192086e-01 -8.95219803e-01 -2.78661162e-01 6.82350993e-01
-8.84569660e-02 -3.77491206e-01 2.65173525e-01 3.06678087e-01
-7.31898844e-02 -4.44320053e-01 8.91271532e-01 2.04815138e-02
-8.76198173e-01 5.21753691e-02 -5.67699730e-01 -2.27891162e-01
1.09696710e+00 -1.43292159e-01 -2.54373819e-01 -8.48642111e-01
-6.88722849e-01 6.13526881e-01 -3.63854691e-02 8.84863913e-01
7.44614661e-01 -1.11837971e+00 -4.68205214e-01 2.73118764e-01
5.16996607e-02 4.70613204e-02 6.63658857e-01 6.50670826e-01
-3.45963329e-01 6.00612700e-01 -1.25940248e-01 -4.47170705e-01
-1.42309666e+00 2.91703045e-01 4.86518145e-01 -6.00001514e-02
-7.49544382e-01 6.82310462e-01 3.11321884e-01 -8.94954264e-01
4.52482253e-01 -4.85299081e-01 -1.19974047e-01 -2.09808603e-01
8.68668318e-01 3.39942902e-01 -1.41446749e-02 -4.88884240e-01
-2.19125450e-01 3.53113174e-01 -2.20247284e-01 -1.82740062e-01
1.11841404e+00 -2.72031605e-01 3.82168978e-01 6.62959874e-01
1.04719150e+00 -2.20372438e-01 -1.32637119e+00 -1.00566119e-01
-4.48831439e-01 -2.62916446e-01 6.31190985e-02 -7.17747211e-01
-7.19137967e-01 1.05783904e+00 8.01010072e-01 6.49761185e-02
8.04093957e-01 -4.66859005e-02 1.07250869e+00 8.44362676e-01
6.71346545e-01 -1.27809453e+00 4.20064420e-01 5.87049961e-01
9.81038392e-01 -1.26348579e+00 -5.14941812e-01 -5.12113571e-01
-7.28704989e-01 9.35896993e-01 8.57451141e-01 2.20976561e-01
5.97744346e-01 1.77592233e-01 2.23584622e-01 2.58539200e-01
-4.77319956e-01 -1.47221074e-01 1.03696890e-01 8.35929155e-01
4.82290298e-01 -1.68625657e-02 -3.26812029e-01 7.31065691e-01
-2.14098141e-01 1.90846443e-01 6.05684996e-01 8.25332165e-01
-6.65174782e-01 -1.40184522e+00 -1.98496252e-01 2.36998573e-01
6.33057626e-03 1.38189390e-01 4.67706136e-02 7.20931172e-01
5.23339175e-02 1.10331666e+00 -1.30751371e-01 -2.28267342e-01
6.37512386e-01 4.78033602e-01 6.16432548e-01 -8.06790471e-01
-9.32294577e-02 -2.30615973e-01 2.85611987e-01 -5.82722485e-01
-2.81899750e-01 -6.42102063e-01 -1.42418051e+00 -3.09721120e-02
-2.69734561e-01 6.68153614e-02 6.06201172e-01 1.27923143e+00
2.22790956e-01 8.92274559e-01 6.01128638e-01 -8.04807723e-01
-6.62170708e-01 -1.31823087e+00 -3.29864085e-01 1.89953864e-01
3.67255360e-01 -8.82736146e-01 -4.81580608e-02 4.21662256e-02] | [14.258269309997559, 6.813375949859619] |
0e4581b6-36b8-4b28-8b8a-f74f355254a5 | voint-cloud-multi-view-point-cloud | 2111.15363 | null | https://arxiv.org/abs/2111.15363v2 | https://arxiv.org/pdf/2111.15363v2.pdf | Voint Cloud: Multi-View Point Cloud Representation for 3D Understanding | Multi-view projection methods have demonstrated promising performance on 3D understanding tasks like 3D classification and segmentation. However, it remains unclear how to combine such multi-view methods with the widely available 3D point clouds. Previous methods use unlearned heuristics to combine features at the point level. To this end, we introduce the concept of the multi-view point cloud (Voint cloud), representing each 3D point as a set of features extracted from several view-points. This novel 3D Voint cloud representation combines the compactness of 3D point cloud representation with the natural view-awareness of multi-view representation. Naturally, we can equip this new representation with convolutional and pooling operations. We deploy a Voint neural network (VointNet) to learn representations in the Voint space. Our novel representation achieves \sota performance on 3D classification, shape retrieval, and robust 3D part segmentation on standard benchmarks ( ScanObjectNN, ShapeNet Core55, and ShapeNet Parts). | ['Bernard Ghanem', 'Silvio Giancola', 'Abdullah Hamdi'] | 2021-11-30 | null | null | null | null | ['3d-classification', '3d-part-segmentation'] | ['computer-vision', 'computer-vision'] | [-4.66316342e-01 -2.81231515e-02 -1.08338639e-01 -6.05202377e-01
-8.15593302e-01 -1.00548935e+00 7.10813940e-01 9.98593122e-03
1.64700329e-01 -3.10653567e-01 2.57244438e-01 -1.10815905e-01
-9.56928765e-04 -8.61039460e-01 -9.68414009e-01 -3.72303337e-01
1.96373999e-01 7.83260524e-01 4.17869717e-01 -1.81607857e-01
2.03560188e-01 1.17958713e+00 -1.59764934e+00 5.29467940e-01
2.90961325e-01 1.25149584e+00 5.16781695e-02 4.78205174e-01
-3.94051999e-01 -1.94470938e-02 -8.32917318e-02 -4.30633456e-01
7.50493824e-01 5.90386271e-01 -7.45601714e-01 4.59385693e-01
9.57161725e-01 -5.35305977e-01 -1.10659450e-01 7.38371611e-01
5.66248059e-01 -1.42005771e-01 5.16498029e-01 -1.20770037e+00
-8.36900830e-01 2.81334966e-01 -8.64977181e-01 1.54094279e-01
4.92739975e-01 1.01459911e-02 1.24316227e+00 -1.54457760e+00
9.53216553e-01 1.52340865e+00 7.87424684e-01 3.40415627e-01
-1.01929140e+00 -2.40546316e-01 4.43304837e-01 -9.00152773e-02
-1.20457101e+00 -1.08205751e-01 8.88761401e-01 -4.20313686e-01
1.45488226e+00 2.40736037e-01 9.84743834e-01 7.13474214e-01
1.77127197e-01 1.13434649e+00 9.88779306e-01 -1.20945666e-02
-4.62997565e-03 -1.22089922e-01 1.41979039e-01 7.69436896e-01
1.25913918e-01 3.01810331e-03 -5.38937449e-01 -1.73408717e-01
9.09384906e-01 4.19905186e-01 -3.44881676e-02 -1.27292728e+00
-1.04259598e+00 5.99014580e-01 7.53505409e-01 -7.37006813e-02
-3.36957991e-01 2.68208593e-01 3.31145078e-01 1.45320863e-01
8.90032887e-01 1.61671907e-01 -8.51526797e-01 1.82558075e-01
-6.12508893e-01 2.17986152e-01 5.91965973e-01 1.28901780e+00
9.66302037e-01 -2.29509309e-01 4.10609961e-01 7.53373444e-01
3.44982743e-01 7.80909359e-01 -5.52708060e-02 -8.93757641e-01
6.23036146e-01 1.19831932e+00 -3.15084875e-01 -7.48633564e-01
-5.74013412e-01 -1.38927013e-01 -4.75436062e-01 4.29272830e-01
2.33778656e-02 3.51034075e-01 -1.21786785e+00 1.16680622e+00
7.09559262e-01 -1.57356903e-01 -2.24574387e-01 8.87985289e-01
1.24061263e+00 2.50988096e-01 -5.40236950e-01 6.07520640e-01
1.27218008e+00 -8.47673595e-01 1.05315447e-01 -8.21170360e-02
3.57731044e-01 -7.34250665e-01 8.06297719e-01 2.22266436e-01
-1.45549726e+00 -5.34411252e-01 -9.52462912e-01 -4.99569535e-01
-6.95468843e-01 -1.86836913e-01 9.13041115e-01 4.54053640e-01
-1.14925933e+00 4.41487491e-01 -9.34341431e-01 -2.75522798e-01
7.97962189e-01 4.72998321e-01 -8.02471340e-01 -3.63851756e-01
-4.31541771e-01 7.84117222e-01 4.70866412e-02 -9.93657624e-04
-7.79282689e-01 -1.02975702e+00 -1.05033553e+00 6.56188726e-02
3.46010625e-01 -1.31106877e+00 1.02086723e+00 -3.11728269e-01
-1.19907320e+00 1.39894581e+00 -5.61170205e-02 1.67733543e-02
3.21026593e-01 -4.69295561e-01 1.92877114e-01 4.61957574e-01
1.02429040e-01 7.68483758e-01 8.02984059e-01 -1.67344272e+00
-3.39814991e-01 -1.13965666e+00 3.74048024e-01 5.35954118e-01
3.82521689e-01 -2.80213565e-01 -6.71114326e-01 -1.93631560e-01
1.04973292e+00 -1.01072824e+00 -2.28485659e-01 3.14034194e-01
-3.62453729e-01 -4.02862251e-01 9.74482059e-01 -1.50487214e-01
1.15499884e-01 -2.17637420e+00 2.64052153e-01 1.87385798e-01
5.12735665e-01 -3.30334976e-02 -3.60032953e-02 3.51019979e-01
-1.82956070e-01 1.72173545e-01 2.32241809e-01 -4.25618857e-01
-5.74712083e-03 2.81071156e-01 -2.32566968e-01 5.31864941e-01
1.71336964e-01 1.33154452e+00 -6.50169909e-01 -1.90833300e-01
5.32644451e-01 5.16599894e-01 -8.84771883e-01 -1.18240695e-02
-3.00366133e-01 1.19993664e-01 -5.57408214e-01 1.02058077e+00
1.09035909e+00 -5.04499614e-01 -3.10136497e-01 -3.18527818e-01
-4.00449941e-03 4.83904816e-02 -1.02640736e+00 2.30749488e+00
-6.43747926e-01 1.22477226e-01 1.63880333e-01 -5.31026959e-01
9.55747724e-01 1.63063124e-01 9.20932114e-01 -1.65618822e-01
-6.43525794e-02 2.52884060e-01 -4.88743961e-01 -2.37792671e-01
4.47961390e-01 1.82176922e-02 -1.09344046e-03 5.17472982e-01
3.87558430e-01 -8.56970847e-01 -5.57296693e-01 2.16813192e-01
7.07778931e-01 3.65118474e-01 2.54614502e-01 -2.88114399e-02
3.11259598e-01 -1.69095382e-01 4.62888516e-02 3.65862250e-01
8.02982301e-02 9.15246367e-01 3.28347415e-01 -9.07514155e-01
-1.00742829e+00 -1.51201522e+00 -1.31199628e-01 9.44033921e-01
2.78318167e-01 -3.69154394e-01 -2.34399378e-01 -9.64259684e-01
5.44321418e-01 1.52684093e-01 -5.07050216e-01 1.22072943e-01
-6.81748867e-01 -1.15220383e-01 2.03742787e-01 8.82160008e-01
3.80027890e-01 -5.69262803e-01 -7.52637744e-01 -1.75852090e-01
3.32601398e-01 -1.31828392e+00 -4.95841950e-01 3.29106033e-01
-1.33144414e+00 -1.19537652e+00 -5.40889621e-01 -5.44992149e-01
6.31802619e-01 9.08208370e-01 1.47561085e+00 -1.49988115e-01
-6.35446534e-02 1.01752150e+00 -3.42435896e-01 -5.65941989e-01
1.30692959e-01 1.09123074e-01 -5.55402339e-02 -2.78676480e-01
5.39762437e-01 -1.02374113e+00 -5.16143322e-01 1.83246061e-01
-7.53871024e-01 3.64631563e-02 4.44870442e-01 5.73607385e-01
1.18798709e+00 -6.89901292e-01 7.68426713e-03 -7.97272325e-01
3.07303127e-02 -2.52538949e-01 -6.43123031e-01 2.30427578e-01
-1.95730150e-01 -4.62669395e-02 2.73592114e-01 -5.46062775e-02
-6.45876288e-01 3.41050923e-01 -2.82151669e-01 -1.23163354e+00
-3.81656587e-01 1.94659084e-01 -6.17237687e-01 -2.50465780e-01
3.04058343e-01 3.74786496e-01 -5.08204587e-02 -7.44160593e-01
1.07717574e+00 3.88505757e-01 1.34478301e-01 -5.72073162e-01
7.48394310e-01 1.10981858e+00 1.04107693e-01 -6.82061136e-01
-1.06265867e+00 -7.20227599e-01 -1.31266046e+00 -1.09856665e-01
1.00392103e+00 -1.14245558e+00 -6.96164906e-01 2.66457349e-01
-1.29570508e+00 2.19152361e-01 -5.23183107e-01 3.21325153e-01
-1.00527620e+00 2.86352664e-01 -2.74919778e-01 -3.87615949e-01
-4.96942908e-01 -1.23392653e+00 1.89900339e+00 9.74131562e-03
2.49025464e-01 -8.86229217e-01 -1.00135311e-01 4.56651807e-01
-1.44991251e-02 3.67510498e-01 1.00941277e+00 -7.15680659e-01
-1.10503221e+00 -2.49764115e-01 -3.33217353e-01 1.72079414e-01
-1.87403373e-02 -2.15148851e-01 -1.27119184e+00 -2.64207006e-01
1.33229032e-01 -2.89181381e-01 9.71090555e-01 6.14963412e-01
1.40904665e+00 2.72174060e-01 -4.48782474e-01 1.29058683e+00
1.51022148e+00 -2.56241232e-01 1.18258037e-01 -5.34578860e-02
9.95065093e-01 2.77549475e-01 5.35291322e-02 3.32421243e-01
6.23228550e-01 6.95829511e-01 9.33746159e-01 -8.17059651e-02
-1.33969188e-01 -4.17869955e-01 -1.79527134e-01 9.96483088e-01
-2.16605723e-01 5.92197254e-02 -1.01955128e+00 4.43115383e-01
-1.45032012e+00 -8.55601847e-01 -3.49719040e-02 1.74216890e+00
-2.88366731e-02 1.89912483e-01 6.46817982e-02 -1.83131382e-01
3.37011755e-01 4.15309846e-01 -8.35032523e-01 -4.60238606e-01
-8.08038190e-02 9.70387384e-02 4.32096481e-01 1.65392667e-01
-1.24716091e+00 8.18270683e-01 5.76718473e+00 3.86550933e-01
-1.10313344e+00 3.92138250e-02 2.39373714e-01 -1.17148980e-01
-6.89339876e-01 -1.64908066e-01 -1.01991820e+00 -1.85506552e-01
2.52539307e-01 1.63206071e-01 1.55552298e-01 1.18296397e+00
-3.05964947e-01 2.78095812e-01 -1.40858221e+00 1.39071012e+00
3.50474089e-01 -1.58901274e+00 4.18315351e-01 3.72656763e-01
7.27673471e-01 7.58031249e-01 2.16979071e-01 1.26017421e-01
1.97943404e-01 -8.18660021e-01 6.72851741e-01 4.64202285e-01
7.45071173e-01 -6.40819132e-01 3.26329947e-01 4.31098759e-01
-1.48942757e+00 1.12553924e-01 -4.85403836e-01 4.27944660e-01
1.66138634e-01 4.46814239e-01 -8.41789901e-01 1.12467694e+00
7.94998109e-01 1.10352123e+00 -6.79937303e-01 9.88899589e-01
1.51501656e-01 -2.17371762e-01 -5.62143683e-01 3.45581651e-01
4.26901460e-01 -8.94353539e-02 8.26374650e-01 6.54360473e-01
2.59117126e-01 1.18992619e-01 3.54361475e-01 8.84220660e-01
-1.21559352e-01 -8.47088322e-02 -1.23237181e+00 3.16668123e-01
1.23641789e-01 1.26541328e+00 -8.84981513e-01 -2.66492933e-01
-7.92541683e-01 6.65274799e-01 4.26414609e-01 1.06331138e-02
-3.92188907e-01 1.44854501e-01 7.80987978e-01 1.26770660e-01
7.86794603e-01 -5.55763781e-01 -6.07359648e-01 -1.46673763e+00
1.22920893e-01 -4.81221914e-01 3.98168325e-01 -1.07184541e+00
-1.63294268e+00 7.27725446e-01 2.93913007e-01 -1.50019002e+00
-1.18608112e-02 -8.84055138e-01 -2.29053542e-01 7.45436370e-01
-1.52164376e+00 -1.70705986e+00 -2.42676094e-01 6.37899458e-01
8.32747281e-01 -4.77752648e-02 8.06789577e-01 -8.34458135e-03
1.07246518e-01 1.42465875e-01 -2.02994928e-01 2.36762874e-02
2.42814556e-01 -1.59795809e+00 8.89167070e-01 2.53256828e-01
6.24664009e-01 6.01558506e-01 -1.08070701e-01 -4.64877337e-01
-1.90076911e+00 -9.60375011e-01 4.13396686e-01 -1.23477852e+00
1.56018078e-01 -5.99910378e-01 -7.81398833e-01 8.33461881e-01
-4.49192151e-02 7.76109517e-01 6.60201490e-01 2.62570471e-01
-8.47365677e-01 7.81405345e-02 -1.08160746e+00 2.89842933e-01
1.51235557e+00 -7.40960956e-01 -8.62014711e-01 3.43496919e-01
8.76118600e-01 -9.35186505e-01 -1.10914469e+00 5.25593579e-01
6.77722335e-01 -1.08515549e+00 1.62964523e+00 -7.77118862e-01
4.13875103e-01 -7.84127340e-02 -5.97983301e-01 -1.34967196e+00
-8.62976387e-02 -1.57968402e-01 -2.64962703e-01 7.00903594e-01
1.17298059e-01 -2.86138624e-01 1.07841825e+00 4.09322798e-01
-6.00716591e-01 -1.30669701e+00 -1.07788992e+00 -7.14817882e-01
3.66609782e-01 -5.70078611e-01 1.07652283e+00 6.38073385e-01
-4.50086355e-01 3.64070415e-01 2.63467699e-01 3.75227988e-01
5.20356059e-01 9.62988734e-01 1.00591981e+00 -1.49719405e+00
-7.75353685e-02 -3.97846162e-01 -6.75215721e-01 -1.33572280e+00
8.24606493e-02 -1.53815544e+00 -4.53527749e-01 -1.72347677e+00
1.49174929e-01 -4.03422683e-01 -4.68086042e-02 1.94139138e-01
1.10806942e-01 1.02251291e-01 5.37602007e-01 1.69756040e-01
-6.96016908e-01 5.40162027e-01 1.75916612e+00 -8.92409906e-02
-2.49381796e-01 1.91356301e-01 -4.36002761e-01 1.05039954e+00
3.65608394e-01 -1.71163678e-01 -3.03663403e-01 -9.77858007e-01
3.23871017e-01 1.61221281e-01 5.55643439e-01 -6.18176639e-01
1.28778234e-01 1.00203097e-01 6.83511078e-01 -1.55991697e+00
9.31608081e-01 -1.03720355e+00 -9.58412215e-02 1.78211778e-02
1.65474489e-01 3.80284846e-01 1.31204829e-01 6.87181592e-01
-7.28643388e-02 3.02994877e-01 4.71626103e-01 -7.82020509e-01
-6.79325938e-01 7.59443521e-01 3.44394654e-01 1.28736871e-03
9.30098355e-01 -5.78838110e-01 -8.96179676e-02 -1.53763056e-01
-9.85159218e-01 2.36408398e-01 7.42109478e-01 6.12091064e-01
1.08473110e+00 -1.39803207e+00 -3.75978887e-01 5.74919999e-01
2.53101408e-01 5.47570586e-01 3.73685718e-01 6.34830892e-01
-4.73349065e-01 4.75908488e-01 -2.82052308e-01 -1.17481256e+00
-1.16967702e+00 7.32889891e-01 3.67651165e-01 -3.93274017e-02
-1.06072068e+00 1.02350771e+00 3.25159669e-01 -9.83282447e-01
-2.67202724e-02 -7.41106391e-01 -7.40020275e-02 6.68881908e-02
7.72097260e-02 1.93502605e-01 4.44314212e-01 -7.26546407e-01
-4.96477813e-01 1.17828357e+00 -3.28953634e-03 8.99976939e-02
1.77527487e+00 -5.94040453e-02 -7.75933489e-02 5.63677132e-01
1.41599393e+00 1.39400035e-01 -1.18222797e+00 -2.63538003e-01
-2.04886291e-02 -6.35244310e-01 -8.28075334e-02 -4.10138518e-01
-1.15184236e+00 1.21367967e+00 4.50635940e-01 1.81545094e-01
6.68279111e-01 5.16865849e-01 8.51842880e-01 5.09899259e-01
7.08614051e-01 -5.06104052e-01 8.24716687e-02 7.42035270e-01
1.13224173e+00 -1.25031078e+00 2.97920048e-01 -6.15001440e-01
-4.66487199e-01 1.24960029e+00 7.14536369e-01 -4.72481787e-01
1.21959710e+00 -2.28548106e-02 6.33863360e-02 -8.78531039e-01
-8.23710382e-01 -2.92726517e-01 5.99793434e-01 6.18225873e-01
-2.41494831e-02 3.51593681e-02 4.10810322e-01 4.46011335e-01
-3.16707820e-01 -3.90681684e-01 2.86120206e-01 8.15931678e-01
-2.48729929e-01 -9.11937773e-01 -2.26377293e-01 5.18876672e-01
-2.64216721e-01 2.49607220e-01 -5.40014982e-01 8.62355173e-01
1.63913861e-01 3.32002610e-01 2.70957768e-01 -4.04920310e-01
6.37419581e-01 2.25277305e-01 7.24625826e-01 -9.63297248e-01
-6.68315589e-01 1.65218472e-01 -3.65266085e-01 -9.43247676e-01
-6.11541927e-01 -6.87853038e-01 -8.98860216e-01 -7.51025602e-02
-3.78725857e-01 -4.14364815e-01 8.07937562e-01 7.80166864e-01
6.02688313e-01 1.25898182e-01 7.05473959e-01 -1.27575648e+00
-5.70968390e-01 -4.27973956e-01 -5.51741302e-01 4.05807644e-01
2.16536373e-01 -7.74376929e-01 -4.25357781e-02 -3.17884624e-01] | [8.123398780822754, -3.4784271717071533] |
94f09e1e-af01-4c52-b3c3-3f300506f33c | lxl-lidar-exclusive-lean-3d-object-detection | 2307.00724 | null | https://arxiv.org/abs/2307.00724v2 | https://arxiv.org/pdf/2307.00724v2.pdf | LXL: LiDAR Excluded Lean 3D Object Detection with 4D Imaging Radar and Camera Fusion | As an emerging technology and a relatively affordable device, the 4D imaging radar has already been confirmed effective in performing 3D object detection in autonomous driving. Nevertheless, the sparsity and noisiness of 4D radar point clouds hinder further performance improvement, and in-depth studies about its fusion with other modalities are lacking. On the other hand, most of the camera-based perception methods transform the extracted image perspective view features into the bird's-eye view geometrically via "depth-based splatting" proposed in Lift-Splat-Shoot (LSS), and some researchers exploit other modals such as LiDARs or ordinary automotive radars for enhancement. Recently, a few works have applied the "sampling" strategy for image view transformation, showing that it outperforms "splatting" even without image depth prediction. However, the potential of "sampling" is not fully unleashed. In this paper, we investigate the "sampling" view transformation strategy on the camera and 4D imaging radar fusion-based 3D object detection. In the proposed model, LXL, predicted image depth distribution maps and radar 3D occupancy grids are utilized to aid image view transformation, called "radar occupancy-assisted depth-based sampling". Experiments on VoD and TJ4DRadSet datasets show that the proposed method outperforms existing 3D object detection methods by a significant margin without bells and whistles. Ablation studies demonstrate that our method performs the best among different enhancement settings. | ['Bing Zhu', 'Yuxuan Xia', 'Qing-Long Han', 'Tao Huang', 'Jianan Liu', 'Weiyi Xiong'] | 2023-07-03 | null | null | null | null | ['3d-object-detection', 'depth-estimation'] | ['computer-vision', 'computer-vision'] | [ 1.97587952e-01 -3.69208395e-01 -1.01437457e-01 -5.02335489e-01
-4.20522541e-01 -3.70897382e-01 7.15712488e-01 -3.76842350e-01
-5.12395859e-01 3.37387353e-01 -3.09795469e-01 -4.64952022e-01
-2.20787134e-02 -8.17015111e-01 -5.08861661e-01 -7.67093360e-01
2.26329505e-01 3.30443174e-01 5.65750659e-01 -2.68834531e-01
4.99423325e-01 8.89522493e-01 -2.15077138e+00 -1.86747327e-01
7.49306321e-01 1.26557362e+00 6.03128910e-01 4.51216131e-01
-1.02914378e-01 2.66965985e-01 -4.35963392e-01 -9.58712623e-02
7.96064258e-01 1.18167959e-02 3.35001260e-01 -1.16484007e-02
6.40217841e-01 -8.12785327e-01 -2.48170525e-01 1.14278591e+00
5.29764116e-01 -1.10177942e-01 6.08501971e-01 -1.29596150e+00
-2.87334859e-01 -3.42359573e-01 -1.07611811e+00 4.48644519e-01
2.41669059e-01 1.38449833e-01 3.83328855e-01 -1.17560661e+00
3.25024962e-01 1.32828486e+00 4.18734342e-01 2.29048669e-01
-8.76493514e-01 -1.16720831e+00 -1.26605183e-01 4.41849977e-01
-1.39402628e+00 -2.30385661e-01 8.60507190e-01 -4.48139101e-01
8.28260064e-01 4.50600386e-01 8.25021088e-01 6.31532669e-01
6.77586019e-01 6.42942429e-01 1.66937494e+00 -3.14468443e-01
4.45989370e-02 4.50401306e-01 1.42336935e-01 4.87602383e-01
6.27525330e-01 8.15200090e-01 -7.22263515e-01 1.75254628e-01
5.27909040e-01 1.35198236e-01 -1.58729523e-01 -5.97643077e-01
-8.74034226e-01 8.17321897e-01 2.75001585e-01 -8.83624926e-02
-2.31670350e-01 -1.81539848e-01 6.31290302e-02 5.28046712e-02
8.13162982e-01 3.85730118e-02 1.05621323e-01 1.62673086e-01
-8.67371202e-01 4.63204443e-01 1.46610886e-01 1.08805466e+00
9.07427192e-01 2.44478077e-01 9.00829956e-02 3.13980371e-01
4.31443930e-01 1.27238154e+00 -5.53758815e-02 -8.08986485e-01
3.36102903e-01 7.32704341e-01 2.61628479e-01 -1.02565670e+00
-4.87752378e-01 -5.39948046e-01 -7.81809688e-01 9.58591402e-01
3.16113830e-02 1.09628655e-01 -1.18201637e+00 1.18035209e+00
5.98703384e-01 4.22494560e-02 2.28468016e-01 1.28027689e+00
7.91379094e-01 5.41802585e-01 -3.64076614e-01 -1.33707091e-01
1.60437179e+00 -3.16926301e-01 -7.16359556e-01 -5.05006850e-01
2.56438792e-01 -7.31370926e-01 5.75110197e-01 5.08967459e-01
-6.95555091e-01 -8.74269605e-01 -1.32056999e+00 1.87840238e-01
-3.45937610e-01 6.72349259e-02 6.10320508e-01 9.46674526e-01
-6.32735848e-01 -3.87191534e-01 -5.86897910e-01 -3.84837806e-01
3.25929552e-01 -6.56376258e-02 -2.55498767e-01 -4.65873003e-01
-1.00504637e+00 1.32122314e+00 2.63528645e-01 1.76237941e-01
-9.03444290e-01 -9.35213625e-01 -7.60284543e-01 -2.89141178e-01
4.85456228e-01 -6.65338159e-01 7.25172043e-01 -7.07859546e-02
-1.10395753e+00 8.74526143e-01 -2.42660657e-01 -8.97663891e-01
3.55321974e-01 -2.93943524e-01 -5.03905833e-01 2.43226871e-01
1.20227404e-01 7.98154593e-01 8.96979749e-01 -1.51679289e+00
-1.03003192e+00 -7.89222598e-01 1.19653270e-01 5.90310454e-01
3.31817329e-01 -2.36622319e-01 -8.12647268e-02 -2.63328869e-02
4.98117924e-01 -6.63180232e-01 -1.86382547e-01 2.91048996e-02
-8.75352994e-02 9.45931748e-02 1.23556900e+00 -7.22016096e-02
7.19403207e-01 -2.27369046e+00 -4.03474748e-01 3.57972011e-02
2.17847466e-01 2.34529316e-01 2.03416690e-01 2.70305991e-01
3.19480330e-01 -3.76871020e-01 -1.65005013e-01 2.83632316e-02
-1.52688980e-01 2.42590234e-01 -4.85282838e-01 6.88881814e-01
-2.30566505e-03 6.24233365e-01 -6.72371387e-01 -4.40151989e-01
9.40583467e-01 5.33547759e-01 -4.35015768e-01 1.09261043e-01
1.36799589e-01 4.38856304e-01 -4.50528502e-01 9.00880814e-01
1.54120827e+00 4.81422514e-01 -4.35712546e-01 -4.51813251e-01
-7.22563684e-01 -2.58751214e-01 -1.00054407e+00 1.24123418e+00
-2.27991641e-01 6.57790422e-01 2.78098464e-01 -8.11729133e-01
1.46297014e+00 -2.07039475e-01 2.60300845e-01 -1.09877849e+00
7.54736811e-02 1.43154457e-01 -1.04960762e-01 -4.67322379e-01
7.99343288e-01 -4.97711033e-01 -2.02031005e-02 -1.35464936e-01
-4.55797076e-01 -6.11148775e-01 -2.79134274e-01 1.01293623e-01
8.06501389e-01 3.40879261e-02 4.68333691e-01 -1.79699048e-01
5.37797153e-01 4.60946858e-01 3.42658550e-01 7.92912185e-01
-2.26658449e-01 4.14855331e-01 -3.45525183e-02 -2.45299071e-01
-7.65636504e-01 -1.14301193e+00 -5.49177229e-01 4.39265937e-01
7.79303908e-01 4.15291041e-02 -3.64251703e-01 -3.21715683e-01
2.03278422e-01 9.97091055e-01 -2.74311483e-01 -6.79987371e-02
-2.24450409e-01 -6.95619941e-01 3.83459002e-01 2.37744197e-01
9.45661843e-01 -3.29717547e-01 -1.43277180e+00 -1.71261951e-02
-2.62227934e-02 -1.41274083e+00 1.34107664e-01 1.17465086e-01
-8.14047575e-01 -8.14098954e-01 -3.11444134e-01 -2.18081281e-01
6.04279339e-01 1.21267533e+00 7.26222754e-01 -2.80585617e-01
-5.16031682e-01 4.07977194e-01 -5.61024249e-01 -9.83304858e-01
1.03957288e-01 -6.12433672e-01 3.49222571e-01 5.97999692e-02
8.65716159e-01 -4.82327372e-01 -7.60839522e-01 4.17671114e-01
-7.21646309e-01 9.10951644e-02 1.04295588e+00 5.71186602e-01
4.89010513e-01 9.53583121e-02 3.22678328e-01 -5.24941802e-01
1.78229719e-01 -7.42077380e-02 -1.06047642e+00 -4.86261606e-01
-7.46672392e-01 -2.95845926e-01 1.18671365e-01 1.46174759e-01
-1.13795984e+00 1.11029662e-01 -6.51469007e-02 -9.14831758e-01
-3.67381603e-01 1.26246274e-01 -2.77050972e-01 -2.98669726e-01
5.96909642e-01 5.29862642e-01 1.82631269e-01 -3.97803813e-01
1.84038654e-01 8.02561224e-01 6.68630719e-01 -1.06331808e-02
1.21110535e+00 1.14547253e+00 2.68725574e-01 -1.13801682e+00
-7.52750516e-01 -7.49781370e-01 -4.93161917e-01 -6.08496010e-01
1.01721215e+00 -1.07918930e+00 -7.81944633e-01 2.42848009e-01
-1.04770911e+00 3.98264766e-01 -2.15170920e-01 8.89578044e-01
-3.77951771e-01 4.84637767e-01 2.47553766e-01 -1.39143133e+00
-5.04810065e-02 -1.13443196e+00 1.22626781e+00 2.21684933e-01
4.24316257e-01 -3.15981358e-01 -1.79241717e-01 8.74690950e-01
3.33565265e-01 1.85956106e-01 5.80175042e-01 -1.83692142e-01
-1.12827253e+00 -3.60118598e-01 -5.62425494e-01 3.77474517e-01
-2.78908581e-01 -4.60917294e-01 -1.05725539e+00 -2.01658711e-01
3.91027808e-01 2.25868016e-01 8.78311157e-01 5.36762595e-01
6.49415195e-01 4.07951027e-01 -4.96407568e-01 7.34429598e-01
1.48122025e+00 5.24583101e-01 7.62484133e-01 3.24834287e-01
3.71152550e-01 5.86535871e-01 1.32576799e+00 4.75019634e-01
4.14749563e-01 6.74071252e-01 1.01632035e+00 -8.19613785e-02
-1.87397286e-01 -1.73848644e-01 4.21898544e-01 2.51818329e-01
-2.39130557e-01 -1.62461832e-01 -7.27843523e-01 2.73622841e-01
-1.37705016e+00 -1.13908005e+00 -3.30567569e-01 2.14350629e+00
-1.19924925e-01 5.42700410e-01 -3.87768865e-01 1.30035654e-01
5.18174589e-01 3.32715869e-01 -6.30066454e-01 -2.26975307e-01
-1.61420032e-02 4.06417251e-02 1.09329534e+00 4.93435204e-01
-8.52307141e-01 7.67491698e-01 5.67206764e+00 8.52708280e-01
-9.10595119e-01 1.03360839e-01 4.48779762e-02 7.61290640e-02
-4.07656252e-01 2.35342965e-01 -1.43346190e+00 2.58099169e-01
4.23241019e-01 2.53056884e-02 -3.62898298e-02 7.95526326e-01
3.72511417e-01 -8.74981105e-01 -6.11808062e-01 1.20964289e+00
3.55303049e-01 -1.12021303e+00 1.02185672e-02 4.53440815e-01
2.69094020e-01 1.50157101e-02 1.34978384e-01 3.60095769e-01
-1.28093317e-01 -7.18490303e-01 7.67761230e-01 5.27948916e-01
6.90760791e-01 -6.98046088e-01 7.84708798e-01 7.81141639e-01
-1.30740380e+00 2.50227638e-02 -5.95506907e-01 -2.29608610e-01
4.91425157e-01 8.96445394e-01 -1.02721334e+00 8.51490974e-01
7.92973101e-01 3.21742147e-01 -4.89201456e-01 8.45574081e-01
5.16446345e-02 2.62589127e-01 -3.80841911e-01 2.91866232e-02
3.90793562e-01 -4.90473330e-01 1.07447350e+00 8.84176970e-01
7.41731703e-01 3.19054335e-01 1.29269764e-01 9.98106778e-01
6.42208934e-01 -3.87961239e-01 -1.18102276e+00 4.85368758e-01
5.54168284e-01 1.35420227e+00 -4.91754860e-01 -1.79930732e-01
-5.62447071e-01 2.93561459e-01 -6.39621973e-01 1.53544620e-01
-9.95730102e-01 -2.29301602e-01 5.96713066e-01 6.97001755e-01
3.59959304e-01 -4.48267072e-01 -2.63915628e-01 -7.08524704e-01
-1.06413476e-01 -4.35954601e-01 -1.34875625e-01 -1.24500906e+00
-1.04274094e+00 3.56095284e-01 6.74419284e-01 -1.78524292e+00
-5.74227888e-03 -7.05465674e-01 -2.18956202e-01 8.75513911e-01
-1.94579101e+00 -1.10290825e+00 -6.31756067e-01 5.31594336e-01
6.66833282e-01 -2.35940635e-01 1.88325837e-01 2.81167686e-01
3.51000787e-03 2.42698826e-02 -2.07127988e-01 -4.39887643e-01
5.20683110e-01 -7.40806460e-01 1.06433062e-02 8.78205359e-01
-1.80424526e-01 3.60656418e-02 9.76064265e-01 -7.30151236e-01
-1.79838824e+00 -8.84969473e-01 4.58354086e-01 -4.01670188e-01
1.65577501e-01 -4.96149898e-01 -6.54023767e-01 3.85963559e-01
2.16356963e-01 -9.36219767e-02 2.25733489e-01 -2.91130424e-01
-1.46820515e-01 -5.86380184e-01 -1.25837982e+00 3.78603816e-01
9.78352785e-01 -1.91208914e-01 -7.38582790e-01 2.29517706e-02
3.60586554e-01 -5.99379838e-01 -3.49218249e-01 9.63586748e-01
6.05776012e-01 -1.46140981e+00 1.26951885e+00 3.27144891e-01
-1.48973465e-01 -6.73801184e-01 -6.97221279e-01 -8.90917361e-01
-8.31881464e-02 -4.94088680e-02 8.61896724e-02 9.34752584e-01
-5.22616552e-03 -7.73685277e-01 8.51736665e-01 -1.59335434e-01
-5.46804130e-01 -4.44717318e-01 -1.36963475e+00 -8.68353486e-01
-4.66220558e-01 -6.73668623e-01 2.81134635e-01 5.09817421e-01
-4.86464649e-01 4.82830405e-01 -3.16973567e-01 6.68782473e-01
1.12067401e+00 3.49833071e-01 1.12468195e+00 -1.34924614e+00
1.20099306e-01 -2.21737381e-02 -6.83621049e-01 -1.35728776e+00
-2.09957272e-01 -7.25214720e-01 2.85072457e-02 -1.45271683e+00
1.07748441e-01 -3.64817619e-01 9.20911580e-02 -1.58433914e-01
1.94786116e-01 3.42138320e-01 2.39007115e-01 1.98606148e-01
-1.63220674e-01 5.50941646e-01 1.28562069e+00 -7.11313263e-02
8.00785199e-02 6.15463071e-02 -4.61422414e-01 7.95130789e-01
7.14664519e-01 -2.68279880e-01 -5.35852134e-01 -1.75220862e-01
9.00667980e-02 2.54501909e-01 8.46896708e-01 -1.42900121e+00
4.09232676e-01 -1.81989208e-01 4.19328332e-01 -1.72747743e+00
9.92807090e-01 -1.10262883e+00 1.57103062e-01 4.51814771e-01
5.02303779e-01 -2.32315436e-02 3.27342898e-01 8.57788503e-01
-2.11476669e-01 -1.17191061e-01 8.86791587e-01 -2.08977535e-01
-1.06402147e+00 1.15024999e-01 -6.33539200e-01 -3.23275834e-01
1.31498539e+00 -9.18632090e-01 -3.43542695e-01 -2.94100106e-01
-4.93579656e-02 1.15769245e-01 2.12843508e-01 3.78916264e-01
9.81977344e-01 -1.16117251e+00 -6.25996828e-01 6.91766500e-01
2.75728017e-01 6.64838105e-02 5.32196045e-01 1.04754806e+00
-3.72708797e-01 6.65370464e-01 -3.29273552e-01 -1.18567693e+00
-1.26919031e+00 4.88001347e-01 7.69260898e-02 1.14302166e-01
-8.01696062e-01 4.64213878e-01 5.79660296e-01 -2.85153031e-01
-1.33166432e-01 -3.05317998e-01 -1.69905096e-01 -4.54863394e-03
4.28222865e-01 1.87310666e-01 1.49230771e-02 -6.60297334e-01
-4.76859987e-01 8.53142858e-01 -3.86511497e-02 -2.14669287e-01
9.68090415e-01 -3.93061996e-01 2.97311068e-01 2.89861172e-01
5.44698358e-01 1.13593571e-01 -1.28893268e+00 -8.99574757e-02
-4.57443953e-01 -8.51103187e-01 4.12684232e-01 -5.16487241e-01
-7.46577621e-01 1.19436789e+00 1.08604157e+00 2.71060139e-01
1.20119727e+00 -3.81320938e-02 5.14602304e-01 2.25585148e-01
7.06803679e-01 -7.94358492e-01 -2.30235904e-01 4.58683461e-01
6.04613423e-01 -1.29104352e+00 3.47973287e-01 -5.11109948e-01
-6.16976738e-01 7.52742708e-01 6.51371956e-01 -1.19096167e-01
5.26888430e-01 6.19150400e-01 6.29344070e-03 -4.73385274e-01
-5.37092805e-01 -6.88362241e-01 3.76981571e-02 8.35935473e-01
-2.41249546e-01 3.90010444e-03 -2.28790179e-01 2.32536748e-01
-1.95343494e-01 -2.20591232e-01 4.68427420e-01 8.16586733e-01
-1.08771694e+00 -6.31294966e-01 -9.13742900e-01 4.60824072e-01
1.43187836e-01 -1.53730074e-02 5.74050508e-02 1.31013489e+00
5.33741534e-01 9.04350698e-01 2.95082092e-01 -6.46937788e-01
6.73756182e-01 -4.40986753e-02 5.55900633e-01 -3.66796732e-01
-1.97460037e-02 1.17028959e-01 -1.54170245e-01 -4.56555188e-01
-6.77142441e-01 -6.36087179e-01 -8.86756361e-01 -3.32366168e-01
-5.33963501e-01 -1.04112580e-01 8.45500767e-01 7.57141113e-01
3.03115278e-01 3.13534588e-01 6.64116800e-01 -9.51132178e-01
-4.31882828e-01 -7.11600721e-01 -8.01748574e-01 -2.98463523e-01
3.33242357e-01 -1.20050633e+00 -6.60772920e-01 -2.78794616e-01] | [7.78403377532959, -1.597353219985962] |
b9189b55-2d30-4be7-ae70-c544c8a0a504 | 190503297 | 1905.03297 | null | https://arxiv.org/abs/1905.03297v3 | https://arxiv.org/pdf/1905.03297v3.pdf | Interpretable Subgroup Discovery in Treatment Effect Estimation with Application to Opioid Prescribing Guidelines | The dearth of prescribing guidelines for physicians is one key driver of the current opioid epidemic in the United States. In this work, we analyze medical and pharmaceutical claims data to draw insights on characteristics of patients who are more prone to adverse outcomes after an initial synthetic opioid prescription. Toward this end, we propose a generative model that allows discovery from observational data of subgroups that demonstrate an enhanced or diminished causal effect due to treatment. Our approach models these sub-populations as a mixture distribution, using sparsity to enhance interpretability, while jointly learning nonlinear predictors of the potential outcomes to better adjust for confounding. The approach leads to human-interpretable insights on discovered subgroups, improving the practical utility for decision support | ['Sara E. Berger', 'Monica Shekhar', 'Kush R. Varshney', 'Chirag Nagpal', 'Subhro Das', 'Dennis Wei', 'Bhanukiran Vinzamuri'] | 2019-05-08 | null | null | null | null | ['subgroup-discovery'] | ['methodology'] | [ 3.07105511e-01 2.80757278e-01 -1.17721426e+00 -4.14348096e-01
-5.80697477e-01 -5.37501872e-01 1.85918838e-01 4.96953189e-01
-1.08542897e-01 9.61586356e-01 8.82411718e-01 -8.82806242e-01
-5.46536565e-01 -6.69491649e-01 -7.35518634e-01 -3.39269996e-01
-3.25386912e-01 6.72983050e-01 -8.45836759e-01 1.71965972e-01
-9.73985940e-02 3.59574199e-01 -7.92981267e-01 5.89012682e-01
1.30764580e+00 5.13818443e-01 -1.38701499e-01 5.07250465e-02
3.80500108e-01 1.07658315e+00 -4.88693267e-02 -2.39385456e-01
4.19227004e-01 -4.71744895e-01 -5.27071841e-02 2.31149271e-02
2.39099219e-01 -5.60216486e-01 -6.15513086e-01 9.63010967e-01
3.62990797e-01 -3.13292861e-01 1.02486634e+00 -7.96255231e-01
-9.38557446e-01 9.83007193e-01 -2.91611791e-01 1.60895094e-01
7.98964873e-02 6.13920093e-01 9.33164299e-01 -5.38949311e-01
5.20566702e-01 1.31004560e+00 5.86961448e-01 3.67662966e-01
-1.69996500e+00 -8.26854765e-01 1.99258298e-01 -1.20821975e-01
-1.12477911e+00 -4.16420966e-01 4.50732321e-01 -1.00457728e+00
3.92724335e-01 1.75819263e-01 5.48956096e-01 1.41038656e+00
5.22797644e-01 4.56191123e-01 1.30511427e+00 2.42796808e-01
3.83215636e-01 -9.49551016e-02 2.96936423e-01 6.23469591e-01
7.57820547e-01 4.36774999e-01 -2.56122261e-01 -1.08207512e+00
9.96600688e-01 9.51219380e-01 -1.09373219e-01 -1.60260886e-01
-9.35148418e-01 1.20091152e+00 1.43735081e-01 -1.16251908e-01
-1.18404925e+00 5.44580780e-02 4.60295491e-02 3.14653777e-02
2.34386936e-01 5.70779741e-01 -4.77761239e-01 1.06857568e-01
-7.71831393e-01 3.91709685e-01 5.61176479e-01 6.84789956e-01
3.12827021e-01 -1.32392719e-01 -4.97779578e-01 5.82956433e-01
4.30368096e-01 6.21657312e-01 2.94486940e-01 -8.77438009e-01
3.30446213e-01 7.88494527e-01 1.98831841e-01 -4.89833087e-01
-5.60868919e-01 -6.90699816e-01 -9.84215915e-01 -1.47325739e-01
2.51162410e-01 -5.31852901e-01 -1.19230819e+00 1.77616560e+00
-9.40693915e-02 3.08436807e-02 8.85150209e-02 4.86015648e-01
5.34169257e-01 1.79163933e-01 6.29171848e-01 -4.28749740e-01
1.53803039e+00 -2.90017128e-01 -8.08159471e-01 -1.82109520e-01
4.40402240e-01 -3.62011939e-01 6.61498725e-01 4.07368720e-01
-1.06580818e+00 -1.38575211e-01 -4.17822301e-01 3.05124223e-01
2.57094383e-01 1.51779041e-01 8.89933169e-01 5.69768012e-01
-3.08324516e-01 8.04719687e-01 -8.25944662e-01 5.86064644e-02
1.14160264e+00 6.97676003e-01 1.77046716e-01 -1.66737050e-01
-9.91378009e-01 5.54644525e-01 2.03939006e-01 -9.30902436e-02
-8.37085009e-01 -1.44944036e+00 -6.63870275e-01 1.83720842e-01
6.25415504e-01 -1.40819848e+00 1.04797328e+00 -4.46398437e-01
-9.05738115e-01 2.78828919e-01 -4.01639253e-01 -6.62469149e-01
3.36958617e-01 -1.31231487e-01 -3.33661526e-01 -2.15648755e-01
2.48123750e-01 -8.72111972e-03 7.11055756e-01 -1.14727449e+00
-6.03174448e-01 -8.81228089e-01 -2.52563536e-01 -9.73601416e-02
2.98673272e-01 -1.31697133e-01 8.17421377e-02 -8.89518440e-01
-1.37181982e-01 -1.19474661e+00 -1.00104594e+00 -4.94035721e-01
-6.11860156e-01 -1.39260128e-01 2.46774256e-02 -7.39508867e-01
1.56468666e+00 -2.01696730e+00 8.64120722e-02 4.19630736e-01
9.01629269e-01 -2.43422031e-01 1.91032141e-01 4.74045247e-01
-6.23901822e-02 3.32722127e-01 -4.49807405e-01 1.14779271e-01
-2.55982786e-01 3.74224007e-01 -3.86262953e-01 4.27235335e-01
2.90632248e-01 1.10736406e+00 -1.01282573e+00 1.43680990e-01
7.09517673e-02 1.16709471e-01 -9.45958078e-01 2.08816692e-01
-3.80470455e-01 8.74545634e-01 -8.23755741e-01 8.09444189e-01
5.47995806e-01 -7.46942639e-01 6.17782354e-01 -4.78316061e-02
4.20429073e-02 2.61438459e-01 -7.99252927e-01 1.06305528e+00
-1.04031064e-01 -6.22939914e-02 -6.66293278e-02 -5.72733462e-01
4.05151188e-01 2.91042894e-01 8.92191470e-01 -3.72123271e-01
2.18526781e-01 3.60738933e-01 6.04916513e-01 -7.05577254e-01
-1.51489198e-01 -7.56409049e-01 -2.09986651e-03 3.38619292e-01
-3.49652976e-01 3.86631489e-01 -2.21657544e-01 -7.58295208e-02
1.07569277e+00 -5.44050932e-01 5.27296901e-01 -3.18417013e-01
-1.13788269e-01 1.08641066e-01 7.74565458e-01 9.77623343e-01
7.80659020e-02 3.08722198e-01 6.12096190e-01 -5.15927672e-01
-9.66427267e-01 -1.27327597e+00 -4.45003569e-01 3.32338899e-01
-3.69820476e-01 -1.81933925e-01 -2.18299516e-02 -4.93837416e-01
6.21248782e-01 7.65498161e-01 -8.73595774e-01 -4.47927892e-01
-4.14106011e-01 -1.31472683e+00 2.12603018e-01 5.62520087e-01
-2.56459326e-01 -6.88778639e-01 -2.03492060e-01 3.29142123e-01
-8.05817172e-03 -7.32027769e-01 -4.50812399e-01 3.71452659e-01
-1.23877573e+00 -1.34637916e+00 -7.00814068e-01 -2.98109382e-01
5.80243289e-01 -3.76188666e-01 9.40796912e-01 -3.38899016e-01
-1.25355378e-01 1.33759677e-01 2.38054559e-01 -6.86949134e-01
-6.06559277e-01 -1.36878863e-01 3.81371051e-01 9.85353440e-03
4.99483794e-01 -5.10694444e-01 -1.11769605e+00 -1.20197777e-02
-8.47349226e-01 -2.34485373e-01 7.80065417e-01 7.24376261e-01
6.67922318e-01 -3.07673246e-01 8.01679671e-01 -1.48990524e+00
9.63780463e-01 -1.34560323e+00 -3.61457616e-01 -7.34151676e-02
-1.17435110e+00 2.64293700e-01 5.84509075e-01 -5.48909783e-01
-8.95730555e-01 1.61107238e-02 1.51507065e-01 -4.09597099e-01
-5.78857660e-02 9.77131784e-01 2.83239275e-01 4.76820022e-01
7.42943645e-01 -4.90760170e-02 1.56985521e-01 -7.45454311e-01
2.83355176e-01 7.50826061e-01 1.57702625e-01 -5.23534596e-01
2.89996505e-01 6.32619500e-01 9.93656442e-02 -5.27792871e-01
-9.46375906e-01 -4.28378373e-01 -7.59939924e-02 5.74035347e-01
9.64421093e-01 -1.24039745e+00 -9.58448470e-01 -6.84549585e-02
-7.05062330e-01 -4.85152215e-01 -4.00992751e-01 1.07629669e+00
-4.37575191e-01 1.04316391e-01 -6.36655033e-01 -9.09104645e-01
-2.24468380e-01 -1.34763527e+00 7.39797354e-01 -2.59223245e-02
-6.71649575e-01 -9.23964262e-01 4.84389335e-01 4.09428626e-01
2.05301151e-01 3.02693605e-01 1.52926779e+00 -8.14944744e-01
-6.07963204e-01 1.59683123e-01 -9.13508013e-02 1.64897725e-01
5.36888003e-01 -3.02570522e-01 -5.38729906e-01 -9.83155221e-02
2.49451786e-01 1.08188279e-01 8.42650831e-01 1.26555252e+00
1.34694564e+00 -7.54532158e-01 -4.93413478e-01 6.81997299e-01
1.22192633e+00 7.71714270e-01 5.34388125e-01 -2.14920074e-01
7.10019171e-01 3.78031880e-01 2.86796931e-02 8.72588873e-01
4.24634814e-01 4.53664601e-01 1.60222515e-01 -2.41892368e-01
2.33575568e-01 -5.54721475e-01 -2.29163785e-02 1.51723549e-01
-3.45283657e-01 -9.45767295e-03 -7.64028788e-01 4.34243262e-01
-1.91678202e+00 -9.16608632e-01 -3.42298388e-01 2.29194164e+00
9.21204686e-01 -2.94446558e-01 5.07832587e-01 -8.27814937e-01
5.82069576e-01 -1.44755900e-01 -9.73915040e-01 -2.08643124e-01
-1.05493523e-01 4.26537454e-01 9.44558918e-01 3.36863458e-01
-7.43141770e-01 4.30904537e-01 7.97816277e+00 4.52923238e-01
-8.74111056e-01 -7.95818567e-02 1.17418408e+00 -1.23886921e-01
-6.67456567e-01 6.62930757e-02 -5.75319231e-01 5.10995090e-01
1.05725396e+00 -2.87839949e-01 4.97419626e-01 4.91372466e-01
8.84544671e-01 2.39716843e-01 -1.23856759e+00 8.18846345e-01
-2.82520115e-01 -1.60895276e+00 2.11035475e-01 7.06824541e-01
9.65468824e-01 2.21468061e-01 4.38907474e-01 7.24963844e-02
8.15086544e-01 -1.38989055e+00 3.39739859e-01 8.61065269e-01
7.61324346e-01 -4.18501437e-01 4.83998686e-01 2.47368515e-01
-3.38397443e-01 -5.46715677e-01 -1.43814698e-01 -2.30808005e-01
2.69936621e-01 7.52262235e-01 -1.14395893e+00 2.76523918e-01
9.19421539e-02 7.37308800e-01 -1.19679488e-01 8.35024834e-01
-1.87586565e-02 8.91213477e-01 -1.13438308e-01 4.15893435e-01
5.07952198e-02 -6.15394413e-01 3.22175950e-01 6.07664049e-01
3.17839712e-01 5.36566854e-01 1.04502454e-01 1.22594035e+00
-3.49412560e-01 6.76670521e-02 -7.99078286e-01 -4.62158412e-01
2.32469901e-01 5.86683750e-01 -1.85389236e-01 -5.20025373e-01
-4.97171849e-01 4.43131715e-01 1.77407116e-02 5.53070307e-01
-6.01769030e-01 6.13723695e-01 8.28014255e-01 3.86287540e-01
-1.45703424e-02 6.41613454e-02 -8.39325845e-01 -1.29161835e+00
-3.35398614e-01 -1.38555801e+00 6.42922103e-01 -2.61334091e-01
-1.66072357e+00 1.36272490e-01 -1.11362979e-01 -8.81535053e-01
-1.04330480e-01 -3.33788097e-01 -2.36359254e-01 1.12106097e+00
-1.23738372e+00 -8.10370684e-01 4.63228345e-01 3.68885696e-01
4.04106349e-01 -1.78175926e-01 6.58578455e-01 2.01404974e-01
-6.66940153e-01 1.81217283e-01 2.27741078e-01 -2.03470320e-01
5.46074152e-01 -1.02251434e+00 2.03367267e-02 1.99926019e-01
-2.13059545e-01 1.31248534e+00 5.41064382e-01 -1.58393443e+00
-1.19930887e+00 -1.19281542e+00 7.01148510e-01 -7.26529121e-01
7.32520521e-01 5.59973083e-02 -7.69898653e-01 8.49504232e-01
-1.15024202e-01 -5.98400533e-01 1.38748372e+00 5.16832948e-01
-2.59440392e-01 3.60484540e-01 -1.09177601e+00 8.38605404e-01
9.97782290e-01 -4.22406197e-01 -6.60508871e-01 5.23779273e-01
5.43059289e-01 4.96229753e-02 -1.22306502e+00 5.44300258e-01
6.93278491e-01 -5.85860133e-01 1.02205384e+00 -1.47754395e+00
8.73197019e-01 3.70521061e-02 6.20246399e-03 -1.15863323e+00
-7.05320954e-01 -7.87656426e-01 -2.27755517e-01 2.12645382e-01
8.87851775e-01 -6.79979742e-01 9.43944156e-01 1.07559514e+00
4.65080477e-02 -1.02256763e+00 -5.73181152e-01 -4.47590292e-01
1.26440063e-01 -7.98146799e-02 6.75587654e-01 1.13028872e+00
1.10864408e-01 4.08441156e-01 -6.31706297e-01 2.81869769e-01
6.38856530e-01 3.44596803e-01 4.75877434e-01 -1.21754515e+00
-7.37860262e-01 -4.44536805e-01 5.06304279e-02 -8.43419909e-01
-2.78963238e-01 -9.35660481e-01 -6.22316182e-01 -1.63903356e+00
9.13512647e-01 -5.26088715e-01 -2.99981982e-01 4.62280393e-01
-5.42365372e-01 -2.49727845e-01 2.07975227e-02 4.25607979e-01
5.17389551e-02 4.40195143e-01 1.22515750e+00 -8.84778947e-02
-5.98260224e-01 4.55737680e-01 -1.29860330e+00 7.44658351e-01
7.07773566e-01 -7.96213031e-01 -4.21149224e-01 -4.61263917e-02
2.42093369e-01 4.00168002e-01 4.08081561e-01 -9.32975635e-02
-3.23298991e-01 -7.86214292e-01 3.57550710e-01 -4.49341506e-01
-8.81888345e-02 -7.81869054e-01 6.33204520e-01 1.19244730e+00
-5.77697575e-01 1.22426853e-01 1.02357626e-01 8.87002826e-01
1.78553388e-01 2.57853866e-02 4.07072753e-01 -2.18955874e-01
1.85850471e-01 6.19636834e-01 -5.90301454e-01 1.06321275e-01
6.31865561e-01 2.49527365e-01 -2.48404473e-01 -4.65285987e-01
-1.11963832e+00 1.03967845e-01 3.46497357e-01 7.21186772e-02
3.85799855e-01 -1.16296029e+00 -8.70493829e-01 9.81697813e-02
1.22289613e-01 -3.78336251e-01 5.16814888e-01 9.21880364e-01
-1.58175901e-01 5.70539415e-01 1.45722002e-01 -2.97468752e-01
-8.09872210e-01 9.37077940e-01 1.83734462e-01 -4.87284154e-01
-5.09633303e-01 1.83278322e-01 6.12284839e-01 -5.29032797e-02
-1.77802622e-01 -6.98031127e-01 -8.25491399e-02 -1.64318368e-01
2.19889179e-01 4.50457513e-01 -3.16377699e-01 -2.07033440e-01
-2.21789762e-01 1.78788099e-02 -8.78378674e-02 2.27395535e-01
1.61067140e+00 9.17197317e-02 -9.76688489e-02 4.27701265e-01
8.87440085e-01 4.05764490e-01 -1.09702909e+00 -2.11961225e-01
-8.67059901e-02 -3.82488906e-01 -1.57035470e-01 -9.90512013e-01
-7.87714243e-01 2.45176032e-01 6.02642834e-01 -5.97110316e-02
5.86570799e-01 3.07913929e-01 6.24401748e-01 8.90308470e-02
-1.01280943e-01 -5.95100999e-01 -4.72835183e-01 -1.15107116e-03
6.83304846e-01 -1.32436180e+00 1.53740808e-01 -5.48666418e-01
-6.34670258e-01 7.11041868e-01 -1.80725202e-01 -2.46388942e-01
7.69076765e-01 4.63650702e-03 -9.44638997e-02 -7.11529493e-01
-5.71208298e-01 -6.03436679e-02 4.55581337e-01 4.33344722e-01
3.60475898e-01 5.87576270e-01 -6.81485891e-01 1.14739168e+00
1.29868239e-01 4.63899136e-01 3.91290814e-01 5.62958777e-01
-3.66027579e-02 -1.21992397e+00 -2.90912271e-01 1.34691501e+00
-7.40242600e-01 -6.71510160e-01 -3.69603157e-01 3.88014525e-01
2.36101389e-01 7.01388359e-01 -1.81575894e-01 3.99333201e-02
6.90769494e-01 1.36217609e-01 3.01643461e-01 -1.00511909e+00
-5.70747197e-01 5.38325846e-01 7.43967295e-02 -3.11786354e-01
-1.02822423e-01 -7.27950037e-01 -9.81295586e-01 9.87270102e-03
1.18463613e-01 4.87090684e-02 5.15200058e-03 8.44102979e-01
8.43649566e-01 7.07324147e-01 5.63969314e-01 4.16282341e-02
-9.99109924e-01 -6.34923816e-01 -6.80572867e-01 6.44854009e-01
4.56186205e-01 -4.94203687e-01 -2.37081185e-01 3.21622728e-03] | [8.051410675048828, 5.440039157867432] |
b4e7e3f1-2ac4-423c-af2b-9f904aaf84e9 | a-coarse-to-fine-multi-stream-hybrid | 1908.10521 | null | https://arxiv.org/abs/1908.10521v1 | https://arxiv.org/pdf/1908.10521v1.pdf | A Coarse-to-Fine Multi-stream Hybrid Deraining Network for Single Image Deraining | Single image deraining task is still a very challenging task due to its ill-posed nature in reality. Recently, researchers have tried to fix this issue by training the CNN-based end-to-end models, but they still cannot extract the negative rain streaks from rainy images precisely, which usually leads to an over de-rained or under de-rained result. To handle this issue, this paper proposes a new coarse-to-fine single image deraining framework termed Multi-stream Hybrid Deraining Network (shortly, MH-DerainNet). To obtain the negative rain streaks during training process more accurately, we present a new module named dual path residual dense block, i.e., Residual path and Dense path. The Residual path is used to reuse com-mon features from the previous layers while the Dense path can explore new features. In addition, to concatenate different scaled features, we also apply the idea of multi-stream with shortcuts between cascaded dual path residual dense block based streams. To obtain more distinct derained images, we combine the SSIM loss and perceptual loss to preserve the per-pixel similarity as well as preserving the global structures so that the deraining result is more accurate. Extensive experi-ments on both synthetic and real rainy images demonstrate that our MH-DerainNet can deliver significant improvements over several recent state-of-the-art methods. | ['Richang Hong', 'Haijun Zhang', 'Yanyan Wei', 'Meng Wang', 'Zhao Zhang'] | 2019-08-28 | null | null | null | null | ['single-image-deraining'] | ['computer-vision'] | [-1.01560145e-03 -3.39829981e-01 2.96811581e-01 -6.17289424e-01
-4.92055297e-01 1.53466966e-02 1.31380588e-01 -4.14032340e-01
-4.32913870e-01 8.69601250e-01 1.04759812e-01 5.38425222e-02
1.81781366e-01 -9.36488569e-01 -7.48789907e-01 -9.37757969e-01
1.99989140e-01 -2.41558641e-01 4.03739631e-01 -4.90542889e-01
-3.04721743e-02 4.83072668e-01 -1.46331859e+00 2.43057773e-01
1.37670016e+00 6.78761005e-01 4.50127691e-01 5.60767591e-01
-2.11054400e-01 7.72782087e-01 -5.21978438e-01 -1.99792594e-01
4.72523093e-01 -7.64799535e-01 -8.08399692e-02 9.62930545e-02
8.33862603e-01 -7.04103172e-01 -4.60782647e-01 1.14242780e+00
5.99585891e-01 7.29888454e-02 2.20342904e-01 -8.46824884e-01
-4.13355738e-01 2.74251550e-01 -9.89654660e-01 6.61495209e-01
-5.80924526e-02 2.23984823e-01 7.23300278e-01 -1.17822099e+00
4.55460221e-01 1.50620866e+00 5.92672408e-01 2.45123252e-01
-9.60205078e-01 -1.03950775e+00 4.70548064e-01 2.41347477e-01
-1.09909618e+00 -2.68557549e-01 6.33265078e-01 2.67593823e-02
3.06737036e-01 3.49988997e-01 7.33507574e-01 6.99224532e-01
2.94308931e-01 8.72926056e-01 1.39065325e+00 -1.54940905e-02
-1.41346917e-01 -5.07419966e-02 2.09411725e-01 7.12577343e-01
4.98827875e-01 3.16919476e-01 -1.82640284e-01 4.11884248e-01
8.03992808e-01 5.57109773e-01 -7.44573653e-01 -8.41321424e-04
-1.02402127e+00 7.77224898e-01 1.05769181e+00 1.90567449e-01
-4.89507139e-01 -2.02040970e-01 -1.21490479e-01 5.70129156e-01
7.74009645e-01 1.32864147e-01 -3.88360828e-01 4.35690194e-01
-1.32930481e+00 6.09012783e-01 5.44799447e-01 5.06175280e-01
1.21465278e+00 3.54829371e-01 -2.45436221e-01 8.51824403e-01
2.43812397e-01 8.87817621e-01 8.49495977e-02 -6.59884989e-01
6.58665836e-01 3.62940073e-01 3.60080823e-02 -9.37416136e-01
-1.68108657e-01 -7.98648953e-01 -1.34634542e+00 5.52277744e-01
-1.34598305e-02 -1.88486204e-01 -1.56482255e+00 1.29018652e+00
3.87989044e-01 5.65023780e-01 1.81593508e-01 1.38995945e+00
1.02701509e+00 1.10863388e+00 -6.70966785e-03 -4.55775023e-01
1.00405502e+00 -1.20101213e+00 -6.88444376e-01 -4.29934114e-01
1.14704750e-03 -8.33278954e-01 8.19460094e-01 4.57456499e-01
-1.02839589e+00 -7.78104365e-01 -1.16567481e+00 7.50605837e-02
-5.72618134e-02 -2.04025000e-01 3.11291069e-01 1.52510196e-01
-6.52052999e-01 6.23036921e-01 -5.73000610e-01 -1.91904038e-01
6.59377337e-01 -2.51513962e-02 -2.80423820e-01 -7.86006510e-01
-1.26059604e+00 7.93044627e-01 3.10255677e-01 7.44814217e-01
-1.13595510e+00 -8.41907620e-01 -8.33851457e-01 -5.22887148e-02
1.65506229e-01 -7.01669753e-01 7.64197528e-01 -1.38704419e+00
-1.19855440e+00 3.92171800e-01 -1.61422670e-01 -3.49201977e-01
4.01893109e-01 -6.00217044e-01 -5.99344552e-01 1.65944085e-01
2.72140536e-03 4.08652276e-01 1.34488428e+00 -1.53314555e+00
-8.07507694e-01 -3.08334917e-01 -8.02627951e-03 6.53811753e-01
3.36185540e-03 -2.16972798e-01 -4.35278833e-01 -1.07154894e+00
1.70934796e-01 -6.01627052e-01 -4.67127919e-01 3.39775272e-02
-2.61690408e-01 5.24111450e-01 1.26127481e+00 -8.24980378e-01
1.16987967e+00 -2.17715478e+00 8.88058618e-02 -3.62032764e-02
4.60715264e-01 7.70244181e-01 -2.77829945e-01 -4.13610451e-02
-2.68875986e-01 -1.49547830e-01 -7.83779502e-01 -1.76694319e-01
-7.11802363e-01 4.64642167e-01 -2.65770525e-01 5.29758215e-01
4.05856162e-01 5.32974482e-01 -8.44334960e-01 -4.35855329e-01
3.67792904e-01 6.07785463e-01 -2.49303222e-01 6.58871889e-01
-9.80329514e-02 4.70416844e-01 -4.53199506e-01 6.92016363e-01
1.51485431e+00 2.81985760e-01 -4.97062117e-01 -3.11803818e-01
-2.37268314e-01 -2.85076827e-01 -1.24911046e+00 1.25222301e+00
-6.08154535e-01 4.10917282e-01 4.35777724e-01 -7.89904475e-01
9.42413568e-01 -2.26950236e-02 -3.99897546e-02 -8.68153691e-01
-6.87389448e-02 2.25861251e-01 -1.95987582e-01 -7.19615161e-01
3.39364558e-01 -5.85644126e-01 5.36456108e-01 -2.97320243e-02
-1.55905932e-01 -8.59390125e-02 -1.49103269e-01 2.18442455e-01
7.18150020e-01 1.37480438e-01 1.93848107e-02 2.13676337e-02
5.87468028e-01 -1.71164036e-01 9.47660327e-01 4.89272714e-01
-1.06558062e-01 1.26103985e+00 5.21035418e-02 -4.86064672e-01
-8.76773417e-01 -1.17967761e+00 -9.52699184e-02 7.86728382e-01
5.19394279e-01 1.07578859e-01 -5.52530050e-01 -7.43257821e-01
-7.78386965e-02 4.92134690e-01 -6.69631183e-01 5.92225604e-02
-9.22782063e-01 -1.05848956e+00 1.93028241e-01 3.73179197e-01
1.14923739e+00 -1.19814932e+00 -2.37552077e-01 3.38511020e-01
-2.13142782e-01 -9.24442053e-01 -4.71421331e-01 1.65470630e-01
-1.01963413e+00 -9.57172751e-01 -1.12920237e+00 -7.85945058e-01
5.59862494e-01 9.30127203e-01 1.14035821e+00 3.01185966e-01
-2.75347590e-01 -4.98763978e-01 -6.49369717e-01 -3.42505485e-01
5.91048785e-02 -1.86033875e-01 -6.13640606e-01 2.23377958e-01
-6.18410148e-02 -8.45607460e-01 -1.06609726e+00 8.00464004e-02
-1.31883097e+00 4.54832166e-02 1.08007455e+00 8.89155626e-01
6.39866889e-01 4.30784971e-02 4.68054056e-01 -1.03337777e+00
4.17935491e-01 -5.34739971e-01 -3.79920185e-01 3.29184532e-02
-5.69134474e-01 1.48995016e-02 7.57033169e-01 -9.14351940e-02
-1.39026892e+00 5.87007147e-04 -2.87892312e-01 -7.78868079e-01
-4.44693752e-02 4.07161772e-01 -2.91202009e-01 -1.02597140e-01
4.00664449e-01 3.06491673e-01 -6.66266158e-02 -5.94826043e-01
4.55037326e-01 4.68598366e-01 5.85241735e-01 1.05160028e-02
1.27023792e+00 6.67246938e-01 -2.18204573e-01 -8.65117490e-01
-1.28104329e+00 -5.64633071e-01 -2.73975372e-01 -1.58254385e-01
9.27130759e-01 -1.33697486e+00 -1.11536734e-01 7.29507267e-01
-1.00918734e+00 -2.07956120e-01 -1.64966598e-01 3.77835125e-01
8.28075185e-02 4.61521447e-01 -5.81528246e-01 -6.92721903e-01
-8.46640229e-01 -9.52770829e-01 8.92337978e-01 7.13180304e-01
6.25317514e-01 -5.79707503e-01 1.78362265e-01 1.31795719e-01
6.81855083e-01 4.31187153e-01 5.18056810e-01 1.12768419e-01
-7.92374969e-01 1.08317301e-01 -6.28707826e-01 8.20261717e-01
3.75298709e-02 1.93713289e-02 -8.67899656e-01 -5.48846722e-01
1.70934737e-01 -1.25818521e-01 1.57443273e+00 2.65114903e-01
7.75416493e-01 -2.38777131e-01 -9.38989315e-03 1.07428026e+00
1.83593178e+00 -1.12124406e-01 1.13066292e+00 4.16622400e-01
9.46939886e-01 3.31702054e-01 9.45475519e-01 1.75801530e-01
1.70444235e-01 3.00087959e-01 7.34012365e-01 -6.95878446e-01
-3.40916365e-01 8.80372748e-02 4.49882746e-01 7.07978308e-01
-1.73785254e-01 -3.36441368e-01 -2.28397310e-01 6.26715660e-01
-1.69107056e+00 -1.07917571e+00 -2.22692072e-01 2.12243700e+00
6.87825263e-01 1.34097353e-01 -1.91798553e-01 -7.07993135e-02
6.82124078e-01 7.07722485e-01 -5.34629524e-01 -1.99618727e-01
-5.13612330e-01 6.97961211e-01 5.95123470e-01 5.95943153e-01
-1.09638059e+00 9.36717331e-01 5.01357698e+00 8.57332647e-01
-1.32431126e+00 1.38291910e-01 6.50336146e-01 -1.45695992e-02
-2.75347769e-01 3.03831790e-02 -8.39695334e-01 5.21185338e-01
3.82002085e-01 2.09713966e-01 2.64479220e-01 5.48980296e-01
5.23406565e-01 -3.26518565e-01 -2.58396626e-01 8.22016180e-01
2.24887282e-01 -1.02412355e+00 9.38547030e-02 -3.91733199e-01
8.84179354e-01 1.86823592e-01 -1.72019750e-01 3.90008003e-01
2.37214401e-01 -1.09159374e+00 5.05321801e-01 7.42892385e-01
7.51332521e-01 -8.58893216e-01 9.91427422e-01 1.69482544e-01
-1.28469813e+00 4.85712402e-02 -5.02691746e-01 1.78747520e-01
1.21100880e-01 1.25076079e+00 -3.96492898e-01 1.08837414e+00
9.61339831e-01 1.03486264e+00 -5.30897498e-01 1.35465622e+00
-4.50507462e-01 7.30020523e-01 -3.64562988e-01 5.33113658e-01
3.62905025e-01 -3.93018365e-01 6.69456065e-01 1.42809188e+00
3.10972601e-01 3.01984310e-01 1.18066199e-01 7.38484621e-01
-4.76766229e-02 -1.74777359e-01 -4.53870833e-01 5.19568861e-01
2.63286196e-02 1.43858135e+00 -3.69172424e-01 -3.77772123e-01
-4.39317226e-01 1.37920940e+00 -4.15286701e-03 6.78448319e-01
-8.14068377e-01 -6.83562219e-01 8.56207430e-01 2.34278217e-01
6.45087779e-01 -1.27353698e-01 -1.30089879e-01 -1.33170736e+00
7.18239397e-02 -1.07821918e+00 1.13729060e-01 -8.19182694e-01
-1.41937947e+00 8.58843863e-01 -2.14197397e-01 -1.45297444e+00
3.23868632e-01 -2.14103721e-02 -1.05013144e+00 1.26446116e+00
-2.21549749e+00 -1.12062359e+00 -1.06058109e+00 8.10068429e-01
7.55035818e-01 3.05878371e-01 1.71071202e-01 5.15135765e-01
-6.51690304e-01 3.78242552e-01 2.48066723e-01 -6.52231127e-02
9.75493968e-01 -1.06075132e+00 1.62860617e-01 1.26741230e+00
-1.75898865e-01 2.10621998e-01 8.31466913e-01 -6.48798764e-01
-9.24975455e-01 -1.67852700e+00 3.91495436e-01 2.42621452e-01
-1.62910856e-02 -5.29959537e-02 -1.42476451e+00 3.93418163e-01
3.19321543e-01 4.37692702e-01 -2.02865526e-01 -3.19315463e-01
-3.17053974e-01 -7.24011719e-01 -1.09370029e+00 4.17884707e-01
8.05571437e-01 6.03111424e-02 -5.94883800e-01 1.22179948e-01
9.05210316e-01 -4.35906768e-01 -3.40745986e-01 6.95385575e-01
2.23690093e-01 -1.42124283e+00 9.58058059e-01 -1.03075288e-01
6.94940031e-01 -6.29152894e-01 7.68655464e-02 -1.53289545e+00
-3.44524354e-01 -4.13745403e-01 -1.00671910e-01 1.12776482e+00
1.83180481e-01 -5.77764511e-01 7.31027961e-01 -3.10453594e-01
-3.44473660e-01 -8.96376967e-01 -5.73987365e-01 -5.42194307e-01
8.59780330e-03 -2.21830122e-02 3.75227928e-01 8.99716198e-01
-8.64062667e-01 3.95395100e-01 -6.77904308e-01 4.28458720e-01
8.31482112e-01 4.25948292e-01 7.55455554e-01 -9.59668875e-01
-1.46649703e-01 -3.49339694e-02 -1.45956516e-01 -1.20141923e+00
-1.84112951e-01 -5.06865323e-01 5.57657838e-01 -1.67868638e+00
2.23462299e-01 -3.42758060e-01 -3.67861629e-01 2.38880962e-01
-7.59509683e-01 3.83798182e-01 3.55983913e-01 3.29758108e-01
-3.89181226e-01 9.91075754e-01 1.77932501e+00 -6.31267577e-03
-4.00222719e-01 8.36988762e-02 -5.92690349e-01 7.84877479e-01
7.28452265e-01 -5.92640817e-01 -1.40833363e-01 -5.53469241e-01
-2.61644423e-01 1.71961650e-01 3.94451022e-01 -1.25642669e+00
1.14120264e-02 -2.52031118e-01 6.10472679e-01 -7.93645501e-01
3.77790987e-01 -6.01410627e-01 -1.16278820e-01 3.14758897e-01
1.20394304e-01 -7.83821046e-02 5.55808060e-02 7.66657948e-01
-6.93258345e-01 1.41520008e-01 1.33631897e+00 -2.29082704e-01
-7.70013094e-01 6.00657165e-01 -8.04229006e-02 -9.36146379e-02
8.87102187e-01 -2.78932571e-01 -3.38533461e-01 -3.69617492e-01
-6.04589283e-01 4.55773264e-01 2.47702733e-01 2.27402791e-01
1.05799627e+00 -9.06708062e-01 -1.14513505e+00 3.73370498e-01
-1.58467114e-01 3.95205051e-01 7.23490238e-01 9.07926857e-01
-9.01558638e-01 -2.18787089e-01 -2.82200783e-01 -2.90793031e-01
-1.40247953e+00 3.95221651e-01 4.54933047e-01 -3.42953861e-01
-1.05489683e+00 8.31810892e-01 6.74841940e-01 -1.22260362e-01
7.98572004e-02 -2.32564777e-01 -1.39118537e-01 -5.47834672e-02
6.89138889e-01 1.30926713e-01 6.60367236e-02 -3.79504561e-01
-6.64089695e-02 5.64008951e-01 -3.26680005e-01 6.78730756e-02
1.56549323e+00 -2.87032813e-01 -2.04754308e-01 2.09674299e-01
1.08864665e+00 -5.38601801e-02 -1.47487617e+00 -3.48782957e-01
-6.92424715e-01 -8.27997983e-01 3.33510250e-01 -7.41091728e-01
-1.85518360e+00 1.02208543e+00 9.33516860e-01 -1.20137475e-01
1.53573012e+00 -5.69462299e-01 1.34975564e+00 1.86177850e-01
-1.00191087e-01 -4.73319918e-01 9.31565538e-02 6.04410052e-01
8.08633327e-01 -1.32337964e+00 2.64898658e-01 -4.54469293e-01
-6.68981969e-01 1.00410390e+00 8.37801099e-01 -6.52538836e-01
9.00486946e-01 1.64763883e-01 3.00223261e-01 -1.99948400e-01
-3.68080944e-01 -4.86012131e-01 -6.31329939e-02 3.85558724e-01
1.90806553e-01 -1.73559174e-01 -4.25964713e-01 2.93409526e-01
2.04413667e-01 1.56379595e-01 5.69159150e-01 8.46249700e-01
-9.32406843e-01 -7.66719639e-01 -7.71300614e-01 5.54291010e-01
-4.32450444e-01 -4.74935621e-01 1.06487185e-01 7.55633414e-01
4.03042018e-01 9.02753472e-01 -5.00019342e-02 -4.42381591e-01
6.31676316e-01 -6.33900881e-01 3.06249648e-01 -5.21467566e-01
-6.50373638e-01 1.97849840e-01 -1.88675374e-01 -5.44546545e-01
-6.61008656e-01 -1.90235421e-01 -1.05200803e+00 -3.76593232e-01
-3.50778401e-01 1.13513447e-01 2.39047125e-01 8.65704596e-01
1.11623704e-01 7.03502357e-01 1.01842141e+00 -1.08721960e+00
-2.36780658e-01 -1.03434706e+00 -8.92414212e-01 3.32329720e-01
7.32894719e-01 -3.56595010e-01 -5.94449639e-01 -2.20565796e-02] | [10.899300575256348, -3.258986473083496] |
c6b933ef-b95b-40db-8c63-3643e63d15ac | mquake-assessing-knowledge-editing-in | 2305.14795 | null | https://arxiv.org/abs/2305.14795v1 | https://arxiv.org/pdf/2305.14795v1.pdf | MQuAKE: Assessing Knowledge Editing in Language Models via Multi-Hop Questions | The information stored in large language models (LLMs) falls out of date quickly, and retraining from scratch is often not an option. This has recently given rise to a range of techniques for injecting new facts through updating model weights. Current evaluation paradigms are extremely limited, mainly validating the recall of edited facts, but changing one fact should cause rippling changes to the model's related beliefs. If we edit the UK Prime Minister to now be Rishi Sunak, then we should get a different answer to Who is married to the British Prime Minister? In this work, we present a benchmark MQuAKE (Multi-hop Question Answering for Knowledge Editing) comprising multi-hop questions that assess whether edited models correctly answer questions where the answer should change as an entailed consequence of edited facts. While we find that current knowledge-editing approaches can recall edited facts accurately, they fail catastrophically on the constructed multi-hop questions. We thus propose a simple memory-based approach, MeLLo, which stores all edited facts externally while prompting the language model iteratively to generate answers that are consistent with the edited facts. While MQuAKE remains challenging, we show that MeLLo scales well with LLMs (up to 175B) and outperforms previous model editors by a large margin. | ['Danqi Chen', 'Christopher Potts', 'Christopher D. Manning', 'Zhengxuan Wu', 'Zexuan Zhong'] | 2023-05-24 | null | null | null | null | ['multi-hop-question-answering'] | ['knowledge-base'] | [ 1.96332365e-01 5.69064081e-01 -6.98628128e-02 -4.09000546e-01
-1.12549269e+00 -7.32021451e-01 7.36645699e-01 5.96313179e-01
-5.28040051e-01 1.06836367e+00 5.03863275e-01 -6.15032911e-01
-3.59084129e-01 -1.16149628e+00 -9.85855639e-01 -1.65126026e-02
2.97521770e-01 8.17921817e-01 6.19952619e-01 -6.28823936e-01
4.20950174e-01 7.36758336e-02 -1.44043088e+00 5.10090709e-01
8.51698041e-01 4.77914065e-01 9.26439464e-02 9.67172742e-01
-6.76302552e-01 1.37735260e+00 -7.41365552e-01 -1.18401086e+00
-1.20760821e-01 -3.70476007e-01 -1.59342456e+00 -5.76469362e-01
7.58551478e-01 -2.42336303e-01 -1.84887439e-01 9.46813881e-01
3.05118650e-01 3.86573851e-01 2.30443940e-01 -9.25003171e-01
-7.81407416e-01 9.84301090e-01 1.83417305e-01 5.25134325e-01
6.67450011e-01 -1.00704268e-01 9.47038531e-01 -8.65053415e-01
7.92041719e-01 1.37943900e+00 8.71726274e-01 6.04744256e-01
-1.20045769e+00 -3.91997099e-01 3.51816267e-01 5.54319382e-01
-1.03127086e+00 -5.02449811e-01 3.92826647e-01 -9.39227864e-02
1.54552710e+00 7.31800616e-01 3.73510361e-01 7.98975646e-01
3.56323063e-01 4.89318043e-01 1.13304520e+00 -6.01686597e-01
1.50782421e-01 3.38251531e-01 2.57032126e-01 8.09949934e-01
1.83504537e-01 1.33311171e-02 -8.51529956e-01 -4.28622633e-01
1.99334383e-01 -4.51295882e-01 1.62310165e-03 6.87169805e-02
-9.98025239e-01 8.35759938e-01 2.00591981e-01 4.71950263e-01
-2.44864285e-01 1.90269738e-01 1.67690352e-01 6.73886061e-01
8.99997428e-02 8.78580034e-01 -7.55437195e-01 -2.07851186e-01
-8.43131602e-01 6.93876266e-01 1.07129335e+00 7.05493927e-01
8.94020081e-01 -2.85343409e-01 -1.41522780e-01 7.15374410e-01
1.20204523e-01 3.96408170e-01 5.71924508e-01 -1.38469863e+00
4.86461073e-01 6.38316095e-01 3.19270521e-01 -1.42010033e+00
-3.23931128e-01 -3.02313715e-01 -4.98535872e-01 -1.85032547e-01
4.52631384e-01 1.35289475e-01 -5.24942100e-01 1.88597965e+00
3.83018345e-01 8.04566070e-02 3.09974819e-01 3.20485860e-01
9.44631636e-01 6.98057711e-01 3.11583191e-01 -1.06718354e-01
1.14496386e+00 -7.60138094e-01 -6.53031349e-01 -7.82152653e-01
7.02052772e-01 -7.19897807e-01 1.10993493e+00 5.64701736e-01
-1.36067355e+00 -4.39830571e-01 -7.38774180e-01 -3.57346833e-01
-6.49500191e-01 -6.41908050e-01 7.48127401e-01 5.65720975e-01
-1.13287592e+00 4.84374613e-01 -3.71143252e-01 -3.30111295e-01
-8.97363573e-02 2.38197029e-01 -3.10199112e-01 -3.34158152e-01
-1.80611110e+00 1.61435938e+00 6.82232976e-01 7.76951835e-02
-4.25409883e-01 -9.31126714e-01 -9.72424328e-01 8.71697962e-02
8.39904904e-01 -9.23078835e-01 1.72973096e+00 -6.67616367e-01
-1.23679495e+00 8.22745860e-01 -4.95909065e-01 -6.25588596e-01
3.12341243e-01 -2.59105712e-01 -7.95192122e-01 -5.80137521e-02
1.64048046e-01 7.75417805e-01 4.72724557e-01 -1.44257712e+00
-8.90453458e-01 -1.53092727e-01 7.11980224e-01 1.93460248e-02
-1.18581727e-01 1.09101094e-01 -4.43473220e-01 -6.08573973e-01
-1.42300222e-02 -4.79673922e-01 -1.34307474e-01 -4.33268636e-01
-2.14037791e-01 -1.52453586e-01 3.97883385e-01 -7.71762848e-01
1.78838229e+00 -1.44591594e+00 -5.29037155e-02 2.94471622e-01
1.78454131e-01 4.37898666e-01 -2.23290265e-01 5.59335291e-01
-4.63724807e-02 1.86361596e-01 -3.86108160e-01 -1.97055712e-02
2.30495468e-01 4.96809781e-01 -8.35535944e-01 -5.09369493e-01
1.50028110e-01 1.21961081e+00 -1.11152005e+00 -6.06551528e-01
-8.19313005e-02 6.51955828e-02 -7.94408262e-01 1.08014839e-02
-6.36371255e-01 -2.90111214e-01 3.57672782e-03 3.13438952e-01
4.84337151e-01 -2.85465240e-01 3.27127904e-01 -4.51693274e-02
3.66416484e-01 5.60631216e-01 -1.34831417e+00 1.40488243e+00
-5.26319027e-01 2.70768583e-01 -1.64219216e-01 -5.66644609e-01
7.90967226e-01 6.94782808e-02 -2.60157794e-01 -9.08434749e-01
-3.38515282e-01 1.60467133e-01 -1.61652878e-01 -7.45015204e-01
9.57835019e-01 -6.90470934e-01 -1.67625770e-01 4.08802420e-01
4.14023176e-02 -5.33996761e-01 4.50783134e-01 7.19811320e-01
1.21500838e+00 -2.71278601e-02 3.77428442e-01 3.92930321e-02
6.51394188e-01 2.70520121e-01 4.20450032e-01 1.15452325e+00
2.39134908e-01 1.11945868e-01 2.83015102e-01 -4.12006170e-01
-5.37505150e-01 -1.06806159e+00 2.10741743e-01 1.32663834e+00
-1.83279499e-01 -8.78494680e-01 -5.51420808e-01 -7.66543984e-01
-5.15364856e-03 1.46464729e+00 -5.87795973e-01 -4.95219529e-01
-7.06822872e-01 -4.02564108e-01 7.72888541e-01 5.52070320e-01
4.26943660e-01 -1.15601873e+00 -6.56469941e-01 6.55590057e-01
-4.93961930e-01 -7.30476737e-01 -3.07892054e-01 1.51783124e-01
-8.58185530e-01 -1.17456889e+00 -9.48375911e-02 -6.87351227e-01
4.18408006e-01 -1.69557944e-01 1.75174892e+00 5.61876297e-01
1.78012818e-01 6.48701012e-01 -2.57923286e-02 -3.36860627e-01
-7.79389083e-01 -5.13159484e-03 -1.63079321e-01 -3.68621707e-01
4.73822415e-01 -4.14029568e-01 -1.29517555e-01 6.42140284e-02
-1.39459801e+00 3.84647511e-02 3.13566655e-01 7.01149523e-01
3.15898746e-01 1.95499778e-01 9.22989726e-01 -1.62211883e+00
8.64349246e-01 -5.12098789e-01 -3.34876657e-01 8.25789571e-01
-6.42602026e-01 3.80550832e-01 6.43507957e-01 -2.59249747e-01
-1.33475506e+00 -4.63603616e-01 -4.47830081e-01 2.61275291e-01
5.06571047e-02 9.82973933e-01 6.78268149e-02 1.57231107e-01
9.00119662e-01 9.60680768e-02 -5.16654968e-01 -2.37573743e-01
6.87951148e-01 2.81626791e-01 9.47050095e-01 -1.04268289e+00
7.80061960e-01 1.41010851e-01 -4.28990036e-01 -3.53899121e-01
-1.30127358e+00 -1.19849958e-01 -1.66873068e-01 -1.67068034e-01
3.01490456e-01 -5.57245493e-01 -5.63439786e-01 2.61115551e-01
-1.29593766e+00 -5.69316328e-01 -3.96001816e-01 -1.24961823e-01
-4.08843368e-01 4.56245422e-01 -5.89403033e-01 -5.55135369e-01
-1.42393425e-01 -6.44357860e-01 6.31588697e-01 4.63159429e-03
-7.91545630e-01 -1.25456560e+00 2.76376128e-01 4.62929964e-01
6.57773852e-01 -5.56209981e-02 1.74469638e+00 -5.63644409e-01
-4.48805869e-01 -8.25375691e-02 1.38270423e-01 2.04402104e-01
-3.44093069e-02 -1.19310439e-01 -8.06449831e-01 3.79493050e-02
3.61447744e-02 -6.12014115e-01 6.36441588e-01 -2.14912921e-01
9.52792943e-01 -6.84153199e-01 -1.06677495e-01 -1.66366711e-01
1.23531318e+00 -6.73101842e-02 8.15315902e-01 4.61080313e-01
2.05804661e-01 5.66089809e-01 4.28856790e-01 4.68596481e-02
9.93281424e-01 4.36632484e-01 3.00368555e-02 5.45120835e-01
-2.46890396e-01 -5.11162937e-01 3.56386811e-01 1.05848682e+00
3.30781341e-01 -1.69774801e-01 -1.01785076e+00 7.43135571e-01
-1.72911084e+00 -1.07644117e+00 -1.63267598e-01 2.14552402e+00
1.57079160e+00 5.34228444e-01 -4.72958535e-01 6.35315850e-02
2.77156383e-01 7.07803220e-02 -4.26803648e-01 -5.72106481e-01
-2.90796101e-01 6.42081559e-01 -1.25902385e-01 1.34506726e+00
-4.58274692e-01 1.05605447e+00 6.89742517e+00 8.07958066e-01
-6.59597218e-01 2.02006608e-01 2.56865948e-01 -2.76020321e-04
-9.54149008e-01 3.32644314e-01 -8.82889032e-01 2.20682725e-01
1.18725777e+00 -4.31525737e-01 4.52843815e-01 3.24156404e-01
-3.28906029e-01 -5.53237438e-01 -1.14452648e+00 6.67547524e-01
3.50960344e-01 -1.58347440e+00 4.48328644e-01 -6.49163008e-01
7.71295846e-01 -5.08651078e-01 -1.90247148e-01 9.50924277e-01
5.53852737e-01 -9.69344020e-01 7.45238662e-01 1.15407193e+00
3.08642834e-01 -7.76277184e-01 5.95018506e-01 6.89838409e-01
-7.87166297e-01 -8.32958147e-02 -3.05523425e-01 -2.38484994e-01
1.37141556e-01 5.24850190e-01 -7.29289949e-01 4.83667880e-01
6.15081131e-01 2.55510379e-02 -1.07084072e+00 6.07870817e-01
-4.02780533e-01 5.54736376e-01 -3.54440004e-01 4.66072559e-02
9.94046703e-02 3.25313419e-01 2.79552847e-01 1.14597499e+00
1.41707793e-01 3.80685627e-01 -2.34376937e-01 7.16474533e-01
-2.09310487e-01 -1.87436149e-01 -5.17281055e-01 4.71855029e-02
4.85784709e-01 9.12504077e-01 -2.49876663e-01 -8.79817545e-01
-3.07066113e-01 6.92840576e-01 6.37292147e-01 3.31355542e-01
-6.00838542e-01 -3.43740344e-01 2.18295634e-01 1.19004332e-01
1.00578576e-01 5.70869818e-02 1.19336732e-02 -1.10683680e+00
1.80483535e-01 -1.36555660e+00 7.14273870e-01 -1.20735824e+00
-1.10730040e+00 4.31989402e-01 1.29648641e-01 -2.54396319e-01
-4.98467624e-01 -3.11701059e-01 -3.13978165e-01 7.96775997e-01
-1.48699927e+00 -7.16709018e-01 1.39566297e-02 4.52956766e-01
4.81666625e-01 2.62833267e-01 1.12679255e+00 2.83964127e-01
-3.67987789e-02 4.92291152e-01 -3.22797596e-01 -2.68864363e-01
7.87361681e-01 -1.25193119e+00 4.97326642e-01 8.75134468e-01
3.30232620e-01 1.11120212e+00 9.10210133e-01 -9.45125818e-01
-1.20356357e+00 -1.01023424e+00 1.87574804e+00 -1.06605947e+00
7.86959708e-01 -4.80827726e-02 -1.51653516e+00 1.02230060e+00
4.36045021e-01 -3.86125982e-01 6.08229518e-01 1.98865440e-02
-5.72370589e-01 -8.39034840e-02 -1.05755949e+00 8.17839742e-01
8.63777041e-01 -7.99401104e-01 -1.48779941e+00 2.74728417e-01
7.17379451e-01 -4.65983897e-01 -6.98894858e-01 3.32636923e-01
1.35434881e-01 -9.23511505e-01 8.94898057e-01 -1.32040024e+00
3.36292386e-01 -4.66606230e-01 -2.14968622e-01 -1.30584168e+00
-2.94759780e-01 -5.40873587e-01 -5.37937522e-01 1.25247765e+00
7.05430031e-01 -6.89752400e-01 5.75647414e-01 1.15898430e+00
7.55056813e-02 -6.46729827e-01 -8.83209467e-01 -5.01055241e-01
1.57480106e-01 -7.81635284e-01 7.62298226e-01 9.86223996e-01
1.37922153e-01 4.31332648e-01 -1.11699544e-01 8.16663578e-02
1.34397060e-01 3.08434330e-02 5.83947241e-01 -9.32778478e-01
-5.29285967e-01 -4.04765785e-01 1.95499420e-01 -1.05759108e+00
1.86326683e-01 -1.02307487e+00 4.88914102e-02 -1.84699857e+00
1.81723401e-01 -3.38990688e-01 -4.73921932e-02 6.53845847e-01
-6.93850875e-01 -5.80185726e-02 2.31230006e-01 -3.51937830e-01
-7.91160047e-01 2.12981582e-01 7.97439098e-01 -2.23761141e-01
2.45440021e-01 -2.73177803e-01 -1.06827962e+00 8.11094403e-01
5.61685860e-01 -6.38643861e-01 -4.51496720e-01 -6.85281873e-01
1.25733829e+00 2.20704898e-01 6.69385791e-01 -9.47235227e-01
7.22124457e-01 -1.41022295e-01 1.84084609e-01 -3.23560894e-01
2.47522756e-01 -7.23983407e-01 4.47299421e-01 3.79621983e-01
-6.92269325e-01 5.07866085e-01 6.35087729e-01 5.00867963e-01
-3.95454109e-01 -5.92552006e-01 3.45688194e-01 -3.76001596e-01
-9.65465546e-01 -2.77753741e-01 -4.90847856e-01 6.40249312e-01
4.94718134e-01 -2.55414825e-02 -6.66209519e-01 -5.02243161e-01
-8.58760297e-01 2.88799912e-01 -1.74246251e-03 5.28963447e-01
6.55089140e-01 -1.30665231e+00 -5.64222217e-01 -8.20767283e-02
1.83187336e-01 -1.41724303e-01 4.01064426e-01 4.64988470e-01
-3.50105047e-01 5.28298795e-01 2.15135857e-01 1.40684322e-01
-1.00354576e+00 6.09213769e-01 3.96838844e-01 -5.86242795e-01
-2.10345924e-01 9.94468033e-01 -2.92947173e-01 -1.05545294e+00
1.42066866e-01 -6.17325842e-01 -2.27743853e-02 1.10005811e-01
7.28520572e-01 3.00170809e-01 6.37910485e-01 -2.36599892e-01
-3.14236283e-01 3.74242753e-01 -2.00942919e-01 -1.85284078e-01
1.09972858e+00 -1.39016822e-01 -2.79668778e-01 5.03686488e-01
6.91372871e-01 3.86985421e-01 -6.86251044e-01 -5.73741138e-01
3.85540456e-01 -4.69435990e-01 -2.82362133e-01 -1.49445856e+00
-5.36780477e-01 6.14383459e-01 -2.14460380e-02 3.34674388e-01
9.51684058e-01 8.64124075e-02 8.87056351e-01 1.08051300e+00
6.26631260e-01 -1.10107160e+00 -5.62834628e-02 9.16892588e-01
1.10880339e+00 -1.01527369e+00 -7.06065446e-02 -9.05392841e-02
-6.20961726e-01 8.45770597e-01 7.83051610e-01 2.58832008e-01
2.39717290e-01 1.58377692e-01 2.49514475e-01 -2.61468768e-01
-1.17234337e+00 9.15715620e-02 1.28571615e-01 3.59049976e-01
2.73907095e-01 -8.88626277e-02 7.59561593e-03 7.53067851e-01
-6.24929965e-01 2.04230204e-01 6.21011257e-01 1.02400744e+00
-6.67680621e-01 -1.31633401e+00 -4.88573819e-01 6.74971700e-01
-2.61471629e-01 -4.48259294e-01 -5.35859048e-01 8.44319344e-01
5.07003181e-02 1.08309579e+00 -3.85452092e-01 -1.01326592e-01
6.72461271e-01 6.65267587e-01 6.98186874e-01 -7.25197732e-01
-1.01999736e+00 -7.39816010e-01 4.74108577e-01 -4.62311149e-01
-3.34731281e-01 -3.36743146e-01 -1.45698893e+00 -6.22266054e-01
-1.30480736e-01 3.88929158e-01 7.88186565e-02 1.21044385e+00
3.66935700e-01 2.53001779e-01 -1.45681933e-01 7.95037150e-02
-5.56729555e-01 -7.40428329e-01 -5.54165654e-02 3.53051871e-01
2.22640246e-01 -4.30239946e-01 -1.65706336e-01 1.78515270e-01] | [10.69953441619873, 8.056760787963867] |
95cd24d3-7e6c-4268-8706-0c87d6e5091a | dual-spls-a-family-of-dual-sparse-partial | 2301.07206 | null | https://arxiv.org/abs/2301.07206v1 | https://arxiv.org/pdf/2301.07206v1.pdf | Dual-sPLS: a family of Dual Sparse Partial Least Squares regressions for feature selection and prediction with tunable sparsity; evaluation on simulated and near-infrared (NIR) data | Relating a set of variables X to a response y is crucial in chemometrics. A quantitative prediction objective can be enriched by qualitative data interpretation, for instance by locating the most influential features. When high-dimensional problems arise, dimension reduction techniques can be used. Most notable are projections (e.g. Partial Least Squares or PLS ) or variable selections (e.g. lasso). Sparse partial least squares combine both strategies, by blending variable selection into PLS. The variant presented in this paper, Dual-sPLS, generalizes the classical PLS1 algorithm. It provides balance between accurate prediction and efficient interpretation. It is based on penalizations inspired by classical regression methods (lasso, group lasso, least squares, ridge) and uses the dual norm notion. The resulting sparsity is enforced by an intuitive shrinking ratio parameter. Dual-sPLS favorably compares to similar regression methods, on simulated and real chemical data. Code is provided as an open-source package in R: \url{https://CRAN.R-project.org/package=dual.spls}. | ['François Wahl', 'Rami El Haddad', 'Clément Marteau', 'Laurent Duval', 'Louna Alsouki'] | 2023-01-17 | null | null | null | null | ['variable-selection'] | ['methodology'] | [ 6.64210379e-01 3.31616774e-02 -4.58603948e-01 -5.23808420e-01
-7.10679591e-01 -6.40286863e-01 6.07060790e-01 3.28047037e-01
-1.69197127e-01 1.07453573e+00 6.49987981e-02 -2.74162650e-01
-4.77284342e-01 -5.63511729e-01 -5.76029420e-01 -1.06363451e+00
4.07867357e-02 5.10232210e-01 -1.19765982e-01 -5.20484298e-02
4.24236387e-01 7.45082140e-01 -1.19986236e+00 8.95935521e-02
7.96985924e-01 8.43465686e-01 2.26027638e-01 1.87306672e-01
-2.62134701e-01 2.61578560e-01 1.89835578e-01 -9.32394862e-02
3.36638182e-01 -4.97717112e-01 -4.00450706e-01 -1.10429630e-01
4.06472422e-02 4.84126836e-01 4.29402977e-01 1.00189412e+00
4.82848138e-01 1.00491673e-01 7.23379552e-01 -1.19652402e+00
-3.88217747e-01 5.55305541e-01 -9.47770238e-01 -2.91686803e-01
5.30905008e-01 -5.32940663e-02 1.07779670e+00 -1.31115234e+00
5.48121929e-01 1.36992383e+00 6.88291132e-01 2.95552015e-01
-2.17497349e+00 -6.01615250e-01 -1.72977462e-01 -1.84223846e-01
-1.19990015e+00 -6.06684327e-01 8.14839602e-01 -8.09945047e-01
4.55835134e-01 7.31803417e-01 6.25804141e-02 1.11595142e+00
8.68323371e-02 5.72690010e-01 1.62589765e+00 -2.96460956e-01
5.98836601e-01 4.03753877e-01 3.93265873e-01 3.58214915e-01
5.03601909e-01 5.64185381e-02 -7.15668142e-01 -7.02478468e-01
3.71955395e-01 3.71790081e-01 -3.29038978e-01 -9.32349980e-01
-1.14457071e+00 1.12372530e+00 6.95857480e-02 1.81094989e-01
-6.72850072e-01 -1.44344509e-01 2.61132658e-01 3.04085106e-01
4.87061322e-01 8.42128694e-01 -4.96942639e-01 3.02803576e-01
-1.08430541e+00 2.95808315e-01 7.94745266e-01 6.73411429e-01
9.93124247e-01 3.73960137e-02 1.23458520e-01 9.10447121e-01
5.16391993e-01 3.70020270e-01 3.00514221e-01 -1.00612569e+00
1.72472626e-01 5.75637877e-01 1.65615186e-01 -9.91957247e-01
-6.05721414e-01 -3.41445208e-01 -8.69850814e-01 3.67086202e-01
3.80904943e-01 -1.06952786e-01 -4.34934944e-01 1.58961117e+00
3.86220545e-01 -1.44901350e-01 -6.11754209e-02 9.56853032e-01
5.36093712e-01 4.54000831e-01 3.14065963e-01 -8.45600545e-01
1.15694594e+00 -5.70707560e-01 -7.32172728e-01 -1.56001583e-01
4.02376294e-01 -7.32192457e-01 8.86317015e-01 8.42054188e-01
-7.67584443e-01 -2.73321509e-01 -7.40635157e-01 5.13690040e-02
-5.85796416e-01 8.10039341e-02 8.43703568e-01 6.40869260e-01
-7.19837368e-01 8.46070290e-01 -7.47568429e-01 -3.70873958e-01
1.73481449e-01 6.30339622e-01 -6.47461474e-01 1.96193039e-01
-8.81002903e-01 4.62018371e-01 2.09784687e-01 5.18135391e-02
-3.78472686e-01 -7.63092339e-01 -8.81885588e-01 -2.38285437e-01
5.10602415e-01 -5.08512497e-01 7.89624095e-01 -9.86606717e-01
-1.53493762e+00 8.65294933e-01 -4.31446552e-01 -2.99965501e-01
5.31174660e-01 -2.86732435e-01 7.63965165e-03 -1.85970128e-01
1.63764566e-01 3.20942253e-01 8.63327980e-01 -1.43640161e+00
9.68996808e-03 -8.30301523e-01 -5.53935111e-01 -1.76256150e-01
-1.61798492e-01 3.08752120e-01 7.13554397e-02 -5.60993791e-01
5.23440957e-01 -1.10786474e+00 -6.56798422e-01 -5.68648539e-02
-6.39451444e-01 -8.64834040e-02 5.45989394e-01 -5.52212596e-01
1.17532182e+00 -1.92622983e+00 5.07003307e-01 5.39076447e-01
2.90057719e-01 -1.11037299e-01 8.85924399e-02 6.80302143e-01
-7.78451502e-01 2.49127001e-02 -7.86001384e-01 -3.76248598e-01
1.93499755e-02 7.42768198e-02 -2.23593891e-01 9.55405474e-01
8.21247622e-02 5.36366224e-01 -9.04640138e-01 -2.64672548e-01
1.76128969e-01 5.15947819e-01 -2.43138537e-01 -3.10012847e-02
-1.19173251e-01 7.49355197e-01 -7.14253664e-01 7.09524930e-01
7.35841215e-01 -2.78920531e-01 3.85720104e-01 -1.30760476e-01
-6.61796510e-01 -1.45569846e-01 -1.45688081e+00 1.58300102e+00
1.67531148e-03 2.33630702e-01 2.10381702e-01 -1.21480727e+00
1.52759635e+00 2.98446745e-01 8.96372437e-01 -2.86005110e-01
-2.80591231e-02 4.93676573e-01 -4.22429383e-01 -4.57247436e-01
2.07611933e-01 -3.62568617e-01 1.25360638e-01 2.25485250e-01
-1.72714159e-01 1.31479144e-01 1.88969180e-01 -2.18971204e-02
9.08960521e-01 5.58660805e-01 1.04047632e+00 -6.08498991e-01
8.69121492e-01 1.30711019e-01 6.86651170e-01 4.31689769e-01
1.45932108e-01 3.98289084e-01 8.84502292e-01 -3.16481292e-01
-1.01246524e+00 -7.33901143e-01 -5.82701802e-01 1.16127217e+00
-2.41332039e-01 -4.47684914e-01 -1.90435812e-01 -3.11454624e-01
3.29615980e-01 5.69372237e-01 -6.75498664e-01 1.42813727e-01
-3.35449994e-01 -9.09016371e-01 7.12521449e-02 1.30771667e-01
-4.07208741e-01 -8.81133378e-01 -1.89798281e-01 2.58482635e-01
3.17952216e-01 -5.57139635e-01 2.28260130e-01 8.16379786e-01
-1.05014372e+00 -8.67337644e-01 -7.59889007e-01 -2.50093192e-01
6.03485942e-01 1.39507174e-01 8.71127427e-01 -5.82961738e-01
-4.38630134e-02 7.68423304e-02 -4.45726216e-01 -3.08406174e-01
-1.25664607e-01 -1.16691515e-01 4.48373705e-01 2.05539569e-01
6.53900504e-01 -9.00319993e-01 -4.05361831e-01 3.60788375e-01
-8.01150680e-01 -2.40882769e-01 7.00189352e-01 9.59011555e-01
1.16993141e+00 -4.09595549e-01 3.76767397e-01 -1.38859582e+00
5.32597005e-01 -7.77316749e-01 -7.26980448e-01 7.31507391e-02
-8.61605346e-01 1.59246802e-01 7.78949797e-01 -2.86997706e-01
-5.58319926e-01 6.06708109e-01 4.30781282e-02 -5.17357051e-01
-3.25184554e-01 7.23443747e-01 -3.33446801e-01 -1.93886817e-01
7.32162535e-01 1.62584260e-01 7.26662725e-02 -9.02637303e-01
1.89628214e-01 4.12707210e-01 9.58248600e-02 -5.99894226e-01
5.54632425e-01 3.92193943e-01 5.41803181e-01 -1.08484662e+00
-3.96265358e-01 -7.18145072e-01 -7.11951196e-01 5.33577204e-02
5.57959080e-01 -5.92253804e-01 -6.41928613e-01 -1.98553741e-01
-6.82448924e-01 -2.11728752e-01 -3.74817938e-01 7.02748716e-01
-7.95404971e-01 4.13123399e-01 -2.45809685e-02 -1.03122377e+00
-1.15174688e-01 -1.07649446e+00 1.06726170e+00 -8.37320760e-02
-4.42952812e-01 -8.71427000e-01 3.96125555e-01 1.67909101e-01
1.99879736e-01 8.07612479e-01 5.30280471e-01 -1.01258647e+00
7.81049058e-02 -7.90117085e-02 -3.50169763e-02 1.42295435e-01
3.40798125e-02 1.39615268e-01 -9.48985398e-01 -3.20749372e-01
1.46140426e-01 -1.11293167e-01 9.80425417e-01 7.02525437e-01
1.04844248e+00 -3.68987709e-01 -4.76583332e-01 7.23829091e-01
1.60614753e+00 -6.06651232e-03 3.63627642e-01 3.01436633e-01
5.66029429e-01 9.22773182e-01 7.14827836e-01 8.14870894e-01
-3.01206529e-01 8.39159012e-01 2.51231194e-01 -2.20847026e-01
4.60888088e-01 -5.12302294e-02 2.76823252e-01 3.93377662e-01
-8.84339288e-02 2.46271938e-01 -1.03277087e+00 -1.28845215e-01
-1.97880781e+00 -8.62843633e-01 -7.88822472e-01 2.43922567e+00
7.46605754e-01 -1.46904022e-01 3.40745687e-01 3.42967957e-01
4.91399586e-01 1.04109585e-01 -5.56556344e-01 -4.82650489e-01
-3.17207634e-01 1.79298446e-01 8.27036381e-01 8.35969329e-01
-1.01573861e+00 5.89440286e-01 5.27478552e+00 6.84545934e-01
-1.05974591e+00 4.65332121e-02 6.26454711e-01 4.06269170e-02
-3.87228310e-01 3.23172539e-01 -8.25552344e-01 2.95811206e-01
8.07705879e-01 -1.47575676e-01 3.24893028e-01 9.56708431e-01
8.02118063e-01 1.47763463e-02 -1.19556451e+00 7.98424780e-01
-2.10279882e-01 -9.14853394e-01 -3.13992530e-01 4.27269280e-01
6.24219656e-01 -1.85503125e-01 7.18696369e-03 -9.81470644e-02
5.77104539e-02 -8.92775297e-01 3.59015137e-01 7.93900907e-01
8.12431276e-01 -4.13471043e-01 3.95504475e-01 1.42079964e-01
-1.00228393e+00 -4.48424518e-02 -3.47915113e-01 5.14865220e-02
1.98183820e-01 1.11966133e+00 -5.64115763e-01 4.70570922e-01
2.67191678e-01 9.75471616e-01 -4.59122479e-01 8.83580327e-01
-1.35179237e-01 5.92939138e-01 -4.38362718e-01 3.67524214e-02
1.85941949e-01 -9.31629658e-01 8.91784668e-01 1.33885264e+00
2.25260973e-01 7.73211643e-02 2.40778714e-01 1.20303667e+00
3.67739707e-01 6.53459787e-01 -7.89540410e-01 -5.47013655e-02
3.19713950e-01 1.35831952e+00 -6.10868394e-01 -7.99791962e-02
-4.51851755e-01 5.98725855e-01 -5.47018722e-02 5.09177685e-01
-4.25583631e-01 -2.49567293e-02 5.67486525e-01 3.06253821e-01
1.89653680e-01 -2.04420313e-01 -4.40645874e-01 -8.85984302e-01
-2.49686539e-01 -1.06650853e+00 3.72603625e-01 -5.99214554e-01
-1.25496447e+00 2.71465242e-01 -1.21567264e-01 -1.11876297e+00
2.42349948e-03 -9.31387544e-01 -1.78540260e-01 1.04304934e+00
-1.43896866e+00 -8.68655562e-01 -1.35588974e-01 6.42638981e-01
3.30471694e-01 -4.72919978e-02 9.31134403e-01 1.82316661e-01
-7.80121803e-01 8.97860602e-02 3.99785608e-01 -5.77576578e-01
9.49218094e-01 -1.62599373e+00 -3.51031899e-01 5.40892482e-01
-2.51214206e-01 9.18450296e-01 1.18866491e+00 -6.69734180e-01
-1.70685709e+00 -7.95761585e-01 9.23412740e-01 -6.97275773e-02
9.21922803e-01 -2.89579064e-01 -9.29238081e-01 5.86127877e-01
-2.95186579e-01 -9.91106927e-02 1.31661451e+00 2.65099943e-01
-3.09397936e-01 -1.79252997e-01 -1.06408930e+00 2.95679957e-01
5.32261431e-01 -2.46849388e-01 -1.82819128e-01 4.99959260e-01
4.46731806e-01 1.19498827e-01 -1.28962231e+00 3.93411845e-01
6.61303401e-01 -9.27162528e-01 1.05694926e+00 -8.82233977e-01
3.59892577e-01 -4.30455923e-01 -3.81836742e-01 -9.05668020e-01
-7.11616158e-01 -9.23674405e-01 -1.00723803e-01 1.11221838e+00
6.68373644e-01 -7.29162276e-01 7.90021896e-01 6.61027014e-01
5.80608696e-02 -7.33677685e-01 -7.52209067e-01 -7.78178513e-01
-7.39895627e-02 -2.01584131e-01 3.44912827e-01 1.34381521e+00
1.98381245e-01 2.94408053e-01 -5.62214971e-01 1.08486200e-02
9.49371874e-01 2.55484164e-01 8.35239470e-01 -1.53095126e+00
-4.72742736e-01 -6.34503126e-01 -3.57735217e-01 -6.10311806e-01
8.05192813e-02 -9.04118776e-01 -2.25250036e-01 -1.09340882e+00
1.11147031e-01 -5.23727953e-01 -4.29451168e-01 6.64213777e-01
-1.36240512e-01 1.08528405e-01 -2.05890089e-02 4.01108772e-01
-1.96220785e-01 2.84852922e-01 8.98669064e-01 1.14426538e-01
-6.20266199e-01 1.37941450e-01 -8.35181832e-01 5.26450157e-01
1.19777322e+00 -7.51616776e-01 -5.24249263e-02 3.19341540e-01
5.88510275e-01 2.90234208e-01 9.21069458e-02 -6.43490434e-01
1.72855616e-01 -5.41527569e-01 1.44207880e-01 -2.36538783e-01
3.49418372e-01 -1.03793073e+00 6.76054418e-01 5.81798494e-01
-6.31279528e-01 -1.00616947e-01 -1.89380571e-01 5.79643071e-01
-1.59439921e-01 -3.72908503e-01 7.61189222e-01 -1.70996323e-01
-4.26876128e-01 1.46454871e-01 -1.86289139e-02 -5.32434464e-01
9.46821153e-01 -3.79584938e-01 5.75989000e-02 -1.11226290e-02
-1.12136805e+00 6.64230809e-02 4.59354639e-01 -3.11413612e-02
4.84744191e-01 -1.18211186e+00 -9.30233598e-01 1.65587068e-01
2.23945156e-01 -3.76441747e-01 -1.59686431e-01 1.53293025e+00
-2.00005323e-01 5.75841784e-01 -2.78179124e-02 -6.31078660e-01
-1.37406206e+00 8.62166524e-01 7.74989724e-02 -2.14170545e-01
-2.98162490e-01 6.32708609e-01 2.71131158e-01 -4.48178321e-01
-2.35344432e-02 1.25027820e-01 -3.66663396e-01 3.70452523e-01
5.51282167e-01 6.39427245e-01 -2.52177358e-01 -6.56893432e-01
-5.81717372e-01 6.05222702e-01 2.70996958e-01 1.22026622e-01
1.89918804e+00 -3.28745134e-02 -5.69599867e-01 7.54058719e-01
1.24383128e+00 3.24097544e-01 -9.75140810e-01 -2.29326636e-01
3.39103848e-01 -4.21050698e-01 3.03160190e-03 -6.42940402e-01
-7.90988743e-01 5.89635372e-01 3.76738787e-01 3.28404516e-01
1.07706177e+00 -1.32373273e-01 -2.06604414e-02 3.39420587e-01
1.34126112e-01 -6.62979662e-01 -3.64212096e-01 6.88074902e-02
1.09408104e+00 -1.28542709e+00 5.47237039e-01 -5.44526637e-01
-6.93383396e-01 1.27263832e+00 3.17874067e-02 -1.82770133e-01
5.63585401e-01 2.95624733e-01 -2.64709413e-01 -3.18280309e-01
-4.76964891e-01 -1.15226554e-02 2.08016828e-01 2.58806825e-01
9.49946225e-01 3.45865279e-01 -1.09074438e+00 7.30254948e-01
1.27857560e-02 4.94371355e-02 2.46152639e-01 7.94805348e-01
-4.83334035e-01 -1.38975453e+00 -6.63621604e-01 4.48445231e-01
-3.46462071e-01 -9.27675664e-02 -6.71191335e-01 5.53623796e-01
-3.53400744e-02 7.32233524e-01 -5.38684249e-01 -8.80671591e-02
2.58349210e-01 1.74290344e-01 8.83111879e-02 -4.10752565e-01
-4.62360740e-01 5.11570096e-01 4.84313443e-02 -8.53360653e-01
-7.08781302e-01 -1.10077047e+00 -8.24504495e-01 -1.87270090e-01
-4.30103749e-01 3.82859200e-01 9.19319868e-01 4.24563646e-01
1.34643599e-01 1.95883453e-01 8.28156114e-01 -8.66270959e-01
-5.87183952e-01 -7.19984114e-01 -1.09966230e+00 2.35096440e-01
3.26419592e-01 -6.57007515e-01 -4.96645331e-01 1.17550001e-01] | [7.734650135040283, 4.478766441345215] |
e5ed0508-1420-42c9-b08e-d0279f1c953c | learning-sentiment-lexicons-in-spanish | null | null | https://aclanthology.org/L12-1645 | https://aclanthology.org/L12-1645.pdf | Learning Sentiment Lexicons in Spanish | In this paper we present a framework to derive sentiment lexicons in a target language by using manually or automatically annotated data available in an electronic resource rich language, such as English. We show that bridging the language gap using the multilingual sense-level aligned WordNet structure allows us to generate a high accuracy (90{\%}) polarity lexicon comprising 1,347 entries, and a disjoint lower accuracy (74{\%}) one encompassing 2,496 words. By using an LSA-based vectorial expansion for the generated lexicons, we are able to obtain an average F-measure of 66{\%} in the target language. This implies that the lexicons could be used to bootstrap higher-coverage lexicons using in-language resources. | ["Ver{\\'o}nica P{\\'e}rez-Rosas", 'Rada Mihalcea', 'Carmen Banea'] | 2012-05-01 | null | null | null | lrec-2012-5 | ['subjectivity-analysis'] | ['natural-language-processing'] | [ 1.63143203e-01 5.23171008e-01 -2.70897210e-01 -2.07779363e-01
-8.83602083e-01 -1.12245297e+00 4.55452561e-01 7.74372935e-01
-8.15483630e-01 1.16697085e+00 4.78095025e-01 -3.64816129e-01
4.09385301e-02 -8.33795905e-01 -3.65835577e-01 -5.37349433e-02
5.37326992e-01 6.46853089e-01 3.69387791e-02 -9.78910506e-01
3.06370676e-01 1.21016227e-01 -1.50322235e+00 2.24167690e-01
9.29158032e-01 6.72777653e-01 2.81794369e-01 2.36490384e-01
-5.30715525e-01 5.41213036e-01 -7.04784155e-01 -1.12257934e+00
3.28770168e-02 -2.14825302e-01 -8.35913181e-01 -4.80750203e-01
-1.15820996e-01 4.40040529e-01 6.55800104e-01 1.16117299e+00
1.71968833e-01 -1.39501393e-01 6.25914335e-01 -6.56063616e-01
-4.53163475e-01 1.01515770e+00 -9.62770283e-02 -1.73695341e-01
8.28445256e-01 -4.10911769e-01 1.54056990e+00 -9.08337235e-01
1.33543527e+00 7.43548632e-01 6.45286083e-01 2.47810006e-01
-1.03394485e+00 -6.35772467e-01 7.72218257e-02 -2.37605363e-01
-1.38477647e+00 -3.02518576e-01 7.34372973e-01 -3.62786323e-01
1.40229154e+00 1.73379704e-01 9.74037290e-01 8.03406835e-01
1.47180751e-01 1.86518192e-01 1.41101599e+00 -1.10403407e+00
-1.38628483e-03 7.39456594e-01 1.45773903e-01 3.95784646e-01
8.38593721e-01 -5.18173039e-01 -5.60325682e-01 -2.54941463e-01
2.42463693e-01 -7.69403160e-01 5.24036027e-02 1.69046536e-01
-1.24552405e+00 1.04189813e+00 -7.62882456e-02 9.28908646e-01
-4.49756652e-01 -6.71554267e-01 7.38389015e-01 4.13401961e-01
7.69656718e-01 1.09762490e+00 -9.73584294e-01 -1.57903016e-01
-8.44461441e-01 2.06549510e-01 1.28822148e+00 1.02183592e+00
7.96344340e-01 -2.39542797e-02 2.60032564e-01 1.01320601e+00
1.34876117e-01 8.51494849e-01 7.55537748e-01 -4.99499470e-01
4.44921076e-01 9.39138710e-01 2.91767508e-01 -9.18755591e-01
-5.21820724e-01 -4.12020147e-01 -4.55108643e-01 -5.13645172e-01
2.43162170e-01 -3.66996139e-01 -5.48403621e-01 1.89133489e+00
9.36848968e-02 -5.91297448e-01 6.14075065e-01 2.80488372e-01
8.68390679e-01 3.91795039e-01 6.72950894e-02 -6.70430481e-01
1.61129463e+00 -3.57898980e-01 -7.16371059e-01 -3.54866415e-01
8.73519778e-01 -1.10784376e+00 1.25366890e+00 4.71420616e-01
-8.92711103e-01 -3.19055468e-01 -1.19721174e+00 7.93061703e-02
-9.21269298e-01 6.02154583e-02 6.18422687e-01 9.66415703e-01
-1.31245816e+00 9.01997741e-03 -2.43158132e-01 -5.29282510e-01
-1.98974118e-01 2.60883749e-01 -5.44423640e-01 1.30826458e-01
-1.63035011e+00 1.33736944e+00 6.45384908e-01 -4.42801833e-01
-5.34620471e-02 -3.48029077e-01 -1.04985118e+00 -3.41193169e-01
2.84961522e-01 -2.83708036e-01 7.54748762e-01 -9.06329811e-01
-1.27684665e+00 1.41736948e+00 -2.80642152e-01 -2.79603124e-01
-5.16489297e-02 8.77804086e-02 -7.94438362e-01 -6.91454485e-02
6.46203339e-01 3.39853913e-01 9.89695415e-02 -1.05092061e+00
-6.20216250e-01 -1.80872872e-01 2.16336604e-02 3.11308503e-01
-6.56277299e-01 3.49921793e-01 -3.17465335e-01 -6.58273876e-01
-4.82224580e-03 -9.17058647e-01 -2.59600788e-01 -1.04319358e+00
-2.26174057e-01 -2.38741077e-02 -3.15388590e-01 -9.55982745e-01
1.41364968e+00 -1.45481682e+00 7.21387714e-02 6.01910770e-01
-9.86105949e-02 2.00966284e-01 -1.83437735e-01 7.99216032e-01
3.18238586e-02 3.20556492e-01 -5.83543107e-02 5.24471551e-02
-5.00413626e-02 1.83301657e-01 -4.14859265e-01 6.22092821e-02
1.12538384e-02 7.84538627e-01 -1.02542901e+00 -4.97022301e-01
1.03499271e-01 4.66610700e-01 -5.16439617e-01 -3.78481776e-01
-4.58889641e-02 1.67362526e-01 -2.92431593e-01 6.02894962e-01
1.32992655e-01 2.47412160e-01 7.74736643e-01 -5.82800955e-02
-2.91188031e-01 5.61363101e-01 -1.03136766e+00 1.63910401e+00
-1.05396056e+00 4.59455162e-01 -2.24939600e-01 -6.49992526e-01
1.27966368e+00 3.01990539e-01 4.53824461e-01 -8.16918552e-01
3.88625264e-01 6.71757579e-01 -2.59963423e-02 -3.99438664e-02
9.18760538e-01 -3.69892538e-01 -7.96363890e-01 2.70279914e-01
4.68185723e-01 -6.39036894e-01 7.12574065e-01 6.58999011e-02
7.74712086e-01 -9.92746204e-02 7.65742600e-01 -8.38881552e-01
9.33063209e-01 5.69454789e-01 4.11843747e-01 3.09829891e-01
1.53892487e-01 1.82322562e-01 4.00219947e-01 -3.01636338e-01
-1.18830395e+00 -6.28096104e-01 -4.48645920e-01 1.00430143e+00
-2.38426358e-01 -9.24413025e-01 -7.55266249e-01 -4.53213155e-01
-3.83969843e-01 9.93137717e-01 -4.60215271e-01 1.52321249e-01
-4.72810864e-01 -9.12646055e-01 6.57487810e-01 5.90225160e-02
-3.87335867e-02 -1.24446654e+00 -1.36652485e-01 4.85958427e-01
-5.74021161e-01 -1.28705609e+00 7.78581947e-02 1.98142901e-01
-3.04297477e-01 -9.03924048e-01 -4.59568083e-01 -9.45683181e-01
3.92358214e-01 -5.18142700e-01 1.62026083e+00 -3.05400044e-01
1.34060875e-01 3.40489112e-02 -7.18825996e-01 -7.83053577e-01
-7.01433003e-01 4.49332863e-01 2.52008289e-01 -4.18525755e-01
8.59151244e-01 -3.20913613e-01 -3.20832580e-02 -1.08661808e-01
-8.83063257e-01 -3.04662526e-01 2.68405229e-01 6.52704835e-01
9.00762677e-01 -3.14903051e-01 9.34145570e-01 -1.15902412e+00
1.13866007e+00 -3.09401751e-01 -6.66584790e-01 1.99488550e-01
-9.58698452e-01 7.43877143e-02 8.02858412e-01 -2.33651221e-01
-8.20872128e-01 4.00308147e-03 -8.00546527e-01 6.40901506e-01
2.99746059e-02 9.33400750e-01 -1.55164525e-02 7.73830414e-02
8.09128463e-01 1.40844166e-01 -4.12910789e-01 -1.48372799e-01
6.26826465e-01 7.54913449e-01 2.26536959e-01 -4.42326337e-01
4.38739449e-01 1.71609595e-01 -3.84552389e-01 -8.83149922e-01
-1.01683223e+00 -6.14420593e-01 -6.10189199e-01 -8.97894800e-02
7.91422904e-01 -1.12384129e+00 -3.84782523e-01 -1.31736279e-01
-8.37248564e-01 4.53695618e-02 -4.53051150e-01 5.23688912e-01
-3.89652312e-01 -7.49521656e-03 -2.50811428e-01 -6.45850003e-01
-7.45001435e-01 -9.65347469e-01 1.01302373e+00 -5.33453412e-02
-9.99866545e-01 -1.13999712e+00 5.35672784e-01 4.15951431e-01
2.21630216e-01 1.49529278e-01 8.90715182e-01 -1.17441809e+00
4.77992833e-01 -3.72020870e-01 9.70202684e-02 3.90922010e-01
1.45436913e-01 -1.82780445e-01 -6.44409657e-01 -1.61202773e-01
-8.20696279e-02 -4.07103181e-01 3.99433464e-01 3.52132320e-02
2.01777905e-01 -1.93465844e-01 -6.55378327e-02 2.28888281e-02
1.74832678e+00 2.06025600e-01 5.08160174e-01 6.42891586e-01
4.32277799e-01 7.69591630e-01 7.54290819e-01 2.25253358e-01
7.24840224e-01 3.94189745e-01 -2.23518834e-01 2.10157618e-01
1.39868215e-01 -2.88199812e-01 4.50728774e-01 1.55218542e+00
-1.11581318e-01 -8.01056325e-02 -1.22551882e+00 1.07417059e+00
-1.31953776e+00 -5.76274335e-01 -7.97169656e-02 2.04865742e+00
1.39851832e+00 3.95534575e-01 -1.62100941e-02 1.40173376e-01
5.60846686e-01 1.14327306e-02 -5.14509715e-02 -7.10810065e-01
-6.51640475e-01 8.11944067e-01 6.95597768e-01 8.56912494e-01
-7.43881702e-01 1.54348075e+00 6.64719343e+00 9.23044205e-01
-8.89618814e-01 1.31246924e-01 2.86947191e-01 1.06944658e-01
-8.50162745e-01 1.82996958e-01 -1.00618589e+00 2.74021119e-01
1.11704469e+00 -6.14360809e-01 2.44114488e-01 7.61976123e-01
-1.48985684e-01 -1.46448061e-01 -3.52493197e-01 9.22863781e-01
4.08515394e-01 -1.36465967e+00 1.50152370e-01 -8.78899619e-02
1.00758255e+00 2.59345055e-01 -3.87678266e-01 4.74565595e-01
5.99260092e-01 -9.12489176e-01 5.50626993e-01 3.32026243e-01
1.31587696e+00 -1.00506556e+00 1.05062532e+00 3.29701304e-01
-1.28587198e+00 4.18566465e-01 -3.56232792e-01 -2.19769135e-01
2.28762686e-01 8.21533501e-01 -8.12927485e-01 6.88612163e-01
6.24552071e-01 3.77272874e-01 -6.70355797e-01 3.28777909e-01
-2.92373121e-01 5.51302671e-01 -2.89801776e-01 -6.35168970e-01
1.90803081e-01 -5.22293508e-01 6.10228717e-01 1.42593801e+00
2.79748708e-01 -3.14243972e-01 1.15402952e-01 2.23844245e-01
-3.56073737e-01 1.24492693e+00 -7.64129460e-01 -9.70312431e-02
4.14298177e-01 1.37994778e+00 -1.05362332e+00 -5.66164374e-01
-4.06374097e-01 8.25300276e-01 2.95666188e-01 1.71996970e-02
-3.57955635e-01 -7.93654859e-01 5.02164304e-01 -2.35404968e-01
-5.48160337e-02 1.53209660e-02 -4.41414356e-01 -1.38793182e+00
2.06528410e-01 -1.04890370e+00 1.54367164e-01 -6.95932388e-01
-1.31489325e+00 1.25771749e+00 -5.84881790e-02 -8.58652771e-01
-5.83197236e-01 -8.34747434e-01 1.97682172e-01 1.23909712e+00
-1.21646547e+00 -1.14962709e+00 2.48074859e-01 3.50931287e-01
1.88324094e-01 -4.27414477e-01 1.28796148e+00 1.92951694e-01
-7.10024163e-02 4.83459711e-01 3.40755321e-02 1.23327039e-01
8.26731861e-01 -1.27938032e+00 1.95525497e-01 5.92877209e-01
4.08437788e-01 8.32095385e-01 8.77282619e-01 -8.86722326e-01
-7.80296445e-01 -9.03887928e-01 1.94385564e+00 -6.60182238e-01
1.01217282e+00 -5.09738028e-01 -5.19177794e-01 5.30126870e-01
5.07975459e-01 -7.30731010e-01 1.17764616e+00 3.78330350e-01
-3.14266950e-01 5.58395647e-02 -1.31933284e+00 7.90075719e-01
8.16645324e-01 -8.38474333e-01 -9.07961309e-01 4.22243834e-01
8.25757563e-01 -2.55172014e-01 -1.11769593e+00 4.54175025e-01
6.06261134e-01 -5.56089163e-01 7.26729453e-01 -3.68139118e-01
2.97751367e-01 -1.09593168e-01 -3.34497780e-01 -1.40119028e+00
1.05366215e-01 -3.79882663e-01 6.02831662e-01 1.21368003e+00
1.11386132e+00 -8.92979741e-01 2.67843992e-01 2.16153502e-01
1.14253610e-01 -6.92327619e-01 -8.40840638e-01 -3.89037013e-01
4.53775585e-01 -8.91783893e-01 6.80211961e-01 1.10218334e+00
4.07030225e-01 7.49357402e-01 -2.71553695e-02 -2.60699183e-01
-1.19844154e-01 -6.32941499e-02 2.51897901e-01 -1.43372321e+00
2.98702657e-01 -3.52902591e-01 -3.32262635e-01 -2.35447660e-01
6.60059035e-01 -1.15821016e+00 -2.39462540e-01 -1.63077438e+00
4.37900051e-02 -4.80529934e-01 -3.59398186e-01 5.20681977e-01
5.90335093e-02 7.17396200e-01 -1.58548936e-01 -1.43274605e-01
-3.64924043e-01 1.43708095e-01 8.51156235e-01 1.97461948e-01
-3.55373979e-01 -6.17699623e-01 -1.38928473e+00 1.01986980e+00
8.72756004e-01 -5.97799122e-01 -2.46507347e-01 1.83031000e-02
1.32781589e+00 -4.36576366e-01 -3.87115926e-01 -6.49035752e-01
-1.90674126e-01 -1.51660413e-01 1.45018935e-01 -2.25351959e-01
1.79696530e-01 -5.70088744e-01 2.16790643e-02 2.75883466e-01
-2.55404323e-01 3.61523539e-01 5.83383024e-01 -1.37766868e-01
-5.06609976e-01 -5.40093362e-01 3.19150418e-01 -3.36559951e-01
-6.08560920e-01 -2.33630776e-01 -5.53840697e-01 4.83808875e-01
8.75300705e-01 -7.70741031e-02 -5.51962527e-04 -3.31150919e-01
-5.86354375e-01 -2.17665270e-01 7.03852534e-01 5.83326101e-01
2.00671300e-01 -1.13724828e+00 -8.62457454e-01 1.22859612e-01
5.98169506e-01 -7.15840816e-01 -1.70987815e-01 3.30027342e-01
-5.97583175e-01 7.38655210e-01 -3.05011153e-01 1.29898921e-01
-9.20276403e-01 2.07780942e-01 -1.50615692e-01 -6.42825007e-01
3.97165399e-03 6.71389461e-01 -5.99326432e-01 -9.82340991e-01
-4.08378214e-01 -1.13643087e-01 -8.22810292e-01 6.50522053e-01
2.92287350e-01 -1.37400225e-01 4.18148458e-01 -1.36020184e+00
-5.89373112e-01 5.91751277e-01 3.15308362e-01 -7.38475204e-01
1.25048995e+00 -1.07764654e-01 -5.37671506e-01 5.81032634e-01
8.84049654e-01 1.02802134e+00 8.72830823e-02 1.01560958e-01
2.44967759e-01 1.40528157e-01 -1.59588203e-01 -9.90252376e-01
-5.80156446e-01 7.83293992e-02 1.76822305e-01 3.08776170e-01
1.11484766e+00 5.38871391e-03 5.42876780e-01 4.70915198e-01
7.26418972e-01 -1.48372149e+00 -6.98405385e-01 1.02299368e+00
8.06431055e-01 -1.07110751e+00 -1.85105965e-01 -5.06153226e-01
-1.01536572e+00 7.69175529e-01 7.66033977e-02 -1.68380886e-01
7.39921451e-01 3.21281463e-01 5.79269886e-01 -2.64789909e-01
-5.32951176e-01 -5.90719998e-01 3.52255523e-01 6.38045013e-01
9.88567948e-01 1.37618616e-01 -1.25662923e+00 1.07932734e+00
-9.53205287e-01 -2.83356160e-01 6.42576993e-01 6.40609443e-01
-1.27818644e-01 -1.65131593e+00 -9.28955600e-02 6.26610100e-01
-8.67520094e-01 -7.45234370e-01 -7.58317232e-01 8.08295906e-01
3.60923469e-01 1.12514603e+00 -2.36702457e-01 -3.08608145e-01
4.53294396e-01 4.08449501e-01 2.40865529e-01 -8.44705760e-01
-8.23987842e-01 -3.25596263e-03 7.10896373e-01 -2.24266931e-01
-7.70597100e-01 -5.65664411e-01 -1.03868616e+00 9.43352208e-02
-3.02845895e-01 4.91360694e-01 7.39845514e-01 1.08564985e+00
1.49046436e-01 2.64481455e-01 1.95663303e-01 -3.79366875e-01
1.23111568e-01 -1.11064100e+00 -5.22321761e-01 3.57071161e-01
-2.97668755e-01 -3.82105440e-01 -2.61876643e-01 7.69709470e-03] | [10.466434478759766, 9.699620246887207] |
55ee6d9f-daa6-42d7-9dfd-ab356786b618 | temp-temporal-message-passing-for-temporal | 2010.03526 | null | https://arxiv.org/abs/2010.03526v1 | https://arxiv.org/pdf/2010.03526v1.pdf | TeMP: Temporal Message Passing for Temporal Knowledge Graph Completion | Inferring missing facts in temporal knowledge graphs (TKGs) is a fundamental and challenging task. Previous works have approached this problem by augmenting methods for static knowledge graphs to leverage time-dependent representations. However, these methods do not explicitly leverage multi-hop structural information and temporal facts from recent time steps to enhance their predictions. Additionally, prior work does not explicitly address the temporal sparsity and variability of entity distributions in TKGs. We propose the Temporal Message Passing (TeMP) framework to address these challenges by combining graph neural networks, temporal dynamics models, data imputation and frequency-based gating techniques. Experiments on standard TKG tasks show that our approach provides substantial gains compared to the previous state of the art, achieving a 10.7% average relative improvement in Hits@10 across three standard benchmarks. Our analysis also reveals important sources of variability both within and across TKG datasets, and we introduce several simple but strong baselines that outperform the prior state of the art in certain settings. | ['William L. Hamilton', 'Jackie Chi Kit Cheung', 'Meng Cao', 'Jiapeng Wu'] | 2020-10-07 | null | https://aclanthology.org/2020.emnlp-main.462 | https://aclanthology.org/2020.emnlp-main.462.pdf | emnlp-2020-11 | ['temporal-knowledge-graph-completion'] | ['knowledge-base'] | [-2.52237003e-02 1.85094208e-01 -8.93806040e-01 -1.12138398e-01
-6.81967914e-01 -3.67866606e-01 7.71392226e-01 3.98609072e-01
-4.88481522e-02 1.04674900e+00 6.42332733e-01 -3.58128846e-01
-4.67552692e-01 -1.07462656e+00 -1.05302823e+00 -2.46992886e-01
-6.85818672e-01 4.99551117e-01 6.10959947e-01 -2.77793020e-01
1.36075065e-01 2.46061403e-02 -1.20308793e+00 4.33512151e-01
7.59561956e-01 8.05879414e-01 -2.64161855e-01 4.92700309e-01
-3.54662836e-01 1.37906766e+00 -3.49540412e-01 -4.20408100e-01
-5.88551573e-02 -7.31207579e-02 -1.05559289e+00 -4.09677625e-01
3.59437048e-01 -1.54830962e-01 -1.09983051e+00 5.72200477e-01
1.92350522e-01 3.41363758e-01 2.75430143e-01 -1.24389112e+00
-9.00849640e-01 1.04410803e+00 -6.78389549e-01 7.25405037e-01
3.81902903e-01 5.96859790e-02 1.09686875e+00 -5.59436500e-01
7.47052312e-01 9.82826173e-01 1.02309525e+00 3.80214825e-02
-1.32773852e+00 -5.73815465e-01 8.70848596e-01 6.49970531e-01
-1.42405796e+00 -3.39401782e-01 6.10439062e-01 -3.01073730e-01
1.65909481e+00 -9.29103792e-02 6.23743236e-01 1.36119246e+00
1.96886018e-01 7.93991566e-01 8.68487239e-01 -2.41673604e-01
2.70562712e-02 -6.11629009e-01 4.80879068e-01 8.22078705e-01
3.87919486e-01 1.21379793e-01 -1.30898726e+00 -2.68419027e-01
6.00303888e-01 1.74192097e-02 -1.95027024e-01 -1.72321275e-01
-1.31237233e+00 6.13921165e-01 4.38577622e-01 1.49983451e-01
-4.41744387e-01 8.18165660e-01 3.42667401e-01 2.70746440e-01
7.94103026e-01 6.71679378e-02 -8.64669979e-01 -3.89536560e-01
-8.91196966e-01 1.67668223e-01 1.03336191e+00 9.99554217e-01
6.60712242e-01 6.66779801e-02 -5.98114848e-01 2.15237632e-01
8.15802664e-02 8.86164680e-02 2.36143470e-01 -6.78254962e-01
6.74930751e-01 6.59083724e-01 1.37172326e-01 -1.00433004e+00
-3.99743557e-01 -5.93983769e-01 -4.97737169e-01 -5.86187959e-01
3.81629080e-01 4.05135825e-02 -1.34725761e+00 1.91272259e+00
3.39790970e-01 1.06620085e+00 -1.46847770e-01 4.32912856e-01
1.00119638e+00 6.51142836e-01 2.46335849e-01 -1.17587119e-01
8.95284593e-01 -1.10915899e+00 -8.50428820e-01 -4.26728815e-01
6.25215352e-01 -2.58083701e-01 8.32279265e-01 1.51682422e-01
-9.88164485e-01 -1.85118876e-02 -8.37188303e-01 1.16608310e-02
-5.87930858e-01 -3.12944800e-01 1.19716179e+00 4.21100259e-01
-1.25922871e+00 8.82494390e-01 -1.36658883e+00 -3.38876843e-01
4.00875688e-01 3.68379563e-01 -1.30574033e-01 -1.18797027e-01
-1.53923714e+00 8.76205266e-01 6.51769698e-01 -5.32413013e-02
-8.89380574e-01 -1.20415354e+00 -8.05721045e-01 1.90057516e-01
9.60036039e-01 -8.99260998e-01 1.14470673e+00 -3.12074870e-01
-1.30029774e+00 2.36464649e-01 -4.21596169e-01 -8.59021723e-01
3.16837072e-01 -4.25592393e-01 -6.99535608e-01 -1.32639706e-01
1.27794743e-02 1.32900164e-01 5.02037108e-01 -8.64439964e-01
-3.76850665e-01 -4.69819084e-02 1.97878510e-01 -3.14645953e-02
-2.38437966e-01 -5.33078492e-01 -9.68211889e-01 -7.39242554e-01
-1.97133973e-01 -6.36005521e-01 -2.46230245e-01 -6.13276482e-01
-5.66550374e-01 -3.50188822e-01 7.32198596e-01 -7.68313527e-01
1.63354099e+00 -1.57664347e+00 5.44782653e-02 2.02325314e-01
3.61712784e-01 -7.87959844e-02 -7.82360360e-02 8.81996870e-01
2.63407290e-01 1.41139150e-01 -5.18385842e-02 -3.96194965e-01
3.17534953e-02 4.65106428e-01 -5.76342642e-01 9.53889936e-02
1.14343993e-01 1.40005839e+00 -1.25822496e+00 -3.73768657e-01
-2.25676466e-02 5.45736492e-01 -4.38708007e-01 -2.92139858e-01
-8.32657337e-01 3.18219095e-01 -5.25258720e-01 7.71154284e-01
2.89343685e-01 -1.05451655e+00 7.34130681e-01 -9.53737795e-02
2.59865850e-01 7.89905131e-01 -8.87754023e-01 1.95097876e+00
-2.64657855e-01 3.31817716e-01 -5.11002183e-01 -9.57531571e-01
3.93121362e-01 1.41573936e-01 5.58376491e-01 -8.10467780e-01
-3.31785083e-01 -1.32259220e-01 -3.46386582e-01 -2.37910286e-01
8.51999402e-01 2.69447595e-01 -4.90607247e-02 5.02940714e-01
2.58206964e-01 5.35597503e-01 1.97171539e-01 7.57313311e-01
1.83092642e+00 3.77076298e-01 1.06908128e-01 3.51378620e-02
-2.15770990e-01 1.00581169e-01 8.62135351e-01 1.07242060e+00
-1.08548999e-02 2.16304779e-01 6.46864474e-01 -5.01780152e-01
-4.65629727e-01 -1.19798887e+00 5.44576526e-01 1.11711669e+00
1.97038829e-01 -8.70644212e-01 6.87064901e-02 -1.01682520e+00
5.04409373e-01 7.09809959e-01 -8.82157087e-01 -2.84271955e-01
-6.23951733e-01 -9.69227672e-01 4.40309793e-01 9.51675594e-01
3.30040991e-01 -8.10013771e-01 6.52602836e-02 6.30854487e-01
-4.88656074e-01 -1.41013026e+00 -2.98471838e-01 1.68873612e-02
-1.07987547e+00 -1.10671020e+00 -1.27970111e-02 -2.95584917e-01
1.66825280e-01 2.70102262e-01 1.68988717e+00 -2.89342310e-02
-1.37473032e-01 6.46946669e-01 -4.33442086e-01 -1.71004787e-01
1.87287152e-01 4.71832544e-01 -1.72703445e-01 -2.08943248e-01
3.63697141e-01 -9.09456670e-01 -7.93671906e-01 -3.92286181e-02
-5.85851729e-01 -2.18239464e-02 5.90613842e-01 7.87814260e-01
6.63176477e-01 1.18498236e-01 9.15335536e-01 -1.23096871e+00
3.38420004e-01 -9.40241456e-01 -4.64121014e-01 3.99794161e-01
-9.69737768e-01 2.99982071e-01 2.76245266e-01 -3.73656601e-01
-1.18533421e+00 -2.95553207e-01 3.68334115e-01 -4.46130216e-01
3.20537716e-01 1.14658618e+00 3.59359115e-01 -6.15743101e-02
5.15761614e-01 3.30857575e-01 -3.02184671e-01 -3.73856992e-01
6.67011023e-01 -2.65090704e-01 5.78212440e-01 -7.34449446e-01
6.82060421e-01 7.10517406e-01 -6.13581762e-02 -2.45425463e-01
-1.27646482e+00 -5.22262990e-01 -3.50933373e-01 -1.79700796e-02
3.30559701e-01 -1.15566778e+00 -7.42062569e-01 2.40121990e-01
-8.32773924e-01 -7.71687090e-01 -3.27933848e-01 3.09135795e-01
-3.16385448e-01 3.86065930e-01 -8.93894255e-01 -6.28382921e-01
-4.22037989e-01 -4.80612218e-01 9.42869663e-01 -5.50983883e-02
-1.70936003e-01 -1.40143907e+00 1.52166009e-01 3.46407980e-01
5.99450350e-01 5.76691806e-01 9.19837177e-01 -4.54483509e-01
-1.32422984e+00 -6.45405147e-04 -1.52116954e-01 -4.78520066e-01
1.48327276e-01 -1.69623122e-01 -7.63375878e-01 -2.65111864e-01
-7.98471808e-01 -3.48482341e-01 1.54352677e+00 5.41057050e-01
1.17810988e+00 -4.37879205e-01 -8.59765410e-01 4.68805820e-01
1.50559092e+00 -2.17155829e-01 5.63278854e-01 1.68643147e-01
9.19565618e-01 8.53378177e-02 5.70302129e-01 4.73545998e-01
1.16036618e+00 5.61131001e-01 4.39315677e-01 2.13024259e-01
-3.97973090e-01 -5.93285918e-01 3.21566582e-01 7.84605503e-01
-1.78428352e-01 -4.83490407e-01 -1.01582396e+00 1.26998329e+00
-2.40700245e+00 -1.12160921e+00 -2.77767658e-01 2.06198835e+00
1.00345063e+00 2.93199211e-01 7.35012293e-02 -5.54952435e-02
4.92917985e-01 5.76156557e-01 -7.06805348e-01 -3.31783295e-03
1.61257163e-02 3.87470156e-01 8.69237006e-01 3.95226449e-01
-1.06465816e+00 1.20212102e+00 7.25973558e+00 7.81656861e-01
-6.64766729e-01 3.24778616e-01 3.07763517e-01 -3.68012905e-01
-4.62688416e-01 3.24584872e-01 -8.39142799e-01 3.87896776e-01
1.23291564e+00 -4.19123828e-01 5.07121384e-01 3.79612714e-01
-8.07039365e-02 -1.55778229e-01 -1.05481803e+00 5.58420539e-01
-1.47539571e-01 -1.88876247e+00 4.21163104e-02 3.39675881e-02
1.14011443e+00 5.29290974e-01 1.04475385e-02 6.69743001e-01
1.18906033e+00 -1.04910135e+00 4.09205675e-01 9.72241640e-01
5.05948722e-01 -6.17475331e-01 4.85550642e-01 1.50593400e-01
-1.65160823e+00 1.34763002e-01 4.87378240e-02 -3.36016893e-01
3.70301455e-01 9.84643877e-01 -9.94414806e-01 1.21058452e+00
8.15067053e-01 1.12060380e+00 -5.08078933e-01 9.35303807e-01
-2.69816756e-01 1.18636334e+00 -4.66002673e-01 1.41879782e-01
1.77235454e-01 3.47578555e-01 3.69433939e-01 1.17720759e+00
1.70214295e-01 -5.48771657e-02 3.61967683e-01 6.68791115e-01
-4.07206178e-01 -3.96542460e-01 -6.19887710e-01 -3.42654020e-01
7.66482890e-01 8.14427733e-01 -5.81956685e-01 -4.45240110e-01
-6.95220768e-01 7.23131001e-01 7.11509943e-01 6.84701562e-01
-1.00399351e+00 -9.35680941e-02 6.34542823e-01 6.73395693e-02
7.15247929e-01 -5.10815322e-01 -2.20779523e-01 -1.32193112e+00
8.97991508e-02 -3.64414752e-01 1.13482857e+00 -4.10472721e-01
-1.59805453e+00 1.18445143e-01 8.48325118e-02 -6.66573763e-01
-2.28075027e-01 -1.44259296e-02 -5.52007496e-01 6.51262105e-01
-1.90886426e+00 -1.50998425e+00 -1.90211058e-01 8.04040551e-01
1.40027508e-01 2.42629156e-01 5.95213473e-01 2.80272424e-01
-4.28509384e-01 5.14065862e-01 2.17954367e-02 9.93012637e-03
6.00862443e-01 -1.42165112e+00 9.19052899e-01 1.00983810e+00
2.76869774e-01 5.71104765e-01 4.56952959e-01 -1.12229550e+00
-1.77087080e+00 -1.45399475e+00 1.04710019e+00 -8.67774248e-01
1.05227983e+00 -2.13538885e-01 -1.08102942e+00 1.41634846e+00
1.33369327e-01 3.50165188e-01 5.14480293e-01 9.63599563e-01
-5.93239248e-01 -2.79386491e-02 -6.08425498e-01 3.95761520e-01
1.59267592e+00 -5.30885518e-01 -5.34922957e-01 4.73329663e-01
1.15815544e+00 -5.97489238e-01 -1.09319246e+00 5.50816774e-01
3.57730627e-01 -5.96604466e-01 1.18923020e+00 -8.77528608e-01
2.94310361e-01 -2.24458903e-01 1.45379215e-01 -1.19402099e+00
-4.58622277e-01 -8.90388787e-01 -1.33332634e+00 1.19870591e+00
4.53346401e-01 -7.36909986e-01 1.20410454e+00 4.13785219e-01
-1.19608395e-01 -7.37845898e-01 -7.57528782e-01 -1.03679264e+00
-2.38566697e-01 -5.51578999e-01 6.13332689e-01 1.31120670e+00
4.95827459e-02 1.96490854e-01 -7.50919402e-01 4.01742131e-01
6.59736037e-01 3.58717471e-01 6.22012138e-01 -1.15457308e+00
-3.78793120e-01 -1.15889609e-01 -3.13733250e-01 -9.89062428e-01
2.28527322e-01 -1.14281583e+00 -3.13478678e-01 -2.20262575e+00
2.00094923e-01 -3.19248408e-01 -7.33518004e-01 9.05276299e-01
-4.50180322e-01 -1.10873915e-01 -3.05459380e-01 2.05113795e-02
-1.19304562e+00 7.51737297e-01 9.18861449e-01 -1.20689392e-01
-2.91042477e-01 -2.85320878e-01 -7.56147444e-01 4.14783269e-01
6.28438354e-01 -5.38718700e-01 -6.75807774e-01 -6.38715029e-01
6.32362902e-01 2.09966347e-01 5.63984215e-01 -7.00718403e-01
7.21716285e-01 -2.84283072e-01 2.60653675e-01 -9.27815497e-01
4.67020780e-01 -3.14347506e-01 3.83784890e-01 2.86283970e-01
-2.24233791e-02 7.41258860e-02 5.05278587e-01 1.53779197e+00
-2.54235268e-01 7.37862706e-01 -2.02875268e-02 -6.10648170e-02
-1.21137857e+00 6.22439325e-01 -1.15107916e-01 3.82523060e-01
7.87829638e-01 2.89070662e-02 -8.20619226e-01 -5.54522753e-01
-9.07626867e-01 6.86615765e-01 -3.25064696e-02 3.42381060e-01
5.32764256e-01 -1.34847736e+00 -4.24564362e-01 -4.55067307e-01
2.16221169e-01 -1.60288990e-01 4.49810505e-01 1.11966884e+00
-1.60620138e-02 3.96920234e-01 3.21813136e-01 -4.12579745e-01
-6.21183693e-01 6.75612628e-01 8.11402872e-02 -8.99421215e-01
-8.91963243e-01 8.22344244e-01 -3.50995809e-01 -2.65264332e-01
2.04575986e-01 -4.59484875e-01 8.48845467e-02 -9.09029692e-02
7.18025640e-02 4.54842865e-01 2.93451637e-01 1.75079610e-02
-5.53029537e-01 1.35304824e-01 -3.14123631e-01 1.92619696e-01
1.52538455e+00 -4.31350023e-02 1.28002554e-01 5.80357015e-01
7.44154334e-01 -8.06660056e-02 -1.16775703e+00 -7.70680547e-01
3.96379977e-01 -3.99143338e-01 3.81189473e-02 -1.13442373e+00
-1.07476795e+00 3.03413093e-01 -1.00331806e-01 2.57622033e-01
9.55005229e-01 2.06004336e-01 1.07131839e+00 4.74773675e-01
6.69945717e-01 -9.20104921e-01 1.40368640e-01 7.93240905e-01
4.99338120e-01 -1.11882675e+00 1.72519192e-01 -6.68363273e-01
-5.27483344e-01 6.43063962e-01 6.08627796e-01 -1.63954988e-01
8.22638214e-01 2.26124004e-01 -4.03114200e-01 -5.87430537e-01
-1.48976219e+00 -4.42501843e-01 3.04979205e-01 4.25698817e-01
3.16105694e-01 -1.69272888e-02 -2.00503141e-01 6.57167614e-01
8.19205344e-02 3.89575332e-01 3.11594456e-01 1.14900005e+00
-6.82024583e-02 -1.10545802e+00 1.40126273e-01 8.18044245e-01
-5.79865277e-01 -3.73316526e-01 -2.67008752e-01 1.02026808e+00
-4.36907858e-01 8.89186561e-01 -3.82819809e-02 -5.17127573e-01
2.45773464e-01 1.29102945e-01 5.92480123e-01 -5.57620525e-01
-4.94814724e-01 -3.15575749e-01 5.53055108e-01 -9.79934454e-01
-4.80554700e-01 -4.08664554e-01 -1.15018880e+00 -5.90629041e-01
-1.70197025e-01 4.86517735e-02 4.71370555e-02 8.42543244e-01
1.00517392e+00 9.80897963e-01 -9.42356512e-02 -4.53906238e-01
-1.25832558e-01 -7.64191747e-01 -3.56670082e-01 1.88336045e-01
2.44229138e-01 -9.34150457e-01 -1.12355769e-01 -8.29594396e-03] | [8.571038246154785, 7.903960227966309] |
78778948-8447-4248-b7f5-dc3ce0895322 | pseudoreasoner-leveraging-pseudo-labels-for | 2210.07988 | null | https://arxiv.org/abs/2210.07988v1 | https://arxiv.org/pdf/2210.07988v1.pdf | PseudoReasoner: Leveraging Pseudo Labels for Commonsense Knowledge Base Population | Commonsense Knowledge Base (CSKB) Population aims at reasoning over unseen entities and assertions on CSKBs, and is an important yet hard commonsense reasoning task. One challenge is that it requires out-of-domain generalization ability as the source CSKB for training is of a relatively smaller scale (1M) while the whole candidate space for population is way larger (200M). We propose PseudoReasoner, a semi-supervised learning framework for CSKB population that uses a teacher model pre-trained on CSKBs to provide pseudo labels on the unlabeled candidate dataset for a student model to learn from. The teacher can be a generative model rather than restricted to discriminative models as previous works. In addition, we design a new filtering procedure for pseudo labels based on influence function and the student model's prediction to further improve the performance. The framework can improve the backbone model KG-BERT (RoBERTa-large) by 3.3 points on the overall performance and especially, 5.3 points on the out-of-domain performance, and achieves the state-of-the-art. Codes and data are available at https://github.com/HKUST-KnowComp/PseudoReasoner. | ['Simon See', 'Ginny Y. Wong', 'Yangqiu Song', 'Hongming Zhang', 'Quyet V. Do', 'Tianqing Fang'] | 2022-10-14 | null | null | null | null | ['knowledge-base-population'] | ['natural-language-processing'] | [-2.24725664e-01 6.12737775e-01 -5.31369269e-01 -4.18988794e-01
-7.65805066e-01 -5.06447375e-01 3.21112752e-01 -3.85225751e-03
-3.31046820e-01 1.14295161e+00 3.04271262e-02 -2.81188309e-01
-9.94781852e-02 -1.06695092e+00 -9.01965380e-01 -3.88255358e-01
3.00493151e-01 8.78727615e-01 6.03573263e-01 -5.10476232e-01
2.90096607e-02 5.60165569e-02 -1.07709050e+00 4.27049428e-01
1.50804973e+00 9.45869148e-01 6.90736324e-02 2.28274658e-01
-4.85875867e-02 1.45920861e+00 -5.95302641e-01 -7.82931507e-01
-7.11819232e-02 -2.98469722e-01 -1.26078165e+00 -4.17049229e-01
2.71879256e-01 -3.46488476e-01 -6.54543877e-01 1.48320854e+00
3.08539063e-01 2.90796280e-01 9.09282327e-01 -1.56372714e+00
-1.32701278e+00 1.40859735e+00 -8.09426755e-02 1.55118957e-01
9.01581272e-02 -3.43928710e-02 1.13326538e+00 -1.01566279e+00
5.16711354e-01 1.33611393e+00 4.36370552e-01 8.89844775e-01
-9.57307220e-01 -1.05471778e+00 -1.02771029e-01 1.00247741e+00
-1.71347380e+00 -4.40699488e-01 7.27035642e-01 -2.37415820e-01
8.08964312e-01 1.62732810e-01 4.41887826e-01 1.19711590e+00
-3.90093863e-01 1.19371498e+00 1.01988339e+00 -3.64072531e-01
5.07416904e-01 4.75820035e-01 5.32618582e-01 8.23315442e-01
3.91869932e-01 -3.43513995e-01 -4.38595861e-01 -4.46445078e-01
5.54631650e-01 -3.60075161e-02 -3.54457945e-01 -1.83589235e-01
-1.06973779e+00 9.61357713e-01 6.24435544e-01 1.20563142e-01
-6.57055676e-02 2.35914722e-01 3.13209355e-01 1.91002205e-01
3.37244123e-01 4.63982165e-01 -8.21015775e-01 -1.37602463e-01
-9.33400214e-01 3.15640599e-01 9.73299921e-01 1.24711931e+00
6.88951612e-01 -1.51709825e-01 -1.41534850e-01 8.47113371e-01
2.11449936e-01 8.34303379e-01 5.49985349e-01 -8.11751068e-01
6.38685465e-01 1.00429738e+00 1.56494826e-02 -8.04839969e-01
-1.65541962e-01 -3.19013447e-01 -8.91536176e-01 -4.61973727e-01
2.57868648e-01 -6.87874556e-02 -1.01750982e+00 1.66541636e+00
3.59988093e-01 5.19669056e-01 4.27654743e-01 6.63919866e-01
1.27013886e+00 6.39125764e-01 1.64078102e-02 6.31356314e-02
1.28612339e+00 -1.19484377e+00 -5.10709643e-01 -3.10259074e-01
7.55600095e-01 -3.94145995e-02 1.17574346e+00 4.68139708e-01
-7.28058696e-01 -2.35652909e-01 -8.86046886e-01 -2.87036330e-01
-6.07567549e-01 2.71190703e-01 5.52419126e-01 3.86215568e-01
-6.12970650e-01 3.80091339e-01 -6.89721346e-01 -3.05420998e-02
7.19481945e-01 1.79742798e-02 -3.49602066e-02 -4.33282465e-01
-1.84538841e+00 1.07694352e+00 1.18145466e+00 -2.74112344e-01
-1.12451243e+00 -7.91573524e-01 -8.82374346e-01 2.82267362e-01
8.27696443e-01 -4.94670510e-01 9.98044610e-01 -2.56822973e-01
-1.24031091e+00 7.81691074e-01 -8.19353852e-04 -5.65721035e-01
5.03163397e-01 -2.30997697e-01 -6.99367464e-01 1.31926641e-01
2.62692988e-01 5.39287865e-01 5.75018108e-01 -1.19603503e+00
-4.80075449e-01 -1.05252057e-01 2.43153811e-01 -1.00510838e-02
-3.46899450e-01 -3.69149774e-01 -3.05392981e-01 -5.92644215e-01
1.19339474e-01 -9.61313128e-01 2.87179835e-02 -4.83019441e-01
-6.95970476e-01 -6.28133178e-01 6.53140962e-01 -5.11827886e-01
1.34147632e+00 -2.04820848e+00 -1.09965198e-01 1.57156989e-01
3.47519636e-01 2.98687011e-01 1.40562356e-01 8.20755884e-02
1.35954199e-02 1.46327764e-01 -2.09051222e-01 1.17351547e-01
2.31264398e-01 3.42259794e-01 -7.07108080e-01 1.73244372e-01
1.51044354e-01 9.72370923e-01 -1.32285976e+00 -7.96434641e-01
-1.36665538e-01 -1.34790331e-01 -5.56311369e-01 2.12021098e-02
-4.63342875e-01 -6.06713519e-02 -5.75157225e-01 8.26526821e-01
6.81446612e-01 -7.34622598e-01 1.03592731e-01 -1.21323004e-01
6.86103821e-01 5.02429664e-01 -1.18929803e+00 1.52681065e+00
-9.84804034e-02 2.99226820e-01 -5.27612209e-01 -1.06756783e+00
9.80596662e-01 2.22067744e-01 -3.80057953e-02 -5.63854501e-02
1.08281821e-01 4.17916507e-01 3.15392390e-02 -4.87217546e-01
5.02121985e-01 -3.91534686e-01 -2.34803021e-01 2.49266520e-01
3.79569471e-01 -1.22033194e-01 4.02646869e-01 5.65446317e-01
1.14013457e+00 -1.78938717e-01 4.39300567e-01 -3.07865381e-01
5.56570113e-01 4.14428055e-01 8.22520733e-01 7.22647667e-01
-1.35907009e-01 -1.23575795e-02 4.23505336e-01 -1.13314837e-01
-6.35129035e-01 -1.16626918e+00 -2.88757205e-01 1.11544251e+00
5.47715664e-01 -4.55179930e-01 -3.28852743e-01 -1.08573091e+00
2.92658925e-01 1.52847600e+00 -6.35395646e-01 -5.55164278e-01
-1.44942909e-01 -5.36962211e-01 1.00274718e+00 7.78796971e-01
8.39009345e-01 -1.25753844e+00 -1.58660077e-02 1.23878634e-02
-4.79058951e-01 -1.16786027e+00 -3.08610965e-02 3.20615560e-01
-6.78233802e-01 -1.27770698e+00 -1.36283174e-01 -6.67712331e-01
7.18876302e-01 -1.98310658e-01 1.23368537e+00 -9.44289863e-02
2.17458587e-02 8.08885414e-03 -6.08749926e-01 -4.33821589e-01
-5.81729114e-01 -5.18787615e-02 2.43045211e-01 -5.02976477e-01
8.59175563e-01 -5.53561330e-01 -1.36206090e-01 1.69827968e-01
-7.80265450e-01 7.60447159e-02 4.12066698e-01 1.13503051e+00
3.77762079e-01 5.09239793e-01 8.39694321e-01 -1.14610291e+00
6.04028821e-01 -8.75650942e-01 -2.04865456e-01 5.44067502e-01
-6.29028380e-01 9.80828255e-02 7.94573367e-01 -6.03975713e-01
-9.98489499e-01 -5.28322637e-01 2.88565606e-01 -7.14137971e-01
-1.55666592e-02 6.25926316e-01 -1.89783454e-01 4.18981999e-01
1.05370498e+00 3.61584693e-01 -4.66070056e-01 -3.10415536e-01
5.23077130e-01 7.82702863e-01 5.53411961e-01 -1.03707147e+00
9.09081340e-01 2.46315420e-01 -4.62237149e-01 -3.26448172e-01
-1.44546366e+00 -4.52784806e-01 -4.66924578e-01 1.85094938e-01
4.73570794e-01 -1.24128437e+00 -6.46522343e-01 2.15527758e-01
-1.05470443e+00 -6.25844419e-01 -3.53752136e-01 4.61372584e-01
-3.64663750e-01 1.53905600e-01 -8.06033671e-01 -7.13808477e-01
-5.01524150e-01 -7.12546468e-01 5.24245501e-01 2.82450289e-01
-8.01014900e-03 -1.05147469e+00 -1.32564932e-01 6.59385145e-01
9.97561663e-02 -1.56364262e-01 1.02904236e+00 -1.42893136e+00
-3.99104744e-01 -6.71317503e-02 -3.69441956e-01 7.83604324e-01
-6.46251664e-02 -3.52956712e-01 -8.84722948e-01 -5.86219281e-02
-1.10204421e-01 -9.78482127e-01 1.14816320e+00 -2.98716761e-02
1.40453815e+00 -3.70499700e-01 -4.78871495e-01 2.33130917e-01
1.26509130e+00 -2.74148844e-02 6.61269903e-01 3.55908632e-01
8.61581087e-01 -1.38683179e-02 8.45704913e-01 5.45773923e-01
8.17065775e-01 1.75111562e-01 3.13686550e-01 3.29858154e-01
5.22567295e-02 -4.98585552e-01 4.61844504e-01 7.49967694e-01
-1.58709720e-01 -1.29287571e-01 -1.19838142e+00 7.63368785e-01
-2.02716994e+00 -1.24647117e+00 1.99960042e-02 1.66055644e+00
1.50123954e+00 2.19177336e-01 -1.30208448e-01 5.09680063e-02
8.21117818e-01 -2.49686375e-01 -8.10889125e-01 7.26512149e-02
-1.38116047e-01 1.66534483e-01 4.25800115e-01 3.57678771e-01
-9.17886972e-01 1.38259625e+00 5.26923466e+00 1.52909184e+00
-7.00546265e-01 2.43180245e-01 4.48003054e-01 1.60791054e-02
-3.15075606e-01 1.39327914e-01 -1.13112557e+00 6.14311576e-01
7.46096551e-01 -4.44147199e-01 5.86412847e-01 1.48553228e+00
-3.62088650e-01 -6.47783279e-03 -1.14726794e+00 1.00900280e+00
2.78984189e-01 -1.36191261e+00 1.27274916e-01 -1.40593141e-01
1.01027489e+00 2.56123483e-01 -3.28249067e-01 1.26752698e+00
7.45203257e-01 -9.69420195e-01 7.86000431e-01 4.10812378e-01
7.13991940e-01 -7.02364504e-01 8.50821972e-01 8.97296250e-01
-7.83693850e-01 -9.95425880e-02 -7.83209920e-01 1.80003092e-01
-1.52259797e-01 8.66795063e-01 -1.12827766e+00 4.45227653e-01
7.31230438e-01 6.71851456e-01 -6.40840471e-01 5.61756909e-01
-8.27399373e-01 1.02588260e+00 -2.77614295e-01 -2.65080154e-01
2.91823093e-02 2.03389004e-01 2.83159107e-01 1.33070350e+00
1.08168103e-01 4.95544910e-01 3.93878251e-01 1.32661629e+00
-5.47960341e-01 -1.12288550e-01 -1.85065672e-01 -1.47406161e-01
1.09930182e+00 1.13993442e+00 -2.86997527e-01 -9.87246215e-01
1.15041904e-01 8.08491766e-01 7.32119322e-01 2.84681886e-01
-1.28915548e+00 -5.17012656e-01 1.33981764e-01 -9.19729918e-02
4.45791930e-01 2.55068630e-01 -6.38878867e-02 -1.63455296e+00
-2.42132582e-02 -9.15406525e-01 6.37335002e-01 -8.99274647e-01
-1.80766547e+00 3.27477038e-01 2.51845747e-01 -9.25856829e-01
-8.95799845e-02 -7.35060573e-01 -5.37173510e-01 6.58853412e-01
-1.51510274e+00 -1.28869152e+00 -4.57101822e-01 8.78181219e-01
2.19792560e-01 -3.45256001e-01 9.04438615e-01 -2.37945635e-02
-5.21444261e-01 6.85215533e-01 2.15232998e-01 5.24699032e-01
8.42568994e-01 -1.39661431e+00 3.38362306e-02 5.06704092e-01
-6.84267879e-02 8.62137914e-01 4.91510481e-01 -8.87483299e-01
-1.04472888e+00 -1.19711697e+00 7.32918501e-01 -7.61237442e-01
1.09574282e+00 -1.64880037e-01 -9.46868896e-01 9.97828126e-01
-1.50239006e-01 2.85989314e-01 9.10331190e-01 5.17753124e-01
-7.86006987e-01 7.87725225e-02 -1.41443264e+00 5.63426852e-01
9.53096628e-01 -5.82474947e-01 -1.25482643e+00 3.30541164e-01
7.39029288e-01 -6.47353411e-01 -9.23401773e-01 3.41901273e-01
9.28345323e-02 -5.50884426e-01 6.64244592e-01 -8.54399443e-01
7.99569607e-01 -4.88335729e-01 -3.94449055e-01 -1.36532712e+00
-4.22218025e-01 -1.52819067e-01 -7.26355314e-01 1.23137987e+00
6.42167270e-01 -6.56005979e-01 8.29493463e-01 7.67424941e-01
-1.99550409e-02 -9.19286847e-01 -8.24845195e-01 -1.07654393e+00
5.80785200e-02 -5.05731165e-01 6.47683620e-01 1.54656112e+00
5.85306704e-01 7.69699514e-01 -1.52986154e-01 2.81844705e-01
6.69640124e-01 3.13915491e-01 6.14764273e-01 -1.12155879e+00
-3.60274285e-01 -1.18969418e-01 -3.31378549e-01 -1.22917223e+00
4.62329626e-01 -1.39247656e+00 -8.00203159e-02 -1.48654640e+00
6.59055114e-01 -7.37403035e-01 -4.86148864e-01 1.08030641e+00
-5.07657409e-01 -1.44018987e-02 2.67111231e-02 1.35223821e-01
-8.41130316e-01 7.59943664e-01 1.22086370e+00 -3.49744439e-01
-6.77238032e-02 -3.27036768e-01 -8.40482473e-01 8.78443956e-01
1.03059280e+00 -6.64916039e-01 -7.33149290e-01 -4.96595874e-02
2.40231946e-01 -2.22998068e-01 6.66279256e-01 -7.65011907e-01
3.60081613e-01 -4.34133619e-01 2.20948726e-01 -6.88810110e-01
3.34127694e-01 -5.29586077e-01 -2.16884464e-01 2.83570111e-01
-4.31567520e-01 -5.89071751e-01 1.77915484e-01 5.82764864e-01
-5.73848411e-02 -6.06779456e-01 6.15080357e-01 -3.13660085e-01
-9.17239547e-01 1.18993305e-01 3.97159040e-01 7.93550551e-01
9.25741136e-01 7.19357431e-02 -9.02419508e-01 -9.24909189e-02
-7.97199309e-01 3.69373262e-01 2.14329436e-01 2.65917759e-02
6.82978809e-01 -1.50217390e+00 -8.83095264e-01 -3.98190439e-01
3.36711556e-01 4.43541318e-01 2.91340858e-01 7.31625974e-01
-3.39952558e-01 2.93283731e-01 1.37187406e-01 -2.45933712e-01
-8.82860482e-01 6.72574222e-01 1.23246230e-01 -3.97691697e-01
-4.51326489e-01 1.22160959e+00 -3.87706086e-02 -7.16832161e-01
3.82116577e-03 -4.06664908e-01 -9.74029675e-02 -2.63524979e-01
4.88002956e-01 5.36564291e-01 -2.73130774e-01 -1.86950296e-01
-5.35704076e-01 -3.03819865e-01 -3.45458090e-01 2.85178900e-01
1.49784589e+00 3.36085945e-01 -3.68756533e-01 3.99207652e-01
8.53728414e-01 1.88863993e-01 -6.65510774e-01 -5.31322360e-01
-1.39440700e-01 -2.20310271e-01 -4.95213941e-02 -1.28129411e+00
-7.40205526e-01 4.58358794e-01 -1.28521100e-01 1.08564757e-01
7.35146046e-01 5.16299129e-01 6.68662071e-01 8.84253144e-01
4.32952106e-01 -1.26484776e+00 1.86840802e-01 7.83778846e-01
8.58594477e-01 -1.28324330e+00 5.40876687e-02 -5.01213253e-01
-9.56296802e-01 9.56333101e-01 8.74214530e-01 -1.35537878e-01
6.78645253e-01 3.25831305e-03 -3.17785352e-01 -4.04221117e-01
-8.42877150e-01 -2.42362022e-02 3.64007562e-01 3.97504061e-01
1.93814397e-01 5.41188955e-01 2.05401648e-02 1.29348052e+00
-3.92603368e-01 1.03153437e-01 4.18589830e-01 6.85929596e-01
-6.34535253e-01 -7.10632563e-01 -2.06636399e-01 6.95472717e-01
4.51123156e-02 -6.89490199e-01 -3.14600319e-01 7.34980881e-01
4.58786935e-01 9.30117846e-01 -5.26833415e-01 -5.47593892e-01
2.47136518e-01 2.71767199e-01 4.19617832e-01 -9.26452756e-01
-3.84913422e-02 -5.95294416e-01 3.24995190e-01 -1.46771714e-01
-2.70024747e-01 -2.69457549e-01 -1.65046215e+00 -5.62022805e-01
-8.20694864e-01 4.37539667e-01 4.69863564e-02 1.11574376e+00
2.38708511e-01 3.71106237e-01 1.94379285e-01 -2.24830471e-02
-1.04498160e+00 -1.30466294e+00 -8.95774782e-01 3.84590089e-01
-1.48817688e-01 -7.88062155e-01 -5.15595913e-01 2.20450297e-01] | [9.875195503234863, 8.15687370300293] |
a68386aa-0a1f-44f8-871d-530cf99072a5 | generative-adversarial-network-for-text-to | null | null | https://ieeexplore.ieee.org/document/9666791 | https://ieeexplore.ieee.org/document/9666791 | Generative Adversarial Network for Text-to-Face Synthesis and Manipulation with Pretrained BERT Model | This work proposes a cyclic generative adversarial network with spatial-wise and channel-wise attention modules for text-to-face synthesis and manipulation. Then, we explore the pre-trained transformer-based BERT model to obtain text embedding. Furthermore, dual-layer perceptual loss and SSIM loss are introduced to reinforce the delicate features and preserve facial identity during the manipulation task. Additionally, we adopt a novel Flickr-Faces-HQ with Text descriptions (FFHQ-Text) dataset with numerous facial attribute annotations to advance the development of the text-to-face task. In particular, by introducing the StyleGAN encoder for learning latent representations to our proposed post-processing method, we demonstrate that even training on a smaller text-to-face dataset can synthesize more realistic images. Experimental results reveal the effectiveness of our approach, which generates photo-realistic facial images, edits the specific facial attribute with the correlated keywords manipulation, outperforms previous state-of-the-art methods both in quality and quantity, and suggests promising future directions. | ['Yutong Zhou，Nobutaka Shimada'] | 2022-01-12 | null | null | null | fg-2022-1 | ['text-to-face-generation'] | ['computer-vision'] | [ 4.62065369e-01 2.77615130e-01 3.15628499e-01 -5.41284859e-01
-8.20380569e-01 -4.53893334e-01 8.44797373e-01 -1.00655138e+00
6.58144429e-02 4.85016167e-01 2.38795117e-01 2.60171175e-01
1.87792808e-01 -7.27499187e-01 -1.01059043e+00 -7.34383404e-01
3.79146457e-01 6.40626997e-02 -5.91651380e-01 -2.47630030e-01
-1.03496477e-01 6.70652568e-01 -1.52888262e+00 4.04237181e-01
6.60011828e-01 1.32048142e+00 -6.06392091e-03 4.06014562e-01
3.66038501e-01 9.38459575e-01 -4.69792128e-01 -1.12387943e+00
5.39084494e-01 -5.30682802e-01 -2.47119710e-01 3.45593244e-01
1.01798189e+00 -8.27231526e-01 -6.35836244e-01 9.17319000e-01
8.01115334e-01 -2.07973309e-02 6.81489229e-01 -1.51295555e+00
-1.25351429e+00 3.00827354e-01 -5.51227629e-01 -6.17044330e-01
3.50463659e-01 6.58197045e-01 7.52601266e-01 -1.32921731e+00
7.58011520e-01 1.71252680e+00 5.29992223e-01 8.43451917e-01
-1.22287190e+00 -1.22782159e+00 -2.45510992e-02 5.47380224e-02
-1.56958270e+00 -9.81653690e-01 1.07288063e+00 -2.49511972e-01
4.56709713e-01 1.22025751e-01 4.87176716e-01 1.52731514e+00
-1.86956506e-02 6.04996622e-01 9.57723200e-01 -1.00830607e-01
-1.83052450e-01 -6.04309365e-02 -1.14504576e+00 8.64971638e-01
-2.65604287e-01 3.11401397e-01 -4.62887555e-01 1.91387102e-01
9.76044893e-01 9.38496441e-02 -2.17599779e-01 -2.41612449e-01
-9.98884559e-01 6.81277394e-01 3.08533221e-01 -8.83984342e-02
-4.25825924e-01 4.96930003e-01 2.83550769e-01 3.40542436e-01
4.86511767e-01 4.18009043e-01 -1.08122237e-01 1.82081595e-01
-1.05213201e+00 3.95839125e-01 3.18684548e-01 1.48792660e+00
6.36647224e-01 5.81870019e-01 -6.23112440e-01 7.89258480e-01
1.35760754e-01 7.61871457e-01 -4.80360258e-03 -1.21387672e+00
4.32814777e-01 2.79355645e-01 -5.36120199e-02 -1.07317770e+00
2.13128000e-01 -1.76536068e-01 -8.97937953e-01 2.05053300e-01
2.19435222e-03 -4.93953824e-02 -9.44606781e-01 1.90309656e+00
2.69315004e-01 2.18262523e-01 -9.80229005e-02 8.26625109e-01
8.01346242e-01 5.17496943e-01 2.58280244e-02 -1.45705715e-01
1.14417636e+00 -1.10352981e+00 -9.97016728e-01 1.76979810e-01
-7.87959900e-03 -8.26405108e-01 1.27663445e+00 2.35165641e-01
-1.42671406e+00 -8.06310475e-01 -7.88108408e-01 -4.32253301e-01
-7.50534311e-02 5.65680563e-01 3.94037336e-01 5.75507522e-01
-1.15618241e+00 5.67073464e-01 -4.26980257e-01 3.18494625e-02
9.24549758e-01 2.68512756e-01 -5.85468173e-01 -3.51949215e-01
-1.12442172e+00 5.15285552e-01 -6.45075664e-02 3.58859509e-01
-1.44703877e+00 -8.69954228e-01 -1.09603333e+00 8.30234289e-02
3.20714206e-01 -7.58765280e-01 8.25906396e-01 -1.25177503e+00
-1.94762778e+00 7.79186904e-01 -2.32206173e-02 1.66667923e-01
8.80357981e-01 -1.85216591e-01 -4.00330991e-01 3.92680347e-01
-1.91625617e-02 1.07543802e+00 1.65402114e+00 -1.32916951e+00
-1.31894693e-01 -1.92396581e-01 2.08170399e-01 -5.60280606e-02
-4.88040626e-01 1.53820604e-01 -6.70705140e-01 -1.18170536e+00
-4.30531740e-01 -8.98561776e-01 1.16205879e-01 6.28114402e-01
-4.85345602e-01 2.21420228e-02 8.70934725e-01 -9.35207963e-01
7.83515394e-01 -2.40111995e+00 3.82743210e-01 -4.52691391e-02
1.95290834e-01 1.63601995e-01 -6.59544945e-01 4.46714848e-01
-2.81898499e-01 2.14097217e-01 -9.83794332e-02 -9.00534153e-01
1.24966837e-01 -1.23826183e-01 -4.38307375e-01 4.90297854e-01
8.59950900e-01 1.17486978e+00 -6.31440818e-01 -4.64991361e-01
2.62755811e-01 8.33635569e-01 -8.63125145e-01 4.21987683e-01
-3.32209677e-01 6.79895401e-01 -3.08587104e-01 8.35032642e-01
9.88243818e-01 1.21417724e-01 -4.92399745e-02 -6.19106174e-01
2.23086044e-01 -2.48317525e-01 -5.44607818e-01 1.99920630e+00
-5.98210394e-01 5.83369911e-01 1.94671735e-01 -6.82540119e-01
8.16311657e-01 3.45767230e-01 3.26291472e-01 -8.04285645e-01
2.57168204e-01 2.74239015e-02 -2.75353968e-01 -6.37317181e-01
4.59537864e-01 -1.89637661e-01 4.87961397e-02 1.54656768e-01
2.32973918e-01 -3.54469895e-01 -9.43901390e-02 1.13741964e-01
7.33021438e-01 6.88965797e-01 -3.60357553e-01 1.45979524e-01
5.13330460e-01 -6.86912358e-01 4.08971459e-01 2.66983882e-02
1.60454772e-02 8.58956218e-01 6.02207422e-01 -7.94562772e-02
-1.40661061e+00 -9.08391356e-01 5.52116148e-02 9.42516148e-01
-1.27322897e-01 -3.14734548e-01 -1.01004565e+00 -6.51007712e-01
2.67555993e-02 7.46184289e-01 -9.13347721e-01 -3.24108005e-01
-5.25267303e-01 -5.06384447e-02 8.67241383e-01 5.48447311e-01
7.38896549e-01 -1.07529545e+00 3.19029503e-02 -6.86042011e-02
-2.10803926e-01 -1.33110714e+00 -9.35193241e-01 -7.00606942e-01
-3.10702324e-01 -7.78198242e-01 -9.25135493e-01 -7.03348637e-01
9.38838780e-01 -1.08035408e-01 7.31788158e-01 1.28029197e-01
-3.17063481e-01 4.01897967e-01 -3.94548476e-01 -1.62570402e-01
-3.74509305e-01 -3.90937328e-01 -1.80510640e-01 5.47998309e-01
-2.12153390e-01 -7.57158875e-01 -8.65335882e-01 2.47768834e-01
-1.11234319e+00 2.23157689e-01 7.04416513e-01 1.00546086e+00
4.04935122e-01 -1.18812889e-01 5.80296099e-01 -6.23102129e-01
3.07918251e-01 -1.44357696e-01 -3.98762524e-01 2.35121623e-01
-3.29521030e-01 -5.30040786e-02 9.20783818e-01 -6.43157661e-01
-1.25077307e+00 1.30747393e-01 -1.89652622e-01 -1.19926691e+00
1.43453151e-01 -1.08461283e-01 -8.40337992e-01 -2.99566001e-01
2.79074669e-01 4.50618237e-01 1.38739824e-01 -1.67339593e-01
7.20224500e-01 5.43971539e-01 6.04139686e-01 -6.12380028e-01
1.25727928e+00 5.68265021e-01 1.61272109e-01 -5.48059583e-01
-4.68971729e-01 3.90196443e-01 -5.00512064e-01 -3.13254327e-01
9.19537783e-01 -1.24546027e+00 -8.33996356e-01 7.00450778e-01
-1.23218119e+00 -3.26188892e-01 -2.08975747e-01 1.81251522e-02
-9.79045808e-01 1.70763806e-01 -6.47587359e-01 -6.99059069e-01
-3.61548603e-01 -1.30874503e+00 1.74546838e+00 -9.52790380e-02
2.09267721e-01 -4.83146876e-01 -5.20702720e-01 6.04117811e-01
5.51627636e-01 5.26511788e-01 6.24389946e-01 1.98838897e-02
-9.00854349e-01 1.08818496e-02 -4.49002713e-01 6.72165692e-01
1.09065861e-01 1.86255172e-01 -1.13235950e+00 -4.58857715e-01
-3.41639400e-01 -6.86662316e-01 7.13363290e-01 -1.27250841e-02
1.58739281e+00 -6.38156772e-01 -6.43553329e-04 1.14200258e+00
1.21656477e+00 3.44160362e-03 1.15737391e+00 -3.31563562e-01
1.04029715e+00 7.34462202e-01 6.55048132e-01 5.14087498e-01
3.35956186e-01 7.56343424e-01 5.44990122e-01 -3.25123399e-01
-4.69307065e-01 -6.30353153e-01 5.61668873e-01 5.48711658e-01
-2.62378067e-01 -2.70915776e-01 -5.62487394e-02 3.55487138e-01
-1.29142916e+00 -1.08421838e+00 4.97032851e-01 1.66775894e+00
9.60765541e-01 -4.53210443e-01 -1.75828919e-01 9.27433185e-03
6.34592175e-01 2.28147238e-01 -6.48950458e-01 -1.01657286e-01
-8.81775469e-02 5.21180511e-01 1.23341173e-01 2.14674473e-01
-9.39108431e-01 1.28737390e+00 5.18705225e+00 1.17239594e+00
-1.33927023e+00 7.70304948e-02 8.47580314e-01 -2.03718051e-01
-5.78790545e-01 -2.96621650e-01 -4.17404681e-01 5.24259150e-01
6.71936333e-01 5.02398647e-02 7.75307834e-01 6.91257477e-01
3.03268969e-01 5.84272742e-01 -1.19180429e+00 1.04707706e+00
4.95656759e-01 -1.23959076e+00 5.60822487e-01 1.19082317e-01
7.58124530e-01 -8.51759553e-01 5.72850227e-01 2.98064977e-01
1.56182544e-02 -1.29216015e+00 1.16905379e+00 7.60532677e-01
1.81543696e+00 -9.07385349e-01 1.46735057e-01 -2.65416712e-01
-1.17688549e+00 -9.99967009e-02 -2.58758873e-01 4.43373501e-01
1.30501434e-01 2.11730480e-01 -4.75156903e-01 7.07884967e-01
5.26989758e-01 8.68359268e-01 -5.07592022e-01 2.85856992e-01
-4.26856160e-01 2.95224816e-01 1.29690036e-01 3.47928315e-01
-1.01272799e-01 -1.45830289e-01 3.58251512e-01 8.35105062e-01
5.90080500e-01 1.97878525e-01 -1.85494602e-01 1.29142976e+00
-6.35435998e-01 -8.29633884e-03 -6.71098828e-01 -4.30327058e-01
5.67323744e-01 1.38042641e+00 -4.92291078e-02 -1.99391156e-01
-2.05259293e-01 1.45948493e+00 1.12741441e-01 4.94782209e-01
-1.25391936e+00 -4.74453866e-01 7.71239817e-01 2.35397816e-01
5.21391749e-01 -1.71449195e-04 1.79733455e-01 -1.09314048e+00
1.49332464e-01 -1.14208913e+00 -3.55254203e-01 -1.24142325e+00
-1.20205534e+00 7.33644962e-01 -2.31485099e-01 -1.24143755e+00
-2.89514333e-01 -3.94068748e-01 -4.21396494e-01 9.10418808e-01
-1.67260623e+00 -2.09034991e+00 -5.36179960e-01 9.61963952e-01
4.99215275e-01 -3.76157701e-01 6.36802912e-01 5.70755422e-01
-5.15951395e-01 1.34021473e+00 -2.41739482e-01 2.68312544e-01
9.72699463e-01 -6.30966246e-01 5.24486482e-01 7.21318901e-01
-3.24413478e-01 6.37391567e-01 4.38219309e-01 -4.98818517e-01
-1.72695971e+00 -1.66021299e+00 3.98694426e-01 -3.22171807e-01
3.22541773e-01 -6.46150291e-01 -5.16080558e-01 7.38081157e-01
3.53413641e-01 1.98429823e-01 4.43930119e-01 -7.04699516e-01
-6.01826906e-01 -3.27646583e-01 -1.47239351e+00 8.20109665e-01
1.47284520e+00 -7.27476597e-01 2.61619370e-02 1.90078497e-01
1.10223234e+00 -3.45654696e-01 -9.79027271e-01 3.92148972e-01
8.76531005e-01 -8.34663332e-01 9.81939137e-01 -4.14488912e-01
1.08352804e+00 -1.61598802e-01 -1.97290078e-01 -1.26315951e+00
-2.98476189e-01 -1.12542784e+00 3.32305938e-01 1.68259299e+00
9.61062983e-02 -3.99713874e-01 5.37564576e-01 3.90946537e-01
-2.86058575e-01 -7.21986473e-01 -7.13174164e-01 -6.35838091e-01
1.38262257e-01 -1.51327640e-01 1.10889506e+00 9.87788141e-01
-6.80303395e-01 -4.48131897e-02 -9.42000568e-01 -2.86415429e-03
6.76552951e-01 -1.28003120e-01 1.11119306e+00 -7.41513193e-01
-9.30551812e-02 -1.67204604e-01 -3.97997886e-01 -8.53618503e-01
5.60778618e-01 -7.00874150e-01 1.09965764e-02 -7.71889269e-01
2.21471358e-02 -1.88576564e-01 2.54382402e-01 5.54345191e-01
-8.62997696e-02 6.09837234e-01 4.07033771e-01 -1.02384597e-01
-2.31956124e-01 1.18294823e+00 1.90929890e+00 -2.36765787e-01
3.75556111e-01 -5.04664540e-01 -8.01086426e-01 3.08159798e-01
3.03004384e-01 -2.25775287e-01 -5.00871301e-01 -4.24710214e-01
4.61833961e-02 1.38772950e-01 6.19841516e-01 -7.82246172e-01
-1.32565469e-01 -2.75131702e-01 6.42645955e-01 -2.47401223e-01
8.37195337e-01 -8.73026192e-01 3.67914766e-01 2.26758663e-02
-5.67410529e-01 -1.26604110e-01 1.34221748e-01 5.22581816e-01
-1.17715172e-01 3.15360814e-01 9.77761328e-01 1.39254853e-01
-2.20431283e-01 7.99617946e-01 2.54660964e-01 -8.20432231e-02
1.20364106e+00 -1.15476273e-01 -2.01541886e-01 -6.78528488e-01
-3.82364839e-01 4.43674736e-02 7.01624513e-01 5.17039299e-01
9.20650959e-01 -1.68328679e+00 -9.61053789e-01 6.37333214e-01
2.15029612e-01 -1.71329886e-01 4.53540266e-01 4.80087489e-01
-4.17240024e-01 1.14885150e-02 -5.09668648e-01 -2.59099603e-01
-1.17717838e+00 6.37062907e-01 1.35994226e-01 2.24735692e-01
-5.05252540e-01 9.00567949e-01 5.05330265e-01 -3.16865444e-01
1.58728287e-01 5.57098761e-02 1.51471168e-01 -1.08453318e-01
4.41665500e-01 1.64715320e-01 -2.00304717e-01 -7.88124144e-01
3.73518514e-03 6.67898536e-01 1.63439736e-02 -2.11862698e-01
1.32457936e+00 -8.37648809e-02 -1.09226599e-01 -2.74228215e-01
1.31901300e+00 1.81265429e-01 -1.73838723e+00 -2.38903556e-02
-8.37479889e-01 -8.79828215e-01 -1.14724137e-01 -6.73270345e-01
-1.68528044e+00 1.04795635e+00 4.59080815e-01 -5.71153998e-01
1.47086561e+00 -2.68430978e-01 9.06048059e-01 2.01085713e-02
1.99720159e-01 -7.57978976e-01 5.39747298e-01 1.18819475e-01
1.59249938e+00 -1.00077915e+00 -1.55710503e-01 -5.76933682e-01
-7.22271621e-01 9.21598434e-01 6.80401802e-01 -1.63303614e-01
4.44171965e-01 1.85340121e-01 -2.05498815e-01 5.24865603e-03
-7.82416224e-01 1.68938890e-01 3.17454070e-01 7.50213981e-01
2.19277233e-01 -2.39664868e-01 6.68336228e-02 5.00626445e-01
-4.34748232e-01 2.19430745e-01 2.91443825e-01 3.98130804e-01
3.17697048e-01 -9.98488605e-01 -7.43046552e-02 1.85620800e-01
-4.69448626e-01 -4.45633948e-01 -3.64991724e-01 7.63424218e-01
3.30651760e-01 7.76008487e-01 2.92928386e-02 -5.80981851e-01
3.34678143e-01 -1.47086143e-01 8.49431992e-01 -2.87058234e-01
-4.95529473e-01 7.13730231e-02 6.72579138e-03 -8.69845510e-01
-3.84978384e-01 -4.89699036e-01 -7.61492133e-01 -5.65789819e-01
-2.83236891e-01 -4.26711023e-01 6.51872575e-01 5.72841108e-01
7.33580351e-01 5.97176850e-01 1.10718274e+00 -1.21794534e+00
-6.33751094e-01 -1.06932867e+00 -5.73864877e-01 6.68995976e-01
2.52510905e-01 -7.91608036e-01 -3.41415197e-01 4.82605487e-01] | [12.531699180603027, -0.19000831246376038] |
ecf2294d-62a5-48d0-ab6a-5cc4e8b2f5d4 | tnn7-a-custom-macro-suite-for-implementing | 2205.07410 | null | https://arxiv.org/abs/2205.07410v2 | https://arxiv.org/pdf/2205.07410v2.pdf | TNN7: A Custom Macro Suite for Implementing Highly Optimized Designs of Neuromorphic TNNs | Temporal Neural Networks (TNNs), inspired from the mammalian neocortex, exhibit energy-efficient online sensory processing capabilities. Recent works have proposed a microarchitecture framework for implementing TNNs and demonstrated competitive performance on vision and time-series applications. Building on these previous works, this work proposes TNN7, a suite of nine highly optimized custom macros developed using a predictive 7nm Process Design Kit (PDK), to enhance the efficiency, modularity and flexibility of the TNN design framework. TNN prototypes for two applications are used for evaluation of TNN7. An unsupervised time-series clustering TNN delivering competitive performance can be implemented within 40 uW power and 0.05 mm^2 area, while a 4-layer TNN that achieves an MNIST error rate of 1% consumes only 18 mW and 24.63 mm^2. On average, the proposed macros reduce power, delay, area, and energy-delay product by 14%, 16%, 28%, and 45%, respectively. Furthermore, employing TNN7 significantly reduces the synthesis runtime of TNN designs (by more than 3x), allowing for highly-scaled TNN implementations to be realized. | ['John Paul Shen', 'Santha Bhasuthkar', 'Prabhu Vellaisamy', 'Harideep Nair'] | 2022-05-16 | null | null | null | null | ['time-series-clustering'] | ['time-series'] | [ 3.51856589e-01 -9.95493978e-02 -1.72063056e-02 -1.81057557e-01
1.56754442e-02 -2.35403091e-01 2.11145476e-01 3.83177623e-02
-8.59240234e-01 3.74409229e-01 -3.03812265e-01 -2.57463574e-01
-4.33222242e-02 -6.69961929e-01 -4.25392091e-01 -6.96616054e-01
-1.59810763e-02 -2.21867353e-01 6.47799313e-01 1.14948437e-01
2.25378096e-01 4.93802935e-01 -1.74010611e+00 1.73714221e-01
4.70460624e-01 1.69732261e+00 2.81912655e-01 2.87709147e-01
-3.72417900e-03 7.22958565e-01 -6.31491125e-01 -2.18114942e-01
3.50915909e-01 -2.26896331e-01 5.88340908e-02 -6.94972575e-01
2.45119154e-01 -3.87900472e-01 -6.40269876e-01 8.61024737e-01
5.72116613e-01 2.94356421e-02 4.54053581e-01 -1.23898649e+00
-9.59194899e-02 9.42888975e-01 -4.10459995e-01 2.85455406e-01
-4.92460370e-01 2.24199772e-01 5.63946664e-01 -5.22090077e-01
4.01124179e-01 9.53841746e-01 6.77185118e-01 8.87587547e-01
-1.19406295e+00 -1.19377184e+00 -3.77328753e-01 1.81164667e-01
-1.28770018e+00 -9.35896039e-01 4.62813199e-01 1.23047881e-01
2.04377627e+00 -1.97622776e-01 9.79301095e-01 1.12340629e+00
9.47359324e-01 5.39160132e-01 9.83561337e-01 -9.58609432e-02
8.90284419e-01 -6.79845095e-01 6.41341601e-03 4.54024464e-01
5.33482671e-01 -7.15032294e-02 -1.02894819e+00 2.36455992e-01
7.45304227e-01 1.86665252e-01 6.31930903e-02 1.84263110e-01
-9.07940030e-01 2.39195392e-01 7.75744796e-01 5.20896614e-02
-1.94843292e-01 1.22089648e+00 8.78306270e-01 3.82668362e-03
-2.50543147e-01 3.15318525e-01 -3.29010397e-01 -6.10306621e-01
-1.01996052e+00 -2.74099022e-01 4.89510685e-01 1.37064230e+00
2.75971919e-01 7.14116037e-01 1.38791263e-01 5.55230677e-01
4.83182698e-01 9.16408956e-01 7.18990862e-01 -1.07433355e+00
1.78656459e-01 8.06881428e-01 -5.64132810e-01 -3.92901748e-01
-8.48170519e-01 -1.23380266e-01 -9.94082808e-01 1.55042082e-01
1.20798506e-01 -3.14308703e-02 -1.52935040e+00 1.60916710e+00
-1.58769697e-01 -2.45378762e-01 1.98375806e-01 5.79131722e-01
7.62415648e-01 7.32559204e-01 2.22805038e-01 -3.48199517e-01
1.25681424e+00 -8.21167827e-01 -9.23090696e-01 -3.79487276e-01
1.85887337e-01 -6.06061995e-01 5.87904274e-01 2.94592470e-01
-1.13776302e+00 -4.91100401e-01 -1.58681417e+00 -3.72263819e-01
-1.68098986e-01 9.14011523e-03 4.78322327e-01 8.44469368e-01
-1.31807005e+00 7.22640336e-01 -1.57009399e+00 -6.92018270e-01
5.69604874e-01 9.92514312e-01 3.03776711e-01 3.00908297e-01
-5.74232340e-01 5.50642967e-01 4.26887214e-01 1.56271890e-01
-1.17582095e+00 -8.52779150e-01 -3.25296730e-01 2.62791932e-01
-9.39103514e-02 -5.41362166e-01 1.45223141e+00 -5.57545424e-01
-1.86173272e+00 3.04545075e-01 -1.12749673e-02 -9.56511259e-01
-1.33510694e-01 1.19287804e-01 -6.75709486e-01 2.92206109e-01
-2.11333677e-01 1.04661441e+00 5.40814817e-01 -2.53573060e-01
-5.26914954e-01 -4.97199118e-01 -4.42712665e-01 -2.83082634e-01
-6.37432635e-01 2.65991781e-02 -3.62451494e-01 -5.03873706e-01
4.80637789e-01 -1.03214991e+00 -4.62291241e-01 2.66370118e-01
1.48554239e-02 3.01721264e-02 9.69936848e-01 1.11405179e-01
1.12591445e+00 -2.33357692e+00 -4.84199047e-01 2.32463881e-01
4.12075549e-01 1.86780155e-01 4.82311882e-02 8.67254883e-02
5.07817805e-01 -3.80149335e-02 -1.02865048e-01 -1.61360726e-01
-3.62259038e-02 2.24649370e-01 -1.57438964e-01 3.17377120e-01
-2.20335752e-01 9.47687030e-01 -3.50217104e-01 6.26206249e-02
1.28548026e-01 4.11911696e-01 -3.41439188e-01 -6.08116984e-01
-9.48474184e-02 -4.09376919e-01 -2.49827709e-02 9.47780371e-01
3.84492338e-01 -3.56604531e-02 5.77692270e-01 -5.24745286e-01
-3.59655291e-01 3.43511462e-01 -5.75760782e-01 1.82766533e+00
-1.91621929e-01 1.07576108e+00 -6.07113130e-02 -2.06176296e-01
1.12906671e+00 2.26981342e-01 4.36929435e-01 -1.48122036e+00
6.73674762e-01 6.64994955e-01 3.83199215e-01 4.11009975e-02
4.13881153e-01 -1.54079735e-01 -2.98341930e-01 2.67701060e-01
5.26581049e-01 3.30383718e-01 1.49832532e-01 -2.02510819e-01
1.59866834e+00 -2.34997302e-01 9.32222977e-02 -8.02718520e-01
-1.56356692e-01 -1.71531469e-01 9.47722435e-01 4.14595634e-01
-5.88598609e-01 -5.83397632e-04 3.95193785e-01 -4.12989408e-01
-1.37497807e+00 -1.35350943e+00 -1.32992506e-01 8.85185719e-01
1.15252972e-01 -5.70373416e-01 -8.39006245e-01 1.98387891e-01
-4.05754745e-01 3.97641838e-01 -1.88282087e-01 -1.18614569e-01
-5.49745798e-01 -6.41423464e-01 1.15743804e+00 1.00208545e+00
1.07262731e+00 -9.59837556e-01 -1.77231085e+00 6.52671695e-01
3.04430783e-01 -1.35326099e+00 -2.16702774e-01 9.17371154e-01
-1.46269977e+00 -3.82187098e-01 4.36566211e-02 -9.52835381e-01
4.80541378e-01 -7.82747865e-02 5.88331103e-01 -5.15440166e-01
-4.70099390e-01 -5.19367270e-02 6.40318543e-03 -6.18495107e-01
2.90489614e-01 -8.91656280e-02 5.99816263e-01 -3.92237067e-01
5.85427523e-01 -1.17076468e+00 -9.90162611e-01 1.41330138e-01
-5.93580782e-01 -4.20135558e-02 5.27582645e-01 5.14122784e-01
9.29663718e-01 1.73596904e-01 5.88774621e-01 -2.94781446e-01
-1.06775291e-01 2.11795330e-01 -9.23174977e-01 -1.61581889e-01
-9.50036883e-01 9.71106887e-02 9.95741904e-01 -5.61135828e-01
-8.04942191e-01 2.88981557e-01 6.39999658e-02 -4.25653756e-01
3.91802162e-01 3.81696597e-02 -8.21613893e-02 -3.02769989e-01
7.35911727e-01 1.55929074e-01 -9.67495367e-02 -1.23169556e-01
-5.19766062e-02 1.95461020e-01 5.78527331e-01 7.80693814e-02
3.74617755e-01 7.16629326e-01 2.30883375e-01 -8.48027468e-01
2.53962815e-01 -5.66614419e-02 -5.75873777e-02 -3.29269528e-01
8.14303517e-01 -1.12718093e+00 -1.11992931e+00 5.60019970e-01
-7.69586444e-01 -5.75570703e-01 -1.09440930e-01 5.48875391e-01
-3.65256041e-01 -4.00107801e-01 -7.65434504e-01 -8.63642752e-01
-1.17499900e+00 -9.65364158e-01 4.01766598e-01 6.48212314e-01
-2.69228369e-01 -4.51395243e-01 -5.69933474e-01 -1.12766847e-01
8.09148133e-01 1.36638656e-01 1.01331580e+00 -1.55748680e-01
-7.53208339e-01 2.59991050e-01 -2.90172428e-01 1.62111351e-03
-1.94734842e-01 1.25217423e-01 -1.25538301e+00 -4.30231154e-01
3.40428501e-01 -2.12969050e-01 8.73680532e-01 6.82333887e-01
8.66619945e-01 -8.65461975e-02 -6.05509758e-01 1.02165782e+00
1.98793542e+00 1.06066048e+00 7.08284318e-01 1.25367790e-01
5.65229535e-01 -7.66465887e-02 -6.06127530e-02 5.74803710e-01
3.55630070e-02 2.39488646e-01 7.19674706e-01 3.82462323e-01
-6.58325553e-02 -1.30366534e-01 8.20389152e-01 1.27252126e+00
7.21066669e-02 -3.79401326e-01 -1.14230275e+00 5.23813248e-01
-1.53701615e+00 -6.44033849e-01 4.18684408e-02 1.87634850e+00
5.29676795e-01 6.84230566e-01 -2.55107075e-01 2.68274933e-01
2.71055460e-01 -1.04217627e-03 -1.32151806e+00 -6.87695920e-01
-3.63631994e-02 6.72497690e-01 1.09691930e+00 -3.40893507e-01
-6.93330705e-01 7.49947071e-01 5.82809830e+00 9.67589855e-01
-1.38594353e+00 1.01331547e-01 3.08673769e-01 -9.13925231e-01
2.94979960e-02 -1.99230239e-01 -1.04489279e+00 3.04886192e-01
1.81354010e+00 -3.82446140e-01 5.30270219e-01 9.25636888e-01
3.10274269e-02 -1.30491108e-01 -1.14542627e+00 1.34258008e+00
-2.02606052e-01 -1.49755192e+00 -9.71032977e-02 1.42536432e-01
6.17289960e-01 4.12613750e-01 3.11208516e-01 3.27976793e-01
-1.07557490e-01 -9.79254246e-01 9.97611761e-01 -3.44109498e-02
1.16130400e+00 -1.20117760e+00 4.64190930e-01 -1.19783454e-01
-1.56293702e+00 -5.89154065e-01 -6.19344652e-01 -2.75102466e-01
-7.26369247e-02 5.76292098e-01 -5.98536074e-01 -1.40224844e-01
1.45921481e+00 4.96530443e-01 -3.93554091e-01 8.88724029e-01
1.14036635e-01 5.56206167e-01 -8.05686355e-01 -8.17841351e-01
2.19304547e-01 1.17270321e-01 1.24460734e-01 9.65578675e-01
4.18898463e-01 3.88732970e-01 -4.89726812e-01 6.86185002e-01
-4.96895850e-01 -2.77222484e-01 -3.44723821e-01 3.71376127e-02
8.57939243e-01 1.38543952e+00 -1.41316319e+00 -4.14435603e-02
-1.10782340e-01 7.92954564e-01 -9.44075063e-02 -1.20808249e-02
-1.03517306e+00 -9.12473083e-01 8.22187901e-01 1.28899636e-02
4.07156944e-01 -5.34372389e-01 -9.48074579e-01 -4.68099982e-01
1.33171886e-01 -3.26586485e-01 -2.00124100e-01 -5.67550123e-01
-5.83347619e-01 9.03850734e-01 -6.14663661e-01 -1.18423343e+00
3.09259444e-01 -8.40820789e-01 -2.56085902e-01 2.52639174e-01
-8.81473720e-01 -4.60073560e-01 -4.10871297e-01 3.22214514e-01
3.78141046e-01 -1.57345057e-01 6.77678943e-01 3.16291004e-01
-7.36772537e-01 8.29472125e-01 4.65422720e-01 3.15518118e-02
3.61780286e-01 -6.99226081e-01 8.50062013e-01 1.03743529e+00
-3.26641977e-01 7.63712466e-01 3.05702955e-01 -3.46996665e-01
-1.93183792e+00 -1.46314502e+00 5.13814390e-01 9.25797448e-02
6.41409099e-01 -9.28110540e-01 -4.18492496e-01 2.36680880e-01
3.77910405e-01 -1.84027329e-01 7.43544221e-01 -4.13252145e-01
-3.23037714e-01 -8.19251537e-01 -1.33096528e+00 1.22517693e+00
1.58678925e+00 -5.85538149e-01 2.36317366e-02 -1.50623739e-01
8.12781334e-01 -1.97323218e-01 -9.96031821e-01 6.94992304e-01
8.18292856e-01 -8.92645240e-01 3.56115669e-01 5.04148364e-01
1.91400886e-01 -4.24125463e-01 -3.96799386e-01 -5.35602987e-01
-2.60142505e-01 -9.34222639e-01 -3.21749687e-01 9.83587146e-01
5.26754677e-01 -5.71235478e-01 1.07181776e+00 4.56312627e-01
-5.05542457e-01 -7.93716729e-01 -1.56560695e+00 -1.04597890e+00
-4.13124204e-01 -6.33475602e-01 2.50185579e-01 2.94281859e-02
1.87664002e-01 5.57391822e-01 2.38931045e-01 -1.34766847e-01
7.29913712e-01 -3.39834929e-01 4.47246954e-02 -9.02264118e-01
1.96232900e-01 -7.22821891e-01 -7.85657048e-01 -1.05211031e+00
-3.44322294e-01 -6.68535769e-01 2.27668285e-01 -1.36424041e+00
-1.55736610e-01 -2.81096667e-01 -3.32682580e-01 5.86576223e-01
7.44349480e-01 5.17026067e-01 2.46981874e-01 -6.16240986e-02
-5.25116920e-01 7.25552440e-01 6.50499940e-01 -1.01705074e-01
-1.27656132e-01 -6.65780067e-01 -1.86190575e-01 7.82108724e-01
8.23090136e-01 -5.12304187e-01 -9.54392135e-01 -7.43499279e-01
1.27297090e-02 -7.05942363e-02 -7.55910128e-02 -2.00410008e+00
9.44341481e-01 2.40612641e-01 7.88317144e-01 -8.26299608e-01
6.06974959e-01 -9.52254593e-01 3.67051035e-01 9.92440224e-01
1.48537219e-01 3.88260573e-01 8.46421778e-01 5.75153649e-01
1.32576957e-01 2.56401122e-01 9.88970399e-01 3.07370603e-01
-1.03014803e+00 3.64170253e-01 -1.05012453e+00 -3.00746679e-01
1.08836031e+00 -7.49770820e-01 -6.50046229e-01 3.25155944e-01
-3.01053584e-01 2.07043160e-02 2.61398017e-01 2.28729006e-02
1.02224636e+00 -1.27663958e+00 1.18145749e-01 3.98249239e-01
-1.46897420e-01 -2.43527353e-01 6.55175984e-01 7.27621198e-01
-9.75454628e-01 7.57697046e-01 -8.96996677e-01 -6.31591439e-01
-9.86397266e-01 4.18335646e-01 4.24757540e-01 2.24683136e-01
-6.40482724e-01 1.10338438e+00 -1.54632345e-01 3.34463954e-01
3.05887282e-01 -9.50219333e-01 1.93922266e-01 -7.78610352e-03
4.78052258e-01 6.59086287e-01 2.42475584e-01 9.06005949e-02
-7.80599475e-01 5.37293315e-01 -8.27839598e-02 -3.26578707e-01
1.25884449e+00 3.47908959e-02 -2.17851102e-01 5.60113788e-01
1.10855603e+00 -6.94567204e-01 -1.43441164e+00 -1.78217739e-01
2.78995603e-01 4.54544455e-01 1.85807154e-01 -6.89008951e-01
-1.24489343e+00 6.62440836e-01 1.08704329e+00 -2.81092346e-01
2.05189037e+00 -4.83007640e-01 1.15994227e+00 4.99948621e-01
7.29510725e-01 -1.43428063e+00 1.09033689e-01 7.07129180e-01
2.83168316e-01 -4.78173524e-01 -7.36724734e-02 6.80110902e-02
-1.16654269e-01 1.37753952e+00 1.01880682e+00 -1.74534738e-01
7.69586146e-01 9.72342730e-01 -9.51202214e-02 -2.14326829e-01
-1.30193079e+00 2.21973449e-01 -1.84985355e-01 4.83343273e-01
5.54990955e-02 1.46052465e-01 -1.97864950e-01 5.80581605e-01
-2.40632132e-01 -9.44096446e-02 4.87646312e-01 1.10948801e+00
-4.78538245e-01 -4.55667347e-01 2.50506401e-01 6.45296216e-01
-3.58797103e-01 -4.57084745e-01 2.41946131e-02 4.00319874e-01
3.01307678e-01 1.03015816e+00 4.44263875e-01 -1.08849549e+00
2.80069202e-01 -1.77979410e-01 5.16132712e-01 -2.25701347e-01
-9.97194588e-01 3.58728796e-01 -9.69390199e-02 -8.08967113e-01
-2.38751873e-01 -3.31823379e-01 -1.77373719e+00 -7.29606271e-01
7.85264745e-02 -4.85681027e-01 1.05788958e+00 5.72911382e-01
6.06612146e-01 9.39959764e-01 2.18338907e-01 -4.81484860e-01
3.21378782e-02 -6.32536829e-01 -5.38666904e-01 -7.38432348e-01
9.88409370e-02 -3.28808784e-01 -7.19341114e-02 -3.70723344e-02] | [8.260205268859863, 2.515133857727051] |
dd6c34de-af9f-490a-9c60-e0b79a6cdb32 | logistic-regression-models-for-aggregated | 1912.03805 | null | https://arxiv.org/abs/1912.03805v2 | https://arxiv.org/pdf/1912.03805v2.pdf | Logistic regression models for aggregated data | Logistic regression models are a popular and effective method to predict the probability of categorical response data. However inference for these models can become computationally prohibitive for large datasets. Here we adapt ideas from symbolic data analysis to summarise the collection of predictor variables into histogram form, and perform inference on this summary dataset. We develop ideas based on composite likelihoods to derive an efficient one-versus-rest approximate composite likelihood model for histogram-based random variables, constructed from low-dimensional marginal histograms obtained from the full histogram. We demonstrate that this procedure can achieve comparable classification rates compared to the standard full data multinomial analysis and against state-of-the-art subsampling algorithms for logistic regression, but at a substantially lower computational cost. Performance is explored through simulated examples, and analyses of large supersymmetry and satellite crop classification datasets. | ['Scott A. Sisson', 'Tom Whitaker', 'Boris Beranger'] | 2019-12-09 | null | null | null | null | ['crop-classification'] | ['miscellaneous'] | [ 6.65931880e-01 1.30632622e-02 -4.68325883e-01 -8.45927894e-01
-1.28514540e+00 -5.39544821e-01 5.91286004e-01 3.55203092e-01
-2.00110361e-01 1.23902309e+00 -1.63278177e-01 -7.22500801e-01
-4.30887699e-01 -1.01680434e+00 -8.22749376e-01 -8.36896956e-01
-3.65951449e-01 6.20695710e-01 -3.75864506e-02 3.87028337e-01
3.33397150e-01 5.19589305e-01 -1.82515550e+00 5.49304560e-02
7.60648251e-01 9.10962641e-01 1.12658463e-01 8.45054805e-01
2.90638208e-01 4.53283995e-01 -5.42071700e-01 7.97557384e-02
-3.52960765e-01 -2.55223900e-01 -3.29333693e-01 -3.58853005e-02
4.01538104e-01 -4.23572630e-01 1.21412754e-01 8.43551278e-01
2.44458362e-01 7.34495744e-02 1.41756546e+00 -1.55382240e+00
-5.95643997e-01 6.48157001e-01 -7.03230321e-01 -3.82061690e-01
4.88888800e-01 -3.22929442e-01 9.41511393e-01 -6.92608893e-01
3.57491672e-01 1.52789938e+00 9.89837110e-01 -4.59719598e-01
-2.29868865e+00 -7.52276540e-01 -3.47864866e-01 3.91016193e-02
-1.57776845e+00 -3.48472930e-02 2.00036168e-01 -3.70626688e-01
1.02886307e+00 4.62650120e-01 3.25633407e-01 8.86265814e-01
3.65313709e-01 5.15059948e-01 1.71655834e+00 -5.92360318e-01
3.89522284e-01 7.15400279e-02 2.27339923e-01 3.82941812e-01
3.45365047e-01 2.41435081e-01 -3.15766394e-01 -8.02277625e-01
5.72954595e-01 -1.47373602e-01 1.38606191e-01 -5.24678111e-01
-8.00971448e-01 1.39572036e+00 6.23071007e-02 -5.04333854e-01
-5.08190632e-01 2.01681435e-01 2.15823180e-03 2.11125258e-02
6.72501266e-01 -2.52720211e-02 -6.01095021e-01 9.99987964e-03
-1.10308075e+00 7.62171268e-01 1.18744433e+00 9.99409735e-01
8.77317667e-01 -1.45954683e-01 7.34669790e-02 8.44938874e-01
2.77961910e-01 1.03577173e+00 -1.27694726e-01 -8.71219158e-01
-1.40183419e-02 1.37016088e-01 1.83903158e-01 -9.67948556e-01
-4.62516129e-01 2.10139692e-01 -6.72235906e-01 3.94968510e-01
6.26546443e-01 -5.39265089e-02 -1.12257779e+00 1.62919581e+00
2.55213678e-01 -1.71704948e-01 -1.71098739e-01 2.85724998e-01
3.01559210e-01 7.88126111e-01 6.05451465e-01 -4.64232504e-01
1.13346076e+00 1.19105298e-02 -5.63707113e-01 4.77946438e-02
4.64440793e-01 -4.98062313e-01 8.34376574e-01 6.48790121e-01
-6.94940805e-01 -3.63032609e-01 -8.16165626e-01 -2.31132880e-02
-7.03783154e-01 1.51376545e-01 8.55417848e-01 8.30926597e-01
-6.94106162e-01 6.00438118e-01 -9.27192688e-01 -2.04849407e-01
3.00985217e-01 4.75226194e-01 -3.33063304e-01 4.00147624e-02
-9.72251713e-01 9.68348503e-01 3.51968348e-01 -6.91420734e-02
-5.85874975e-01 -5.33084273e-01 -1.13063252e+00 3.33823189e-02
1.36897594e-01 1.23244703e-01 1.16417801e+00 -3.32366407e-01
-9.99023080e-01 4.36526597e-01 -4.89314854e-01 -2.44165674e-01
-1.77305564e-02 4.48321998e-02 -1.09062605e-01 -1.34140760e-01
9.03174728e-02 7.51805007e-01 6.98038816e-01 -1.00765896e+00
-8.47568512e-01 -4.41403836e-01 -5.95789015e-01 -1.47599563e-01
1.00918310e-02 9.75342169e-02 2.73534089e-01 -3.98952395e-01
2.12880582e-01 -9.56356525e-01 -2.62939095e-01 -4.21191901e-02
-3.90693396e-01 -3.07199061e-01 5.52968025e-01 -9.75285232e-01
1.27600169e+00 -1.87917566e+00 1.30382359e-01 5.02449393e-01
-2.25565121e-01 -4.81367797e-01 1.53906867e-01 2.13306934e-01
-1.10782318e-01 7.05505088e-02 -6.23847008e-01 2.03434657e-02
1.76551208e-01 2.75415808e-01 -6.08337283e-01 7.19109893e-01
4.70674038e-01 4.89110529e-01 -6.02012455e-01 -6.53748274e-01
3.05155873e-01 4.44100916e-01 -3.63678247e-01 1.77217990e-01
-3.41671824e-01 -1.79983094e-01 -1.38765231e-01 9.28000271e-01
9.41899002e-01 -6.68291897e-02 4.45614338e-01 -4.05891836e-02
-2.05348462e-01 1.19091347e-01 -1.09609127e+00 8.94919813e-01
-8.19158927e-02 6.22616470e-01 -1.00078665e-01 -1.12179255e+00
1.17524219e+00 -8.53693783e-02 8.81551653e-02 -2.30575681e-01
-1.20794542e-01 8.50236565e-02 -4.35885817e-01 -9.39634964e-02
6.30055010e-01 -1.75629407e-01 -8.06806266e-01 1.85062647e-01
5.88595867e-02 -5.56694925e-01 9.55193788e-02 -2.84360647e-01
7.76998699e-01 6.76213264e-01 8.65131855e-01 -2.69001245e-01
-1.80460155e-01 4.36526597e-01 2.92808771e-01 1.07739818e+00
1.96792185e-01 3.67757797e-01 9.75379288e-01 -1.33882523e-01
-1.20807076e+00 -1.31160510e+00 -9.36400235e-01 1.40529561e+00
-2.55872995e-01 -1.69731960e-01 -3.38635743e-01 -2.00662613e-01
7.04240203e-01 1.21951377e+00 -8.10787320e-01 2.87014246e-01
-4.34469692e-02 -1.39225805e+00 6.36753440e-01 6.22762918e-01
-1.78981870e-01 -8.92453194e-01 -6.65834367e-01 2.29068056e-01
-5.24662016e-03 -7.50869393e-01 6.10188723e-01 9.23830330e-01
-7.41768897e-01 -8.28489244e-01 -3.30787897e-01 -6.03433490e-01
4.95055228e-01 -2.03441292e-01 9.69813287e-01 -5.46826124e-01
-5.63361645e-01 -9.35897138e-03 -2.07391918e-01 -4.08815444e-01
-4.53433722e-01 -5.86423278e-02 -5.22211045e-02 -7.28956461e-01
9.20383573e-01 -5.61568260e-01 6.20260425e-02 2.95668632e-01
-7.56986082e-01 -1.09598553e-03 5.36415637e-01 1.17701054e+00
8.22903395e-01 2.00797424e-01 5.63513696e-01 -8.52809072e-01
5.11578918e-01 -8.85389745e-01 -1.04234803e+00 4.40873861e-01
-6.18105173e-01 6.49781674e-02 5.96442521e-01 -6.27760231e-01
-8.58408749e-01 3.45679045e-01 2.91547835e-01 9.05413851e-02
-5.83171666e-01 9.42682981e-01 1.65775791e-01 1.11052036e-01
6.21551573e-01 1.35525092e-01 1.13286428e-01 -3.93348515e-01
4.72375125e-01 1.08099890e+00 5.99441230e-01 -5.67802131e-01
3.83409500e-01 3.11299413e-02 5.48986852e-01 -9.49401319e-01
-3.87185127e-01 4.48926128e-02 -8.16072345e-01 5.74213937e-02
6.34985805e-01 -7.99981952e-01 -8.50198150e-01 3.53209794e-01
-7.26472199e-01 -4.84659255e-01 1.04804367e-01 7.23659873e-01
-8.15356910e-01 1.21181406e-01 -3.50030720e-01 -1.20036924e+00
2.24546477e-01 -1.01200104e+00 1.29836941e+00 1.85192600e-01
-4.11250383e-01 -9.11680341e-01 1.09964803e-01 -1.57680497e-01
2.27230996e-01 5.37699223e-01 1.20768237e+00 -7.56880999e-01
-3.51699233e-01 -4.29389030e-01 -5.03648043e-01 -1.11917190e-01
-4.31213677e-02 4.64094847e-01 -8.99664998e-01 -1.73208565e-01
-4.50729728e-01 -8.22044909e-01 8.08207273e-01 8.58534753e-01
1.39335799e+00 -5.57757653e-02 -5.85840166e-01 5.88742673e-01
1.48980641e+00 -8.86435062e-02 4.97147918e-01 1.16574831e-01
2.88303941e-01 8.03944111e-01 1.07901788e+00 7.42802620e-01
4.80827808e-01 4.70629424e-01 7.51721561e-02 -1.10997679e-02
4.36479211e-01 -3.14363003e-01 2.26189852e-01 1.52346581e-01
3.77310812e-01 -1.96417838e-01 -8.79995406e-01 4.88787919e-01
-1.78425622e+00 -9.33109164e-01 -3.21892112e-01 2.31956029e+00
1.17008805e+00 -2.94611286e-02 2.92607099e-01 8.70213136e-02
6.86881959e-01 -7.22962245e-02 -5.27176440e-01 -4.66393828e-01
-2.37489522e-01 4.14003432e-01 9.58029568e-01 5.92138410e-01
-1.38243139e+00 7.48115301e-01 8.52656937e+00 1.14974773e+00
-5.33603370e-01 -3.90685052e-01 6.69492662e-01 1.09739691e-01
-2.35271826e-01 4.03224766e-01 -1.04451799e+00 3.15546662e-01
1.35812259e+00 -1.22902334e-01 4.75680947e-01 1.09026659e+00
7.37883570e-03 -8.52632105e-01 -9.90138173e-01 4.39450890e-01
-5.24371527e-02 -7.55611360e-01 -2.09968224e-01 1.53800875e-01
3.27546000e-01 -2.77486622e-01 4.52766716e-02 5.30309141e-01
7.39922523e-01 -1.44757402e+00 3.68481755e-01 5.29567122e-01
1.15255249e+00 -8.82792890e-01 4.58620220e-01 3.46070915e-01
-1.00897586e+00 -4.97857518e-02 -6.85952246e-01 -3.65350455e-01
4.60691638e-02 4.86279517e-01 -9.51604247e-01 1.41452923e-01
6.26541138e-01 3.86616081e-01 -6.99590087e-01 7.82888114e-01
-3.14612091e-02 8.44741106e-01 -8.02649498e-01 -5.44462502e-01
-1.54109508e-01 -7.47764483e-02 1.14337429e-01 1.43901968e+00
4.08875078e-01 1.17081687e-01 3.06403309e-01 9.60789382e-01
6.24578774e-01 -2.21075758e-01 -7.42365241e-01 1.34441912e-01
9.11896884e-01 1.10330772e+00 -8.16341937e-01 -2.86063462e-01
-3.81857187e-01 2.98575789e-01 3.44671160e-01 2.98664123e-01
-9.21221554e-01 -7.25328386e-01 1.35686263e-01 -2.92100161e-01
5.85243940e-01 -3.15086633e-01 -4.03160006e-01 -5.39794147e-01
-4.51765001e-01 -9.41307247e-01 3.58843952e-01 -9.01410639e-01
-1.35454226e+00 -6.36214986e-02 9.53046262e-01 -8.44680309e-01
-9.37882185e-01 -8.66431892e-01 1.70564856e-02 1.17077661e+00
-1.13522804e+00 -1.15212107e+00 -2.50810683e-01 2.28634968e-01
4.32674848e-02 1.15203544e-01 1.42518389e+00 -3.32640916e-01
-2.89601684e-01 4.95546192e-01 5.45647085e-01 -3.13167274e-01
5.61703563e-01 -1.55084550e+00 3.12846661e-01 3.82090271e-01
-4.99937564e-01 5.25154769e-01 1.02770722e+00 -1.06881022e+00
-1.40530384e+00 -7.01558113e-01 7.37632215e-01 -4.19090509e-01
8.54536176e-01 -4.56555814e-01 -9.88042772e-01 7.86544800e-01
-3.57238293e-01 -3.28398407e-01 8.89400601e-01 4.57465202e-01
-2.69943893e-01 1.83365360e-01 -1.32525933e+00 3.79526585e-01
4.24928814e-01 -6.02905631e-01 -5.75246096e-01 2.60220349e-01
4.67935473e-01 -1.34057254e-01 -1.34622395e+00 6.30886376e-01
1.23463845e+00 -5.27799666e-01 9.08569813e-01 -5.55390537e-01
3.54784817e-01 -2.21331939e-01 -7.95303166e-01 -1.07962978e+00
-4.03702050e-01 -3.67143959e-01 1.40589580e-01 1.08253002e+00
3.56231749e-01 -5.08143365e-01 5.70839882e-01 7.21584857e-01
5.29859364e-01 -5.12250960e-01 -8.89203429e-01 -6.15033984e-01
2.62221187e-01 -5.97075582e-01 4.81364101e-01 6.61418200e-01
2.76201904e-01 7.41349906e-02 -5.05309224e-01 3.10825706e-01
1.28566921e+00 3.65532726e-01 8.67977142e-01 -1.28013611e+00
-3.05589318e-01 -1.58396766e-01 -5.54048598e-01 -7.85232902e-01
3.57961535e-01 -5.62138975e-01 3.23735386e-01 -1.11405754e+00
4.83585626e-01 -5.86346984e-01 1.41552001e-01 6.72801256e-01
-1.97543263e-01 2.05168486e-01 -3.11522216e-01 -1.22213691e-01
-1.05707563e-01 3.15300584e-01 3.98758829e-01 2.05003232e-01
-2.01167598e-01 -6.74116313e-02 -3.17336559e-01 6.90570354e-01
6.15322530e-01 -8.06033075e-01 -1.14744931e-01 2.52685487e-01
1.24973819e-01 5.68223178e-01 7.02071846e-01 -5.27916789e-01
4.87775616e-02 -6.45517886e-01 1.00191772e+00 -1.27220774e+00
4.07122254e-01 -6.42302513e-01 4.38909262e-01 1.61106408e-01
-6.46240950e-01 -2.90106330e-02 3.13903034e-01 5.56110203e-01
6.97754025e-02 -2.48469144e-01 5.24520636e-01 3.89634788e-01
-3.77443463e-01 -3.55098039e-01 -5.09519458e-01 -4.41725373e-01
8.79418135e-01 -7.02372491e-02 -4.81656522e-01 -4.30730134e-01
-7.76461661e-01 1.18888043e-01 4.72547024e-01 -3.69739756e-02
5.65751553e-01 -1.11607015e+00 -9.01576519e-01 4.39964056e-01
-5.18249534e-02 -2.71888673e-01 -2.55310893e-01 5.95092475e-01
-4.45004255e-01 4.47211564e-01 -3.48396182e-01 -9.10255969e-01
-1.29646206e+00 4.45110589e-01 -2.16260418e-01 -1.40499204e-01
-2.38427252e-01 4.37766522e-01 -3.30718532e-02 -6.95633233e-01
3.23620200e-01 -5.50464094e-01 -1.86894313e-02 2.12500900e-01
3.58540416e-01 4.54806715e-01 -5.16892314e-01 -4.76264298e-01
-3.63334924e-01 4.85164553e-01 2.30918795e-01 -4.76542950e-01
1.47162759e+00 3.09304725e-02 -1.98586181e-01 8.33267033e-01
1.03048682e+00 -3.03873330e-01 -1.34846294e+00 1.48522347e-01
6.36046845e-03 -6.68429792e-01 2.04330888e-02 -7.22936749e-01
1.07555971e-01 6.10819876e-01 5.07172763e-01 6.23344600e-01
1.09620333e+00 -1.63257703e-01 -1.08054042e-01 4.40974206e-01
3.75209451e-01 -8.88478339e-01 -7.53653646e-01 1.46824390e-01
6.66470408e-01 -1.11990476e+00 7.34029472e-01 -3.86345625e-01
-4.51085985e-01 1.21105266e+00 9.42968950e-02 -2.30264664e-01
8.68203044e-01 5.81159890e-01 -5.38138449e-01 2.10125878e-01
-8.73946130e-01 2.73747556e-02 1.57262281e-01 9.63519037e-01
5.25183916e-01 6.46322191e-01 -2.70484328e-01 5.65582573e-01
-3.19123387e-01 8.03532377e-02 4.27447051e-01 9.85957623e-01
-5.96113563e-01 -8.79572630e-01 -8.05113614e-01 9.77683067e-01
-3.61279368e-01 -1.64037615e-01 -9.24442187e-02 1.14415693e+00
-2.95865506e-01 9.54122126e-01 3.03762376e-01 -2.94366896e-01
-2.77269911e-02 3.40310454e-01 5.30240238e-01 -1.92889616e-01
8.99325609e-02 3.60714734e-01 3.45243394e-01 -3.02957207e-01
-3.49663585e-01 -1.18410361e+00 -8.88677478e-01 -5.03034115e-01
-7.28030622e-01 -2.72770286e-01 9.77805734e-01 5.69421470e-01
-1.17638126e-01 1.75201654e-01 6.94622397e-01 -1.15836644e+00
-1.11207390e+00 -1.17982888e+00 -8.95006776e-01 2.99241897e-02
3.04514319e-01 -9.12158668e-01 -7.17996776e-01 2.56189369e-02] | [7.420284748077393, 4.2454423904418945] |
873ad3ff-d988-4267-a587-955d8e537554 | random-forests-versus-neural-networks-whats | 1609.05797 | null | http://arxiv.org/abs/1609.05797v3 | http://arxiv.org/pdf/1609.05797v3.pdf | Random Forests versus Neural Networks - What's Best for Camera Localization? | This work addresses the task of camera localization in a known 3D scene given
a single input RGB image. State-of-the-art approaches accomplish this in two
steps: firstly, regressing for every pixel in the image its 3D scene coordinate
and subsequently, using these coordinates to estimate the final 6D camera pose
via RANSAC. To solve the first step, Random Forests (RFs) are typically used.
On the other hand, Neural Networks (NNs) reign in many dense regression tasks,
but are not test-time efficient. We ask the question: which of the two is best
for camera localization? To address this, we make two method contributions: (1)
a test-time efficient NN architecture which we term a ForestNet that is derived
and initialized from a RF, and (2) a new fully-differentiable robust averaging
technique for regression ensembles which can be trained end-to-end with a NN.
Our experimental findings show that for scene coordinate regression,
traditional NN architectures are superior to test-time efficient RFs and
ForestNets, however, this does not translate to final 6D camera pose accuracy
where RFs and ForestNets perform slightly better. To summarize, our best
method, a ForestNet with a robust average, which has an equivalent fast and
lightweight RF, improves over the state-of-the-art for camera localization on
the 7-Scenes dataset. While this work focuses on scene coordinate regression
for camera localization, our innovations may also be applied to other
continuous regression tasks. | ['Alexander Krull', 'Philip H. S. Torr', 'Eric Brachmann', 'Carsten Rother', 'Daniela Massiceti'] | 2016-09-19 | null | null | null | null | ['camera-localization'] | ['computer-vision'] | [ 3.90721768e-01 -4.66247052e-01 -6.91530406e-02 -4.93471563e-01
-9.96268332e-01 -7.85865963e-01 4.24199998e-01 -5.66758990e-01
-6.33204162e-01 4.71692383e-01 -1.05149828e-01 -3.55921447e-01
-2.61863787e-02 -4.59483355e-01 -1.01329422e+00 -8.36923659e-01
2.68203795e-01 4.63257134e-01 7.93036446e-03 2.97298074e-01
1.82203919e-01 8.72083962e-01 -1.25284958e+00 -3.35054874e-01
5.83302796e-01 1.34299636e+00 1.97374612e-01 7.81534612e-01
1.31526321e-01 9.66427505e-01 -3.48679692e-01 -2.76643276e-01
4.78806645e-01 -1.29153505e-01 -5.45822918e-01 5.46404719e-02
9.59413230e-01 -3.61636937e-01 -1.49412811e-01 8.88104141e-01
4.26017612e-01 2.25555241e-01 5.90045452e-01 -1.00981593e+00
-2.72489399e-01 2.62340158e-01 -6.70376301e-01 -2.14931309e-01
3.51924390e-01 5.25629297e-02 6.42093301e-01 -1.04956055e+00
3.90230745e-01 9.75813031e-01 1.21580446e+00 3.57877791e-01
-1.32922232e+00 -7.73569226e-01 3.56681108e-01 -1.64036751e-01
-1.58197021e+00 -4.01628256e-01 8.09668303e-01 -2.92580426e-01
1.05307841e+00 2.49596797e-02 6.87317371e-01 1.24618936e+00
2.45459452e-02 6.38171434e-01 1.34393382e+00 -3.20412457e-01
2.22877309e-01 -1.82607859e-01 -8.74226689e-02 9.00531113e-01
1.44222900e-01 3.66495401e-02 -4.96353805e-01 1.58898413e-01
9.41628397e-01 2.33758554e-01 -3.21137518e-01 -8.20284486e-01
-1.39339697e+00 8.82919729e-01 8.11962128e-01 5.74989431e-02
-4.95336354e-01 5.59547782e-01 -1.61261968e-02 6.44356385e-02
6.31247878e-01 6.34347022e-01 -5.84282398e-01 -1.02658443e-01
-1.30542922e+00 1.26895800e-01 8.88595164e-01 1.05252397e+00
1.04445231e+00 -6.02142885e-02 2.18659461e-01 6.60206079e-01
3.39249462e-01 8.76047313e-01 1.87842950e-01 -1.08715665e+00
4.78518754e-01 4.20777768e-01 -6.91926628e-02 -8.44298303e-01
-5.13915539e-01 -6.46137178e-01 -9.17157650e-01 1.80875286e-01
5.24676323e-01 -8.58466253e-02 -1.11835289e+00 1.61894965e+00
1.63505077e-01 4.99135345e-01 -9.42317918e-02 1.02410352e+00
6.57445550e-01 2.97241271e-01 -1.37016550e-01 5.14772236e-02
9.54011798e-01 -1.29597473e+00 -1.55590639e-01 -6.09970927e-01
2.88592190e-01 -8.14261913e-01 8.26114237e-01 2.83713847e-01
-7.28447497e-01 -6.23075306e-01 -1.11068809e+00 -7.44960085e-02
-5.45897961e-01 4.30884629e-01 7.65032411e-01 5.45240819e-01
-1.26787305e+00 4.23297673e-01 -9.45467651e-01 -5.90812564e-01
3.18084180e-01 6.54868662e-01 -5.67979515e-01 -3.28535289e-01
-5.52390933e-01 1.23626792e+00 -1.20283425e-01 4.55060542e-01
-9.64524686e-01 -4.52145606e-01 -9.44175601e-01 -3.64460915e-01
5.07632315e-01 -1.04339862e+00 1.17864203e+00 -8.88664246e-01
-1.68381965e+00 8.02396297e-01 -4.33467209e-01 -5.62949181e-01
5.50063193e-01 -4.91211534e-01 3.71819101e-02 -5.39101735e-02
3.13599885e-01 6.68846726e-01 1.05218351e+00 -1.20964944e+00
-5.12063384e-01 -5.31594336e-01 -1.11273333e-01 3.14191282e-01
1.97932646e-01 -2.52502352e-01 -6.95595086e-01 -4.11277235e-01
5.37438154e-01 -1.26757073e+00 -4.65028256e-01 -1.34586245e-01
-4.38938230e-01 6.19724803e-02 7.56875396e-01 -4.51985240e-01
8.23901176e-01 -1.94516587e+00 2.81614304e-01 2.26391017e-01
1.21091917e-01 -1.94641836e-02 -2.14714974e-01 -1.55072302e-01
-7.17336610e-02 2.22057030e-02 -3.96398187e-01 -6.36608243e-01
-2.51586467e-01 5.55932745e-02 -3.79260898e-01 8.24499428e-01
1.55149415e-01 8.80293429e-01 -7.16762960e-01 -3.10201675e-01
5.93343854e-01 7.11213708e-01 -3.49380583e-01 1.70476422e-01
7.73841292e-02 5.84212542e-01 -3.40698332e-01 8.40566278e-01
8.46384645e-01 -1.88752979e-01 -2.76722997e-01 -3.10830861e-01
-2.02338696e-01 5.18836603e-02 -1.15637565e+00 2.03445673e+00
-7.12133884e-01 6.18287325e-01 5.94924688e-02 -8.06470871e-01
9.91931558e-01 -1.70131400e-01 4.93125975e-01 -4.00800109e-01
1.26132760e-02 1.73495770e-01 -5.48471570e-01 -1.51509792e-01
4.53462243e-01 9.75101367e-02 -1.43492237e-01 1.76651642e-01
2.94995934e-01 -6.46277010e-01 -1.01238489e-01 -2.16295585e-01
1.32844841e+00 8.04988444e-01 2.61281013e-01 1.17143683e-01
6.26465559e-01 1.24393411e-01 3.11385512e-01 9.86104012e-01
-4.10708971e-02 9.18002665e-01 1.13299161e-01 -5.54425299e-01
-8.13275218e-01 -1.15167594e+00 1.01165064e-01 9.44012105e-01
2.56301224e-01 -2.79824227e-01 -8.09405506e-01 -9.62281346e-01
4.07358594e-02 5.09358704e-01 -4.66918707e-01 2.38623723e-01
-7.09563971e-01 -4.59333777e-01 4.26310152e-01 6.90247595e-01
8.04886878e-01 -6.80269539e-01 -6.96506679e-01 1.41114295e-02
-1.91908360e-01 -1.34332180e+00 -4.73062575e-01 6.90024912e-01
-9.87319350e-01 -1.14123487e+00 -7.43264079e-01 -5.53087831e-01
8.45587194e-01 7.41803885e-01 1.20983148e+00 -3.18921536e-01
8.66878703e-02 7.08451331e-01 -1.68511614e-01 -3.21755677e-01
1.06013782e-01 3.23566526e-01 1.26444578e-01 -1.29465804e-01
4.72540319e-01 -6.67880774e-01 -5.92564940e-01 4.89778101e-01
-4.50498998e-01 -3.93075868e-02 1.03292370e+00 6.54924691e-01
1.03595054e+00 -5.26863746e-02 -3.08239222e-01 -8.03033471e-01
2.53370076e-01 -1.41483068e-01 -9.14459050e-01 1.84829980e-01
-4.54770863e-01 1.77623853e-01 7.40001142e-01 -4.33634132e-01
-6.87576175e-01 9.52448964e-01 -9.53684375e-02 -7.86284447e-01
-3.63948613e-01 2.71970928e-01 1.77214071e-01 -5.78059971e-01
6.26736104e-01 1.95529521e-01 -5.87052256e-02 -3.53429735e-01
5.82604825e-01 3.41102511e-01 7.33761549e-01 -3.90187263e-01
1.31614673e+00 5.13326526e-01 3.64129841e-01 -6.89231277e-01
-1.16411614e+00 -7.49037504e-01 -1.18177009e+00 -1.85392499e-01
1.10745287e+00 -1.29990029e+00 -6.22911274e-01 5.00924826e-01
-1.10163975e+00 -4.32372242e-01 -1.84118897e-01 6.59418523e-01
-7.97966421e-01 4.40412723e-02 -1.54382288e-01 -7.73194790e-01
-2.16126546e-01 -1.20623255e+00 1.52524638e+00 2.30258897e-01
-3.64520564e-03 -6.25195682e-01 6.71992451e-02 2.86727905e-01
4.92000043e-01 2.77860910e-01 3.27358991e-01 -2.29569763e-01
-9.16972995e-01 -4.15022761e-01 -3.47251087e-01 4.30295974e-01
-8.75675231e-02 -6.79524764e-02 -1.15403581e+00 -2.47778043e-01
9.70423296e-02 -2.05242708e-01 1.14854312e+00 6.86530530e-01
9.95813549e-01 6.05102219e-02 -2.48166099e-01 1.30524039e+00
1.68071759e+00 -1.31199732e-01 3.93924624e-01 5.67784548e-01
1.01185250e+00 1.72798410e-01 5.57454467e-01 -1.17456608e-01
5.45742273e-01 6.24329627e-01 7.33815432e-01 -3.36036026e-01
-1.45271778e-01 -4.97734874e-01 5.52676558e-01 6.86374366e-01
-5.19559920e-01 1.17920838e-01 -7.70970702e-01 9.75775197e-02
-1.87640262e+00 -7.74481833e-01 1.04479212e-02 2.28053761e+00
1.55585706e-01 -1.14432521e-01 8.32044333e-02 -1.04342818e-01
5.48798263e-01 3.37989181e-01 -7.28271663e-01 1.12319045e-01
-2.08696246e-01 3.24213386e-01 1.27037978e+00 2.39362195e-01
-1.56006062e+00 1.13217568e+00 6.16281462e+00 5.55871129e-01
-1.36298943e+00 1.47161588e-01 4.39343035e-01 -1.06892325e-01
2.81185567e-01 2.97773570e-01 -1.02064931e+00 1.12088054e-01
6.93149328e-01 6.37058973e-01 9.57481980e-01 1.26327968e+00
-8.78153816e-02 -2.88098633e-01 -1.08080041e+00 1.31097031e+00
4.83118117e-01 -1.01207387e+00 -3.83159488e-01 8.68983939e-02
6.19755805e-01 5.89521408e-01 2.48686355e-02 3.41681480e-01
5.07003784e-01 -1.13829494e+00 8.65462661e-01 6.23357892e-01
8.80548358e-01 -6.61386490e-01 8.65925550e-01 3.06008279e-01
-1.27373993e+00 -1.84744865e-01 -5.03762007e-01 7.59618431e-02
9.37415510e-02 5.25003016e-01 -8.25342119e-01 4.97656941e-01
9.20725346e-01 8.95110071e-01 -9.38643038e-01 1.10354650e+00
-6.56304598e-01 4.14434940e-01 -5.34905195e-01 -9.05367285e-02
3.81280214e-01 -3.81830186e-01 3.55086386e-01 1.10449398e+00
5.52830219e-01 -3.78158540e-01 1.51461706e-01 7.88367391e-01
-5.61012030e-02 -2.24159285e-01 -9.71947193e-01 3.79256248e-01
3.41436416e-01 1.42971444e+00 -8.75310242e-01 1.21717550e-01
-4.78911251e-01 1.06932223e+00 4.88899618e-01 4.94825572e-01
-9.60991383e-01 -1.16153024e-01 4.95182276e-01 -7.49270767e-02
5.42100906e-01 -6.59630477e-01 -2.12190658e-01 -1.26057100e+00
6.36628717e-02 -6.24967575e-01 7.17319399e-02 -1.05546010e+00
-1.33620191e+00 6.06050313e-01 -1.90079674e-01 -1.32668698e+00
-4.62829351e-01 -8.90870035e-01 -2.06467539e-01 7.25422800e-01
-1.62547624e+00 -1.56380391e+00 -6.92552865e-01 8.01099718e-01
5.03024399e-01 5.23977401e-03 8.81412268e-01 -3.71780768e-02
-4.02328819e-01 2.17786267e-01 -4.09035571e-03 2.70521432e-01
7.76946843e-01 -1.29720581e+00 4.13124681e-01 1.02348173e+00
5.08204401e-01 7.30835378e-01 2.68030822e-01 -3.22328836e-01
-1.79895532e+00 -1.24253106e+00 6.14699602e-01 -9.05251741e-01
4.80534047e-01 -5.68226635e-01 -2.12410793e-01 9.40515697e-01
-1.31500065e-01 2.95742571e-01 2.37842485e-01 2.92951435e-01
-3.61126930e-01 -4.63319838e-01 -9.87967134e-01 4.33855265e-01
1.25085604e+00 -7.03742206e-01 -1.49906188e-01 2.73623973e-01
5.72140574e-01 -7.16613710e-01 -6.82318628e-01 5.56011260e-01
6.75335407e-01 -1.08397353e+00 1.19507039e+00 -2.51456201e-02
1.97518840e-01 -6.69920385e-01 -4.55699980e-01 -1.25410712e+00
-1.72467560e-01 -4.67452407e-01 8.99468362e-02 1.01762187e+00
4.07408088e-01 -6.15766346e-01 9.43673253e-01 3.81472737e-01
1.54323559e-02 -6.77359760e-01 -1.06237292e+00 -6.21656716e-01
-2.27551147e-01 -8.13453138e-01 5.04304886e-01 5.30478776e-01
-9.96007323e-01 5.29815197e-01 -4.63741243e-01 2.78066456e-01
6.64845943e-01 6.30363673e-02 1.29145062e+00 -1.18060851e+00
-6.07430153e-02 -4.00183022e-01 -4.20503736e-01 -1.47558320e+00
3.22313905e-01 -6.61256552e-01 4.15479213e-01 -1.38649499e+00
1.53169185e-01 -4.63236928e-01 -1.04108326e-01 2.89919704e-01
5.22233248e-02 4.01020497e-01 3.63066345e-01 3.15429777e-01
-8.37863863e-01 2.32357815e-01 9.68424916e-01 -1.47103712e-01
-1.17871530e-01 4.58561242e-01 -5.59516609e-01 9.54334795e-01
2.85894901e-01 -4.41630036e-01 -2.37763166e-01 -5.07748485e-01
1.99852273e-01 -1.55687645e-01 6.81692243e-01 -1.29777253e+00
6.20336175e-01 2.03568131e-01 8.76591086e-01 -8.10065210e-01
5.14531851e-01 -1.20060301e+00 9.95655730e-02 9.32566375e-02
2.83053108e-02 3.25838596e-01 -1.11293323e-01 6.33646011e-01
-1.25350043e-01 -1.05665259e-01 6.81993484e-01 -4.27427620e-01
-7.48949409e-01 3.97823304e-01 -1.71892911e-01 -2.93060809e-01
7.63830662e-01 -5.28896451e-01 -1.51245594e-01 -5.61806798e-01
-4.24936354e-01 -1.11377507e-01 6.41058743e-01 2.55163401e-01
5.82981229e-01 -1.10547113e+00 -4.46903378e-01 3.26166630e-01
-9.45030898e-02 5.48430562e-01 -2.16127649e-01 1.06983268e+00
-7.06897318e-01 4.41125005e-01 2.78579891e-01 -1.11958265e+00
-1.05451870e+00 4.24341977e-01 4.13866103e-01 -4.25709963e-01
-2.83234417e-01 9.90416467e-01 -2.43016351e-02 -9.92999613e-01
4.20872688e-01 -4.45856750e-01 1.26429141e-01 -1.34031773e-01
6.20076209e-02 1.69071227e-01 1.13741525e-01 -7.99278915e-01
-5.78714550e-01 1.35869014e+00 3.46995085e-01 2.10060067e-02
1.60056889e+00 -1.04791321e-01 -1.47970691e-01 3.11322451e-01
1.16301990e+00 4.06084768e-02 -1.50590575e+00 -9.84791759e-03
-1.10760033e-01 -5.00131726e-01 1.72215000e-01 -8.83474767e-01
-1.22997952e+00 7.50511408e-01 7.10857511e-01 -1.82379827e-01
1.33959103e+00 1.65061988e-02 3.56947660e-01 6.48744285e-01
5.89206278e-01 -8.21446776e-01 -1.62735507e-01 7.29257703e-01
6.31715953e-01 -1.33095419e+00 1.85557976e-01 -2.12640539e-01
-5.36641300e-01 1.30881548e+00 5.15944779e-01 -3.63671601e-01
5.32835543e-01 4.27948087e-01 1.94246277e-01 -4.87541296e-02
-2.15231687e-01 -4.75176573e-01 1.76247045e-01 7.20081568e-01
3.74475598e-01 -1.44497186e-01 5.15695512e-01 2.11024389e-01
-4.14333254e-01 -5.89316674e-02 -8.02492499e-02 6.19832397e-01
-1.72877859e-03 -7.04284787e-01 -4.47184563e-01 3.04908067e-01
-1.99735135e-01 -1.72750562e-01 -6.01198494e-01 1.02346325e+00
1.82197854e-01 8.99824679e-01 5.59075177e-03 -5.79023778e-01
4.78528053e-01 -9.63409767e-02 6.30798459e-01 -3.74484122e-01
-4.23861176e-01 2.63351686e-02 -1.06208600e-01 -9.17649329e-01
-8.73267233e-01 -6.56149447e-01 -6.22206807e-01 -2.12077156e-01
-6.98670924e-01 -1.73886091e-01 1.30962026e+00 9.16775763e-01
1.15729719e-01 3.00236940e-01 6.76401556e-01 -1.41377652e+00
-4.67156589e-01 -9.66690481e-01 -3.11783940e-01 -2.64091901e-02
4.93524134e-01 -7.05789149e-01 -6.02083564e-01 -1.98646516e-01] | [7.738099575042725, -2.1276705265045166] |
c11d6859-92b8-4396-9132-03ffbab5dee3 | equity-beyond-bias-in-language-technologies | null | null | https://aclanthology.org/W19-4446 | https://aclanthology.org/W19-4446.pdf | Equity Beyond Bias in Language Technologies for Education | There is a long record of research on equity in schools. As machine learning researchers begin to study fairness and bias in earnest, language technologies in education have an unusually strong theoretical and applied foundation to build on. Here, we introduce concepts from culturally relevant pedagogy and other frameworks for teaching and learning, identifying future work on equity in NLP. We present case studies in a range of topics like intelligent tutoring systems, computer-assisted language learning, automated essay scoring, and sentiment analysis in classrooms, and provide an actionable agenda for research. | ['Alan W. black', "Ezekiel Dixon-Rom{\\'a}n", 'Shrimai Prabhumoye', 'Michael Madaio', 'Brittany McLaughlin', 'David Gerritsen', 'Elijah Mayfield'] | 2019-08-01 | null | null | null | ws-2019-8 | ['automated-essay-scoring'] | ['natural-language-processing'] | [-9.87236053e-02 4.24486160e-01 -1.10224593e+00 -7.56862164e-01
-1.58418998e-01 -5.02224743e-01 2.52704382e-01 9.11824763e-01
-5.87894499e-01 8.74190032e-01 6.72705591e-01 -9.28609073e-01
-2.53582299e-01 -7.20792472e-01 -2.85662591e-01 4.69183996e-02
6.17757976e-01 5.30033484e-02 -2.47448772e-01 -4.96225268e-01
1.08371711e+00 2.40588054e-01 -1.67655182e+00 -1.43500656e-01
1.75090206e+00 2.67101917e-02 -3.69028032e-01 4.26609695e-01
-6.19539142e-01 1.98684227e+00 -6.95735514e-01 -1.34520984e+00
-3.33842963e-01 -8.40503752e-01 -1.26576412e+00 -5.89054644e-01
1.24351418e+00 -7.23103404e-01 2.06202865e-01 1.32154942e+00
7.16591239e-01 1.54422432e-01 5.11925757e-01 -1.42516673e+00
-1.39468086e+00 7.29526818e-01 -5.36807120e-01 2.97693819e-01
5.75253069e-01 -2.26160735e-01 9.25860703e-01 -3.44743967e-01
5.03222883e-01 1.19940829e+00 6.18125618e-01 1.11311162e+00
-9.16710436e-01 -1.14989150e+00 1.02080993e-01 2.13543892e-01
-1.78572208e-01 -3.58822286e-01 2.92987198e-01 -8.15868795e-01
3.01590860e-01 3.51120263e-01 1.55071628e+00 8.80340993e-01
4.32655126e-01 1.06837583e+00 1.76588058e+00 -6.92358851e-01
4.91154529e-02 7.87497401e-01 6.39145970e-01 9.10342753e-01
4.94816631e-01 -3.05026341e-02 -1.15935326e+00 2.14513138e-01
4.59810734e-01 -8.90758857e-02 1.24511540e-01 -1.79977179e-01
-5.80758512e-01 1.50711274e+00 5.51379169e-04 -3.97631861e-02
3.13906878e-01 1.39836043e-01 3.22807938e-01 6.03531361e-01
6.04528069e-01 7.85525024e-01 -1.96392581e-01 -6.65503204e-01
-9.96526301e-01 4.25764143e-01 1.05433249e+00 8.41295719e-01
1.68734580e-01 1.57148242e-02 -2.32293740e-01 9.01192546e-01
6.24513209e-01 4.94635403e-01 6.21163785e-01 -1.46024251e+00
1.26558930e-01 4.26158696e-01 -2.87613720e-01 -8.25343668e-01
9.20520872e-02 -1.52832195e-01 1.48212701e-01 4.87091690e-01
6.06732249e-01 -2.27483109e-01 -3.12420309e-01 1.59015739e+00
5.54123998e-01 -2.41169557e-01 -1.40916631e-01 5.20285010e-01
1.22192490e+00 -1.16298787e-01 6.68246388e-01 -5.65375052e-02
1.34750712e+00 -1.09187353e+00 -9.10466254e-01 -1.56100139e-01
8.56442869e-01 -1.11814737e+00 1.03034794e+00 4.32716429e-01
-1.46370983e+00 8.77474919e-02 -5.70477843e-01 -6.49796963e-01
-4.42211539e-01 -5.28814733e-01 8.03606570e-01 1.63458991e+00
-1.11432672e+00 6.18356645e-01 -2.72765398e-01 -5.12703955e-01
1.00595248e+00 -9.71522778e-02 9.98322591e-02 -1.15375243e-01
-9.86500740e-01 1.19438577e+00 -2.90041029e-01 -5.93756974e-01
-5.15013576e-01 -1.35647166e+00 -9.43753600e-01 4.43458222e-02
1.59632061e-02 -6.85476959e-01 1.58858681e+00 -1.21616685e+00
-1.89324510e+00 1.50505531e+00 1.25487432e-01 5.30611351e-02
3.41042519e-01 -5.71656168e-01 6.68800101e-02 -8.30705687e-02
5.47098398e-01 7.06517756e-01 1.72725514e-01 -6.32340491e-01
-7.63684571e-01 -4.35337365e-01 9.49034244e-02 3.77917230e-01
-8.32224369e-01 8.09921801e-01 1.13898492e+00 -4.87071067e-01
-3.32697600e-01 -4.48943466e-01 1.41689911e-01 1.79587573e-01
3.33282739e-01 -7.16434658e-01 2.87734300e-01 -6.82698250e-01
1.30838382e+00 -1.64719009e+00 -6.44516945e-01 -9.97540280e-02
7.05804944e-01 7.10842162e-02 8.92830417e-02 4.38730687e-01
3.15398304e-03 5.44310510e-01 5.84753871e-01 2.60669678e-01
4.63438243e-01 -1.04891896e-01 -8.07779208e-02 3.31363827e-01
-3.38803291e-01 5.09649575e-01 -1.61934114e+00 -9.16789889e-01
7.94743225e-02 1.40524477e-01 -8.41157019e-01 2.83718646e-01
1.74513742e-01 3.98777463e-02 -3.21812093e-01 5.08701026e-01
3.09932321e-01 2.11345285e-01 -9.98569839e-03 1.08257174e+00
-6.50647283e-01 9.89264846e-01 -7.47690856e-01 1.33161986e+00
-4.58015233e-01 1.04982305e+00 -8.14408883e-02 -6.68631136e-01
7.99382091e-01 2.06594735e-01 2.00524047e-01 -9.21649754e-01
-7.31903315e-02 2.17763394e-01 2.73851603e-01 -7.91454673e-01
9.07872617e-01 -2.54840672e-01 1.65005341e-01 1.06220472e+00
2.13985518e-01 -7.00435579e-01 1.69841290e-01 3.57846409e-01
5.18392503e-01 4.00204331e-01 3.84952605e-01 -8.57293487e-01
1.72326371e-01 2.08382737e-02 5.52141190e-01 5.00378311e-01
-9.32760656e-01 -1.05444565e-01 5.63571870e-01 -3.80146563e-01
-5.81483781e-01 -6.75522268e-01 -4.47697312e-01 2.06105590e+00
-1.52112037e-01 -1.98174551e-01 -1.05264783e+00 -7.33273089e-01
9.46998745e-02 1.06669533e+00 -4.45026875e-01 9.64831561e-02
-8.56664106e-02 -3.25612009e-01 4.41367626e-01 2.68420517e-01
4.10290182e-01 -7.62286365e-01 -9.06980872e-01 -1.10348091e-01
2.03081459e-01 -5.58524966e-01 -3.79950047e-01 -8.00531730e-02
-8.95681798e-01 -1.08588243e+00 -4.03472483e-01 -1.18891835e+00
2.93874204e-01 2.81458676e-01 1.53462160e+00 7.09107518e-01
-8.66490975e-02 9.08724129e-01 7.04321489e-02 -1.42731261e+00
-4.53826755e-01 -1.59323007e-01 -8.76981616e-02 -1.13822007e+00
1.26051748e+00 -4.41823095e-01 -6.31040573e-01 -2.91233540e-01
-4.35269862e-01 5.40755577e-02 -9.16092247e-02 8.29566777e-01
-2.68904239e-01 -4.44613874e-01 1.02733767e+00 -1.25793910e+00
1.04806411e+00 -5.83152592e-01 -4.17836457e-01 2.15935022e-01
-1.27084672e+00 -4.13086981e-01 1.60816759e-01 9.77269188e-02
-1.13312352e+00 -9.62821305e-01 -1.87439933e-01 6.21356308e-01
-1.21162280e-01 2.55034596e-01 2.97122806e-01 -5.44346511e-01
7.92331040e-01 -3.62103730e-01 1.18478060e-01 -1.12101445e-02
2.80313730e-01 5.15895009e-01 1.98639315e-02 -1.15600777e+00
2.44021788e-01 -1.93374902e-01 -2.43379384e-01 -8.78311038e-01
-1.16258407e+00 -1.61399189e-02 -1.97965309e-01 -6.18121564e-01
8.20349634e-01 -1.05166793e+00 -1.26301610e+00 1.67218119e-01
-6.91377103e-01 -4.93449628e-01 -5.41397572e-01 1.02008247e+00
-2.38789827e-01 -1.47809714e-01 -8.70850027e-01 -8.74879479e-01
-2.34347388e-01 -9.56682384e-01 3.02837849e-01 1.21882844e+00
-7.01467097e-01 -1.57993996e+00 5.85952461e-01 1.22040153e+00
3.54695499e-01 5.94738126e-03 1.23548698e+00 -1.02950370e+00
-1.68712214e-01 6.41661227e-01 2.77082533e-01 2.79364139e-01
-4.85273689e-01 5.31776547e-01 -9.68173862e-01 3.04040372e-01
3.89576964e-02 -1.01412201e+00 3.38490009e-01 3.85325104e-01
9.36971605e-01 -5.36465526e-01 3.32320005e-01 1.08402170e-01
1.33287299e+00 3.95245552e-02 3.13510209e-01 8.52659762e-01
4.82345998e-01 1.29443061e+00 4.99683291e-01 2.29565620e-01
1.14824212e+00 -2.52631634e-01 -2.80234199e-02 2.99089879e-01
3.58533785e-02 -5.56849539e-01 3.58654290e-01 1.40479338e+00
-4.26764674e-02 3.78603727e-01 -1.06946778e+00 8.10757935e-01
-1.28798509e+00 -1.43122506e+00 -2.88596421e-01 2.09050679e+00
1.46626770e+00 -2.43671939e-01 2.84548372e-01 -1.68404669e-01
5.17038584e-01 1.28032997e-01 -8.04824010e-02 -1.51007462e+00
4.69606906e-01 6.68472826e-01 1.47880673e-01 9.23036516e-01
-4.40778226e-01 9.82927144e-01 7.32558012e+00 6.14652455e-01
-9.34604943e-01 -1.15032513e-02 1.04948258e+00 -2.58220345e-01
-1.24583471e+00 -3.05240095e-01 -8.09154630e-01 3.51126045e-01
8.86990547e-01 -5.49942434e-01 4.95914035e-02 9.04782295e-01
7.30119571e-02 -4.20462310e-01 -1.09064305e+00 3.01727831e-01
1.29344836e-01 -9.86370206e-01 -5.46430588e-01 -3.13809304e-03
1.48052704e+00 -2.77024746e-01 4.78133500e-01 3.97523254e-01
1.05391359e+00 -1.39727151e+00 6.93428516e-01 1.94337610e-02
3.07944566e-01 -8.03873897e-01 5.65566897e-01 -7.03723505e-02
4.15731035e-02 2.08294280e-02 -3.68323654e-01 -8.27661514e-01
-9.11105216e-01 3.64562571e-01 -5.40131390e-01 -1.71379521e-01
8.28364193e-01 8.78849685e-01 -4.41079587e-01 1.16118944e+00
-6.45675361e-01 9.21227872e-01 3.61045510e-01 -7.21921921e-01
2.95550525e-02 -5.78173935e-01 -1.22968191e-02 1.12572992e+00
-5.15262224e-02 4.02065814e-01 -1.95794716e-01 6.07480705e-01
-2.89305359e-01 6.99953675e-01 -5.81898391e-01 -5.98117448e-02
8.89981687e-01 1.21682811e+00 -5.46696842e-01 1.07309204e-02
-7.09569275e-01 2.09340230e-01 3.89666080e-01 1.77937433e-01
-4.29589003e-01 -2.10386544e-01 1.18510163e+00 1.07262433e-02
-5.10825396e-01 1.61128357e-01 -9.66144681e-01 -1.05973339e+00
-7.66749084e-01 -1.42468095e+00 2.76484638e-01 -3.49184364e-01
-1.23184037e+00 -6.84695244e-01 -2.80493826e-01 -4.48120296e-01
-1.67589188e-01 -4.99216735e-01 -7.92668462e-01 6.96151435e-01
-1.59242117e+00 -6.22450650e-01 -1.78502217e-01 1.58161759e-01
2.24365249e-01 -1.00311004e-01 8.24392378e-01 1.57024935e-01
-6.33539557e-01 8.18689704e-01 3.17902833e-01 -1.26740828e-01
1.31873608e+00 -1.64123440e+00 -1.80964723e-01 5.40717125e-01
-7.55764991e-02 5.08878648e-01 6.52363658e-01 -4.55414712e-01
-1.19214272e+00 -3.25787693e-01 1.50307631e+00 -4.83889341e-01
7.29400933e-01 1.54507324e-01 -6.47428572e-01 7.97949672e-01
1.01494277e+00 -8.85702670e-01 1.77343571e+00 4.82259661e-01
-4.13507313e-01 1.65272310e-01 -1.44515169e+00 6.89671934e-01
9.82621312e-01 -8.29047143e-01 -6.79731011e-01 5.01078546e-01
3.85467589e-01 -7.89846420e-01 -1.19538391e+00 -4.90250260e-01
9.43614423e-01 -1.26331353e+00 6.68572605e-01 -9.04662549e-01
1.63383842e+00 7.39986598e-01 4.41962391e-01 -1.28924477e+00
-1.90340668e-01 -7.66491354e-01 3.59812498e-01 1.63024294e+00
-5.95368482e-02 -6.37002409e-01 1.16919196e+00 1.47389758e+00
-8.43034387e-02 -1.09611535e+00 -4.99798149e-01 9.31426659e-02
1.12454712e+00 6.32361248e-02 7.02285230e-01 1.84810090e+00
5.09799421e-01 3.30463022e-01 2.23770842e-01 -5.88844955e-01
7.52925754e-01 7.29546603e-03 5.62073290e-01 -1.37617493e+00
3.68605793e-01 -1.06551647e+00 -2.55394101e-01 -5.89224696e-01
4.86573488e-01 -7.96842515e-01 -3.29946458e-01 -1.39699495e+00
4.89493102e-01 -3.34788054e-01 1.16992004e-01 8.48199055e-02
-5.79407811e-01 6.33533373e-02 1.59503534e-01 -6.41452670e-01
-5.82651079e-01 3.45844120e-01 1.56844389e+00 1.28843755e-01
1.19046584e-01 -3.58431429e-01 -1.58536041e+00 9.94992375e-01
9.35914040e-01 -5.72688103e-01 -5.04811883e-01 -5.78375459e-01
5.73728144e-01 -2.44901687e-01 -5.87698780e-02 -2.57377714e-01
1.31805375e-01 -1.17708647e+00 6.30866230e-01 2.62838781e-01
-5.45576215e-01 -6.99904323e-01 -7.29176879e-01 3.48829180e-01
-8.56455803e-01 2.12197065e-01 1.03053488e-01 -4.36681569e-01
-3.02418470e-02 -9.91917670e-01 8.22929919e-01 -1.31506890e-01
-2.70599097e-01 -1.30252481e-01 -7.39692390e-01 9.03499961e-01
1.02838969e+00 -3.03571075e-01 -9.49939668e-01 -6.84723020e-01
-7.42435530e-02 7.19970703e-01 4.85559851e-01 2.92887628e-01
4.72557962e-01 -1.37275028e+00 -7.90123940e-01 -1.18870683e-01
-2.82334864e-01 -4.60127562e-01 -7.41332397e-02 4.15218681e-01
-8.97849083e-01 4.09965664e-01 -6.25529826e-01 3.72613757e-03
-1.52427208e+00 -9.43562388e-02 4.96937066e-01 4.94192308e-03
1.09300852e-01 1.27594817e+00 -1.00192748e-01 -7.39475667e-01
3.86177599e-01 -6.59402534e-02 -6.22346282e-01 3.95315140e-01
7.17369378e-01 1.04662180e+00 -5.17169535e-01 -2.36274645e-01
6.74218163e-02 4.14142847e-01 3.59988324e-02 2.29795538e-02
1.31252861e+00 -1.30526647e-01 -5.21779656e-01 4.66223866e-01
8.80774438e-01 5.83308816e-01 -5.63109517e-01 2.49206439e-01
2.06744701e-01 -9.58575606e-01 -1.34784384e-02 -9.39931393e-01
-7.25712359e-01 9.65350449e-01 4.42732543e-01 7.56559297e-02
7.16120780e-01 -5.20727336e-01 2.20058635e-01 1.28809651e-02
-1.70877695e-01 -1.75924468e+00 8.73167515e-02 2.91988850e-01
2.90673465e-01 -1.31930530e+00 6.02510154e-01 -8.02127123e-02
-5.09285510e-01 1.25851595e+00 9.66643870e-01 2.57143587e-01
7.59854615e-01 -1.08104743e-01 3.88932884e-01 -1.97688848e-01
-5.49032629e-01 2.60221153e-01 2.79969126e-02 6.77193165e-01
1.67452371e+00 1.87839717e-01 -9.63025212e-01 4.81277704e-01
-1.09861898e+00 1.73810527e-01 1.16881740e+00 9.91293311e-01
-8.74335349e-01 -1.24119496e+00 -4.55962956e-01 5.83196104e-01
-1.28427279e+00 -5.06128311e-01 -6.76879704e-01 5.30220985e-01
6.02036715e-02 9.83841598e-01 8.30101445e-02 6.00732863e-02
-8.76601711e-02 3.17887723e-01 9.72357929e-01 -7.11136222e-01
-1.44957900e+00 -7.20937431e-01 2.24625185e-01 -3.30637306e-01
-7.62075067e-01 -8.44796360e-01 -8.39862943e-01 -1.31939387e+00
-3.74288023e-01 3.62983018e-01 5.65471351e-01 7.89130211e-01
-2.46904984e-01 3.24534148e-01 2.04020292e-01 2.06824169e-01
-4.50100690e-01 -5.19675672e-01 -6.85486138e-01 -2.36036301e-01
3.01565200e-01 -1.65212139e-01 -2.03164667e-01 -2.61575699e-01] | [11.235286712646484, 9.119438171386719] |
778dff11-791f-4bfc-b911-211efff2b727 | an-efficient-recurrent-adversarial-framework | 2012.13033 | null | https://arxiv.org/abs/2012.13033v2 | https://arxiv.org/pdf/2012.13033v2.pdf | An Efficient Recurrent Adversarial Framework for Unsupervised Real-Time Video Enhancement | Video enhancement is a challenging problem, more than that of stills, mainly due to high computational cost, larger data volumes and the difficulty of achieving consistency in the spatio-temporal domain. In practice, these challenges are often coupled with the lack of example pairs, which inhibits the application of supervised learning strategies. To address these challenges, we propose an efficient adversarial video enhancement framework that learns directly from unpaired video examples. In particular, our framework introduces new recurrent cells that consist of interleaved local and global modules for implicit integration of spatial and temporal information. The proposed design allows our recurrent cells to efficiently propagate spatio-temporal information across frames and reduces the need for high complexity networks. Our setting enables learning from unpaired videos in a cyclic adversarial manner, where the proposed recurrent units are employed in all architectures. Efficient training is accomplished by introducing one single discriminator that learns the joint distribution of source and target domain simultaneously. The enhancement results demonstrate clear superiority of the proposed video enhancer over the state-of-the-art methods, in all terms of visual quality, quantitative metrics, and inference speed. Notably, our video enhancer is capable of enhancing over 35 frames per second of FullHD video (1080x1920). | ['Radu Timofte', 'Luc van Gool', 'Danda Pani Paudel', 'Zhiwu Huang', 'Dario Fuoli'] | 2020-12-24 | null | null | null | null | ['video-enhancement'] | ['computer-vision'] | [ 2.96348214e-01 -3.74110848e-01 -3.60026583e-02 -2.63581667e-02
-8.27216923e-01 -5.05683601e-01 4.81033862e-01 -1.92202032e-01
-5.91802955e-01 8.19324613e-01 2.20583186e-01 -9.24707353e-02
-1.74753219e-02 -7.25024581e-01 -9.10419226e-01 -8.57451916e-01
-1.59313977e-01 -4.49679375e-01 2.87491143e-01 -2.77624547e-01
-1.18720889e-01 2.69559711e-01 -1.34976232e+00 1.28244340e-01
9.10143852e-01 1.25368428e+00 3.77777755e-01 6.84342802e-01
2.70752639e-01 1.19221723e+00 -4.12498176e-01 -4.23320234e-01
3.25347483e-01 -4.69115108e-01 -2.11423799e-01 3.45520787e-02
5.15213013e-01 -8.76815915e-01 -9.99476731e-01 9.80701685e-01
7.33189404e-01 2.56976843e-01 4.52043235e-01 -1.12200511e+00
-6.75883651e-01 1.44596383e-01 -7.05155313e-01 3.99739832e-01
1.23949395e-02 4.17793334e-01 8.37851048e-01 -7.83061028e-01
5.24351716e-01 8.23461771e-01 6.75167799e-01 4.77517992e-01
-9.73866522e-01 -7.50233233e-01 2.07208171e-01 3.75014633e-01
-1.32005990e+00 -5.43269694e-01 9.30779517e-01 -3.16624105e-01
5.26354373e-01 -1.95862129e-02 4.78806376e-01 1.10982728e+00
9.53117087e-02 7.58950770e-01 8.69538903e-01 -9.37422067e-02
1.53996199e-01 -9.59196016e-02 -4.46458757e-01 7.27833807e-01
-1.42227560e-01 4.14489269e-01 -4.84642476e-01 2.76534051e-01
1.21856713e+00 2.77249575e-01 -4.54114228e-01 -2.09835321e-01
-1.12889504e+00 6.39095247e-01 5.74804783e-01 2.59770334e-01
-3.62078130e-01 3.49009812e-01 5.60150802e-01 2.76455402e-01
3.73590380e-01 1.28294066e-01 -2.00351864e-01 -6.31856397e-02
-1.11976206e+00 1.89912781e-01 1.26313642e-01 1.05063868e+00
5.02772808e-01 3.98175001e-01 -3.26823562e-01 6.28585219e-01
3.19493786e-02 4.86568123e-01 2.97362357e-01 -1.17844820e+00
7.05539167e-01 1.73363328e-01 1.27478451e-01 -1.27039194e+00
1.44224730e-03 -6.94851756e-01 -1.29161310e+00 2.66069740e-01
3.56728226e-01 -3.28123331e-01 -6.93922460e-01 2.04351330e+00
2.19112828e-01 5.60408294e-01 -1.36165600e-03 9.77655411e-01
5.90211511e-01 7.49117196e-01 1.34478047e-01 -3.63659024e-01
1.12456453e+00 -1.10924065e+00 -8.22353482e-01 -5.49517991e-03
3.01001936e-01 -5.67565024e-01 8.67589414e-01 6.70121387e-02
-1.31827283e+00 -9.14061427e-01 -1.09468412e+00 -1.92712262e-01
-1.33485183e-01 1.87689975e-01 4.20120031e-01 1.67828903e-01
-1.05091393e+00 6.21411800e-01 -7.73402691e-01 1.83042511e-01
5.93256652e-01 3.68176103e-01 -3.44510376e-01 -1.17933229e-01
-1.29400814e+00 3.83265972e-01 3.03254485e-01 1.69705257e-01
-1.09298909e+00 -9.29440916e-01 -9.83731925e-01 1.32165015e-01
3.15554738e-01 -6.87442482e-01 8.93894851e-01 -1.00677538e+00
-1.38415813e+00 3.24711084e-01 4.57419045e-02 -4.31911439e-01
7.79937208e-01 -2.63478547e-01 -4.86824036e-01 4.78841692e-01
-8.13738108e-02 6.55337811e-01 1.11686516e+00 -9.90878880e-01
-7.02967167e-01 -9.83474106e-02 2.45571136e-01 1.81446165e-01
-5.82820415e-01 -1.70001388e-01 -7.05788851e-01 -1.21771932e+00
-2.71017373e-01 -5.82199216e-01 -2.19649717e-01 4.32953238e-01
5.80677874e-02 3.02945882e-01 1.06617665e+00 -9.44463611e-01
1.26482582e+00 -2.43529844e+00 1.41531244e-01 -2.32351393e-01
3.98883790e-01 4.13222075e-01 -3.25229973e-01 1.07271403e-01
5.30542322e-02 -1.39522672e-01 -3.17168176e-01 -2.72658199e-01
-1.30977318e-01 5.02098724e-02 -3.72143805e-01 3.77396613e-01
4.00285602e-01 9.94087875e-01 -1.09630978e+00 -5.91931462e-01
3.83199394e-01 8.79271448e-01 -6.78314686e-01 4.64914501e-01
-1.06969297e-01 8.09931457e-01 -4.12554801e-01 3.95859957e-01
8.25785697e-01 -3.25437903e-01 2.94242296e-02 -4.97468829e-01
-9.01476666e-02 1.48878153e-02 -1.06069624e+00 1.89560509e+00
-5.02557993e-01 6.80237055e-01 1.28629848e-01 -1.06379831e+00
5.45303106e-01 5.03373742e-01 5.10049224e-01 -9.78165090e-01
6.14453666e-03 1.10147245e-01 -2.35422611e-01 -5.23701787e-01
3.71568084e-01 -1.53111562e-01 5.76109700e-02 3.21890503e-01
2.94926107e-01 2.19139636e-01 2.13109985e-01 1.68421268e-01
1.11185312e+00 1.75907299e-01 1.07493296e-01 2.07382485e-01
5.29704034e-01 -6.48191929e-01 7.37809479e-01 6.89159870e-01
-3.23606223e-01 6.31903768e-01 3.32145870e-01 -2.17868254e-01
-1.29942906e+00 -1.20896208e+00 1.13935344e-01 9.58696067e-01
3.19123209e-01 -1.53904781e-01 -4.54900295e-01 -6.65571630e-01
-3.13828140e-01 6.57366589e-02 -5.06955862e-01 -8.82593691e-02
-7.93041050e-01 -3.71233791e-01 5.89079618e-01 7.93242931e-01
8.91663611e-01 -8.59585404e-01 -4.66431856e-01 3.29585433e-01
-2.76367813e-01 -1.38525665e+00 -8.58184993e-01 -1.57873765e-01
-1.02521563e+00 -8.24366987e-01 -1.07644284e+00 -8.77252460e-01
5.93792319e-01 3.67404044e-01 9.36343551e-01 2.81917155e-02
-2.13041350e-01 1.69054389e-01 -2.29732245e-01 2.72895128e-01
-2.81217992e-01 -7.00244755e-02 -1.67227149e-01 9.16092098e-02
-2.32733786e-01 -9.99338627e-01 -1.11797595e+00 2.77374774e-01
-1.23747218e+00 1.37950748e-01 7.33290017e-01 1.13694537e+00
5.23383915e-01 2.55494803e-01 6.47580683e-01 -5.07734001e-01
2.14722425e-01 -4.24894750e-01 -6.06695294e-01 9.90949050e-02
-1.61881804e-01 -1.50845258e-03 9.32162225e-01 -4.33996022e-01
-1.10559916e+00 -5.08341417e-02 -2.24003211e-01 -6.30587935e-01
6.07100921e-03 3.09255242e-01 -1.48121759e-01 -1.56130001e-01
2.95591563e-01 4.60102409e-01 1.30157635e-01 -2.72993207e-01
3.43169659e-01 3.79882097e-01 8.80144417e-01 -4.71999943e-01
9.02769268e-01 5.17646611e-01 -1.11854866e-01 -6.76771104e-01
-5.78437865e-01 -2.30524167e-01 -4.02045876e-01 -2.95887083e-01
7.91033447e-01 -1.28411400e+00 -6.53935194e-01 6.96807504e-01
-1.07948697e+00 -3.52183580e-01 -2.51917660e-01 4.91546452e-01
-5.67022800e-01 5.20942152e-01 -1.05674696e+00 -4.55724210e-01
-3.72843534e-01 -1.14235187e+00 9.30538237e-01 2.33576462e-01
2.78439432e-01 -1.03939223e+00 -2.35764638e-01 2.36267716e-01
6.62720978e-01 3.82989466e-01 7.41523266e-01 1.00169070e-01
-8.91303539e-01 -4.39530564e-03 -4.56920773e-01 7.13737071e-01
2.25406423e-01 -2.75958031e-01 -8.47736001e-01 -5.34909129e-01
6.05199635e-02 -4.05249119e-01 8.40725183e-01 3.91381174e-01
1.37249207e+00 -4.25771594e-01 4.44786623e-03 8.33781779e-01
1.62747800e+00 1.41771138e-01 7.64660954e-01 2.55857091e-02
7.34057903e-01 3.12382609e-01 5.23428679e-01 5.69038928e-01
1.07973710e-01 7.52405345e-01 4.44052160e-01 -3.80355209e-01
-1.74736679e-01 -2.21644521e-01 4.82382268e-01 8.86646330e-01
-2.20105186e-01 -2.20414221e-01 -3.87558162e-01 6.09521210e-01
-1.81008494e+00 -1.27001059e+00 4.52122062e-01 2.09026814e+00
9.42419052e-01 5.98418154e-02 8.29277039e-02 2.10642323e-01
7.22778857e-01 4.30201620e-01 -5.90233147e-01 2.60351360e-01
-1.66437477e-01 1.81336865e-01 2.89245963e-01 4.24076110e-01
-1.23009670e+00 4.80647087e-01 5.96142673e+00 9.51777220e-01
-1.24573648e+00 2.55007774e-01 8.17608118e-01 -2.37925276e-01
-1.67176560e-01 -4.18497533e-01 -2.94960260e-01 8.27541053e-01
7.03736484e-01 5.27610257e-03 4.04553920e-01 4.79679823e-01
3.16093743e-01 3.02619725e-01 -9.55691457e-01 1.06359613e+00
3.59831639e-02 -1.53478062e+00 6.40366003e-02 -9.93714407e-02
8.50523591e-01 -1.08407296e-01 4.09983426e-01 2.47351095e-01
-6.76046684e-02 -8.73864770e-01 7.58759379e-01 3.64940733e-01
1.04356861e+00 -8.47004712e-01 6.32526875e-01 1.04881555e-01
-1.35899866e+00 -1.87785223e-01 -1.73534930e-01 9.52058583e-02
2.65460104e-01 4.52674240e-01 -1.44834623e-01 5.49958408e-01
7.11028695e-01 9.39479589e-01 -2.22038120e-01 8.17645192e-01
-1.17050067e-01 3.77662331e-01 -1.42814085e-01 5.15950978e-01
2.89755255e-01 -1.38909906e-01 4.32352453e-01 1.26763117e+00
4.99041229e-01 1.34528369e-01 2.91498266e-02 7.14055538e-01
-3.52018476e-01 -2.62910038e-01 -5.95504522e-01 8.99033397e-02
3.58638823e-01 1.17436421e+00 -2.45968178e-01 -2.57437080e-01
-6.13443911e-01 1.11802018e+00 1.92923725e-01 5.73702276e-01
-1.20599711e+00 -3.35010409e-01 6.15953326e-01 1.09481588e-01
5.91109097e-01 -2.18200505e-01 9.95628387e-02 -1.17618501e+00
4.36309516e-01 -1.00469708e+00 2.31120437e-01 -6.20785177e-01
-1.21914804e+00 6.84708059e-01 -3.98516268e-01 -1.60340798e+00
-2.78482258e-01 -4.01284933e-01 -5.60089350e-01 6.60928547e-01
-1.78120089e+00 -1.03234875e+00 -6.66327059e-01 9.57837760e-01
4.10276085e-01 -1.26526758e-01 4.28700149e-01 9.76084352e-01
-4.58634317e-01 1.00763631e+00 1.96845666e-01 3.33294690e-01
7.99998641e-01 -9.56595123e-01 1.48358718e-01 1.11507785e+00
-1.65038720e-01 3.23295176e-01 5.06456137e-01 -2.48726264e-01
-1.27987587e+00 -1.42046249e+00 3.20299596e-01 5.27740717e-02
6.53824508e-01 -2.56878734e-01 -9.06074107e-01 3.42891008e-01
2.63722956e-01 5.96793830e-01 4.74251300e-01 -4.90458280e-01
-3.92498583e-01 -4.04287457e-01 -9.81390715e-01 6.73765838e-01
1.04551291e+00 -7.72312999e-01 -5.68695478e-02 4.91899252e-02
9.39869583e-01 -4.55623716e-01 -9.95563567e-01 4.85133588e-01
4.77958530e-01 -9.63972807e-01 1.03952527e+00 -2.69917130e-01
9.53800857e-01 -6.16293192e-01 -1.57947198e-01 -9.90517914e-01
-2.39917159e-01 -7.40792513e-01 -4.79162723e-01 1.29382181e+00
1.26944557e-01 -3.69148850e-01 7.12202966e-01 2.20687002e-01
-1.53431281e-01 -8.40900719e-01 -8.98585856e-01 -7.36584246e-01
-1.76655605e-01 -2.77161270e-01 3.74358952e-01 8.83570373e-01
-2.97014803e-01 1.54141158e-01 -8.01694334e-01 2.65772492e-01
7.31296241e-01 2.97996076e-03 5.09522974e-01 -4.20058221e-01
-8.23616683e-01 -1.81243986e-01 -6.90138042e-01 -1.47743309e+00
-4.40704376e-02 -4.74909276e-01 -1.31144049e-02 -1.11633778e+00
2.03822300e-01 -5.03091514e-01 -5.79157710e-01 2.41615415e-01
-3.51433009e-01 5.72216153e-01 2.57055968e-01 6.56357259e-02
-7.09571481e-01 8.79111409e-01 1.48728848e+00 -2.80696929e-01
3.07798386e-02 -4.01692539e-01 -5.44672728e-01 4.89292741e-01
5.57002842e-01 -1.96548790e-01 -6.39264524e-01 -7.49511361e-01
-5.53077757e-02 4.05049205e-01 6.28580213e-01 -1.31090224e+00
3.35470676e-01 5.54897673e-02 5.09951711e-01 -4.26477015e-01
4.34553623e-01 -8.05865526e-01 1.00941367e-01 3.07296902e-01
-2.92537898e-01 7.27835074e-02 2.53069192e-01 8.09871793e-01
-6.77945793e-01 1.65821150e-01 9.48896229e-01 1.37341768e-01
-7.76630044e-01 6.98874295e-01 -4.26797867e-02 1.50241882e-01
1.05708528e+00 -1.20777771e-01 -2.45855585e-01 -5.38154960e-01
-6.29370451e-01 6.73978869e-03 4.49292392e-01 2.74507642e-01
7.50263751e-01 -1.52758360e+00 -7.37684906e-01 1.81783035e-01
-1.22758105e-01 3.09341773e-02 8.21936131e-01 9.06054080e-01
-4.46013361e-01 1.60298139e-01 -3.85951281e-01 -4.94664460e-01
-9.07098293e-01 6.64209425e-01 3.27613622e-01 -4.38153148e-01
-7.35823274e-01 6.71012878e-01 5.81251919e-01 1.07078321e-01
3.12437773e-01 -5.38202040e-02 2.17112526e-02 -1.97250545e-01
7.49642730e-01 3.18066180e-01 -2.22761631e-01 -3.91142607e-01
-4.69983742e-02 4.48491961e-01 -1.75826743e-01 -2.26040371e-02
1.36355245e+00 -2.18063742e-01 1.99417174e-01 -1.77351795e-02
1.40083909e+00 -7.69016221e-02 -1.94595551e+00 -3.83336782e-01
-6.14431083e-01 -5.53490877e-01 8.09282884e-02 -5.52053511e-01
-1.54364598e+00 8.68417025e-01 7.92003930e-01 -1.92682147e-01
1.60625494e+00 -3.19368541e-01 1.12969482e+00 -9.37964916e-02
1.85052291e-01 -9.53773439e-01 3.70040059e-01 3.04704726e-01
7.57414877e-01 -1.32931924e+00 -2.02108234e-01 -3.46451372e-01
-3.76969427e-01 8.51323426e-01 5.57596922e-01 -2.56256372e-01
4.69343930e-01 3.73510718e-01 -1.40227869e-01 1.74113512e-01
-7.28674114e-01 2.91592889e-02 2.32682541e-01 4.71551865e-01
4.42876726e-01 -3.04530680e-01 -7.25620016e-02 4.04798895e-01
4.31136727e-01 1.27069309e-01 2.09366873e-01 8.34235132e-01
3.66831012e-02 -9.57094610e-01 2.16887835e-02 2.67013729e-01
-6.10093176e-01 -2.29935929e-01 4.51282710e-01 7.30941534e-01
2.54635394e-01 7.46406198e-01 1.97235689e-01 -4.28775340e-01
1.39907166e-01 -5.52701890e-01 4.97205764e-01 -1.23016359e-02
-3.40613395e-01 1.18403442e-01 -2.79202670e-01 -6.23365939e-01
-6.16030991e-01 -4.12790626e-01 -9.75536644e-01 -4.74267066e-01
-9.14171711e-02 -1.47771451e-03 2.18520522e-01 8.42215478e-01
5.67519665e-01 7.87613809e-01 9.59891260e-01 -8.71889830e-01
-4.66667295e-01 -6.60362959e-01 -4.81652826e-01 6.03035927e-01
6.27834678e-01 -5.81563175e-01 -2.91259795e-01 3.24305147e-01] | [11.016698837280273, -1.6350346803665161] |
d6f09f35-a6a6-485e-99b1-3791109de944 | temporal-logic-guided-safe-reinforcement | 1903.09885 | null | http://arxiv.org/abs/1903.09885v1 | http://arxiv.org/pdf/1903.09885v1.pdf | Temporal Logic Guided Safe Reinforcement Learning Using Control Barrier Functions | Using reinforcement learning to learn control policies is a challenge when
the task is complex with potentially long horizons. Ensuring adequate but safe
exploration is also crucial for controlling physical systems. In this paper, we
use temporal logic to facilitate specification and learning of complex tasks.
We combine temporal logic with control Lyapunov functions to improve
exploration. We incorporate control barrier functions to safeguard the
exploration and deployment process. We develop a flexible and learnable system
that allows users to specify task objectives and constraints in different forms
and at various levels. The framework is also able to take advantage of known
system dynamics and handle unknown environmental dynamics by integrating
model-free learning with model-based planning. | ['Calin Belta', 'Xiao Li'] | 2019-03-23 | null | null | null | null | ['safe-exploration'] | ['robots'] | [-5.77252358e-02 2.80447572e-01 -3.37212235e-01 6.84409514e-02
-5.02075195e-01 -9.67232764e-01 5.65266013e-01 6.39742389e-02
-4.56653386e-01 1.23917198e+00 -2.95054436e-01 -6.44628346e-01
-6.88257992e-01 -6.80635691e-01 -6.55549586e-01 -5.18637121e-01
-7.25459158e-01 2.85419583e-01 6.08913302e-01 -4.38672215e-01
1.53242350e-01 7.56280839e-01 -1.46829426e+00 -7.97502622e-02
7.05915332e-01 7.57748127e-01 4.16237175e-01 9.77400780e-01
3.40733558e-01 7.05714047e-01 -3.65494788e-01 7.36394405e-01
4.80196983e-01 -1.07147114e-03 -7.07664967e-01 3.64656970e-02
-4.00807858e-01 -2.17905730e-01 -1.73875783e-03 7.44710088e-01
1.60232633e-01 4.83787000e-01 3.01370263e-01 -1.51078868e+00
2.39587054e-01 5.76809525e-01 -3.04847378e-02 -3.58488411e-02
1.09480627e-01 7.61931479e-01 4.74862039e-01 -3.47209070e-03
3.74309629e-01 1.22754502e+00 2.06986025e-01 6.83974624e-01
-1.27601206e+00 -4.91831690e-01 7.40664780e-01 -1.25296367e-02
-1.38810360e+00 -2.70883262e-01 3.84153426e-01 -6.00678086e-01
1.21130943e+00 2.38789141e-01 8.53449404e-01 8.00970018e-01
7.41671145e-01 4.41013962e-01 1.29292047e+00 -5.27737141e-01
6.88688099e-01 -9.31960344e-02 -3.10802042e-01 6.37697518e-01
1.21406093e-01 7.11146832e-01 -1.16743259e-01 -7.43409470e-02
8.29685450e-01 -1.60493717e-01 -1.65720940e-01 -5.96789360e-01
-1.15157890e+00 6.19720221e-01 1.39845207e-01 -2.89776642e-02
-3.41412783e-01 6.88127697e-01 3.47755432e-01 5.51380575e-01
-3.58495384e-01 1.05560732e+00 -7.72273123e-01 -3.14666122e-01
-4.05069113e-01 6.87470257e-01 1.01605308e+00 9.63296711e-01
5.10784090e-01 4.79349017e-01 -2.70017982e-01 9.75936651e-02
1.85899109e-01 4.50756669e-01 -1.35022283e-01 -1.51455915e+00
2.66050309e-01 3.55647117e-01 8.63811255e-01 -2.49148920e-01
-4.77555573e-01 -2.85796613e-01 -1.55780613e-01 1.26223040e+00
2.06312791e-01 -4.86912817e-01 -1.04730499e+00 1.67170405e+00
4.38881338e-01 -4.16849144e-02 1.96165636e-01 6.97968364e-01
-6.49053991e-01 6.25281692e-01 5.88450767e-02 -5.33004045e-01
9.52324748e-01 -5.02049446e-01 -8.92925262e-01 -3.71221066e-01
3.86020929e-01 -1.79123655e-02 1.08162034e+00 9.08019662e-01
-1.18620646e+00 -8.98488313e-02 -1.40850854e+00 5.72403789e-01
-3.89868081e-01 -2.94156045e-01 4.85318959e-01 3.83933485e-01
-1.08284187e+00 7.73072004e-01 -1.58726585e+00 -6.98175398e-04
1.10121317e-01 7.86675096e-01 5.74005507e-02 4.24604088e-01
-1.11561799e+00 1.43965971e+00 7.58983016e-01 1.83928423e-02
-1.44300818e+00 -3.98801982e-01 -1.04520273e+00 -2.81089209e-02
1.22846818e+00 -3.46243590e-01 1.75328565e+00 -2.68147200e-01
-1.94893479e+00 -2.62627453e-01 5.20759523e-01 -7.03908980e-01
6.26713097e-01 -2.74714082e-01 -1.56546220e-01 -1.76990420e-01
-2.71280408e-01 4.07354832e-01 8.72746348e-01 -1.13074434e+00
-6.50969386e-01 6.71095401e-02 5.92325985e-01 3.15370053e-01
-8.62602815e-02 -3.92343849e-01 -3.31344567e-02 -6.02733009e-02
-4.06292439e-01 -1.31929028e+00 -8.60439777e-01 -3.35346721e-02
-7.84575269e-02 1.28913090e-01 7.48483479e-01 -2.56109327e-01
1.27642703e+00 -1.68781912e+00 4.98512208e-01 4.21190381e-01
-2.15854853e-01 -1.06525635e-02 -1.62304416e-02 5.78605592e-01
3.78527820e-01 5.36978394e-02 -2.83591479e-01 1.36957422e-01
2.53803253e-01 6.42331064e-01 -4.69624698e-01 2.08994776e-01
3.13820750e-01 6.61681950e-01 -8.76529098e-01 -3.96908790e-01
3.89792830e-01 7.49258772e-02 -5.33950567e-01 2.05529198e-01
-1.04324508e+00 6.24118805e-01 -9.32837129e-01 4.14007097e-01
2.22817603e-02 2.93029964e-01 7.04421759e-01 6.36589408e-01
-7.59627879e-01 1.64580584e-01 -1.54678440e+00 1.39397728e+00
-9.16117013e-01 1.69004619e-01 6.59415841e-01 -5.70537150e-01
5.91114044e-01 2.24517301e-01 2.88796514e-01 -4.94548053e-01
1.59839228e-01 1.98129639e-02 6.60054460e-02 -4.12447393e-01
4.03462023e-01 -8.83363709e-02 -3.97573560e-01 4.38654155e-01
-3.45118165e-01 -8.85179043e-01 2.05153003e-01 -3.42680275e-01
1.17056596e+00 5.63285470e-01 5.43637335e-01 -5.64025640e-01
4.03164983e-01 3.06971610e-01 7.94264257e-01 6.18374825e-01
5.93834594e-02 -3.52038413e-01 8.38466704e-01 -2.65957922e-01
-7.96421826e-01 -9.31842387e-01 1.51922137e-01 9.27678585e-01
7.58657753e-02 -3.33472520e-01 -4.60420966e-01 -4.63747591e-01
1.01136774e-01 9.90583479e-01 -5.69078445e-01 -2.42575154e-01
-7.12526381e-01 -1.14272699e-01 -1.09582990e-01 6.09268248e-01
1.08564226e-02 -9.85964119e-01 -1.42775536e+00 4.51474905e-01
4.82411712e-01 -6.97846770e-01 -4.98833001e-01 6.82718515e-01
-6.63449585e-01 -1.05843711e+00 1.33083453e-02 -3.39207381e-01
3.93934071e-01 -3.70822310e-01 6.00617170e-01 -8.04123208e-02
-3.51027161e-01 5.94934642e-01 4.76741232e-02 -6.34322524e-01
-6.23460293e-01 2.08095014e-02 2.48813912e-01 -6.68328106e-01
-6.93282068e-01 -3.70409369e-01 -2.34359533e-01 3.85880321e-01
-1.03674293e+00 1.38264121e-02 1.15846284e-01 8.36387634e-01
7.11491585e-01 5.88785231e-01 3.68436724e-01 -3.09420377e-01
9.42355096e-01 -1.29849404e-01 -1.62288761e+00 4.93694812e-01
-9.16065574e-01 5.97614050e-01 6.59803569e-01 -7.71633446e-01
-8.80221486e-01 3.47595125e-01 3.34058195e-01 -3.08206320e-01
7.25186393e-02 5.67035615e-01 -2.66447991e-01 -1.67687133e-01
5.32945573e-01 -1.03611037e-01 2.89221227e-01 -9.61687192e-02
2.82654732e-01 1.57543793e-01 8.39102790e-02 -1.04218614e+00
7.05147505e-01 1.66529387e-01 3.38643074e-01 -3.68469238e-01
-4.49642241e-01 2.31703937e-01 -6.21470928e-01 -3.79319876e-01
6.65858090e-01 -5.41625023e-01 -1.24387205e+00 4.60666381e-02
-7.74058044e-01 -1.37675297e+00 -4.99450743e-01 2.41679907e-01
-9.93465304e-01 -2.59947628e-01 -2.42182374e-01 -1.35619187e+00
1.19409919e-01 -1.26649284e+00 6.94657326e-01 2.99025327e-01
-2.72744894e-01 -9.44658518e-01 1.83606386e-01 -6.23857200e-01
5.72929442e-01 6.65099502e-01 6.75347388e-01 -1.16465613e-01
-9.19651091e-01 1.08471006e-01 5.16331851e-01 5.83064184e-03
1.25107691e-01 5.24279848e-02 -4.33519602e-01 -6.70904100e-01
-5.94081283e-02 -3.51442575e-01 2.92242348e-01 3.27975720e-01
9.68920708e-01 -7.87203968e-01 -4.40825433e-01 1.14907876e-01
1.44745815e+00 6.22991741e-01 3.18510830e-01 6.96329534e-01
1.87003210e-01 7.75145769e-01 1.16926122e+00 7.39502549e-01
3.78959328e-01 8.73653531e-01 8.05746734e-01 5.60365975e-01
5.35612524e-01 -1.17304832e-01 6.47907794e-01 -8.00164342e-02
2.14358419e-01 -3.45353596e-02 -1.20828807e+00 5.36850393e-01
-2.13277006e+00 -7.91410863e-01 4.60547715e-01 2.47090030e+00
1.09995925e+00 6.10377371e-01 2.79879063e-01 -1.45246416e-01
2.88293272e-01 -2.38696411e-01 -8.51717889e-01 -5.92173934e-01
6.23386264e-01 1.54572830e-01 7.96356440e-01 1.12157202e+00
-9.81918097e-01 8.97750020e-01 7.19060707e+00 4.33089077e-01
-1.10026574e+00 -2.18323275e-01 -3.70971188e-02 -3.01158577e-01
-1.25769973e-01 1.66421980e-01 -6.61381125e-01 1.34129614e-01
1.12921512e+00 -4.91703719e-01 9.12889063e-01 8.59123409e-01
5.46986938e-01 -4.08690929e-01 -1.14904821e+00 2.48787269e-01
-7.85239637e-01 -1.11373985e+00 -6.92804396e-01 1.89408548e-02
5.38820684e-01 -2.21035466e-01 2.96780728e-02 4.95976061e-01
9.11997557e-01 -1.04761469e+00 9.23772991e-01 6.73595309e-01
6.70280874e-01 -1.03569555e+00 1.96218699e-01 7.58325398e-01
-1.16407108e+00 -5.95826983e-01 1.84783995e-01 -4.94809538e-01
4.94056821e-01 9.07325000e-03 -9.71587539e-01 2.12757751e-01
3.38753074e-01 3.08756858e-01 -1.19714305e-01 1.00583303e+00
-4.31195974e-01 1.57764792e-01 -5.63109219e-01 -3.60504955e-01
3.82291853e-01 -1.08705282e-01 6.01952553e-01 7.33924747e-01
3.30014735e-01 -3.13205570e-02 8.70426774e-01 9.91143227e-01
1.00721204e+00 -6.66858196e-01 -8.61486495e-01 -1.31559715e-01
4.73154515e-01 9.36795056e-01 -6.23834789e-01 -4.30683270e-02
2.54942685e-01 3.98472458e-01 1.59598008e-01 3.91083509e-01
-9.39945817e-01 -3.61548334e-01 9.25981462e-01 1.10334232e-02
2.10152313e-01 -1.14153481e+00 -6.97682276e-02 -6.58825815e-01
-1.91406488e-01 -7.88889527e-01 3.88469577e-01 -4.89009917e-01
-6.81825578e-01 3.76128495e-01 6.79167330e-01 -1.08414888e+00
-8.51791143e-01 -6.98136508e-01 -5.09986520e-01 8.10802102e-01
-1.54092669e+00 -6.55783772e-01 1.88326016e-01 5.35538077e-01
4.78441298e-01 9.00712013e-02 6.68561101e-01 -4.60578233e-01
-5.26203632e-01 -2.15559617e-01 -3.14382017e-01 -5.83623052e-01
3.05414736e-01 -1.59042525e+00 1.07278191e-01 7.49829113e-01
-6.85789049e-01 6.30980134e-01 1.12355244e+00 -8.41307938e-01
-1.81369650e+00 -9.67778206e-01 -5.77473566e-02 -4.40866947e-01
9.31366801e-01 -3.95304650e-01 -9.15342927e-01 7.19039619e-01
2.41847098e-01 -1.81991890e-01 -4.46343571e-02 -1.46508828e-01
1.35705888e-01 -1.29606366e-01 -1.02837455e+00 8.04409146e-01
6.24437928e-01 -2.70015001e-01 -4.04564649e-01 2.08927765e-01
1.11902332e+00 -7.44905174e-01 -8.54564071e-01 4.75367874e-01
4.00674701e-01 -3.62361729e-01 6.86092615e-01 -6.26396775e-01
-4.60829407e-01 -7.06915915e-01 7.12203085e-02 -1.35383630e+00
-2.30292603e-01 -1.21163177e+00 -3.92050952e-01 7.49857843e-01
5.19227445e-01 -6.91079140e-01 6.15453482e-01 1.01816702e+00
-2.20159590e-01 -5.23921490e-01 -9.56578135e-01 -1.24662077e+00
1.02487333e-01 -5.87526023e-01 5.37789106e-01 5.49223125e-01
4.03789014e-01 -1.12979554e-01 -4.14301217e-01 5.57125568e-01
4.14509863e-01 1.73090713e-03 3.64641309e-01 -6.77896619e-01
-5.74317098e-01 -2.95857519e-01 3.35892767e-01 -3.95757169e-01
3.53477508e-01 -3.40860546e-01 5.94018042e-01 -1.42556345e+00
-5.58875442e-01 -6.15445197e-01 -2.49443308e-01 8.73932481e-01
3.35277021e-01 -7.71558940e-01 2.22068891e-01 -2.57177681e-01
-6.21196687e-01 6.99117899e-01 1.34637260e+00 9.18162987e-03
-7.63289332e-01 2.09615007e-01 -1.98985860e-01 4.73119885e-01
1.19994140e+00 -2.80089051e-01 -8.69450569e-01 -2.50679702e-01
3.98316264e-01 4.73817170e-01 2.17469454e-01 -1.05648363e+00
2.29897380e-01 -1.13641453e+00 -3.45000237e-01 -3.16102356e-01
3.86928380e-01 -1.11693263e+00 3.78445446e-01 1.06726909e+00
-5.97759485e-01 4.02299821e-01 7.54970789e-01 7.67334461e-01
2.66602039e-01 -2.74351649e-02 6.77208781e-01 -1.40161946e-01
-8.85385215e-01 2.11299375e-01 -7.89550602e-01 -2.31788546e-01
1.52269340e+00 1.15552559e-01 -1.61271036e-01 -2.66708732e-01
-9.53750491e-01 1.04246366e+00 5.65772057e-01 4.17425394e-01
5.36116302e-01 -9.12512004e-01 -1.51912808e-01 2.91904304e-02
8.11321214e-02 -8.10954049e-02 -8.33361819e-02 5.02058744e-01
-4.20222104e-01 3.54425132e-01 -5.35215855e-01 -2.80079693e-01
-8.85247231e-01 7.23287880e-01 6.52318060e-01 -3.98586333e-01
-3.89315575e-01 2.95001686e-01 -1.82556301e-01 -4.72627133e-01
3.00248623e-01 -6.59147203e-01 -2.24434435e-02 -2.19597504e-01
7.02055931e-01 2.30853423e-01 -2.18773246e-01 5.46817899e-01
-5.64224064e-01 2.06309229e-01 1.95404232e-01 -6.93913519e-01
1.24873865e+00 -9.19397101e-02 -6.70926198e-02 6.50373042e-01
1.65467024e-01 -3.54460567e-01 -2.10099316e+00 1.54032275e-01
4.60194498e-01 -2.46634886e-01 2.25344390e-01 -1.09202695e+00
-3.05515736e-01 6.24771178e-01 3.91005129e-01 2.26943195e-01
8.95532131e-01 -4.98544633e-01 1.04899637e-01 6.18657768e-01
8.92820239e-01 -1.43845320e+00 1.47205293e-01 8.40664148e-01
1.04846907e+00 -7.21803844e-01 8.54279771e-02 -1.25425264e-01
-7.84688652e-01 1.23408830e+00 9.69888151e-01 1.63432788e-02
4.48745221e-01 1.05832100e+00 -2.09534138e-01 8.48765150e-02
-1.36188388e+00 -3.01163137e-01 -2.00078469e-02 7.29030371e-01
-1.93891451e-01 5.25033511e-02 -1.16963558e-01 2.36108601e-01
3.94483805e-01 1.51337907e-01 7.65817583e-01 1.60687411e+00
-8.69729578e-01 -1.50925827e+00 -6.19616330e-01 6.95623159e-02
-2.35132128e-02 3.59232068e-01 -1.28993737e-02 1.14272654e+00
-8.87603313e-02 6.19791985e-01 -2.21693426e-01 -1.37742788e-01
3.07525158e-01 4.28136364e-02 6.99862719e-01 -1.06039166e+00
-4.14963752e-01 2.06187367e-01 3.86399955e-01 -9.28102136e-01
1.03261918e-01 -8.13484073e-01 -1.58168173e+00 1.99261218e-01
-1.73830003e-01 1.85876667e-01 7.57659793e-01 8.30468357e-01
3.57236624e-01 9.05307055e-01 5.63782334e-01 -8.47577810e-01
-1.16462660e+00 -3.51813287e-01 -6.20422065e-01 -7.07759738e-01
6.91874862e-01 -9.24436867e-01 -3.78712974e-02 -1.30093232e-01] | [4.590867042541504, 2.0818052291870117] |
0500f57f-e71e-4239-be55-377405a1b71f | on-training-targets-and-activation-functions | 2201.06426 | null | https://arxiv.org/abs/2201.06426v1 | https://arxiv.org/pdf/2201.06426v1.pdf | On Training Targets and Activation Functions for Deep Representation Learning in Text-Dependent Speaker Verification | Deep representation learning has gained significant momentum in advancing text-dependent speaker verification (TD-SV) systems. When designing deep neural networks (DNN) for extracting bottleneck features, key considerations include training targets, activation functions, and loss functions. In this paper, we systematically study the impact of these choices on the performance of TD-SV. For training targets, we consider speaker identity, time-contrastive learning (TCL) and auto-regressive prediction coding with the first being supervised and the last two being self-supervised. Furthermore, we study a range of loss functions when speaker identity is used as the training target. With regard to activation functions, we study the widely used sigmoid function, rectified linear unit (ReLU), and Gaussian error linear unit (GELU). We experimentally show that GELU is able to reduce the error rates of TD-SV significantly compared to sigmoid, irrespective of training target. Among the three training targets, TCL performs the best. Among the various loss functions, cross entropy, joint-softmax and focal loss functions outperform the others. Finally, score-level fusion of different systems is also able to reduce the error rates. Experiments are conducted on the RedDots 2016 challenge database for TD-SV using short utterances. For the speaker classifications, the well-known Gaussian mixture model-universal background model (GMM-UBM) and i-vector techniques are used. | ['Zheng-Hua Tan', 'Achintya kr. Sarkar'] | 2022-01-17 | null | null | null | null | ['text-dependent-speaker-verification'] | ['speech'] | [-1.09402591e-03 -2.36758307e-01 -3.77003849e-02 -6.31187201e-01
-9.04659152e-01 -2.46807516e-01 6.33569300e-01 -4.16358039e-02
-5.83786011e-01 4.91877079e-01 1.84329927e-01 -7.33532906e-01
-4.76916544e-02 -1.19179383e-01 -3.44056457e-01 -9.80883837e-01
-1.68070570e-02 1.00423016e-01 -1.71979651e-01 -1.80015370e-01
1.29848465e-01 4.15952981e-01 -1.46184719e+00 -1.25498474e-02
9.05600965e-01 1.31840301e+00 -2.67258845e-02 6.25456274e-01
-2.54914165e-02 7.72584796e-01 -9.25520778e-01 -4.56134260e-01
9.31318626e-02 -3.65798831e-01 -6.72156930e-01 -3.95055532e-01
2.57583618e-01 5.38321622e-02 -2.27143973e-01 8.52427006e-01
1.01039827e+00 4.17262226e-01 8.38640511e-01 -1.13874698e+00
-4.69204903e-01 7.96301603e-01 -2.74204195e-01 2.84637004e-01
1.61480054e-01 3.73255350e-02 7.93127835e-01 -1.11731422e+00
2.22221151e-01 1.46384835e+00 7.68546820e-01 8.19256306e-01
-1.10958672e+00 -7.53605902e-01 9.30920169e-02 5.03966033e-01
-1.61179972e+00 -9.48576808e-01 6.40993834e-01 -3.96313697e-01
1.10254645e+00 2.53965497e-01 -1.52742684e-01 1.30537415e+00
4.43190746e-02 9.14023936e-01 1.06611371e+00 -7.57542074e-01
2.86153287e-01 5.24251521e-01 4.20826793e-01 3.64690214e-01
-2.22378522e-01 3.98312896e-01 -5.98958373e-01 -9.62780565e-02
-4.12041470e-02 -4.94770110e-01 -3.94314885e-01 1.47701874e-01
-7.06535876e-01 8.76029015e-01 1.45281404e-01 5.31628668e-01
-2.38028899e-01 -8.07485357e-02 5.97731769e-01 5.37416935e-01
6.23737037e-01 2.94512846e-02 -5.49308479e-01 -1.56384811e-01
-1.12151933e+00 -8.26069787e-02 7.53563523e-01 4.99408305e-01
3.38494420e-01 6.79333806e-01 -6.02346838e-01 1.28331506e+00
5.29286802e-01 3.96712035e-01 8.93280327e-01 -3.92450899e-01
4.74668294e-01 2.16154814e-01 -2.69737303e-01 -3.69350791e-01
-3.24359775e-01 -6.88548744e-01 -8.66671562e-01 3.48806083e-01
4.30747092e-01 -3.95482749e-01 -9.78219450e-01 1.81269157e+00
1.79975167e-01 1.38837487e-01 3.60579282e-01 6.49469852e-01
1.02992296e+00 7.25815058e-01 2.32296765e-01 -2.43765593e-01
1.29636681e+00 -8.08231771e-01 -8.75052333e-01 -1.33741856e-01
6.23633742e-01 -7.50529528e-01 7.71713376e-01 3.16828668e-01
-9.23424184e-01 -7.06798851e-01 -1.20436478e+00 9.94969085e-02
-6.00364566e-01 3.77012104e-01 -3.33928689e-02 1.09747934e+00
-1.17659461e+00 5.77142596e-01 -6.64131343e-01 -1.09224930e-01
2.31864959e-01 5.30854464e-01 -1.96193635e-01 2.01474562e-01
-1.45137584e+00 1.31647408e+00 1.86447054e-01 1.64285913e-01
-7.94493556e-01 -6.61269307e-01 -8.23611081e-01 2.67533571e-01
-1.75691500e-01 -2.87786126e-01 1.26016319e+00 -8.84475827e-01
-2.12275410e+00 7.59631038e-01 -1.91180259e-01 -9.65770006e-01
4.98838305e-01 -1.32236823e-01 -7.49795675e-01 -3.75733614e-01
-4.73729819e-01 4.91431832e-01 1.07713223e+00 -9.84430611e-01
-4.34457153e-01 -2.74301529e-01 -5.07231295e-01 1.09540276e-01
-4.43553120e-01 4.42853600e-01 2.03310832e-01 -4.18371320e-01
-1.20389670e-01 -7.38561869e-01 8.95117819e-02 -6.02581024e-01
-4.80652690e-01 -6.69788778e-01 9.99245644e-01 -1.13077307e+00
1.51158822e+00 -2.53305507e+00 1.93630815e-01 1.61418334e-01
-1.81325421e-01 5.20878255e-01 3.29334289e-02 1.03130750e-01
-1.94164336e-01 1.85996795e-03 -2.04256192e-01 -8.92715871e-01
3.45841765e-01 -5.20073064e-03 -1.56912114e-02 5.41985750e-01
1.06671914e-01 5.40311217e-01 -2.15017214e-01 -3.57288599e-01
2.88824439e-01 9.74802256e-01 -3.07018787e-01 1.32581383e-01
1.52496651e-01 2.33179018e-01 4.19628136e-02 4.12555516e-01
7.74094820e-01 3.09649140e-01 -1.70533642e-01 5.59386145e-03
-2.04017475e-01 7.75079906e-01 -1.08522117e+00 1.26212311e+00
-7.98544645e-01 8.44854653e-01 2.53247976e-01 -8.32135856e-01
1.25393295e+00 6.93807006e-01 1.23806089e-01 -5.36661446e-01
3.38927001e-01 3.94451112e-01 2.15988949e-01 -1.59985527e-01
3.83696169e-01 -2.69485772e-01 2.30612278e-01 1.47283971e-01
3.92704964e-01 3.26786935e-01 -1.36205301e-01 -1.60526961e-01
7.71477461e-01 -2.37751484e-01 2.02837154e-01 -2.27209061e-01
9.16357458e-01 -7.34537780e-01 5.79913020e-01 4.40792918e-01
-4.94468331e-01 5.08360863e-01 3.71511161e-01 1.29122138e-01
-6.01930976e-01 -9.15210724e-01 -3.92373502e-01 1.25125921e+00
-3.43450606e-01 -7.09620565e-02 -6.84008837e-01 -6.76898420e-01
-3.77269997e-03 1.26534128e+00 -4.56326067e-01 -2.68123150e-01
-5.92677891e-01 -7.20261872e-01 8.77877712e-01 4.28839654e-01
2.38764659e-01 -1.02197540e+00 -1.42753989e-01 1.59925118e-01
-2.39191353e-02 -1.00463581e+00 -5.84483445e-01 6.06303990e-01
-5.84242582e-01 -4.40598190e-01 -9.43680644e-01 -7.76447535e-01
2.77256370e-01 -1.94633886e-01 7.54339099e-01 -3.65958154e-01
1.19987540e-01 2.12685138e-01 -2.75077820e-01 -4.64985639e-01
-7.84748316e-01 1.12955183e-01 2.81186938e-01 3.76723439e-01
5.36310196e-01 -3.08233827e-01 -3.25674176e-01 2.71926075e-01
-4.47167188e-01 -4.71156746e-01 4.12832737e-01 9.72718537e-01
1.25040784e-01 -1.00354657e-01 7.45921314e-01 -3.81623983e-01
7.97446370e-01 -4.12597060e-01 -5.34672439e-01 2.56056368e-01
-7.54729450e-01 1.82632551e-01 4.35399532e-01 -6.16699338e-01
-1.17141283e+00 -1.92978084e-01 -5.47473907e-01 -6.04351521e-01
-2.09172145e-01 5.32844543e-01 -2.58696347e-01 2.61873249e-02
8.02347600e-01 4.19794083e-01 9.21128541e-02 -6.22346640e-01
1.13562264e-01 1.09431028e+00 2.94095039e-01 -2.38947019e-01
3.90100747e-01 -9.56788436e-02 -4.80672896e-01 -1.04774749e+00
-3.54268491e-01 -4.50207919e-01 -3.64377618e-01 -1.45634592e-01
8.04337561e-01 -7.73104787e-01 -8.19760203e-01 6.05266571e-01
-1.13066530e+00 -2.41996318e-01 -2.29788035e-01 7.12253630e-01
-1.69637293e-01 2.00756267e-01 -6.51401937e-01 -1.35407269e+00
-7.63264894e-01 -1.38865554e+00 8.07108462e-01 2.64754951e-01
-8.28656405e-02 -9.84698415e-01 -8.69900361e-02 3.08268428e-01
9.27288473e-01 -6.62828162e-02 7.63027608e-01 -1.15035164e+00
1.40112147e-01 -1.47126302e-01 1.31919950e-01 1.00011849e+00
-2.52546761e-02 1.66542269e-02 -1.54741096e+00 -3.65132749e-01
2.43471652e-01 4.84593660e-02 1.00070143e+00 7.74998665e-01
9.51070368e-01 -3.56915593e-01 -1.16488799e-01 5.93672991e-01
9.86889541e-01 5.32020211e-01 5.16528845e-01 2.25909397e-01
3.74842942e-01 6.71511054e-01 1.18608885e-01 2.76632786e-01
3.79099250e-01 8.90216827e-01 1.11302346e-01 4.85814027e-02
-1.36261001e-01 9.92643535e-02 9.75295722e-01 9.97539103e-01
-1.75196230e-01 -3.77914459e-01 -8.71436417e-01 2.22395211e-01
-1.40467346e+00 -1.08852112e+00 8.53393506e-03 2.47802901e+00
5.77486753e-01 1.56849906e-01 2.78985560e-01 6.26549423e-01
8.66062462e-01 1.10692263e-01 -3.92072529e-01 -7.25972891e-01
-3.82114947e-01 2.45499879e-01 3.56796414e-01 7.05590367e-01
-1.18823540e+00 6.72286332e-01 5.17381048e+00 9.49992716e-01
-1.60552645e+00 4.84746456e-01 6.25599325e-01 6.31910376e-03
8.89177322e-02 -4.11712080e-01 -1.18708932e+00 6.38419449e-01
1.60520279e+00 -2.59711016e-02 3.15349489e-01 8.24218452e-01
1.62843868e-01 2.49623939e-01 -1.01041269e+00 1.01101732e+00
1.45475566e-01 -7.40014315e-01 -5.26269019e-01 3.15804593e-02
3.75669420e-01 3.01497042e-01 3.55705678e-01 7.59120643e-01
7.71180317e-02 -9.88382995e-01 1.00778532e+00 1.27557889e-01
6.98936582e-01 -6.91351891e-01 9.48177457e-01 7.84760416e-02
-9.92527425e-01 -3.14459532e-01 -1.60443678e-01 4.50306088e-01
1.02792352e-01 5.13836861e-01 -1.11867714e+00 5.04058957e-01
4.63699222e-01 3.19896817e-01 -2.71341681e-01 9.97914553e-01
-1.07574075e-01 1.07034552e+00 -4.06015068e-01 -1.72129467e-01
1.06735446e-01 6.46293461e-02 6.56720877e-01 1.55464578e+00
3.83421600e-01 -3.73954058e-01 -3.97304893e-01 6.86934173e-01
-2.44107414e-02 1.48927003e-01 -1.35706335e-01 2.20312238e-01
4.72694278e-01 9.24901128e-01 -6.52243048e-02 -2.30780974e-01
-2.00363845e-01 7.99053431e-01 8.05211961e-02 4.89379883e-01
-8.58660281e-01 -3.19258332e-01 8.78953218e-01 2.05440186e-02
4.18692082e-01 -9.12138596e-02 -3.13737065e-01 -9.03997421e-01
-9.33122784e-02 -7.93110788e-01 4.33517575e-01 -2.56467730e-01
-1.15974295e+00 9.59487677e-01 -9.27876830e-02 -1.16992927e+00
-3.76299500e-01 -6.19422793e-01 -7.74111569e-01 1.26311743e+00
-1.72441614e+00 -8.53421807e-01 1.06085695e-01 6.15968764e-01
6.07620776e-01 -5.25251567e-01 7.82778978e-01 7.52803802e-01
-8.94489169e-01 1.14908242e+00 3.01382959e-01 7.88066238e-02
5.93591094e-01 -1.18257987e+00 3.91576946e-01 6.52231753e-01
2.45327018e-02 4.17547643e-01 6.35884821e-01 -4.31730151e-02
-1.03698564e+00 -9.25145268e-01 1.12015200e+00 -2.98732594e-02
2.02825800e-01 -4.65357721e-01 -1.09684432e+00 4.29293931e-01
1.72815636e-01 -5.28390370e-02 7.32640147e-01 2.01586068e-01
-4.99598563e-01 -2.19783947e-01 -1.18030167e+00 1.82085618e-01
3.25502276e-01 -6.41554534e-01 -2.79143333e-01 8.40940922e-02
5.01641035e-01 -3.28246385e-01 -7.80659974e-01 4.51383054e-01
4.55156654e-01 -9.47526097e-01 1.04810715e+00 -3.87802601e-01
-6.50729463e-02 -1.37739122e-01 -2.41558179e-01 -1.60103178e+00
-1.71921283e-01 -5.19709349e-01 -2.51485884e-01 1.58196306e+00
7.96454012e-01 -9.32525516e-01 5.90538561e-01 4.40448910e-01
-3.60977083e-01 -8.22348475e-01 -1.44526875e+00 -1.00469267e+00
2.29549348e-01 -5.64950824e-01 5.46216726e-01 8.38922918e-01
-1.21101603e-01 4.75653023e-01 -4.48007256e-01 6.77584950e-03
3.79675567e-01 -4.50745702e-01 4.51993585e-01 -1.07525063e+00
-2.19650567e-01 -8.72593343e-01 -4.76901650e-01 -9.05206680e-01
2.99698263e-01 -9.56387222e-01 6.90779388e-02 -1.16442442e+00
-3.92409801e-01 -4.23116237e-01 -6.53042793e-01 4.00267005e-01
-2.04540089e-01 -2.25075930e-01 2.72533685e-01 -1.28694832e-01
-1.38870910e-01 7.66792655e-01 6.72626972e-01 -2.92865396e-01
-4.98069406e-01 5.32711744e-01 -2.88817644e-01 3.66982490e-01
9.82891738e-01 -3.96176070e-01 -2.42515411e-02 -1.37049973e-01
-5.30443013e-01 1.46474734e-01 1.50588518e-02 -1.02080071e+00
3.65024984e-01 4.16882783e-01 1.41348734e-01 -4.89342362e-01
7.97955453e-01 -4.89571363e-01 -1.81355119e-01 4.63585526e-01
-3.50081086e-01 -2.95199662e-01 2.67656803e-01 1.17301531e-01
-4.38991278e-01 -1.86950594e-01 1.16858077e+00 5.57808161e-01
-3.70986342e-01 -7.62775764e-02 -4.49118495e-01 -1.75178543e-01
6.23442531e-01 -2.46218309e-01 -9.92538258e-02 -4.52984482e-01
-7.12639391e-01 7.50984848e-02 -3.77544582e-01 5.81962883e-01
5.59492290e-01 -1.22396457e+00 -1.17419946e+00 3.94711703e-01
-7.30782449e-02 -5.19645333e-01 3.84524345e-01 1.14680111e+00
-3.67934778e-02 5.87591648e-01 1.89585045e-01 -6.32958651e-01
-1.55009675e+00 2.29717195e-01 6.46048248e-01 -2.82204211e-01
-7.97083080e-02 1.24935389e+00 5.32155447e-02 -6.69223607e-01
7.57985055e-01 -4.33028370e-01 -3.08692127e-01 9.14364532e-02
5.32475531e-01 6.26756012e-01 4.80410367e-01 -9.03541148e-01
-5.78229010e-01 1.65964112e-01 -1.10089324e-01 -9.23692361e-02
1.22808599e+00 -1.17826991e-01 3.68797034e-01 4.62458730e-01
1.36624610e+00 -2.20888719e-01 -9.10388410e-01 -4.64090884e-01
1.69706255e-01 1.50867641e-01 4.51112866e-01 -8.76410425e-01
-1.06366682e+00 1.16204965e+00 1.00446236e+00 3.27707469e-01
9.89671171e-01 -3.17976832e-01 5.73707640e-01 -1.03065409e-01
-1.73605792e-02 -9.62586641e-01 -5.16614199e-01 5.77188194e-01
9.62931395e-01 -1.36096716e+00 -3.13639969e-01 -2.54667755e-02
-6.84894800e-01 1.06981146e+00 4.88870621e-01 3.78344357e-01
9.91814494e-01 2.72056639e-01 1.85273990e-01 3.06874633e-01
-7.78155565e-01 -9.87340137e-02 5.79300404e-01 3.72067690e-01
6.23834908e-01 2.59776145e-01 -2.17765868e-01 6.91177785e-01
-3.60923320e-01 -3.60919803e-01 -4.29235213e-03 5.00170171e-01
-2.09301487e-01 -1.01102769e+00 -7.18085468e-01 3.67822230e-01
-6.30046844e-01 -2.22899377e-01 -2.25911617e-01 4.45681691e-01
6.27314067e-03 1.26781559e+00 -1.07257232e-01 -5.88779449e-01
3.72308612e-01 5.13455093e-01 2.14853808e-01 -2.48261392e-01
-1.13386476e+00 -3.05893645e-02 1.52306020e-01 4.53157946e-02
-5.41626550e-02 -7.21206367e-01 -9.62084055e-01 -3.10072571e-01
-8.66639435e-01 2.86767691e-01 1.38299322e+00 9.39070821e-01
2.46671259e-01 5.78848541e-01 7.46983826e-01 -7.01882720e-01
-9.62124527e-01 -1.59613001e+00 -4.19135898e-01 -3.48814055e-02
5.69904447e-01 -6.98478460e-01 -8.20121825e-01 -1.16210923e-01] | [14.327168464660645, 6.07444429397583] |
7cdddb45-468c-4644-9fca-6d5fc33ffdb8 | knowledge-enhanced-agents-for-interactive | 2305.05091 | null | https://arxiv.org/abs/2305.05091v1 | https://arxiv.org/pdf/2305.05091v1.pdf | Knowledge-enhanced Agents for Interactive Text Games | Communication via natural language is a crucial aspect of intelligence, and it requires computational models to learn and reason about world concepts, with varying levels of supervision. While there has been significant progress made on fully-supervised non-interactive tasks, such as question-answering and procedural text understanding, much of the community has turned to various sequential interactive tasks, as in semi-Markov text-based games, which have revealed limitations of existing approaches in terms of coherence, contextual awareness, and their ability to learn effectively from the environment. In this paper, we propose a framework for enabling improved functional grounding of agents in text-based games. Specifically, we consider two forms of domain knowledge that we inject into learning-based agents: memory of previous correct actions and affordances of relevant objects in the environment. Our framework supports three representative model classes: `pure' reinforcement learning (RL) agents, RL agents enhanced with knowledge graphs, and agents equipped with language models. Furthermore, we devise multiple injection strategies for the above domain knowledge types and agent architectures, including injection via knowledge graphs and augmentation of the existing input encoding strategies. We perform all experiments on the ScienceWorld text-based game environment, to illustrate the performance of various model configurations in challenging science-related instruction-following tasks. Our findings provide crucial insights on the development of effective natural language processing systems for interactive contexts. | ['Kaixin Ma', 'Jonathan Francis', 'Filip Ilievski', 'Jiarui Zhang', 'Prateek Chhikara'] | 2023-05-08 | null | null | null | null | ['instruction-following', 'text-based-games'] | ['natural-language-processing', 'playing-games'] | [ 2.38256752e-01 5.20519257e-01 8.58187210e-03 -1.60441652e-03
-2.01819256e-01 -8.80851150e-01 8.82207930e-01 5.71875870e-01
-6.35490358e-01 7.08131015e-01 1.65391058e-01 -6.23357058e-01
-3.54024947e-01 -1.34490478e+00 -8.05008292e-01 -2.46754169e-01
-2.28586212e-01 8.09271574e-01 7.97644317e-01 -7.88207650e-01
3.68485600e-01 3.69900823e-01 -1.86695981e+00 2.70343989e-01
1.09260750e+00 4.83236730e-01 4.79678065e-01 9.18210983e-01
-4.81124252e-01 1.76900852e+00 -5.19958138e-01 -2.18303785e-01
-7.35160569e-03 -4.41562176e-01 -1.37119639e+00 -1.49854079e-01
-1.21911265e-01 -4.70007896e-01 -3.89586598e-01 9.45682406e-01
1.48747191e-02 5.79156339e-01 4.78890628e-01 -1.27205086e+00
-5.22783160e-01 9.07343328e-01 2.32534967e-02 1.01348780e-01
8.78159225e-01 5.27568638e-01 7.00454175e-01 -1.84776351e-01
6.94461346e-01 1.38140273e+00 1.08987466e-01 7.69565582e-01
-8.66526544e-01 -3.17829430e-01 4.10758227e-01 4.89905477e-01
-9.85515594e-01 -2.29505301e-01 5.71222007e-01 -5.60106695e-01
1.24734128e+00 -9.48438495e-02 8.22011650e-01 8.58203351e-01
1.06275685e-01 1.08402359e+00 1.17412531e+00 -7.42339313e-01
4.04781491e-01 1.86157063e-01 2.41160810e-01 1.11131167e+00
8.17808062e-02 4.65012193e-01 -1.08694339e+00 5.25862090e-02
1.20983934e+00 -2.87402034e-01 3.98216546e-02 -7.25492358e-01
-1.16526973e+00 8.43622386e-01 7.77313113e-02 3.32228392e-01
-2.38688260e-01 1.61187813e-01 3.24677438e-01 7.24723577e-01
6.85121194e-02 7.17664182e-01 -4.68962163e-01 -4.48226720e-01
-3.07340268e-02 4.13823336e-01 1.25612128e+00 1.16820550e+00
9.23723996e-01 1.09741732e-01 5.06676398e-02 3.87806505e-01
3.82846534e-01 4.25711215e-01 4.16816205e-01 -1.16120696e+00
4.14566159e-01 9.55529630e-01 1.19494669e-01 -9.54489827e-01
-4.64747727e-01 -6.57132715e-02 -2.45996296e-01 3.25645715e-01
3.70960832e-01 -2.49156699e-01 -7.30549395e-01 1.90016472e+00
5.68079948e-01 2.67768890e-01 5.21367311e-01 6.08662963e-01
9.70884085e-01 6.27636790e-01 3.53045672e-01 -6.51110336e-02
1.42190981e+00 -9.57278311e-01 -5.41328490e-01 -4.47484463e-01
9.43596601e-01 -5.86193874e-02 1.12249136e+00 3.82119179e-01
-1.18967378e+00 -4.98940676e-01 -8.73348236e-01 -8.20684433e-02
-8.32913339e-01 -5.78158259e-01 9.43806648e-01 5.03002346e-01
-1.23553693e+00 3.70785683e-01 -7.17983007e-01 -4.41424131e-01
7.48915002e-02 4.95252281e-01 -2.12590933e-01 -7.30882585e-02
-1.28497756e+00 1.11088133e+00 8.04377556e-01 -2.71470785e-01
-1.28196764e+00 -5.51131546e-01 -1.17216206e+00 -1.38652362e-02
9.85556722e-01 -8.78211498e-01 1.57421839e+00 -9.28004801e-01
-2.09790015e+00 8.23029459e-01 3.16177994e-01 -5.00663400e-01
1.77501246e-01 -2.56576657e-01 8.07316974e-03 3.03962618e-01
-5.32791279e-02 6.58065617e-01 5.19851863e-01 -1.28529000e+00
-7.98923671e-01 -4.58300918e-01 1.07302141e+00 6.68156087e-01
7.64541775e-02 -3.95486541e-02 -2.68539906e-01 -2.22286165e-01
-1.00333221e-01 -7.53342986e-01 -3.64702135e-01 -3.16558450e-01
1.10526986e-01 -4.67376709e-01 3.63324195e-01 -1.85360819e-01
9.03995275e-01 -1.84903932e+00 4.20027226e-01 8.52977261e-02
3.05817366e-01 2.57562280e-01 -3.20570350e-01 6.83480620e-01
3.31331432e-01 -6.21184483e-02 -2.08914150e-02 2.47704834e-01
2.49502495e-01 6.78592682e-01 -2.35797301e-01 -3.06413025e-01
4.53938209e-02 1.10136950e+00 -1.28693628e+00 -5.18505692e-01
4.18002725e-01 -5.30762784e-02 -8.29838455e-01 6.76525831e-01
-9.70061123e-01 4.98435348e-01 -9.07572031e-01 2.00291842e-01
-1.57627016e-02 -1.21962458e-01 3.46440881e-01 4.45203036e-01
-1.15000665e-01 5.22404969e-01 -1.23236334e+00 1.98966897e+00
-6.08871818e-01 2.64311969e-01 1.81421429e-01 -8.77173424e-01
4.73809004e-01 3.27525854e-01 7.92184770e-02 -8.10392857e-01
1.43818129e-02 -2.20560178e-01 2.62090415e-01 -8.06471527e-01
5.42662740e-01 -6.58690557e-02 -1.68245763e-01 8.07918310e-01
2.65565127e-01 -7.36733317e-01 4.22440410e-01 6.39644325e-01
1.03332853e+00 6.47380531e-01 4.51111704e-01 -1.38807103e-01
4.83961791e-01 4.37874019e-01 2.39690654e-02 1.16412389e+00
-1.30465493e-01 -1.76049948e-01 4.38925415e-01 -4.48370874e-01
-3.43353868e-01 -8.62603128e-01 6.12626076e-01 1.76357210e+00
3.48828137e-01 -5.91150522e-01 -8.50527525e-01 -5.21507680e-01
-4.78823811e-01 9.00116622e-01 -4.92076069e-01 -3.32242250e-01
-6.53583109e-01 -2.64302284e-01 5.09381056e-01 4.74161416e-01
6.66531503e-01 -1.70159614e+00 -1.28204226e+00 3.18872064e-01
-1.18781567e-01 -1.13017809e+00 3.59437056e-02 3.95329863e-01
-8.42594683e-01 -1.37422597e+00 2.20971763e-01 -8.87746751e-01
4.70475346e-01 2.12164462e-01 1.36125755e+00 4.96728927e-01
2.55125109e-02 1.25813842e+00 -6.10811532e-01 -7.21384287e-01
-6.07519209e-01 3.00341286e-04 -1.62289411e-01 -4.82108444e-01
2.99845994e-01 -5.67648172e-01 -3.08373481e-01 4.24131826e-02
-1.09670258e+00 5.69150150e-01 4.77622807e-01 7.19074965e-01
1.96134955e-01 1.59954280e-01 3.23645025e-01 -1.09315979e+00
1.06433439e+00 -3.45276088e-01 -5.75105906e-01 3.65937561e-01
-1.67099327e-01 3.51795793e-01 5.36878407e-01 -3.41591597e-01
-1.38963437e+00 -1.44316688e-01 -9.88501385e-02 2.34734312e-01
-6.09594405e-01 8.57887864e-01 -1.06982112e-01 -2.13684887e-01
8.15364420e-01 4.97250289e-01 5.45909181e-02 1.07269317e-01
7.23119378e-01 1.72581926e-01 4.02330160e-01 -1.52788901e+00
6.53145313e-01 -5.69743058e-03 -3.08418334e-01 -8.77556920e-01
-7.62487113e-01 -3.66746336e-01 -5.86456358e-01 -1.14411660e-01
8.17198455e-01 -6.28890455e-01 -1.27741039e+00 5.06683767e-01
-1.03196549e+00 -1.12847698e+00 -5.27718246e-01 2.41584495e-01
-1.01715112e+00 2.09639534e-01 -7.98906684e-01 -8.11450243e-01
8.53246152e-02 -1.31323683e+00 5.88428795e-01 3.72390181e-01
-1.42752215e-01 -1.21837044e+00 2.86281943e-01 3.45229477e-01
4.02651966e-01 -1.14786774e-01 1.36133635e+00 -8.69533360e-01
-7.46179998e-01 3.07980776e-01 7.71336108e-02 -2.18873546e-01
-1.38925448e-01 -3.47005248e-01 -7.80530334e-01 7.14511424e-02
-5.68419918e-02 -9.67434645e-01 2.74513036e-01 -5.37121631e-02
8.98320198e-01 -2.64747292e-01 -2.03049317e-01 1.79530699e-02
1.13189328e+00 5.53558528e-01 5.31273544e-01 3.77205700e-01
4.06897783e-01 8.54348421e-01 5.47700286e-01 1.89954981e-01
8.66473794e-01 5.27180195e-01 4.36538368e-01 3.12162936e-01
1.31545633e-01 -4.95690405e-01 3.23352516e-01 8.80620599e-01
-4.25779104e-01 -8.00390244e-02 -1.11343765e+00 4.33745712e-01
-2.11510015e+00 -9.69611824e-01 3.56957108e-01 1.71605909e+00
1.23949909e+00 5.78978807e-02 4.18554693e-02 -2.03842908e-01
1.68322966e-01 9.47752893e-02 -6.38617218e-01 -4.32308048e-01
3.47005010e-01 4.41941142e-01 -1.87330469e-01 8.50818813e-01
-8.99489462e-01 1.50625920e+00 6.45969868e+00 6.34288788e-01
-5.42834044e-01 -2.06029676e-02 1.56038627e-01 4.90249515e-01
-2.23179594e-01 -9.86531004e-02 -4.43504065e-01 -2.52786905e-01
1.02709079e+00 -1.98104247e-01 7.73095846e-01 7.26537645e-01
-5.82709983e-02 -5.36365569e-01 -1.30381560e+00 5.64419031e-01
-2.19818484e-02 -1.32419419e+00 4.36506599e-01 -2.51736850e-01
4.27378982e-01 -1.80978194e-01 -1.86979309e-01 9.07255113e-01
1.18169427e+00 -9.22415972e-01 6.88133657e-01 3.10632080e-01
2.11558342e-01 -7.11457491e-01 4.22799975e-01 9.40067053e-01
-9.93112743e-01 -8.47130790e-02 -6.27831966e-02 -6.02073193e-01
-2.50093997e-01 -3.59516621e-01 -8.35280061e-01 6.15721643e-01
4.65939283e-01 3.43681782e-01 -3.90863985e-01 6.03495240e-01
-5.58986187e-01 4.18845683e-01 -2.95692027e-01 -4.06456262e-01
2.06265852e-01 -4.03362989e-01 2.77101934e-01 8.61266494e-01
-2.77091920e-01 9.46972609e-01 6.08192027e-01 7.52021372e-01
2.64730513e-01 4.77284752e-03 -8.81288230e-01 -5.31163663e-02
3.21214765e-01 8.93199086e-01 -8.21294248e-01 -5.62162638e-01
-5.46672165e-01 5.02777398e-01 7.34576464e-01 4.50881183e-01
-5.79214633e-01 -1.76631987e-01 4.99920577e-01 -2.91408226e-02
-1.29167467e-01 -5.65112829e-01 3.45570818e-02 -1.20573521e+00
-4.31907028e-01 -1.36485195e+00 4.59853977e-01 -8.74886274e-01
-7.72632360e-01 3.78899783e-01 4.24806178e-01 -4.82105404e-01
-5.62611818e-01 -8.30237150e-01 -5.09452522e-01 3.60629827e-01
-1.40903032e+00 -1.00014460e+00 -2.06557781e-01 1.01586521e+00
7.04091787e-01 -2.43407845e-01 1.16522610e+00 -3.82534057e-01
-1.44459248e-01 6.33836305e-03 -5.83126664e-01 8.60918015e-02
4.83540893e-02 -1.37973464e+00 1.41219527e-01 5.63630044e-01
2.52071261e-01 8.05111766e-01 6.93188667e-01 -5.72191894e-01
-1.87236142e+00 -6.80372238e-01 2.67244339e-01 -6.55253172e-01
8.08731854e-01 -3.07485104e-01 -9.09227908e-01 9.81242299e-01
5.07155597e-01 -4.98626381e-01 6.78779244e-01 1.38694644e-01
-1.65833935e-01 3.24123353e-01 -8.90184522e-01 1.01324069e+00
1.29272687e+00 -7.03751922e-01 -1.21548438e+00 3.48163366e-01
1.05308127e+00 -6.53078020e-01 -6.31703794e-01 8.08098763e-02
1.81483030e-01 -9.32925344e-01 9.65002120e-01 -1.00326157e+00
3.99649173e-01 -3.82167220e-01 7.00388700e-02 -1.42040932e+00
-1.49179652e-01 -7.08650231e-01 -2.06470132e-01 7.83210576e-01
3.01676214e-01 -6.58613622e-01 8.08019757e-01 8.53362679e-01
-2.15194955e-01 -4.42757428e-01 -4.83239681e-01 -2.81802446e-01
1.96017683e-01 -7.59416997e-01 5.07083893e-01 8.50412965e-01
6.33187056e-01 6.44270122e-01 1.44397661e-01 1.44605324e-01
2.26551980e-01 2.12761879e-01 1.00126076e+00 -1.24970281e+00
-5.97370207e-01 -3.62203151e-01 6.01378677e-04 -1.36684084e+00
4.44349080e-01 -6.83233917e-01 2.19571501e-01 -1.68873751e+00
5.31938858e-02 -3.92975837e-01 6.62560761e-02 7.15248764e-01
-3.61839011e-02 -4.39204961e-01 1.69014782e-01 -1.15918152e-01
-1.11075783e+00 3.94474238e-01 1.58190382e+00 -1.84391126e-01
-3.15144747e-01 -1.78969607e-01 -5.43836832e-01 1.01373494e+00
7.80545115e-01 -6.32323697e-02 -1.05206048e+00 -6.60748482e-01
6.71444952e-01 3.63510400e-01 3.11626226e-01 -9.46336269e-01
8.32393587e-01 -7.79095411e-01 -1.61303625e-01 -1.83844358e-01
2.61826843e-01 -8.94937038e-01 -2.32453361e-01 4.54535812e-01
-5.67096055e-01 2.47609876e-02 5.22021234e-01 5.33088744e-01
-2.80815572e-01 -3.80752176e-01 2.85614908e-01 -6.71713948e-01
-1.28127527e+00 -2.64876708e-02 -9.81288075e-01 3.61710429e-01
1.16545677e+00 -2.97421366e-01 -4.64868605e-01 -5.58499217e-01
-7.97515154e-01 4.39928591e-01 2.24926472e-01 2.90067494e-01
6.70516431e-01 -6.78219855e-01 -3.24726909e-01 2.08119735e-01
2.33728200e-01 3.27041000e-01 1.15190811e-01 3.18889946e-01
-7.22382009e-01 4.98780519e-01 -4.05642569e-01 -2.24943236e-01
-9.02390182e-01 6.98319376e-01 4.22434270e-01 -4.57265109e-01
-4.25534427e-01 6.35308385e-01 6.70598090e-01 -8.91495049e-01
3.86656642e-01 -5.03396332e-01 -6.79099500e-01 -3.13195765e-01
5.94015658e-01 -9.54018440e-04 -1.13765404e-01 -1.37200937e-01
2.64850520e-02 4.00726974e-01 -6.79124817e-02 -2.50026047e-01
1.17919099e+00 1.39972903e-02 -1.01015195e-01 4.22478348e-01
1.12929761e-01 -2.28689820e-01 -9.92584646e-01 -4.45997059e-01
1.74816132e-01 7.17826188e-02 -2.32009187e-01 -9.27965462e-01
-5.45463979e-01 8.47194970e-01 2.47861333e-02 4.25368994e-01
8.29533994e-01 5.09835631e-02 1.16726466e-01 1.02473545e+00
1.08149767e+00 -1.05165052e+00 5.10184765e-01 1.11246896e+00
7.81621993e-01 -1.11118138e+00 -1.44719779e-01 -4.88117635e-01
-5.83526611e-01 1.02284765e+00 1.10658026e+00 1.74216613e-01
3.88093710e-01 3.50760072e-01 -4.76395991e-03 -5.19451678e-01
-1.11555827e+00 -5.26976943e-01 -1.55264571e-01 1.06548071e+00
3.53050351e-01 -5.84066333e-03 1.38938516e-01 5.31636059e-01
-3.55147094e-01 8.74713361e-02 6.23378992e-01 1.47000384e+00
-7.28389740e-01 -1.03642464e+00 -2.92426527e-01 2.00224042e-01
-5.98905310e-02 -1.32198200e-01 -4.99824643e-01 1.03257120e+00
-5.87199815e-02 1.12168574e+00 -1.10979185e-01 -8.81889537e-02
5.25052249e-01 1.13052957e-01 9.16737258e-01 -1.19176579e+00
-9.56025958e-01 -5.38269997e-01 2.88903892e-01 -6.44255459e-01
-5.93794525e-01 -3.16743553e-01 -1.58621550e+00 -2.08574831e-01
-9.27964523e-02 4.06840563e-01 1.15495220e-01 1.25632954e+00
1.44202858e-01 6.10177875e-01 -1.95580691e-01 -6.83692813e-01
-2.84245044e-01 -7.33063996e-01 -2.73608685e-01 3.14249456e-01
2.09020004e-02 -7.07843661e-01 2.95708235e-02 1.92826480e-01] | [3.916210412979126, 1.2467976808547974] |
a2c1a8a7-e900-4cb4-97f3-c4e61f706dcf | ris-assisted-device-activity-detection-with | 2206.06805 | null | https://arxiv.org/abs/2206.06805v1 | https://arxiv.org/pdf/2206.06805v1.pdf | RIS Assisted Device Activity Detection with Statistical Channel State Information | This paper studies reconfigurable intelligent surface (RIS) assisted device activity detection for grant-free (GF) uplink transmission in wireless communication networks. In particular, we consider mobile devices located in an area where the direct link to an access point (AP) is blocked. Thus, the devices try to connect to the AP via a reflected link provided by an RIS. Therefore, a RIS phase-shift design is desired that covers the entire blocked area with a wide reflection beam because the exact locations and times of activity of the devices are unknown in GF transmission. In order to study the impact of the phase-shift design on the device activity detection, we derive a generalized likelihood ratio test (GLRT) based detector and present an analytical expression for the probability of detection. Assuming knowledge of statistical CSI, we formulate an optimization problem for the phase-shift design for maximization of the guaranteed probability of detection for all locations within a given coverage area. To tackle the non-convexity of the problem, we propose two different approximations of the objective function. The first approximation leads to a design that aims to reduce the variations of the end-to-end channel while taking system parameters such as transmit power, noise power, and probability of false alarm into account. The second approximation can be adopted for versatile RIS deployments because it only depends on the line-of-sight component of the end-to-end channel and is not affected by system parameters. For comparison, we also consider a phase-shift design maximizing the average channel gain and a baseline analytical phase-shift design for large blocked areas. Our performance evaluation shows that the proposed approximations result in phase-shift designs that guarantee high probability of detection across the coverage area and outperform the baseline designs. | ['Robert Schober', 'Vahid Jamali', 'Friedemann Laue'] | 2022-06-14 | null | null | null | null | ['activity-detection'] | ['computer-vision'] | [ 5.81425726e-01 6.90939069e-01 -7.96376169e-02 1.43457025e-01
-5.36242366e-01 -4.70935524e-01 -5.42192422e-02 2.21147195e-01
-2.46670887e-01 6.09082818e-01 -4.30810660e-01 -7.25415230e-01
-6.03315353e-01 -1.09505439e+00 -6.81944132e-01 -1.06250608e+00
-3.68345171e-01 7.94934705e-02 1.69247597e-01 -1.96660832e-01
-2.39301428e-01 5.52198887e-01 -1.11717868e+00 -4.60872591e-01
5.62831402e-01 1.16542697e+00 1.99248821e-01 4.31259245e-01
3.17502797e-01 -2.02158019e-02 -9.34030175e-01 1.09168984e-01
1.35457993e-01 -3.59939307e-01 -4.79501896e-02 -9.44615304e-02
-4.15930331e-01 -2.86249757e-01 -1.86655074e-01 6.94974184e-01
6.03516042e-01 -3.52225423e-01 8.32774580e-01 -1.05826831e+00
4.39091057e-01 5.34447491e-01 -7.38175809e-01 2.47190762e-02
4.14682806e-01 -4.87905860e-01 5.77140570e-01 -1.66352630e-01
2.00181603e-01 6.12995148e-01 5.44435859e-01 -7.53561333e-02
-8.58137727e-01 -5.81979752e-01 -1.27111286e-01 -2.76950806e-01
-1.75497353e+00 -2.78251082e-01 5.22729158e-01 -1.42088339e-01
3.61574382e-01 5.68420053e-01 5.63951969e-01 5.28220713e-01
2.83003777e-01 3.18557948e-01 3.60863268e-01 -9.13325071e-01
4.92305130e-01 1.96770713e-01 -9.52481627e-02 5.67083120e-01
1.00238526e+00 5.72182164e-02 -6.61359504e-02 -3.96109551e-01
6.54653728e-01 -3.61871362e-01 -5.87409735e-01 -4.96537060e-01
-9.35120404e-01 2.96845555e-01 3.29338700e-01 6.19007468e-01
-5.97634435e-01 3.09853852e-02 -3.57602894e-01 2.77528763e-01
7.04554617e-02 3.09792101e-01 -1.75984487e-01 1.54951543e-01
-7.67389536e-01 -1.68187976e-01 1.05232716e+00 1.27293682e+00
4.86675978e-01 -1.61636382e-01 -5.66452086e-01 3.11564893e-01
6.42144680e-01 1.14209473e+00 -3.82419527e-01 -3.21143627e-01
5.45807004e-01 2.56407052e-01 4.77370411e-01 -9.56835985e-01
-6.65335119e-01 -1.40391135e+00 -7.32656121e-01 -1.60258219e-01
5.49635351e-01 -7.10422158e-01 -4.38662946e-01 1.63838041e+00
2.30584487e-01 -4.13773395e-02 3.94229412e-01 5.51177919e-01
9.25148204e-02 5.95392406e-01 -4.85307902e-01 -8.85926545e-01
1.29260290e+00 -1.71829522e-01 -6.76196516e-01 -1.80813819e-01
6.72126532e-01 -5.10076821e-01 3.54678988e-01 4.34125483e-01
-9.39400554e-01 7.98536316e-02 -1.68148720e+00 8.91482234e-01
2.44013742e-01 5.01107037e-01 -1.37163356e-01 1.33134270e+00
-6.47290647e-01 -2.40047961e-01 -5.35470188e-01 -3.47020119e-01
1.15064509e-01 4.76176083e-01 3.14277351e-01 -4.53403145e-02
-9.42096889e-01 1.58821017e-01 -1.45704925e-01 1.67778194e-01
-2.81898975e-01 -5.55497169e-01 -4.37856585e-01 2.97035784e-01
4.34506804e-01 -5.78139305e-01 1.05787194e+00 -6.91202581e-01
-1.49537301e+00 2.34905705e-01 -1.96988761e-01 -2.44064435e-01
3.38682860e-01 2.06751958e-01 -4.41960603e-01 1.52890354e-01
-7.69443586e-02 -4.54350471e-01 4.51609969e-01 -1.18276811e+00
-7.25661635e-01 -3.20941448e-01 1.77687123e-01 1.65828481e-01
-3.13060164e-01 -4.41203117e-01 -4.31708902e-01 -2.14695245e-01
4.03357714e-01 -1.03038454e+00 -2.18473777e-01 -3.77267569e-01
-7.29503512e-01 3.37175459e-01 6.15478873e-01 -2.41893262e-01
1.15430319e+00 -2.32865620e+00 -1.90752640e-01 8.24862540e-01
-3.57585698e-01 6.80664778e-02 1.62757084e-01 6.85887814e-01
3.71577829e-01 -1.53315768e-01 -1.09109402e-01 -1.55310318e-01
-4.59002912e-01 -1.64859608e-01 1.02048434e-01 7.59752214e-01
-3.92996252e-01 2.39573732e-01 -7.30404377e-01 -1.84315182e-02
2.02997122e-02 4.57589865e-01 -4.35586005e-01 1.37270495e-01
2.56643742e-01 5.40105939e-01 -9.61215615e-01 3.68467927e-01
9.60978091e-01 -2.76486333e-02 5.09555161e-01 -1.82447225e-01
-4.68608737e-01 -3.63103822e-02 -1.20874798e+00 1.22951269e+00
-7.94456542e-01 2.24225000e-01 3.05809081e-01 -1.08702815e+00
9.15947556e-01 3.90790731e-01 7.43352652e-01 -8.61178696e-01
3.11620384e-01 2.65728712e-01 2.67301984e-02 -1.94236562e-01
2.33104955e-02 9.96450558e-02 -3.37177575e-01 2.99469382e-01
-3.09124380e-01 5.00174999e-01 -3.05943161e-01 8.19978416e-02
1.44960070e+00 -3.19085628e-01 4.35765266e-01 -5.59872925e-01
6.62666380e-01 -2.95144111e-01 5.11906147e-01 8.52776349e-01
3.08736295e-01 9.73938107e-02 5.01117945e-01 2.96494842e-01
-4.43064004e-01 -1.05552149e+00 -2.09048912e-01 3.45841020e-01
8.79631877e-01 -1.62220001e-03 -6.43678427e-01 -2.17509478e-01
-5.35558118e-03 8.32817852e-01 -1.59002706e-01 -2.65589476e-01
-3.54721606e-01 -8.49270046e-01 5.16822934e-01 -1.69631422e-01
6.09634042e-01 2.81174754e-04 -8.60967755e-01 1.55650854e-01
-1.52890801e-01 -1.13538039e+00 8.54988545e-02 1.32984996e-01
-4.03327495e-01 -1.01883006e+00 -7.07931578e-01 -4.81631339e-01
8.72634947e-01 6.21343970e-01 4.72126782e-01 -7.17133284e-02
5.65380044e-03 7.35214055e-01 -4.13006783e-01 -5.01434267e-01
-1.52994558e-01 1.18616380e-01 -2.34948136e-02 1.54424310e-01
-1.21428452e-01 -3.30863088e-01 -8.27717364e-01 9.59607065e-01
-4.28606331e-01 -4.30314124e-01 6.69554174e-01 2.80748457e-01
4.28205311e-01 5.05400002e-01 8.46899331e-01 -2.46843770e-01
2.20572725e-01 -6.43523097e-01 -8.62675488e-01 3.35832983e-01
-3.23238790e-01 5.36951013e-02 3.12091887e-01 -9.58611295e-02
-1.05618525e+00 1.24623198e-02 -1.86148077e-01 4.31736320e-01
1.67372182e-01 3.35728407e-01 -7.83419013e-01 -2.60040253e-01
5.13286650e-01 -3.54544371e-02 -2.52904624e-01 -1.12096585e-01
-1.38069242e-01 9.24396515e-01 -2.61866711e-02 -1.86212972e-01
9.08214211e-01 6.89689875e-01 5.17540038e-01 -1.55304801e+00
-4.45957690e-01 -5.80907285e-01 1.42855510e-01 -3.56896967e-01
3.04375082e-01 -1.02885616e+00 -8.95274580e-01 1.89061418e-01
-9.43729222e-01 -7.39213899e-02 -8.67162552e-03 6.94191396e-01
-2.68970013e-01 3.10630441e-01 2.04001918e-01 -1.41410625e+00
-2.90633321e-01 -8.97179723e-01 1.07364953e+00 2.00750932e-01
-6.59852698e-02 -5.77965260e-01 -2.44843900e-01 -2.41691947e-01
4.16609824e-01 4.40607071e-01 7.67114341e-01 -1.77679807e-01
-4.92005140e-01 -5.60377717e-01 2.64362320e-02 -2.09423885e-01
2.02076241e-01 -6.42683506e-01 -6.17102325e-01 -5.59883893e-01
1.34216830e-01 6.98722243e-01 1.52832910e-01 8.84392977e-01
7.28924811e-01 -3.05276036e-01 -1.14896750e+00 4.70576733e-01
1.46945643e+00 3.70037049e-01 8.69272411e-01 1.56332284e-01
7.93882757e-02 4.12423193e-01 1.01404428e+00 6.10850632e-01
8.11103210e-02 9.96848106e-01 8.04997265e-01 -7.74659291e-02
2.39301115e-01 4.47725169e-02 1.16968714e-01 -7.13547170e-02
2.35470608e-01 -1.24527466e+00 -4.10679549e-01 3.29799801e-01
-1.58734870e+00 -4.64020580e-01 -5.45150638e-01 2.95379043e+00
-7.69354403e-02 1.82594731e-01 1.27810553e-01 4.79955584e-01
7.40641177e-01 -2.39886120e-01 -2.21917495e-01 -1.41262757e-02
1.25642372e-02 -1.40730396e-01 1.29248440e+00 5.81694841e-01
-5.93749762e-01 -5.84962778e-02 5.12951756e+00 7.52577245e-01
-9.51557815e-01 9.00292248e-02 2.21181527e-01 -7.77401822e-03
-4.73246276e-01 -5.70150055e-02 -8.04736197e-01 3.57614100e-01
6.98177576e-01 1.04112297e-01 -1.04249157e-01 4.34958696e-01
5.19928992e-01 -5.87380469e-01 -7.65127182e-01 8.05752695e-01
-2.25505322e-01 -7.25469232e-01 -4.70051229e-01 6.31464183e-01
3.36358398e-01 -4.16316569e-01 -9.61276069e-02 -2.01885462e-01
-4.75781798e-01 -3.02339673e-01 6.02752507e-01 6.48033023e-01
7.53125131e-01 -8.82193089e-01 7.29641020e-01 5.80236435e-01
-1.17505038e+00 -3.79795313e-01 2.21501127e-01 -7.88407102e-02
4.67244238e-01 1.20475996e+00 -8.17958415e-01 9.39230919e-01
2.64101803e-01 -2.17230245e-01 2.54176766e-01 1.28223586e+00
-2.62226522e-01 6.74079537e-01 -8.28955591e-01 -4.68726128e-01
-1.95362747e-01 -6.00767918e-02 1.04389060e+00 7.36143529e-01
1.01297891e+00 5.53678162e-02 -3.03659558e-01 4.47349638e-01
1.90111592e-01 1.10662088e-01 -5.49665272e-01 6.69398010e-01
8.40993464e-01 8.12767446e-01 -6.80492580e-01 4.70295906e-01
-4.30225790e-01 6.74311638e-01 -4.12427634e-01 6.13487661e-01
-6.72685504e-01 -4.97469068e-01 4.92853433e-01 7.36270189e-01
3.87915105e-01 -5.68964183e-01 -3.19958568e-01 -3.24941635e-01
2.98575252e-01 -2.25378424e-01 2.99345199e-02 -9.01729241e-02
-3.99680883e-01 1.86764732e-01 6.36339709e-02 -1.41640496e+00
-1.74304590e-01 -1.36359379e-01 -4.23164070e-01 7.44315267e-01
-1.43627679e+00 -7.76078880e-01 -3.68034065e-01 2.47256726e-01
-1.42358124e-01 4.85902466e-02 6.58159494e-01 5.80067933e-01
-4.07657415e-01 9.34333503e-01 4.33667183e-01 -1.84898913e-01
5.40319234e-02 -5.10967553e-01 -2.55468607e-01 9.55514789e-01
-3.55321378e-01 5.21343648e-01 8.92069221e-01 -5.93671739e-01
-1.65671957e+00 -8.63242805e-01 4.94780093e-01 3.22893232e-01
-1.33915991e-03 -5.04300058e-01 -7.88315386e-02 6.35426119e-02
-1.52613401e-01 -2.28765517e-01 6.54609621e-01 -7.45434463e-02
4.19885725e-01 -4.14788842e-01 -1.45403004e+00 5.34783661e-01
9.58598435e-01 2.63247758e-01 3.46082687e-01 1.82508767e-01
2.09755823e-01 -2.37417370e-01 -5.11669874e-01 6.88305378e-01
8.29827785e-01 -5.14809787e-01 7.68999338e-01 4.52307552e-01
-6.76393032e-01 -3.48748386e-01 -2.62106538e-01 -1.02012801e+00
-2.71377921e-01 -9.22430158e-01 3.18399742e-02 1.14267516e+00
6.55527711e-01 -7.07471848e-01 7.61833251e-01 8.23826995e-03
1.40596539e-01 -6.38698816e-01 -1.20087624e+00 -9.72256362e-01
-4.57748502e-01 -2.94476151e-01 4.89649147e-01 1.02349676e-01
1.62162244e-01 4.47669595e-01 -2.27612883e-01 1.10875356e+00
7.48974741e-01 -4.15952951e-01 7.59788215e-01 -1.21809947e+00
-4.49911445e-01 1.04506938e-02 -4.03126895e-01 -1.66779554e+00
-4.43467408e-01 -2.70389736e-01 8.81892666e-02 -1.69057131e+00
-4.65419352e-01 -1.05235338e+00 -5.39149232e-02 9.89764184e-02
3.93571764e-01 -2.25501098e-02 -4.25391674e-01 -1.27935410e-01
-1.57422617e-01 3.55165482e-01 8.74589860e-01 -1.45033747e-02
-5.13091445e-01 1.02120233e+00 -6.12868607e-01 4.23951775e-01
8.01121175e-01 -4.38143790e-01 -6.46340668e-01 -1.47848189e-01
6.35001183e-01 4.12994742e-01 2.13042989e-01 -1.32332921e+00
9.46312621e-02 1.46768421e-01 -1.56981319e-01 -4.84717131e-01
4.37539876e-01 -1.38514936e+00 3.46710861e-01 7.16285586e-01
2.69169420e-01 -7.34485269e-01 6.66495562e-02 9.94843781e-01
4.54395145e-01 -4.32687402e-01 6.43336713e-01 7.22095191e-01
1.89385056e-01 -1.95935220e-01 -8.79687726e-01 -6.72127008e-01
1.43893003e+00 -3.91482145e-01 -2.98464745e-01 -7.29198873e-01
-1.91111133e-01 3.08399349e-01 9.88741294e-02 1.75073683e-01
1.66044638e-01 -7.81095266e-01 -3.75426054e-01 8.66330191e-02
1.24626480e-01 -2.50435859e-01 2.61302948e-01 1.01519620e+00
-2.13442102e-01 7.59710073e-01 3.27188253e-01 -7.88201332e-01
-1.21336627e+00 5.07032238e-02 6.07347012e-01 -9.31452289e-02
-2.37785473e-01 4.77191120e-01 1.19351007e-01 2.50839323e-01
3.16470712e-01 -2.69618064e-01 -2.53126353e-01 -1.96596399e-01
3.44627410e-01 4.70570832e-01 3.69065315e-01 -5.47872901e-01
-4.65395391e-01 6.82741463e-01 5.23894131e-01 -1.50721252e-01
7.34965265e-01 -5.60664952e-01 4.08267736e-01 -2.05150936e-02
8.55681241e-01 6.45510674e-01 -7.35583246e-01 -9.53268185e-02
-3.78666878e-01 -2.21493825e-01 2.81841397e-01 -5.84288001e-01
-8.81137669e-01 9.02649462e-02 9.26507831e-01 5.50334096e-01
1.14577675e+00 -1.11592628e-01 3.29143047e-01 3.34297061e-01
1.17182600e+00 -7.06233621e-01 -1.65405408e-01 1.36101633e-01
3.68062198e-01 -3.82139802e-01 2.78399978e-02 -9.98113394e-01
2.23505259e-01 8.11942995e-01 2.10636511e-01 2.81137645e-01
9.37267840e-01 5.04984438e-01 -2.96051264e-01 -1.99681610e-01
1.24920532e-01 -2.53332019e-01 -1.19735807e-01 7.62076735e-01
2.32934341e-01 1.10621139e-01 -8.06641936e-01 6.94356859e-01
6.38221055e-02 -2.48858809e-01 4.85280663e-01 1.08265412e+00
-6.72652066e-01 -1.20108116e+00 -5.23809254e-01 4.59133387e-01
-5.14029443e-01 3.73933643e-01 1.20323800e-01 6.29823625e-01
1.33906677e-01 1.50213659e+00 7.43197203e-02 -1.19247787e-01
8.59388769e-01 -5.26439190e-01 4.61468697e-01 -3.85004044e-01
1.82460964e-01 1.09639034e-01 4.47271407e-01 -2.01590389e-01
-2.26273268e-01 -5.02835512e-01 -1.29429305e+00 1.61275834e-01
-7.48784065e-01 5.01537859e-01 8.65862906e-01 1.07727468e+00
4.48084474e-01 8.59531283e-01 9.46226239e-01 -3.73344392e-01
-1.70439363e-01 -5.57705998e-01 -9.26227629e-01 -6.63281083e-01
3.48886669e-01 -8.40247035e-01 -3.80445987e-01 -7.85776556e-01] | [6.254181861877441, 1.2242292165756226] |
8c8f703b-845d-48ee-817c-7f7257b91b7e | self-sufficient-framework-for-continuous-sign | 2303.11771 | null | https://arxiv.org/abs/2303.11771v1 | https://arxiv.org/pdf/2303.11771v1.pdf | Self-Sufficient Framework for Continuous Sign Language Recognition | The goal of this work is to develop self-sufficient framework for Continuous Sign Language Recognition (CSLR) that addresses key issues of sign language recognition. These include the need for complex multi-scale features such as hands, face, and mouth for understanding, and absence of frame-level annotations. To this end, we propose (1) Divide and Focus Convolution (DFConv) which extracts both manual and non-manual features without the need for additional networks or annotations, and (2) Dense Pseudo-Label Refinement (DPLR) which propagates non-spiky frame-level pseudo-labels by combining the ground truth gloss sequence labels with the predicted sequence. We demonstrate that our model achieves state-of-the-art performance among RGB-based methods on large-scale CSLR benchmarks, PHOENIX-2014 and PHOENIX-2014-T, while showing comparable results with better efficiency when compared to other approaches that use multi-modality or extra annotations. | ['Joon Son Chung', 'In So Kweon', 'Dong-Jin Kim', 'Myungchul Kim', 'Jae Won Cho', 'Youngtaek Oh', 'Youngjoon Jang'] | 2023-03-21 | null | null | null | null | ['sign-language-recognition', 'pseudo-label'] | ['computer-vision', 'miscellaneous'] | [ 3.39888394e-01 -1.82829767e-01 -5.09161036e-04 -5.53650856e-01
-1.01353228e+00 -4.65655833e-01 5.98878562e-01 -6.40914857e-01
-7.60724545e-01 6.11316800e-01 4.84552830e-01 -9.07120258e-02
3.22733223e-01 -2.55164981e-01 -6.47581875e-01 -5.86754024e-01
1.69139579e-01 1.50174975e-01 4.41801190e-01 -3.83038595e-02
-1.98047966e-01 4.67409283e-01 -1.88969147e+00 3.98114592e-01
6.23481810e-01 1.07867026e+00 -1.64876640e-01 9.13517416e-01
-1.88610360e-01 1.18556833e+00 -1.60791114e-01 -1.78982452e-01
2.58221179e-01 -4.12763208e-01 -7.01287270e-01 5.03606498e-02
1.16196859e+00 -5.87653160e-01 -6.60144031e-01 6.84862614e-01
7.63085246e-01 2.68031228e-02 3.42532486e-01 -1.20291150e+00
-2.97395051e-01 -8.78626332e-02 -4.26451325e-01 -2.37226918e-01
3.68797988e-01 6.73259199e-01 7.71519363e-01 -8.90585780e-01
9.60324228e-01 1.15889847e+00 8.92261147e-01 1.02864099e+00
-8.77124906e-01 -8.12781334e-01 4.29907620e-01 2.70814002e-01
-1.18824828e+00 -6.79930508e-01 4.21475261e-01 -4.00996178e-01
9.21302140e-01 9.45996195e-02 9.03171003e-01 1.29483867e+00
-3.60978335e-01 1.13731825e+00 1.51114213e+00 -5.12171328e-01
1.23121694e-01 -4.49160576e-01 3.24044555e-01 1.16015923e+00
-7.51830935e-02 2.58040488e-01 -8.47611606e-01 -4.86070989e-03
6.45784616e-01 -1.08525321e-01 -4.76425648e-01 -1.60481647e-01
-1.35465932e+00 1.16951443e-01 1.11500889e-01 9.83310193e-02
-3.24290872e-01 7.10418999e-01 3.41184974e-01 -2.22314633e-02
9.95672941e-02 -2.53592134e-01 -6.85407162e-01 -4.06256765e-01
-1.10721910e+00 -5.83413877e-02 6.19179368e-01 1.01122499e+00
5.35257101e-01 -8.00850540e-02 -3.12332034e-01 5.25781691e-01
5.97398520e-01 7.91687250e-01 2.05615014e-01 -1.28371942e+00
2.72502840e-01 4.40122187e-01 1.94488257e-01 -8.55203643e-02
-6.63191617e-01 -1.62442937e-01 -5.57254851e-01 6.37743592e-01
8.31902981e-01 -1.36888415e-01 -1.77385581e+00 1.84445536e+00
2.76173383e-01 8.00146401e-01 -3.77898328e-02 1.08527339e+00
1.32390881e+00 8.30217749e-02 4.09832418e-01 2.28548542e-01
1.21244943e+00 -1.32107162e+00 -6.18370414e-01 -9.69793350e-02
6.28838241e-01 -4.00184155e-01 1.16551828e+00 2.84589887e-01
-8.45717907e-01 -3.39824617e-01 -4.53701973e-01 -4.13750440e-01
-2.98515677e-01 5.05874813e-01 6.88160717e-01 6.41282618e-01
-1.23557234e+00 2.15098217e-01 -1.05168080e+00 -6.47984326e-01
5.85170925e-01 2.93321729e-01 -6.53795362e-01 -3.95655900e-01
-6.25483811e-01 7.84360766e-01 -1.74391389e-01 3.14531416e-01
-7.04714835e-01 -7.63745368e-01 -8.77117157e-01 -4.42764640e-01
1.66036308e-01 -5.92725575e-01 1.21940064e+00 -8.20371926e-01
-1.89142621e+00 9.99032199e-01 -5.31298399e-01 -7.88105726e-02
9.82779205e-01 -3.40566874e-01 -1.59399852e-01 1.91722006e-01
-3.32395226e-01 1.04487085e+00 7.02323616e-01 -1.19586349e+00
-8.75678360e-01 -2.76866108e-01 -9.94320586e-03 4.30244990e-02
1.26794443e-01 1.76128730e-01 -8.17733705e-01 -3.95970613e-01
2.34778255e-01 -1.10006523e+00 -6.48453385e-02 6.03197992e-01
-3.29822183e-01 -3.06652814e-01 9.25499439e-01 -1.14538789e+00
5.57105660e-01 -1.93999207e+00 -8.36055900e-04 6.40567616e-02
3.12686503e-01 4.81922984e-01 -5.49195766e-01 -1.61918327e-01
2.10045710e-01 -1.99567869e-01 -2.42200747e-01 -6.95735276e-01
-4.02803384e-02 4.01215076e-01 -4.75371107e-02 7.05515265e-01
1.08227164e-01 1.29146004e+00 -9.73316193e-01 -5.10589898e-01
5.69344223e-01 8.90317738e-01 -3.61806929e-01 -3.90207544e-02
-2.63558418e-01 8.50872993e-01 -1.11287713e-01 1.24396288e+00
6.07587993e-01 -4.42485750e-01 -4.23121080e-02 -4.32934999e-01
-1.43957362e-01 2.01060578e-01 -1.19454706e+00 2.02552772e+00
-2.24974379e-01 8.32176566e-01 2.05087498e-01 -4.20784473e-01
4.21043813e-01 2.97965199e-01 7.27264881e-01 -6.70409381e-01
2.93567121e-01 1.63966671e-01 -3.61737102e-01 -4.63198781e-01
6.03260100e-02 1.06731094e-01 2.35282078e-01 4.62663770e-01
2.39565238e-01 2.59322196e-01 9.83737037e-02 -9.85905603e-02
1.46592426e+00 6.80909455e-01 -1.15338206e-01 2.12440431e-01
5.64212382e-01 -1.82174280e-01 5.17009258e-01 8.44271123e-01
-6.75692976e-01 7.21026421e-01 7.14794695e-02 -3.43706846e-01
-7.73829341e-01 -1.06084239e+00 2.85318736e-02 1.15754068e+00
1.41965881e-01 -2.55291224e-01 -4.47055757e-01 -9.54419136e-01
-7.06415623e-03 3.34265858e-01 -6.32737458e-01 3.50216627e-01
-8.25901031e-01 -2.76226580e-01 1.03404713e+00 8.14202011e-01
7.12569177e-01 -1.03368270e+00 -8.43818247e-01 -1.47120412e-02
-3.58843684e-01 -1.52686322e+00 -4.88010466e-01 3.78450304e-02
-4.76217270e-01 -1.20234203e+00 -1.02584863e+00 -6.99964523e-01
6.10337138e-01 -1.12774923e-01 7.34547734e-01 4.84738462e-02
-3.54596496e-01 8.67609799e-01 -4.46141899e-01 -5.89410625e-02
-2.66869776e-02 -2.51615822e-01 -2.01853365e-01 1.54765369e-03
4.61435914e-01 -3.37875128e-01 -5.83580077e-01 2.60594010e-01
-6.32576883e-01 2.33603731e-01 5.36731064e-01 7.99399197e-01
6.50480628e-01 -9.12979424e-01 2.04110548e-01 -2.41715014e-01
1.92340225e-01 2.77352631e-01 -5.91528833e-01 7.16484487e-01
-3.00275296e-01 2.05345556e-01 2.51581609e-01 -5.88798642e-01
-9.46065903e-01 5.91549039e-01 -2.48880073e-01 -4.14281070e-01
-4.73891795e-01 -1.74458966e-01 2.42936239e-02 -6.39555514e-01
3.70507032e-01 4.59521621e-01 8.22989270e-02 -5.68429291e-01
7.52957880e-01 7.87111938e-01 9.29999173e-01 -5.10895550e-01
4.87737209e-01 1.02852368e+00 2.06301101e-02 -7.65073657e-01
-9.60330606e-01 -7.44965672e-01 -9.19503331e-01 -4.10004437e-01
9.58377302e-01 -9.37691033e-01 -9.20668066e-01 1.04917192e+00
-1.13115561e+00 -6.97869956e-01 -4.96194392e-01 4.76789057e-01
-6.91583931e-01 5.70159316e-01 -6.66818380e-01 -8.32060456e-01
-4.72401112e-01 -9.55147386e-01 1.57033694e+00 2.57630944e-01
6.17799722e-02 -4.04969335e-01 8.45440477e-02 4.62446392e-01
4.02245909e-01 3.68560672e-01 1.48407683e-01 -3.37427445e-02
-9.16654408e-01 -7.36054704e-02 -7.62573481e-01 5.07670939e-01
6.80986643e-02 -1.37259483e-01 -1.34323263e+00 -1.24178559e-01
-9.08643842e-01 -6.28322184e-01 1.22320831e+00 4.77016777e-01
8.57715905e-01 1.70182005e-01 -2.17158437e-01 9.55874681e-01
1.03604531e+00 -2.30442762e-01 6.62386417e-01 1.20047420e-01
8.68677914e-01 3.18823159e-01 2.08056375e-01 3.63981456e-01
8.03477347e-01 6.35320187e-01 4.35670093e-02 -2.54666716e-01
-9.00350094e-01 -1.52544528e-01 3.35507005e-01 5.27803600e-01
-5.31790674e-01 -4.49441858e-02 -9.45479631e-01 6.08728170e-01
-2.04371428e+00 -8.30378294e-01 -2.00692654e-01 1.80377817e+00
7.47103930e-01 -4.86882240e-01 2.14561790e-01 5.71833290e-02
4.17106986e-01 6.24363497e-02 -7.31825650e-01 3.18339318e-01
-4.57772255e-01 3.83337379e-01 7.83461928e-01 3.66271466e-01
-1.25798070e+00 1.21386731e+00 6.32971239e+00 3.83104920e-01
-1.31472242e+00 4.61086303e-01 1.10530891e-01 -1.72800541e-01
2.42255047e-01 -2.07899749e-01 -8.20772171e-01 3.03992182e-01
7.14586258e-01 6.84734225e-01 5.39739668e-01 5.88719308e-01
1.21667080e-01 -1.78665638e-01 -9.31020021e-01 1.12925291e+00
2.00027660e-01 -1.19133377e+00 -4.20153379e-01 1.55620696e-02
7.44668126e-01 7.27725983e-01 -3.90302867e-01 2.15220049e-01
5.78601480e-01 -8.80357563e-01 9.45015311e-01 1.00074887e+00
1.41806400e+00 -9.74431336e-02 6.25558197e-01 6.40151352e-02
-1.40156376e+00 7.19608068e-02 2.96725035e-01 2.37655789e-01
5.81172764e-01 1.09585851e-01 -3.18455696e-01 1.24180689e-01
8.07803750e-01 8.82918715e-01 -4.84924406e-01 1.21774244e+00
-5.98274052e-01 6.74051225e-01 -8.51593614e-01 1.04390442e-01
2.17488006e-01 2.81516165e-01 2.39811391e-01 1.44197476e+00
3.34221125e-02 1.38211831e-01 1.16898529e-01 6.49169683e-01
3.68728563e-02 -2.01021165e-01 -4.09037247e-02 6.98422939e-02
1.09992012e-01 8.62398565e-01 -5.65578103e-01 -4.58052695e-01
-6.73848569e-01 1.32210612e+00 1.78215876e-01 6.61012232e-01
-7.45268226e-01 -5.93910553e-02 9.10044551e-01 1.00921005e-01
3.84691030e-01 -4.86187011e-01 -3.16351831e-01 -1.44540393e+00
2.68663853e-01 -6.46372855e-01 4.22469616e-01 -8.39500785e-01
-1.04983950e+00 2.65591651e-01 -4.86258119e-01 -1.07906961e+00
-1.46901682e-01 -8.02293241e-01 -4.38052975e-02 7.59686232e-01
-2.18829751e+00 -1.86941886e+00 -9.92204845e-01 9.30223942e-01
3.68448049e-01 1.42158225e-01 7.42660880e-01 4.78313923e-01
-2.22023338e-01 8.45345736e-01 -3.09810005e-02 5.57323456e-01
7.23653793e-01 -9.45399106e-01 4.04589117e-01 8.86376321e-01
1.44142553e-01 1.10695563e-01 1.66928783e-01 -6.85061395e-01
-1.16843641e+00 -1.07420897e+00 1.01494718e+00 -6.91962063e-01
4.88580763e-01 -3.03434610e-01 -3.29146564e-01 7.02376723e-01
-3.99723113e-01 7.05561936e-01 2.44752869e-01 -2.02594042e-01
-7.03665316e-01 1.37083143e-01 -1.22058928e+00 3.23835135e-01
1.71668625e+00 -6.89281642e-01 -2.89041102e-01 3.53957474e-01
4.35807675e-01 -6.87164783e-01 -5.30591249e-01 5.70591807e-01
1.17917216e+00 -7.21611142e-01 1.01890481e+00 -5.90569735e-01
1.67879537e-01 -5.35478473e-01 -3.40602934e-01 -6.06314421e-01
5.93237812e-03 -5.46656430e-01 -4.23475236e-01 8.70702446e-01
-9.75639280e-03 -5.22404075e-01 1.04757094e+00 8.17250669e-01
-6.90371394e-02 -5.90181351e-01 -1.29924858e+00 -7.55949616e-01
-3.37415248e-01 -7.96056449e-01 3.00750941e-01 7.31076300e-01
-4.10252690e-01 -3.08998346e-01 -5.51132202e-01 1.08451232e-01
9.39802587e-01 -1.05894662e-01 9.83328938e-01 -1.19594252e+00
-7.41452947e-02 -2.37314716e-01 -6.75891697e-01 -1.31498814e+00
3.37541074e-01 -6.47389770e-01 2.94749290e-01 -1.94930494e+00
-1.26693204e-01 -4.64421928e-01 -3.50849509e-01 1.11914182e+00
1.75535440e-01 5.03890097e-01 3.40742558e-01 2.43767112e-01
-9.78306949e-01 4.98623490e-01 1.28027070e+00 -1.21112101e-01
-1.27214581e-01 -2.72734523e-01 -2.68574297e-01 7.28265107e-01
2.11916193e-01 -2.61416078e-01 2.23153666e-01 -4.07719642e-01
-4.52301115e-01 -1.82886854e-01 7.75837958e-01 -1.00105941e+00
3.57086480e-01 -1.72133632e-02 4.11866382e-02 -6.90807700e-01
4.97873098e-01 -7.50143766e-01 -1.65042162e-01 4.43249077e-01
-1.47310734e-01 -4.50458735e-01 -4.35861126e-02 5.95245838e-01
9.07807052e-03 3.44597220e-01 8.74021113e-01 -9.54295024e-02
-1.26323104e+00 2.18926892e-01 -1.59174085e-01 4.73896116e-02
7.49115944e-01 -4.13288683e-01 -5.43659568e-01 -3.62791419e-01
-6.78424180e-01 2.59700626e-01 3.84675145e-01 6.05214953e-01
5.42422593e-01 -1.14888895e+00 -6.92431450e-01 2.87703216e-01
2.73274839e-01 2.74841562e-02 1.88511848e-01 1.08752239e+00
-5.81734300e-01 5.08497775e-01 6.77204179e-03 -8.85468245e-01
-1.50773478e+00 -2.95553684e-01 5.68714023e-01 -5.00236601e-02
-1.24573541e+00 1.09670281e+00 -2.42242917e-01 -7.99200773e-01
6.78774655e-01 -7.90373206e-01 2.58500934e-01 -2.19292253e-01
6.91442490e-01 3.59765619e-01 3.33443633e-03 -8.90827298e-01
-6.15103126e-01 8.41744661e-01 3.53023112e-01 -1.95249438e-01
1.30899513e+00 3.63277178e-03 1.15158588e-01 1.88656420e-01
8.71831417e-01 -3.58655125e-01 -1.73237336e+00 -4.64302719e-01
-1.78050399e-01 -4.19966400e-01 3.11425179e-01 -1.36258912e+00
-1.00567710e+00 7.12916493e-01 1.14105833e+00 -9.20755923e-01
1.13350224e+00 2.31022358e-01 1.10160625e+00 4.63884383e-01
6.65703416e-01 -1.13523459e+00 -1.34266332e-01 6.67691290e-01
5.42874813e-01 -1.33924854e+00 -3.28452975e-01 -2.07889766e-01
-3.54627281e-01 1.00819468e+00 4.98979509e-01 6.46189675e-02
6.10917866e-01 5.08015096e-01 3.84893775e-01 -9.44140255e-02
-3.05742979e-01 -8.65491509e-01 5.56619704e-01 6.50476754e-01
1.41460285e-01 1.10310223e-02 -1.55689999e-01 2.63773024e-01
2.17738390e-01 7.48806655e-01 5.31744994e-02 1.17468202e+00
-1.76534638e-01 -9.49039519e-01 -1.88236445e-01 4.97919261e-01
6.15086146e-02 -1.55478731e-01 -4.51563179e-01 7.69309103e-01
4.40918386e-01 7.36370265e-01 -1.66371539e-01 -3.78247678e-01
2.71352500e-01 4.01147991e-01 8.43723834e-01 -2.36552820e-01
-3.49292070e-01 1.24937408e-02 2.52003133e-01 -9.73399699e-01
-8.73306870e-01 -8.59493375e-01 -1.46891844e+00 8.58396466e-04
-3.74148488e-01 -7.70092785e-01 8.69189024e-01 1.31623673e+00
4.54897255e-01 3.12177032e-01 -3.33029360e-01 -9.58396792e-01
-1.56378910e-01 -9.72714126e-01 -3.48814905e-01 4.48284328e-01
5.64606547e-01 -7.13467062e-01 -2.92851031e-01 2.85129249e-01] | [9.17138671875, -6.485856533050537] |
c46cc823-63be-4687-8025-3e3e1b6fd916 | pemp-leveraging-physics-properties-to-enhance | 2211.01978 | null | https://arxiv.org/abs/2211.01978v1 | https://arxiv.org/pdf/2211.01978v1.pdf | PEMP: Leveraging Physics Properties to Enhance Molecular Property Prediction | Molecular property prediction is essential for drug discovery. In recent years, deep learning methods have been introduced to this area and achieved state-of-the-art performances. However, most of existing methods ignore the intrinsic relations between molecular properties which can be utilized to improve the performances of corresponding prediction tasks. In this paper, we propose a new approach, namely Physics properties Enhanced Molecular Property prediction (PEMP), to utilize relations between molecular properties revealed by previous physics theory and physical chemistry studies. Specifically, we enhance the training of the chemical and physiological property predictors with related physics property prediction tasks. We design two different methods for PEMP, respectively based on multi-task learning and transfer learning. Both methods include a model-agnostic molecule representation module and a property prediction module. In our implementation, we adopt both the state-of-the-art molecule embedding models under the supervised learning paradigm and the pretraining paradigm as the molecule representation module of PEMP, respectively. Experimental results on public benchmark MoleculeNet show that the proposed methods have the ability to outperform corresponding state-of-the-art models. | ['Yanyan Lan', 'Wei-Ying Ma', 'ZhiMing Ma', 'Kang Liu', 'Wenhao Huang', 'Weizhi Ma', 'Yimeng Chen', 'Yuancheng Sun'] | 2022-10-18 | null | null | null | null | ['molecular-property-prediction'] | ['miscellaneous'] | [ 2.97071844e-01 -2.26524502e-01 -5.24546504e-01 -3.18812728e-01
-5.60506344e-01 -2.24759370e-01 5.60665667e-01 5.96337557e-01
-3.68087143e-01 1.18956995e+00 5.78257516e-02 -3.31780106e-01
-2.34446555e-01 -1.08120251e+00 -9.02292073e-01 -9.91947234e-01
7.00395703e-02 1.39605612e-01 4.74524081e-01 -1.18207641e-01
3.42395306e-01 4.37769771e-01 -1.21170437e+00 3.49824309e-01
1.22004533e+00 9.77845728e-01 9.65846777e-02 2.21785083e-01
-2.82656640e-01 8.60004961e-01 4.55768593e-02 -2.80326039e-01
-9.23424885e-02 -2.90959388e-01 -6.69613421e-01 -7.26693451e-01
1.16323143e-01 1.88338514e-02 -5.15719056e-01 8.86718929e-01
7.70481050e-01 9.17362720e-02 7.96186030e-01 -9.65554357e-01
-1.03727245e+00 3.84611905e-01 -4.54428762e-01 5.86977564e-02
5.56084454e-01 2.02014267e-01 9.16460037e-01 -7.58100927e-01
3.22482765e-01 1.18561149e+00 7.04672337e-01 4.78100091e-01
-1.06553495e+00 -1.04563916e+00 1.40612587e-01 6.26949966e-01
-1.20233953e+00 -2.76212394e-01 1.01283312e+00 -4.71823841e-01
1.14586532e+00 -2.19505578e-01 4.35523331e-01 1.04164040e+00
5.34207821e-01 6.16537154e-01 1.13783908e+00 -1.08440556e-01
1.25871435e-01 4.14343625e-01 2.06147686e-01 9.05877173e-01
2.06828490e-01 3.16136360e-01 -5.32198727e-01 -4.21711802e-01
4.51622069e-01 1.62429959e-01 -4.99084741e-01 -4.77837682e-01
-1.12993145e+00 8.19455266e-01 7.03943968e-01 1.83825687e-01
-5.57529509e-01 4.26959768e-02 4.71485764e-01 1.09094687e-01
6.49251640e-01 4.47075635e-01 -8.89442921e-01 2.14425296e-01
-5.48751414e-01 1.74034879e-01 8.06214571e-01 6.47136152e-01
8.49296749e-01 -2.45418340e-01 -1.47055820e-01 5.87822258e-01
4.26492661e-01 1.96940407e-01 6.41352534e-01 1.97204068e-01
3.48181844e-01 7.59787619e-01 -7.85936490e-02 -8.39175403e-01
-4.73748356e-01 -3.18496078e-01 -7.32924223e-01 -8.59377086e-02
-1.49655014e-01 3.14483643e-02 -9.12128747e-01 1.65118051e+00
5.31293929e-01 6.82108760e-01 4.01632518e-01 4.39402610e-01
1.31300211e+00 1.06671894e+00 8.14300418e-01 -1.85298949e-01
1.23489392e+00 -1.14333189e+00 -6.68431759e-01 4.13112611e-01
7.97295511e-01 -6.62236452e-01 8.44907999e-01 1.63913205e-01
-6.84954643e-01 -7.70681083e-01 -1.37557900e+00 7.50221759e-02
-9.02695596e-01 -1.49798458e-02 1.13452744e+00 7.13672519e-01
-3.64468068e-01 1.22976470e+00 -8.05188000e-01 -7.29091540e-02
7.17940748e-01 7.17510164e-01 -6.47427559e-01 1.04604838e-02
-1.58991039e+00 1.03793323e+00 9.21750426e-01 -1.03508197e-01
-9.02487516e-01 -1.23708892e+00 -7.67556131e-01 1.78948343e-01
2.82488286e-01 -7.40873992e-01 7.33079672e-01 -6.43594980e-01
-1.90079641e+00 3.55807990e-01 1.39998302e-01 -8.73269960e-02
1.41128097e-02 -1.30598843e-01 -6.28571391e-01 9.86991823e-02
-2.54318327e-01 3.20454478e-01 3.99814993e-01 -1.08275747e+00
-3.74876201e-01 -2.02680513e-01 -2.82654236e-03 1.56954244e-01
-6.35689259e-01 -2.71310598e-01 6.44021435e-03 -2.91284144e-01
-3.70569944e-01 -6.23702288e-01 -2.66173810e-01 -2.67001037e-02
-4.01640892e-01 -5.22930980e-01 6.67719364e-01 -3.63065302e-01
1.09063101e+00 -1.83359945e+00 2.67946601e-01 1.65797979e-01
2.57485837e-01 5.93990564e-01 -4.00437832e-01 6.83617949e-01
-3.62734050e-01 9.33385417e-02 -1.49501488e-01 -2.40640584e-02
-1.42012373e-01 -3.06381714e-02 -2.06565648e-01 4.95476633e-01
3.54304135e-01 8.96288574e-01 -1.03617668e+00 -5.36255181e-01
3.63951653e-01 6.87207580e-01 -5.16586721e-01 3.65166634e-01
-4.53533083e-01 6.71357512e-01 -9.99467552e-01 7.31732845e-01
9.42209125e-01 -3.72403800e-01 1.98593482e-01 -5.37308455e-01
-1.95779309e-01 4.32773590e-01 -6.37658477e-01 1.88167620e+00
-2.09889039e-01 -2.93403357e-01 -6.55693650e-01 -1.15146947e+00
8.43816757e-01 5.00800490e-01 6.32705390e-01 -4.59370941e-01
5.27869090e-02 4.55186874e-01 4.12440598e-01 -5.77623963e-01
1.15452394e-01 -4.90029812e-01 6.20587468e-01 -4.05765958e-02
3.63710076e-01 2.40513667e-01 -1.22599974e-01 -1.97881743e-01
9.31268096e-01 5.68296015e-01 5.12842119e-01 -2.52072334e-01
1.14978862e+00 -2.32724547e-01 5.95654368e-01 2.25983322e-01
-4.77601960e-02 -1.06529230e-02 3.28028738e-01 -4.45777833e-01
-8.57676864e-01 -7.50900269e-01 -5.31663358e-01 9.99743283e-01
2.94541925e-01 -5.80151200e-01 -4.78718251e-01 -7.91935265e-01
2.04464987e-01 1.48646668e-01 -7.90376961e-01 -5.66475928e-01
-4.52593952e-01 -1.13155043e+00 3.66495401e-01 4.64809716e-01
7.10307419e-01 -1.27570927e+00 2.05217376e-01 4.78247881e-01
4.71838146e-01 -1.00076413e+00 6.47906810e-02 2.82347590e-01
-8.87639225e-01 -1.06533778e+00 -7.17918873e-01 -6.80868447e-01
2.44133756e-01 5.05502969e-02 6.45289421e-01 -4.96790046e-03
-6.10868521e-02 -1.25743821e-01 -5.33789337e-01 -5.84375441e-01
-2.96639830e-01 3.82723153e-01 1.59014966e-02 2.05626681e-01
4.08490837e-01 -1.04945874e+00 -9.29476440e-01 -4.37280163e-02
-7.70188153e-01 6.60426691e-02 1.20072544e+00 7.61622846e-01
8.27702641e-01 -8.27278420e-02 9.40822184e-01 -1.19861734e+00
5.16774654e-01 -7.41160989e-01 -2.39562094e-01 4.20327723e-01
-7.86238134e-01 3.50341797e-01 1.00456679e+00 -4.95625883e-01
-9.73738313e-01 6.35737404e-02 -4.17687565e-01 -2.79470652e-01
-1.19311072e-01 9.92570758e-01 -5.15977800e-01 -6.17725551e-01
3.63085896e-01 5.30428112e-01 -3.35399628e-01 -6.88817322e-01
3.45994741e-01 4.80311871e-01 6.83078766e-02 -7.27716386e-01
7.27043271e-01 1.65869430e-01 4.41514164e-01 -5.18575490e-01
-8.18918467e-01 -5.12947798e-01 -6.98949993e-01 4.15854514e-01
9.26823497e-01 -8.89789760e-01 -9.99319375e-01 3.36027950e-01
-1.17083371e+00 7.77890980e-02 2.07731664e-01 7.61828244e-01
-4.85126466e-01 6.62596643e-01 -3.76954496e-01 -4.18221444e-01
-7.50097036e-01 -1.16161335e+00 8.17830086e-01 3.57217759e-01
4.35020864e-01 -1.18454313e+00 5.29386520e-01 2.39651445e-02
3.31853926e-01 4.44714308e-01 1.51725352e+00 -9.82354581e-01
-5.27078450e-01 -1.47073522e-01 -4.58084583e-01 1.86214253e-01
4.24747169e-01 -1.70858264e-01 -1.18145204e+00 -2.59810656e-01
-3.10113370e-01 -3.84033263e-01 1.22263992e+00 2.06509158e-01
1.44602370e+00 -5.32606244e-02 -7.95892537e-01 7.56350458e-01
1.61072862e+00 4.38104987e-01 8.63480210e-01 2.33317912e-01
9.93799984e-01 3.66505712e-01 3.85266542e-01 2.59416044e-01
1.82174712e-01 4.95998025e-01 2.39924699e-01 -1.27447382e-01
5.46909980e-02 -6.26318693e-01 1.66029125e-01 8.18653166e-01
-4.38163906e-01 -1.88435122e-01 -7.02323735e-01 4.65516523e-02
-2.06805134e+00 -8.00156891e-01 5.47989234e-02 2.10202813e+00
1.20028412e+00 8.88682604e-02 -9.73395258e-02 -1.06102869e-01
3.81654143e-01 2.52492398e-01 -7.02264071e-01 -1.55210301e-01
1.44987658e-01 6.73004270e-01 5.29768288e-01 1.51677206e-01
-1.32724285e+00 1.12940955e+00 5.50579882e+00 1.17415762e+00
-1.13820362e+00 1.52947292e-01 4.08370435e-01 4.88075733e-01
-2.38741919e-01 7.24374130e-02 -8.22327077e-01 4.94931877e-01
1.10688770e+00 4.45016809e-02 -6.94296807e-02 8.70154738e-01
-3.88009707e-03 2.52246141e-01 -1.43971217e+00 1.04570889e+00
-9.60099921e-02 -1.60257947e+00 5.52333057e-01 -2.28781831e-02
5.35717845e-01 -1.71773732e-01 8.10512304e-02 5.81534266e-01
-4.51656401e-01 -1.23277545e+00 -1.31954523e-02 7.23925591e-01
7.11210668e-01 -7.76832700e-01 8.50366950e-01 1.71182930e-01
-1.46630347e+00 1.78188235e-01 -8.03669572e-01 3.66427451e-02
-8.12020972e-02 3.16003650e-01 -3.77291560e-01 1.27660561e+00
1.90515637e-01 1.34502363e+00 -5.17074645e-01 1.21698463e+00
-1.78750739e-01 4.66946721e-01 -1.75901055e-02 -3.78466576e-01
3.18675727e-01 -4.03345853e-01 1.57393739e-02 1.17837334e+00
-7.81576857e-02 2.81854093e-01 4.25942391e-01 8.53451550e-01
-3.12057644e-01 6.92269802e-01 -4.99191910e-01 -4.21487421e-01
2.54491508e-01 1.31631827e+00 -2.95146346e-01 -4.67248201e-01
-7.12195456e-01 8.19958270e-01 4.02796060e-01 2.13403955e-01
-1.04213130e+00 -7.17510164e-01 6.60560012e-01 -2.12435752e-01
2.59291559e-01 5.29085249e-02 8.30805004e-02 -1.20032310e+00
-2.32005447e-01 -6.60799921e-01 4.78785932e-02 -2.95530230e-01
-1.60452354e+00 5.04661202e-01 -1.03565224e-01 -1.18034363e+00
6.87336326e-01 -1.12522018e+00 -8.16958606e-01 1.09326684e+00
-2.44975877e+00 -1.47953534e+00 -8.41438994e-02 6.01109147e-01
1.45494580e-01 -3.19923759e-01 1.15402281e+00 6.16319418e-01
-6.61096931e-01 5.09303093e-01 2.31413454e-01 -2.43053332e-01
8.42580020e-01 -1.02624416e+00 -5.61446100e-02 1.05927442e-03
-2.37575069e-01 8.09652388e-01 2.59779513e-01 -5.21775007e-01
-1.66525817e+00 -1.16599798e+00 4.89853710e-01 -2.40249440e-01
7.76565909e-01 -1.52380899e-01 -1.11236870e+00 2.49328196e-01
8.97512361e-02 1.45654649e-01 1.41106653e+00 2.49484435e-01
-5.14964938e-01 -8.39645565e-02 -9.67453480e-01 2.15162843e-01
8.27122033e-01 -6.29627764e-01 -6.07361555e-01 6.71612918e-01
8.99866283e-01 -1.18696816e-01 -1.21388686e+00 8.24872017e-01
7.62376964e-01 -4.17408168e-01 1.25124407e+00 -1.14874804e+00
6.29514635e-01 -2.56502509e-01 -3.36598158e-02 -1.11901021e+00
-5.42601585e-01 -3.00524384e-01 -2.69430876e-01 9.90547001e-01
5.95837474e-01 -7.25369036e-01 7.98684180e-01 3.10171723e-01
-1.59438834e-01 -1.30617082e+00 -7.70059168e-01 -7.59250343e-01
4.49506015e-01 1.53151929e-01 5.51554918e-01 1.15853035e+00
1.17591821e-01 9.12844181e-01 -4.47537988e-01 1.13974147e-01
3.10122371e-01 3.22349906e-01 5.70131242e-01 -1.43140256e+00
-5.32368183e-01 -4.73311990e-01 -5.96112072e-01 -1.03854334e+00
4.41652387e-01 -1.14615595e+00 -4.98700470e-01 -1.58556855e+00
7.15814829e-01 -4.35979009e-01 -1.06278121e+00 5.20940483e-01
-5.04169703e-01 -2.12186903e-01 -5.22915870e-02 1.97363958e-01
-4.78040367e-01 1.15907097e+00 1.46164548e+00 -3.25643629e-01
-3.65900129e-01 -1.05970770e-01 -5.64531565e-01 5.47324300e-01
8.30266595e-01 -4.63602096e-01 -5.14122546e-01 1.51195392e-01
3.55782211e-02 -3.05510998e-01 2.14399695e-01 -1.04698551e+00
9.22705308e-02 -2.70824194e-01 5.69118440e-01 -5.48369467e-01
4.94862527e-01 -7.62078106e-01 -8.11739564e-02 6.47823513e-01
-1.45088091e-01 -4.81691092e-01 4.03137743e-01 9.88940477e-01
-1.70729846e-01 -1.27455324e-01 8.80681098e-01 -4.12228592e-02
-6.19716048e-01 9.43891227e-01 6.55028298e-02 -6.04302049e-01
1.03307569e+00 -7.02153891e-02 -4.11611885e-01 3.64388824e-01
-4.27119344e-01 1.48037210e-01 -4.58285026e-02 1.02786466e-01
8.58979642e-01 -1.31074011e+00 -5.46626091e-01 -1.22350752e-01
4.35061812e-01 -5.51307738e-01 2.10279018e-01 8.78231108e-01
-4.84528750e-01 6.29098177e-01 -3.60236764e-01 -1.25893787e-01
-1.04069149e+00 1.12248194e+00 3.54529679e-01 -4.73106235e-01
-3.79195780e-01 7.06020653e-01 4.86811727e-01 -6.16429985e-01
1.36604048e-02 -3.37433904e-01 -6.48370504e-01 -1.39381230e-01
4.45346534e-01 7.59373158e-02 -6.05626218e-02 -3.02431464e-01
-5.32074094e-01 9.11190391e-01 -5.22772312e-01 5.94275236e-01
1.48076367e+00 6.42890573e-01 -1.88942224e-01 2.91650981e-01
1.46913934e+00 -2.15878680e-01 -8.13254118e-01 -3.80608886e-01
-1.19431198e-01 -2.05008939e-01 1.56627432e-01 -9.77569699e-01
-7.45919764e-01 1.14169109e+00 6.67964876e-01 -3.09181839e-01
8.75396907e-01 -3.19443107e-01 1.05996633e+00 7.38287151e-01
2.34919950e-01 -6.84377372e-01 -9.06254724e-03 2.61083990e-01
5.60312986e-01 -1.42610109e+00 4.51755702e-01 -7.77918637e-01
-2.24800929e-01 1.36003208e+00 6.86464906e-01 -1.16335444e-01
1.11991155e+00 -4.02469009e-01 -4.84242886e-01 -2.68788457e-01
-6.03546977e-01 -1.53634727e-01 4.30915028e-01 4.86784518e-01
8.88644397e-01 -9.15506482e-02 -6.39067411e-01 7.59563029e-01
3.37993234e-01 1.89427540e-01 -1.61395356e-01 9.03745592e-01
-6.25901878e-01 -1.62233973e+00 2.61669248e-01 2.94525743e-01
-3.72798294e-01 -5.56837738e-01 -3.88924271e-01 6.59162462e-01
3.25488538e-01 6.32820964e-01 -5.76061308e-01 -6.00534678e-01
2.59110212e-01 2.27592990e-01 7.70710409e-01 -6.87676728e-01
-5.68227470e-01 -2.47271806e-01 -8.64860341e-02 -4.03747410e-01
-8.95335436e-01 2.17861049e-02 -1.30675423e+00 -2.95698225e-01
-5.74273765e-01 4.49681729e-01 4.25018728e-01 8.80161107e-01
5.24967372e-01 7.46430397e-01 5.80640852e-01 -7.88917184e-01
-5.00463784e-01 -1.08163917e+00 -5.87010920e-01 2.16943696e-01
1.84642762e-01 -8.86371315e-01 1.40371069e-01 -1.04357988e-01] | [5.111632823944092, 5.909398555755615] |
2c5bf9ed-b4b0-4277-9f52-4daab38060a8 | mesh-r-cnn | 1906.02739 | null | https://arxiv.org/abs/1906.02739v2 | https://arxiv.org/pdf/1906.02739v2.pdf | Mesh R-CNN | Rapid advances in 2D perception have led to systems that accurately detect objects in real-world images. However, these systems make predictions in 2D, ignoring the 3D structure of the world. Concurrently, advances in 3D shape prediction have mostly focused on synthetic benchmarks and isolated objects. We unify advances in these two areas. We propose a system that detects objects in real-world images and produces a triangle mesh giving the full 3D shape of each detected object. Our system, called Mesh R-CNN, augments Mask R-CNN with a mesh prediction branch that outputs meshes with varying topological structure by first predicting coarse voxel representations which are converted to meshes and refined with a graph convolution network operating over the mesh's vertices and edges. We validate our mesh prediction branch on ShapeNet, where we outperform prior work on single-image shape prediction. We then deploy our full Mesh R-CNN system on Pix3D, where we jointly detect objects and predict their 3D shapes. | ['Georgia Gkioxari', 'Justin Johnson', 'Jitendra Malik'] | 2019-06-06 | mesh-r-cnn-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Gkioxari_Mesh_R-CNN_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Gkioxari_Mesh_R-CNN_ICCV_2019_paper.pdf | iccv-2019-10 | ['3d-shape-modeling'] | ['computer-vision'] | [ 3.49564314e-01 5.34061015e-01 1.84940830e-01 -2.82409102e-01
-3.61639261e-01 -5.19473970e-01 6.61240518e-01 2.47200295e-01
2.34561771e-01 -4.71028462e-02 -4.26099170e-03 -2.47237667e-01
4.29564357e-01 -1.20508885e+00 -1.02213240e+00 1.64596945e-01
-3.28332067e-01 1.12970769e+00 8.83557141e-01 -1.76741719e-01
2.08570898e-01 1.22358477e+00 -1.78375554e+00 6.32618666e-01
8.69707689e-02 1.22555184e+00 1.22897185e-01 8.44506979e-01
-2.14751527e-01 3.86717618e-01 -1.34072393e-01 -1.66367188e-01
5.95330417e-01 3.07094485e-01 -7.99245358e-01 3.62744749e-01
9.81977701e-01 -3.36756349e-01 -5.55609345e-01 7.30855584e-01
3.08966458e-01 -3.87533784e-01 6.97030485e-01 -1.04707408e+00
-7.92418063e-01 4.26715970e-01 -7.72599638e-01 -1.37568161e-01
2.86568940e-01 2.87939191e-01 1.02399206e+00 -1.39075911e+00
9.24340189e-01 1.66944516e+00 1.09445536e+00 3.16923946e-01
-1.65865874e+00 -4.64874506e-01 9.47074145e-02 -3.73282850e-01
-1.53365290e+00 -2.21948400e-01 8.11183214e-01 -6.71594083e-01
1.49091220e+00 2.71736503e-01 8.00081193e-01 6.52198315e-01
2.50121415e-01 7.20144987e-01 7.04059243e-01 3.02025657e-02
-5.14027514e-02 -3.92248929e-01 -4.25193489e-01 1.08037639e+00
8.43532830e-02 1.38073117e-01 -2.17731763e-02 5.92991784e-02
1.44757533e+00 -1.05994850e-01 1.05356827e-01 -6.14022076e-01
-1.40084529e+00 2.85082161e-01 8.00752461e-01 -2.14194208e-01
-3.80777657e-01 5.28726816e-01 1.70450628e-01 2.15000749e-01
8.33097994e-01 3.98864657e-01 -6.33686304e-01 4.37397242e-01
-9.10917580e-01 5.21463513e-01 6.35752201e-01 1.16204488e+00
7.58327484e-01 -8.45695101e-03 -6.49081916e-02 6.42412245e-01
2.84961283e-01 5.52948534e-01 -3.34159672e-01 -1.04573941e+00
1.24908708e-01 1.18557918e+00 -6.76411949e-03 -1.32537353e+00
-6.76535785e-01 -4.73627210e-01 -7.00742006e-01 6.35820985e-01
1.21077292e-01 3.88074636e-01 -1.09268725e+00 1.07385147e+00
5.36801457e-01 4.78305221e-01 -5.96124113e-01 8.82234275e-01
1.27952409e+00 4.41380322e-01 -1.79984391e-01 6.54726505e-01
1.13499498e+00 -8.96761239e-01 3.71035308e-01 -7.72822648e-02
3.99185717e-01 -8.14868152e-01 9.69073415e-01 2.39488855e-01
-1.41604877e+00 -8.69991124e-01 -8.22835565e-01 -2.70803690e-01
-2.54294038e-01 7.87427798e-02 4.11232442e-01 1.29435450e-01
-1.34237075e+00 8.22406471e-01 -8.96982968e-01 -2.91493535e-01
8.92911077e-01 5.72378516e-01 -2.28208050e-01 1.77718356e-01
-3.87228668e-01 7.66837358e-01 2.83839405e-01 -2.25806758e-01
-9.66208339e-01 -1.04354870e+00 -8.65432441e-01 4.33114916e-02
1.44393876e-01 -8.95304084e-01 1.30964065e+00 -6.65886402e-01
-1.06368923e+00 1.49033022e+00 6.98033348e-02 -3.19242924e-01
4.82347578e-01 2.42988184e-01 -1.16683170e-01 2.61561642e-03
1.75638217e-02 1.23356426e+00 6.97459877e-01 -1.69173825e+00
-6.43209994e-01 -2.48686999e-01 1.30011775e-02 3.34657393e-02
5.43312430e-01 -2.77407795e-01 -4.76090252e-01 -6.12911999e-01
5.54885924e-01 -7.37721384e-01 -5.32581925e-01 6.00579143e-01
-6.53500080e-01 -3.41882020e-01 8.93210292e-01 -1.55908585e-01
4.94434237e-01 -1.91770816e+00 -8.56869593e-02 4.64375317e-01
6.55696273e-01 -1.34034380e-01 -5.53272188e-01 1.56245455e-01
-1.31714165e-01 1.69878870e-01 -9.70416367e-02 -7.63375163e-01
9.01464075e-02 1.85006633e-01 -4.90246326e-01 3.61943364e-01
7.53441632e-01 1.23883021e+00 -9.29468870e-01 -4.94982690e-01
4.41849053e-01 5.79449594e-01 -9.05570805e-01 1.29884884e-01
-8.29478025e-01 3.32963705e-01 -3.20482284e-01 9.11715925e-01
8.18585634e-01 -7.18873143e-01 -1.06395073e-01 -4.29125398e-01
-1.97158426e-01 2.09953964e-01 -1.05326819e+00 1.55378723e+00
-3.18767101e-01 5.43378890e-01 2.04959825e-01 -6.55402839e-01
1.14066541e+00 -5.69382273e-02 7.92257249e-01 -5.19100964e-01
1.16839454e-01 2.22373217e-01 -2.32911974e-01 -1.93102136e-01
5.40149152e-01 1.21737286e-01 3.10431957e-01 3.81104112e-01
-1.22334763e-01 -7.76072860e-01 -3.34413677e-01 -7.68167991e-03
1.30493784e+00 4.49931264e-01 -1.09274559e-01 -2.62766182e-01
1.80954691e-02 3.87932479e-01 1.99152008e-02 7.20254838e-01
2.19080195e-01 1.04875588e+00 2.92495817e-01 -1.02794230e+00
-1.52215457e+00 -1.44779384e+00 -2.63237089e-01 8.08404386e-01
5.03899395e-01 -3.75527531e-01 -2.38238037e-01 -6.07772887e-01
4.87383336e-01 2.65522838e-01 -7.18113840e-01 3.12714964e-01
-8.71118784e-01 -9.43974853e-02 3.26069027e-01 5.41843891e-01
9.77469236e-03 -1.33529544e+00 -7.62053013e-01 2.59032607e-01
3.85309756e-01 -1.16100514e+00 -3.91820759e-01 2.68207081e-02
-9.03983295e-01 -1.27555370e+00 -3.30306470e-01 -1.03857398e+00
8.52379322e-01 2.87905604e-01 1.87275207e+00 5.94096780e-01
-6.24554992e-01 4.92254704e-01 -5.57381064e-02 -4.31083530e-01
-4.83910978e-01 9.31610093e-02 -1.08241290e-01 -2.61127323e-01
-1.21384859e-02 -8.06620479e-01 -7.55366683e-01 2.74690241e-01
-5.19844115e-01 5.43540537e-01 5.62204063e-01 3.96701813e-01
1.27291572e+00 -1.28909722e-01 2.33541220e-01 -9.65155959e-01
9.20044482e-02 -5.18602550e-01 -6.28077745e-01 -1.24911917e-02
-2.11287886e-01 -9.23042521e-02 3.39901358e-01 -4.05554384e-01
-5.33044338e-01 5.27872562e-01 -3.52775544e-01 -8.35596621e-01
-4.66103286e-01 9.83340293e-02 2.40678191e-01 -3.11624825e-01
7.08009243e-01 7.16286674e-02 -8.55607241e-02 -5.57528675e-01
3.88404459e-01 2.54500210e-01 6.51310742e-01 -6.81180596e-01
9.24950242e-01 7.01931000e-01 3.40707541e-01 -7.88347244e-01
-7.54566014e-01 -1.08675241e-01 -8.47381294e-01 -2.97373861e-01
6.77761555e-01 -9.35029149e-01 -9.12954688e-01 1.88500553e-01
-1.45188379e+00 -6.32322967e-01 -6.05075240e-01 -1.06133118e-01
-8.55878115e-01 1.21020162e-02 -6.62327528e-01 -4.85035300e-01
-2.70575911e-01 -1.10609329e+00 1.82749116e+00 -2.63382733e-01
-2.44374573e-01 -6.18266881e-01 -1.32706929e-02 -1.37665033e-01
3.05704951e-01 5.39821148e-01 1.04571736e+00 -2.42578462e-01
-1.10079515e+00 -2.74468288e-02 -7.45967925e-01 -2.39360273e-01
-9.18705016e-03 2.52936453e-01 -8.41647863e-01 -3.37317511e-02
-6.89737856e-01 -3.01496953e-01 7.53134906e-01 5.20057499e-01
1.70312870e+00 4.41141706e-03 -6.56183004e-01 7.84794152e-01
1.31973660e+00 -3.72916043e-01 2.75355369e-01 -1.93893164e-01
9.26594257e-01 4.47654635e-01 1.20547675e-01 2.44513497e-01
4.80108559e-01 5.90114832e-01 8.46279442e-01 -6.25270784e-01
-7.32003808e-01 -4.86564934e-01 -2.94834942e-01 5.65682352e-01
-1.79415017e-01 -2.23650336e-02 -1.12988043e+00 5.32876909e-01
-1.48748934e+00 -5.72411954e-01 -3.76782447e-01 1.81614673e+00
6.50552034e-01 5.68326712e-01 1.78514630e-01 -3.86680454e-01
6.27028525e-01 -2.73899920e-03 -7.89954424e-01 -2.45756000e-01
-1.25715956e-01 6.33329272e-01 4.92734939e-01 4.28946376e-01
-1.18467891e+00 1.08770275e+00 7.12175560e+00 4.76824939e-01
-1.17073011e+00 -3.35482776e-01 8.02547932e-01 7.23953918e-02
-5.53124070e-01 -2.41148680e-01 -4.64440703e-01 -1.55543029e-01
2.77464867e-01 2.13340104e-01 3.20392609e-01 9.35371101e-01
-3.28471735e-02 2.19223380e-01 -1.46938360e+00 1.09149921e+00
-1.69947743e-01 -1.78654790e+00 2.64558464e-01 1.79855138e-01
8.96971583e-01 6.42730951e-01 -5.67465499e-02 6.53030500e-02
6.79700017e-01 -1.42034483e+00 1.05804420e+00 7.17116952e-01
1.27419460e+00 -4.88848984e-01 2.33568549e-01 3.66382688e-01
-1.60110700e+00 3.86898041e-01 -6.39247656e-01 -1.53414503e-01
-2.93500759e-02 4.98761266e-01 -1.28945339e+00 8.59186947e-02
7.17060566e-01 9.18195128e-01 -6.50451362e-01 1.14799643e+00
1.76960930e-01 3.58024687e-01 -4.91978377e-01 1.72153547e-01
-8.27018619e-02 2.06790745e-01 5.81854999e-01 1.25245106e+00
2.68407732e-01 2.55002856e-01 6.02915823e-01 1.41481817e+00
-3.76949012e-01 -1.57960597e-02 -8.41450632e-01 2.52022654e-01
6.17811978e-01 1.11947632e+00 -1.30569088e+00 -4.32406455e-01
-3.98755759e-01 5.91049910e-01 5.31457186e-01 -6.07459359e-02
-6.43146813e-01 -1.09126803e-03 7.80234516e-01 6.67470872e-01
4.99635398e-01 -4.49862063e-01 -7.27909088e-01 -8.24017406e-01
-2.28486747e-01 -4.43150699e-01 -2.03062311e-01 -9.95050728e-01
-1.53326523e+00 6.27843916e-01 -3.58282864e-01 -1.17977178e+00
5.11992455e-01 -7.21895218e-01 -4.12286758e-01 6.96750045e-01
-1.25777781e+00 -1.50317049e+00 -3.24654698e-01 1.68019041e-01
6.99595034e-01 1.48466527e-01 8.53304923e-01 5.10147922e-02
-3.34838070e-02 1.91958547e-01 -5.29316187e-01 2.62235254e-01
1.67136267e-01 -1.29965901e+00 1.39049137e+00 6.40394762e-02
3.53252977e-01 1.33895725e-01 4.04095620e-01 -1.04749584e+00
-1.41554689e+00 -1.57852602e+00 5.25041699e-01 -9.19465482e-01
4.99817282e-01 -6.46046400e-01 -7.76489794e-01 8.73342037e-01
-2.97467500e-01 5.57901859e-01 -1.29254296e-01 -1.73433833e-02
-5.36283791e-01 3.31322849e-01 -1.02738297e+00 6.51178420e-01
1.75330412e+00 -3.82457316e-01 -4.12364185e-01 3.10213447e-01
9.41702485e-01 -8.52246404e-01 -1.09591341e+00 8.00734341e-01
6.58571005e-01 -8.63410056e-01 1.48630035e+00 -7.07968295e-01
7.73106098e-01 -3.75905067e-01 -7.76025057e-02 -1.11644900e+00
-6.17923677e-01 -1.80985930e-03 -1.62725851e-01 2.98873365e-01
3.39580774e-01 -5.98248579e-02 1.25607443e+00 2.89587051e-01
-5.71982682e-01 -1.15839338e+00 -7.22719550e-01 -4.23158407e-01
7.18970224e-02 -7.07959592e-01 9.93908226e-01 7.46694863e-01
-6.03434980e-01 1.18966654e-01 1.30623743e-01 3.25275123e-01
5.94777822e-01 7.08761930e-01 9.34469223e-01 -1.71490514e+00
-9.51828361e-02 -7.32330382e-01 -6.36278987e-01 -1.47209454e+00
4.42251004e-02 -1.22974777e+00 1.49033129e-01 -1.61949956e+00
-1.93894953e-02 -8.56692076e-01 7.80448243e-02 6.08723044e-01
2.72842139e-01 9.01940227e-01 6.98112175e-02 1.65100709e-01
-6.06776178e-01 3.84836107e-01 1.73914063e+00 -3.99662137e-01
-1.67595446e-01 -1.88056618e-01 -1.16181284e-01 9.75991488e-01
4.99850243e-01 -2.93651074e-01 -1.33568481e-01 -7.74777651e-01
4.98347580e-01 1.88707504e-02 7.30566740e-01 -1.05754471e+00
5.77280559e-02 -2.25857392e-01 8.96447122e-01 -1.21770334e+00
4.80695963e-01 -8.98626566e-01 2.38858789e-01 3.62365484e-01
-3.04932714e-01 8.94837603e-02 2.53856897e-01 3.77467602e-01
3.08266491e-01 4.64702457e-01 7.34415233e-01 -5.46432018e-01
-4.66861337e-01 7.87825406e-01 4.38624993e-02 -1.14356279e-01
8.83877397e-01 -5.93262613e-01 -7.58730248e-02 -5.98438680e-02
-1.10926700e+00 2.29612634e-01 8.90337765e-01 4.36992824e-01
1.17233384e+00 -1.44092929e+00 -8.77263963e-01 4.43984628e-01
3.30729373e-02 5.46407163e-01 4.77422811e-02 4.26678538e-01
-8.47681284e-01 1.04329593e-01 -1.72207713e-01 -9.91143405e-01
-1.06149912e+00 5.28471351e-01 5.54698825e-01 1.92762226e-01
-1.07568705e+00 1.01496267e+00 4.09953654e-01 -8.41486692e-01
2.75170922e-01 -7.95341492e-01 1.43128589e-01 -5.29616117e-01
1.34126589e-01 6.67192712e-02 1.57992259e-01 -7.81607509e-01
-2.89220423e-01 7.81943381e-01 3.40631396e-01 4.31182325e-01
1.61404622e+00 1.26584172e-01 -3.06391358e-01 3.08579266e-01
1.04311347e+00 5.01733012e-02 -1.42202938e+00 -3.35681200e-01
-3.40174250e-02 -4.73414958e-01 -7.07787201e-02 -5.71984649e-01
-1.17205775e+00 7.74437010e-01 4.13180351e-01 4.72589374e-01
6.56593323e-01 5.41102409e-01 9.52694416e-01 3.37489724e-01
6.07448220e-01 -5.80220580e-01 2.22961083e-01 7.61678338e-01
1.17128491e+00 -1.11039853e+00 1.88363381e-02 -8.17331672e-01
7.48866498e-02 1.34060228e+00 8.40322852e-01 -7.97082961e-01
1.03405309e+00 5.35367727e-01 -3.24084520e-01 -8.70241880e-01
-8.20924819e-01 -2.60160744e-01 7.02165961e-01 5.55213332e-01
3.85703236e-01 3.98671061e-01 5.45688629e-01 2.59431005e-01
-4.39933538e-01 -1.26033589e-01 3.03128690e-01 4.88654286e-01
-5.77174008e-01 -8.18833470e-01 -4.04955477e-01 8.62179995e-01
5.33577390e-02 -2.62277114e-04 -6.17963850e-01 6.13816559e-01
5.15763521e-01 3.35749865e-01 6.69837236e-01 -6.62604988e-01
6.02314353e-01 -4.28948045e-01 7.00355768e-01 -1.17162609e+00
-4.60179985e-01 -7.47850910e-02 -8.83629248e-02 -8.62770438e-01
-1.32395372e-01 -4.21094984e-01 -1.43597043e+00 -2.96653867e-01
1.39791220e-01 -5.21980584e-01 4.87950832e-01 5.70575237e-01
6.83518529e-01 4.77296740e-01 5.27011812e-01 -1.64058352e+00
-1.58739418e-01 -5.52751303e-01 -3.73232275e-01 3.26012254e-01
3.24776679e-01 -6.44931793e-01 2.16041282e-01 1.15888827e-01] | [8.326817512512207, -3.7125790119171143] |
6c4ddc82-b1b4-4cb1-b5cb-daf3213a531d | learning-semantic-representations-for | null | null | https://icml.cc/Conferences/2018/Schedule?showEvent=1961 | http://proceedings.mlr.press/v80/xie18c/xie18c.pdf | Learning Semantic Representations for Unsupervised Domain Adaptation |
It is important to transfer the knowledge from label-rich source domain to unlabeled target domain due to the expensive cost of manual labeling efforts. Prior domain adaptation methods address this problem through aligning the global distribution statistics between source domain and target domain, but a drawback of prior methods is that they ignore the semantic information contained in samples, e.g., features of backpacks in target domain might be mapped near features of cars in source domain. In this paper, we present moving semantic transfer network, which learn semantic representations for unlabeled target samples by aligning labeled source centroid and pseudo-labeled target centroid. Features in same class but different domains are expected to be mapped nearby, resulting in an improved target classification accuracy. Moving average centroid alignment is cautiously designed to compensate the insufficient categorical information within each mini batch. Experiments testify that our model yields state of the art results on standard datasets.
| ['Zibin Zheng', 'Shaoan Xie', 'Liang Chen', 'Chuan Chen'] | 2018-07-01 | null | null | null | icml-2018-7 | ['learning-semantic-representations'] | ['methodology'] | [ 1.89236194e-01 3.87786776e-02 -6.07921004e-01 -9.79899228e-01
-8.40539098e-01 -7.34385729e-01 5.09959102e-01 -1.10753022e-01
-2.65297383e-01 8.41503024e-01 8.41186345e-02 3.02933484e-01
-6.19508140e-02 -7.74103403e-01 -8.38738382e-01 -6.81263983e-01
3.32065701e-01 7.31932402e-01 4.47749943e-01 1.94436967e-01
1.43974572e-01 4.12765622e-01 -1.35178590e+00 3.49933088e-01
8.35571885e-01 1.10264742e+00 3.89665902e-01 7.49317324e-03
-6.09162390e-01 5.60800850e-01 -6.51741922e-01 -2.06231430e-01
3.74548167e-01 -5.41103899e-01 -7.40420759e-01 1.37351215e-01
5.77912033e-01 -2.02199705e-02 -1.34604931e-01 1.33771372e+00
-2.57409047e-02 2.09039092e-01 1.36157668e+00 -1.62246013e+00
-1.16000080e+00 4.08764422e-01 -6.62540078e-01 1.23470947e-01
-2.98509419e-01 -3.61010820e-01 7.00940609e-01 -9.35845315e-01
5.70832968e-01 1.29053915e+00 6.23954058e-01 5.47296345e-01
-1.12318861e+00 -9.37753022e-01 3.94285291e-01 4.07390982e-01
-1.52804315e+00 -4.19348776e-01 9.16892529e-01 -4.72983420e-01
3.50372076e-01 -2.30677739e-01 4.75477576e-02 1.07726061e+00
-6.80308789e-02 7.64168084e-01 8.29420805e-01 -2.56537765e-01
4.75716025e-01 7.44387805e-01 1.52236879e-01 2.69509822e-01
3.31504732e-01 8.46834257e-02 -3.94268781e-01 1.29579097e-01
5.82546353e-01 2.36805916e-01 1.98891908e-01 -1.11856008e+00
-1.02362990e+00 1.07030606e+00 6.48477435e-01 3.11919153e-01
-7.69804195e-02 -1.02576390e-01 2.72242397e-01 1.85715064e-01
7.95551002e-01 1.73477277e-01 -7.89271593e-01 3.16258848e-01
-7.32902586e-01 4.18862514e-02 5.07915437e-01 1.62770963e+00
1.24300051e+00 -2.02163294e-01 -1.78138137e-01 1.00036538e+00
3.68106991e-01 7.46140122e-01 6.00541413e-01 -8.63733470e-01
4.76531982e-01 7.07473159e-01 1.96367130e-01 -7.17409730e-01
-1.44893020e-01 -4.15403306e-01 -5.50985694e-01 2.75613200e-02
6.14503026e-01 -7.27409273e-02 -1.12572515e+00 1.81145668e+00
4.91054982e-01 2.30683431e-01 2.61513352e-01 7.78823197e-01
6.50397360e-01 5.51124334e-01 3.08813423e-01 2.55928963e-01
1.24496031e+00 -1.06169760e+00 -4.56190050e-01 -5.35668135e-01
6.77147865e-01 -6.42023027e-01 1.01717556e+00 -3.33168030e-01
-3.51177573e-01 -7.91191161e-01 -8.15256536e-01 9.23882723e-02
-7.97056258e-01 3.90194118e-01 5.06078601e-01 6.62219703e-01
-6.05429888e-01 4.65539336e-01 -6.17284358e-01 -6.95082605e-01
8.15061629e-01 2.56540954e-01 -4.56619233e-01 -3.26460660e-01
-1.05802071e+00 8.87404621e-01 6.13859832e-01 -4.00762975e-01
-9.39235449e-01 -1.04182243e+00 -1.06561899e+00 1.42616689e-01
2.32818350e-01 -1.71869904e-01 1.31338763e+00 -1.25146270e+00
-1.30528808e+00 9.17263865e-01 -3.37347835e-01 -1.68313354e-01
2.12398797e-01 -1.32955080e-02 -6.01979852e-01 -1.81581333e-01
5.57812572e-01 1.08379292e+00 8.66250217e-01 -1.38440263e+00
-1.00194037e+00 -5.40059149e-01 -3.78412396e-01 3.17670971e-01
-4.33283865e-01 -3.19902569e-01 -2.79967755e-01 -5.17667830e-01
6.41435236e-02 -8.49094629e-01 -1.39588285e-02 4.42121401e-02
-4.63293977e-02 -3.58593434e-01 1.16536450e+00 -2.66579151e-01
5.89853525e-01 -2.40035057e+00 -3.11073899e-01 9.06517729e-02
-1.42659903e-01 6.44579753e-02 -3.52038234e-01 5.27594751e-03
-1.28407791e-01 -1.90228984e-01 -2.58361436e-02 5.81652485e-02
1.68639943e-01 2.29044408e-01 -5.04671454e-01 5.71042418e-01
4.13148582e-01 8.57108295e-01 -1.07330763e+00 -5.64620614e-01
2.19775140e-01 5.87089136e-02 -2.83683687e-01 1.23720959e-01
-2.89664060e-01 6.54147148e-01 -5.71297765e-01 7.29768217e-01
1.00958216e+00 -3.83607715e-01 -6.97803870e-02 -3.03387761e-01
3.51547271e-01 1.23772196e-01 -9.95319009e-01 1.70751274e+00
-2.91136652e-01 4.85896081e-01 -1.47561505e-01 -1.39215219e+00
1.19357574e+00 7.96553213e-03 5.18209696e-01 -7.48956919e-01
1.33653119e-01 4.39528488e-02 -2.38078162e-01 -1.44009054e-01
4.70610082e-01 -3.23903710e-01 -2.55324423e-01 2.51911849e-01
2.51762927e-01 -9.04077142e-02 -7.18679652e-02 -1.27356127e-01
5.22038996e-01 3.69369328e-01 1.91839904e-01 -3.95754308e-01
2.30667666e-01 4.93078738e-01 6.93433404e-01 6.20582342e-01
-3.55486035e-01 5.87868392e-01 2.19283342e-01 -2.55546600e-01
-1.02670288e+00 -1.45861137e+00 -3.17743152e-01 1.46823967e+00
4.95686978e-01 2.38279268e-01 -5.87615967e-01 -1.32394743e+00
8.84299502e-02 8.50105584e-01 -7.10823715e-01 -4.69916910e-01
-2.90753543e-01 -6.26851857e-01 2.75866151e-01 8.76927018e-01
4.57963437e-01 -7.97173858e-01 -1.93911135e-01 1.09945886e-01
-6.98987395e-02 -1.06744432e+00 -5.87998688e-01 2.62112379e-01
-9.52100277e-01 -1.05145967e+00 -9.76535082e-01 -1.08141530e+00
1.01380968e+00 4.29462761e-01 1.06457233e+00 -7.83666372e-01
-8.87483284e-02 1.23963036e-01 -5.36119759e-01 -6.65683746e-01
-3.20596308e-01 2.20572367e-01 -2.93913893e-02 4.52131405e-02
1.21972048e+00 -7.76073560e-02 -3.34733099e-01 7.43901014e-01
-6.18208885e-01 -3.05789798e-01 5.23131609e-01 8.69707823e-01
4.46119189e-01 8.17144513e-02 8.32611501e-01 -1.13720787e+00
3.68399024e-02 -9.36165988e-01 -6.30816817e-01 3.61292630e-01
-4.55294102e-01 4.14285064e-02 6.41898036e-01 -7.28411615e-01
-1.23638427e+00 3.00587147e-01 4.65665251e-01 -5.49335957e-01
-4.93768841e-01 1.22273356e-01 -5.12411118e-01 2.78328478e-01
9.13849354e-01 -1.17781134e-02 8.53727683e-02 -4.99024630e-01
4.99783427e-01 7.23947406e-01 4.52995896e-01 -5.70516407e-01
8.27816784e-01 5.91883600e-01 -2.68383503e-01 -3.78198028e-01
-1.17220950e+00 -5.97127080e-01 -8.09509456e-01 8.69240239e-02
7.67800868e-01 -1.07206273e+00 -7.28461444e-02 2.47990265e-01
-7.91706502e-01 -3.25425297e-01 -5.28142154e-01 7.07019687e-01
-4.66409624e-01 1.95445251e-02 -9.21047702e-02 -3.89808625e-01
2.77953178e-01 -9.36071754e-01 1.08693349e+00 3.58651519e-01
-2.08391264e-01 -1.13417554e+00 -1.44742131e-01 3.94966006e-02
4.30344880e-01 -2.31117338e-01 9.24293995e-01 -1.15957403e+00
-4.58016872e-01 -2.66560107e-01 -6.61050677e-01 4.79557514e-01
6.89196706e-01 -4.24777627e-01 -1.10785675e+00 -2.19392553e-01
-1.68734387e-01 -3.19830090e-01 7.34843731e-01 5.29399276e-01
1.02765834e+00 1.79759171e-02 -8.43984723e-01 3.24173987e-01
1.18936205e+00 3.66469502e-01 3.50321174e-01 7.90329427e-02
6.30044639e-01 9.87587929e-01 1.27107108e+00 2.28566244e-01
3.58161896e-01 5.98184586e-01 1.78911641e-01 -9.61301178e-02
-2.94175357e-01 -4.20287251e-01 3.58712077e-01 1.95934087e-01
6.92203104e-01 -1.43760502e-01 -8.32768202e-01 8.96897197e-01
-1.78564942e+00 -7.43237317e-01 1.11158073e-01 2.06951356e+00
6.61610365e-01 -5.29364273e-02 4.40538190e-02 -4.35762852e-01
1.15523934e+00 -2.74133056e-01 -9.34735179e-01 1.77063286e-01
1.86030775e-01 -6.87808394e-02 8.63022506e-01 3.27164054e-01
-1.33951652e+00 1.23737335e+00 6.39961243e+00 1.15277874e+00
-9.73237753e-01 3.04653615e-01 6.42025530e-01 9.73083153e-02
-1.16227612e-01 -2.03588493e-02 -1.14202178e+00 6.81399941e-01
8.08506846e-01 -3.26665580e-01 5.62838577e-02 1.35124111e+00
-3.44903797e-01 -4.54768278e-02 -1.41205084e+00 8.62220526e-01
9.77038816e-02 -1.04469299e+00 -3.00254356e-02 1.14715599e-01
9.49924350e-01 6.63662255e-02 2.15336263e-01 6.86248600e-01
7.61639178e-01 -9.43083227e-01 6.04339540e-01 2.90402621e-01
9.28119957e-01 -6.89234614e-01 8.28782737e-01 2.78292358e-01
-1.18772268e+00 -7.57782767e-03 -8.93423676e-01 2.09704190e-01
-9.51660350e-02 6.10232949e-01 -1.21191001e+00 3.76442343e-01
7.31976986e-01 9.56808209e-01 -4.93002713e-01 7.77934372e-01
1.17022730e-01 3.49981010e-01 -2.78470159e-01 1.22610316e-01
1.28730118e-01 -2.66658455e-01 1.27286687e-02 9.01158094e-01
6.71010375e-01 -3.99920076e-01 3.09027314e-01 1.01420248e+00
-1.48775548e-01 -2.14795098e-02 -9.02816355e-01 -2.36811265e-02
7.96840250e-01 1.11659122e+00 -9.28744316e-01 -5.75319231e-01
-4.71503615e-01 1.00349963e+00 2.29269207e-01 4.92552966e-01
-1.08087695e+00 -3.29303920e-01 8.25102389e-01 -2.71721538e-02
4.81913120e-01 1.91446885e-01 -5.06805658e-01 -1.01241791e+00
-2.18216896e-01 -1.95930392e-01 4.95203674e-01 -4.70657587e-01
-1.85707557e+00 2.07340986e-01 1.82916120e-01 -1.62211955e+00
-2.29727328e-01 -7.03491509e-01 -2.84741133e-01 9.03004825e-01
-1.52968109e+00 -1.15681207e+00 -3.39406341e-01 6.92943156e-01
6.42497241e-01 -4.77342099e-01 6.92528546e-01 4.78922099e-01
-7.04628527e-02 7.50005305e-01 4.77944940e-01 1.60388172e-01
1.28236365e+00 -1.22854912e+00 2.11846024e-01 4.04797286e-01
-3.55410716e-03 3.37936074e-01 3.97744715e-01 -6.04821026e-01
-5.90503097e-01 -1.62539256e+00 7.64441371e-01 -8.34903419e-01
4.91239339e-01 -3.89571965e-01 -8.93241942e-01 9.32240605e-01
-2.29573831e-01 3.09209377e-01 8.63495946e-01 1.16465226e-01
-5.98714828e-01 -4.23871130e-01 -1.39077115e+00 3.32281798e-01
1.07542193e+00 -5.41738510e-01 -6.32180750e-01 3.58195275e-01
5.96731424e-01 -2.82851458e-01 -6.37089074e-01 4.14554745e-01
1.77693024e-01 -5.48291028e-01 1.06212103e+00 -5.50818443e-01
2.30398268e-01 -5.22340238e-01 -2.72133410e-01 -1.50951266e+00
-4.21692640e-01 4.45259303e-01 1.46704972e-01 1.60664761e+00
3.68958443e-01 -4.96839195e-01 1.11999476e+00 6.11856163e-01
-2.07249925e-01 3.43671557e-03 -8.86915028e-01 -1.19614041e+00
4.18395609e-01 -1.10723376e-01 6.58885717e-01 1.21025705e+00
-2.02273905e-01 5.04450679e-01 -6.00200258e-02 2.85242021e-01
6.18669987e-01 3.04710567e-01 7.36919343e-01 -1.44882417e+00
1.58160016e-01 -2.53121167e-01 -3.66416335e-01 -1.11103582e+00
5.88868141e-01 -1.22016478e+00 2.40473628e-01 -1.26284838e+00
3.72580439e-01 -8.50662291e-01 -3.98303628e-01 5.05927563e-01
-4.26987112e-02 3.77777517e-01 -3.60064544e-02 1.64017573e-01
-7.21677423e-01 6.86271906e-01 1.07736611e+00 -3.26428831e-01
-1.47650480e-01 -7.32403770e-02 -8.43467593e-01 7.62360990e-01
8.73438537e-01 -9.01021063e-01 -8.12356293e-01 -4.70590502e-01
-4.70483929e-01 -4.69484031e-01 3.49950701e-01 -1.03643882e+00
1.28162831e-01 -3.81368726e-01 8.43956292e-01 -7.40182102e-01
2.15138137e-01 -1.32466912e+00 2.21534967e-02 6.72774836e-02
-4.26096886e-01 -5.14249086e-01 5.23823015e-02 7.44872868e-01
-3.94343644e-01 -3.50523651e-01 1.11186230e+00 -2.97717471e-02
-1.14061701e+00 3.03620994e-01 -1.35005265e-01 1.79210767e-01
1.36952233e+00 -3.81905407e-01 -2.80928463e-01 -1.51906133e-01
-5.90292335e-01 2.38569900e-01 6.32146657e-01 7.26588547e-01
1.82376593e-01 -1.59992445e+00 -4.42167550e-01 2.35091746e-01
6.25982344e-01 1.20324716e-01 3.18640649e-01 3.82491320e-01
-1.17526948e-01 4.64380324e-01 -3.25266182e-01 -8.27788055e-01
-8.68232667e-01 9.64286029e-01 1.74504608e-01 2.26575863e-02
-4.30212200e-01 8.29430580e-01 7.68688202e-01 -8.51254642e-01
2.66780376e-01 -2.35939622e-01 -1.70557186e-01 1.52458996e-01
5.06285369e-01 2.96514392e-01 -8.93486515e-02 -6.17122233e-01
-4.80608344e-01 7.14918256e-01 -3.07430327e-01 1.86517432e-01
1.02268219e+00 -3.10403824e-01 2.60808617e-01 4.20673728e-01
1.46820724e+00 -3.02951127e-01 -1.60618246e+00 -5.31190753e-01
3.71443510e-01 -5.75874805e-01 -4.22316760e-01 -7.92433918e-01
-1.01907969e+00 8.10259938e-01 8.80112290e-01 -1.90719396e-01
9.05912042e-01 6.25601053e-01 5.15786529e-01 3.88096333e-01
5.13604701e-01 -1.34682465e+00 1.14073485e-01 6.68871999e-01
4.06695694e-01 -1.54099786e+00 -3.72582674e-01 -5.91645360e-01
-9.63199019e-01 8.15187156e-01 9.62202609e-01 -1.38953447e-01
6.72788143e-01 -4.45689522e-02 2.95970440e-01 -3.98775004e-02
-2.74135768e-01 -1.78657979e-01 2.00957477e-01 1.15349710e+00
2.56755531e-01 1.08073622e-01 2.35120252e-01 7.25809693e-01
2.53148600e-02 -1.16613298e-03 1.32844403e-01 7.75811791e-01
-5.30006528e-01 -1.17091680e+00 -5.65391123e-01 4.31822121e-01
-1.88827604e-01 4.55439128e-02 -2.81698644e-01 7.41415679e-01
2.62972116e-01 9.32422698e-01 5.50231159e-01 7.66471587e-03
1.92128912e-01 3.75712663e-01 3.10232669e-01 -9.10040855e-01
1.54015720e-01 2.15846244e-02 -2.31005162e-01 -3.61833274e-01
-4.48527247e-01 -4.56574500e-01 -1.36603141e+00 -2.70912163e-02
-1.71492770e-01 2.92515635e-01 6.14088893e-01 7.83648908e-01
5.61725855e-01 4.19139892e-01 5.90824544e-01 -7.17582643e-01
-6.89736784e-01 -1.11515772e+00 -9.45542753e-01 6.80497527e-01
1.80544600e-01 -1.05993247e+00 -2.25677654e-01 3.57017726e-01] | [10.33004379272461, 2.8516030311584473] |
c8de20a5-05c1-4a2b-9854-f13996136f0d | towards-implicit-content-introducing-for | null | null | https://aclanthology.org/D17-1233 | https://aclanthology.org/D17-1233.pdf | Towards Implicit Content-Introducing for Generative Short-Text Conversation Systems | The study on human-computer conversation systems is a hot research topic nowadays. One of the prevailing methods to build the system is using the generative Sequence-to-Sequence (Seq2Seq) model through neural networks. However, the standard Seq2Seq model is prone to generate trivial responses. In this paper, we aim to generate a more meaningful and informative reply when answering a given question. We propose an implicit content-introducing method which incorporates additional information into the Seq2Seq model in a flexible way. Specifically, we fuse the general decoding and the auxiliary cue word information through our proposed hierarchical gated fusion unit. Experiments on real-life data demonstrate that our model consistently outperforms a set of competitive baselines in terms of BLEU scores and human evaluation. | ['Yaoyuan Zhang', 'Rui Yan', 'Lili Yao', 'Dongyan Zhao', 'Yansong Feng'] | 2017-09-01 | null | null | null | emnlp-2017-9 | ['short-text-conversation'] | ['natural-language-processing'] | [ 4.63452190e-01 1.77473396e-01 3.39264631e-01 -5.64312458e-01
-1.16604745e+00 -5.67287028e-01 7.59291291e-01 -2.16340512e-01
-4.97687727e-01 1.09454632e+00 9.01976109e-01 -2.06574246e-01
4.77087587e-01 -5.43274701e-01 -3.45624715e-01 -7.17296422e-01
7.18530476e-01 4.43099469e-01 2.34235868e-01 -7.33284116e-01
3.49897027e-01 -1.92535520e-01 -1.10110867e+00 7.25665748e-01
1.05563450e+00 5.92089474e-01 3.86350930e-01 8.99859786e-01
-3.86987478e-01 9.06950712e-01 -9.92545903e-01 -8.23180020e-01
-1.43554047e-01 -1.00371218e+00 -1.00168765e+00 -1.97042212e-01
-1.54876754e-01 -3.02122682e-01 -1.13454178e-01 8.12603295e-01
9.59995151e-01 3.23056936e-01 4.09090430e-01 -8.04733932e-01
-4.32606339e-01 8.91743183e-01 -1.85681924e-01 1.00762747e-01
8.59007299e-01 3.59376490e-01 9.04480159e-01 -8.81660283e-01
4.36346501e-01 1.52836514e+00 3.33077133e-01 7.81974614e-01
-9.88823354e-01 -4.59796906e-01 1.08018063e-01 1.13411792e-01
-9.63932037e-01 -5.26525617e-01 6.71183944e-01 -1.07490838e-01
7.59721100e-01 3.98943812e-01 7.90843442e-02 1.58445680e+00
2.51296490e-01 9.63076413e-01 1.05439103e+00 -3.80664170e-01
2.09313948e-02 2.07409933e-01 1.67301968e-01 2.54105777e-01
-4.29428309e-01 -2.55618304e-01 -5.04974067e-01 -2.69446880e-01
3.25325310e-01 -1.99369356e-01 -4.03169274e-01 4.27578866e-01
-1.08814526e+00 1.08319879e+00 2.75298685e-01 2.90394634e-01
-5.80622852e-01 -3.62438969e-02 4.34815198e-01 2.22789243e-01
4.97230440e-01 4.59429085e-01 -2.37592414e-01 -3.50047350e-01
-6.19549096e-01 4.37791228e-01 1.16338873e+00 9.19196069e-01
4.10130948e-01 -4.82453108e-02 -8.51019025e-01 1.06857383e+00
2.11463377e-01 2.81616926e-01 6.58522844e-01 -8.85814071e-01
5.96192539e-01 3.72897714e-01 1.92692548e-01 -7.59999812e-01
-1.07014909e-01 -4.89461511e-01 -9.74687874e-01 -4.44461226e-01
2.39367634e-01 -5.69667697e-01 -5.51045179e-01 1.68951094e+00
1.57542259e-01 -2.67682821e-02 2.53396451e-01 9.07171428e-01
1.18981600e+00 9.52728093e-01 2.43201956e-01 -2.68514842e-01
1.28069282e+00 -1.28701401e+00 -1.00329673e+00 -2.23427221e-01
5.60853183e-01 -9.90234554e-01 1.04533899e+00 2.84634203e-01
-1.24563277e+00 -6.96352422e-01 -6.28444195e-01 -1.66823432e-01
-4.79825810e-02 -6.80444092e-02 2.68477380e-01 5.44217825e-01
-9.76321697e-01 1.14443116e-01 -1.51739866e-01 -2.62265295e-01
-2.59569108e-01 2.14489281e-01 -1.90480441e-01 -9.17515457e-02
-1.64576459e+00 8.63222122e-01 4.63992834e-01 2.10561648e-01
-5.89496911e-01 -2.53818005e-01 -7.12106705e-01 7.77103379e-02
3.94854635e-01 -8.70464802e-01 1.77323675e+00 -8.27493846e-01
-2.00324082e+00 3.97056282e-01 -4.33738708e-01 -3.28069478e-01
4.95386422e-01 -3.18872720e-01 -1.39986113e-01 1.44642651e-01
-2.13121861e-01 7.47066200e-01 5.96310794e-01 -1.18873715e+00
-4.30792958e-01 1.13024287e-01 2.13560358e-01 2.54209012e-01
-1.03803732e-01 2.07636327e-01 -3.93897653e-01 -6.49384260e-01
-4.59511459e-01 -1.00929701e+00 -3.81557405e-01 -9.62560833e-01
-4.01184976e-01 -6.47268713e-01 3.95682275e-01 -1.03418446e+00
1.55389786e+00 -1.75518537e+00 1.96474746e-01 -5.02327830e-02
1.11164860e-02 5.43692052e-01 -4.17980462e-01 8.92721891e-01
1.40243784e-01 -7.30391964e-02 -3.29143405e-01 -3.64775389e-01
1.08180093e-02 -5.54792397e-02 -3.40035230e-01 -1.96495488e-01
2.93001235e-01 1.11437893e+00 -1.04985249e+00 -5.46772301e-01
-1.44564271e-01 4.32789445e-01 -8.01480472e-01 8.58028710e-01
-4.03462201e-01 7.31689751e-01 -5.27972043e-01 1.59806326e-01
5.81998587e-01 -9.32057351e-02 2.46178925e-01 4.50257510e-02
1.87336933e-02 6.32073820e-01 -6.75756335e-01 1.79293966e+00
-5.50991654e-01 2.22717032e-01 4.24395390e-02 -7.01364279e-01
1.08743584e+00 5.88825464e-01 6.83435276e-02 -5.93226194e-01
2.60976702e-01 1.47577405e-01 1.53558806e-01 -7.10323036e-01
9.99451101e-01 -1.29054427e-01 -5.33190966e-01 5.10106862e-01
1.33396149e-01 -1.09652221e-01 1.38465852e-01 4.29440558e-01
1.03009081e+00 1.40212253e-01 3.39290261e-01 -6.80923834e-03
9.04514611e-01 -3.55321497e-01 4.34767574e-01 7.09405184e-01
-6.29473627e-02 7.43646383e-01 5.24863660e-01 3.82981971e-02
-8.03576052e-01 -6.79567158e-01 4.30256099e-01 1.32718182e+00
-2.09575146e-01 -3.38533163e-01 -1.08052087e+00 -7.88670003e-01
-4.73714530e-01 9.05422986e-01 -3.98496360e-01 -1.65567547e-01
-7.97819614e-01 -6.67595088e-01 5.79057395e-01 3.48423660e-01
5.83918571e-01 -1.42131424e+00 -1.70274287e-01 6.03283525e-01
-9.88574684e-01 -1.07368243e+00 -9.38283324e-01 -2.28343263e-01
-5.50293982e-01 -4.73613024e-01 -1.03219783e+00 -6.97512329e-01
3.06606054e-01 3.76277864e-01 1.24046814e+00 1.38948604e-01
3.92986327e-01 1.56574976e-02 -7.56263554e-01 -1.97156325e-01
-7.02902317e-01 4.79686826e-01 -5.34298956e-01 2.38250077e-01
3.68261427e-01 -3.69500548e-01 -7.05996692e-01 3.54942709e-01
-9.24343526e-01 1.09798133e-01 9.04162169e-01 9.63337541e-01
-9.91656561e-04 -6.16954744e-01 1.19546509e+00 -1.09831965e+00
1.52399766e+00 -7.50917733e-01 3.06734852e-02 3.20290208e-01
-2.55599380e-01 1.97032019e-01 7.18339205e-01 -3.37476790e-01
-1.53085756e+00 -1.45504296e-01 -6.82164967e-01 2.93174148e-01
-2.23738313e-01 4.84504402e-01 -3.92866790e-01 2.96919733e-01
5.83786070e-01 4.85602081e-01 -1.23813540e-01 -5.87070763e-01
3.97285193e-01 1.21482587e+00 6.16278827e-01 -6.49681091e-01
2.88814932e-01 -2.88223684e-01 -6.24258399e-01 -5.58743358e-01
-7.86641538e-01 -5.71615517e-01 -3.15673441e-01 -3.54763776e-01
9.14748371e-01 -6.93647981e-01 -9.96324122e-01 3.83803338e-01
-1.70097029e+00 -1.20211422e-01 2.65074521e-01 2.35347241e-01
-4.09127891e-01 5.52535236e-01 -7.81548321e-01 -9.72307920e-01
-7.55669177e-01 -1.18239903e+00 1.05342066e+00 4.74248379e-01
-4.93624449e-01 -6.94604695e-01 2.57653892e-01 7.74222732e-01
6.27554059e-01 1.46421418e-02 5.93362093e-01 -9.40424085e-01
-4.12863463e-01 -1.60341114e-01 -2.00070843e-01 4.49818373e-01
-1.14291288e-01 -4.41786110e-01 -8.89782727e-01 -9.42251459e-02
1.28415301e-01 -5.35837471e-01 9.01817620e-01 -6.60190731e-02
1.17011869e+00 -5.94987094e-01 5.19967005e-02 -1.52730187e-02
9.91937160e-01 3.63405108e-01 9.42622960e-01 -1.53128907e-01
4.43019301e-01 9.36207294e-01 4.45248902e-01 3.55378687e-01
6.40787125e-01 5.88373542e-01 6.47193640e-02 3.14394057e-01
1.08912317e-02 -4.37713385e-01 4.67550874e-01 1.42003763e+00
6.68886751e-02 -8.20170581e-01 -6.72788441e-01 4.34790254e-01
-1.93725240e+00 -1.10508382e+00 -2.07572594e-01 1.87058961e+00
1.04506207e+00 1.12180457e-01 1.36809155e-01 -1.91099644e-01
7.99559772e-01 7.46289194e-02 -7.46404529e-02 -6.46960378e-01
-9.88084003e-02 2.19303936e-01 -1.17563449e-01 7.46100843e-01
-5.77966511e-01 9.31292057e-01 6.31642294e+00 7.48840153e-01
-8.23248565e-01 1.62861153e-01 5.37348807e-01 2.04831019e-01
-4.41311270e-01 -1.32285625e-01 -9.01248753e-01 7.61547565e-01
1.24338412e+00 -1.68009594e-01 1.82316899e-01 4.63644028e-01
4.59261537e-01 -1.03731692e-01 -7.28675425e-01 6.93370938e-01
2.45282352e-01 -1.06519234e+00 2.05646947e-01 -1.84811711e-01
6.36743248e-01 -4.00036514e-01 -2.87871361e-01 6.44914150e-01
4.91900831e-01 -8.66286457e-01 3.21492761e-01 8.50263596e-01
4.04062033e-01 -6.87525868e-01 1.14042282e+00 6.35058224e-01
-7.09879220e-01 1.04266435e-01 -7.01606497e-02 -3.63101065e-02
6.86534703e-01 2.92683035e-01 -1.11217272e+00 7.65658140e-01
1.13258861e-01 2.12657675e-01 -2.83318013e-01 1.03657496e+00
-3.23746353e-01 8.10376942e-01 9.31001902e-02 -5.26254833e-01
4.16130006e-01 -7.83517510e-02 4.08612430e-01 1.55736911e+00
2.33212620e-01 5.68332076e-01 1.54307544e-01 5.42190969e-01
-3.31254989e-01 3.01213920e-01 -3.48630339e-01 -1.13335298e-02
3.76477957e-01 1.19596064e+00 -1.90330833e-01 -4.28448230e-01
-2.62400478e-01 1.24595726e+00 3.02032799e-01 3.25006992e-01
-7.13449359e-01 -5.39042234e-01 1.99942455e-01 -1.98873460e-01
1.08090073e-01 -8.32634568e-02 -1.16356775e-01 -1.08257699e+00
-4.93934266e-02 -1.29382205e+00 3.67227793e-01 -8.68522823e-01
-1.36206901e+00 9.16582584e-01 -1.50289352e-03 -9.79070485e-01
-7.51853168e-01 -1.23531029e-01 -7.59282351e-01 1.25270116e+00
-1.40564573e+00 -8.97466779e-01 -4.39396292e-01 3.65218729e-01
1.06479716e+00 -1.64132323e-02 8.60346138e-01 3.36525440e-01
-4.41198736e-01 7.97914386e-01 -1.69681683e-01 9.61147770e-02
9.21233118e-01 -9.93185222e-01 4.98988539e-01 6.97925091e-01
-3.65681350e-01 8.74222279e-01 9.57337141e-01 -6.18210018e-01
-9.89905894e-01 -9.06456888e-01 1.32219601e+00 -2.93804973e-01
2.80953199e-01 -3.94555748e-01 -1.15935636e+00 3.80854040e-01
1.03560734e+00 -8.71971905e-01 8.11613262e-01 -2.04610646e-01
-1.13029331e-01 1.71071038e-01 -9.97376263e-01 6.55675411e-01
7.67952681e-01 -4.34767514e-01 -8.30602884e-01 2.37899333e-01
1.07327223e+00 -3.30842257e-01 -5.89361191e-01 3.01384658e-01
4.93306309e-01 -8.62901568e-01 6.24652147e-01 -6.62427425e-01
5.49527049e-01 -8.80567431e-02 -4.79738228e-02 -1.54216707e+00
-4.14217234e-01 -1.16407418e+00 -8.35530367e-03 1.44881535e+00
4.81038392e-01 -5.55508137e-01 4.93113071e-01 4.88897532e-01
-3.22230637e-01 -5.90436041e-01 -5.51666081e-01 -3.38310212e-01
4.79999632e-02 2.91917380e-02 7.32172310e-01 4.98551726e-01
1.92433432e-01 9.78592932e-01 -8.63009512e-01 -4.05444980e-01
1.97639465e-01 -1.08989000e-01 9.53584790e-01 -8.98389697e-01
-3.99307460e-01 -4.56609637e-01 2.59277076e-01 -1.68470407e+00
2.36476570e-01 -6.23973727e-01 5.88298619e-01 -1.59659648e+00
2.53854901e-01 5.02921604e-02 -2.86947340e-02 -2.93546468e-02
-8.17615688e-01 -4.51075695e-02 2.12755024e-01 -1.05129898e-01
-6.70879602e-01 8.53333712e-01 1.34729087e+00 1.07591130e-01
-1.53275713e-01 1.86798587e-01 -9.71451938e-01 2.29657367e-01
8.88026774e-01 -3.66993845e-01 -2.98578858e-01 -4.50611919e-01
1.95100680e-02 5.14956832e-01 7.91076012e-03 -4.68934506e-01
2.85644293e-01 -6.06842153e-02 -1.22792296e-01 -7.24295795e-01
3.70446652e-01 -3.47253263e-01 -1.20864406e-01 3.01634967e-01
-7.62654126e-01 2.73391485e-01 -5.38440868e-02 5.81698477e-01
-4.26932395e-01 -3.05765390e-01 4.97647047e-01 -2.55626768e-01
-4.55159187e-01 -6.45752549e-02 -4.48419362e-01 9.94394198e-02
5.79652011e-01 1.94003373e-01 -4.21230316e-01 -1.15056407e+00
-3.01618338e-01 3.52175474e-01 -8.96428525e-02 7.46748805e-01
5.88998616e-01 -1.29124451e+00 -1.25595832e+00 -1.89344376e-01
1.36530831e-01 -3.70662898e-01 3.89427483e-01 6.53024673e-01
-2.13992566e-01 7.85658300e-01 9.68277734e-03 -2.76266903e-01
-1.16183245e+00 2.92589128e-01 1.55779123e-01 -5.80699921e-01
-1.51866317e-01 8.75748277e-01 3.06143373e-01 -5.07551372e-01
1.77152842e-01 9.65358615e-02 -5.49911022e-01 5.73074929e-02
7.05832422e-01 2.95537949e-01 1.29874283e-02 -6.75256312e-01
-2.59389486e-02 -8.35361704e-02 -1.91216514e-01 -4.28887755e-01
1.05055618e+00 -2.88112402e-01 -7.86658302e-02 2.28592575e-01
1.22040856e+00 2.20015179e-03 -9.09004986e-01 -4.26936239e-01
9.78185907e-02 -2.85506010e-01 -4.80993867e-01 -8.79667997e-01
-5.66122353e-01 1.00805378e+00 6.93633547e-03 2.91609824e-01
9.89746988e-01 -2.52907753e-01 1.33650386e+00 3.61159593e-01
6.33701682e-02 -9.41033542e-01 3.25539500e-01 8.41195703e-01
1.18048072e+00 -1.19337392e+00 -6.26727939e-01 -2.64981776e-01
-9.33900535e-01 9.69399393e-01 8.07440817e-01 1.13429196e-01
1.28647178e-01 1.08644934e-02 3.18525821e-01 2.75264353e-01
-1.09361327e+00 -1.51194528e-01 2.12934688e-01 2.65581340e-01
8.14037740e-01 -3.70678529e-02 -9.45928097e-01 7.96392798e-01
-2.84783989e-01 -2.06981283e-02 6.33365154e-01 7.52865672e-01
-5.01642227e-01 -1.41896176e+00 -2.50084341e-01 1.07240148e-01
-6.15819633e-01 -2.57370085e-01 -6.35564625e-01 1.17091544e-01
-3.35749120e-01 1.53493369e+00 -4.00455564e-01 -5.64677179e-01
4.41185713e-01 4.69617128e-01 1.50370181e-01 -6.71537280e-01
-1.19037986e+00 1.25762939e-01 4.39300239e-01 -2.65797317e-01
-4.18171197e-01 -5.39576471e-01 -9.01598275e-01 -2.73693442e-01
-3.69963437e-01 5.73308945e-01 4.94262040e-01 1.05289471e+00
3.80157858e-01 6.11723840e-01 8.80599558e-01 -6.25207663e-01
-6.97177410e-01 -1.59781039e+00 1.30617842e-01 4.02365297e-01
2.15893596e-01 -1.57772973e-01 -1.65047824e-01 1.62395909e-02] | [12.524234771728516, 8.28912353515625] |
d86baef4-2e8b-4653-829c-1d027cb109c5 | mailex-email-event-and-argument-extraction | 2305.13469 | null | https://arxiv.org/abs/2305.13469v1 | https://arxiv.org/pdf/2305.13469v1.pdf | MAILEX: Email Event and Argument Extraction | In this work, we present the first dataset, \dataset, for performing event extraction from conversational email threads. To this end, we first proposed a new taxonomy covering 10 event types and 76 arguments in the email domain. Our final dataset includes $\sim$4K emails annotated with $\sim$9K event instances. To understand the task challenges, we conducted a series of experiments comparing two commonly-seen lines of approaches for event extraction, i.e., sequence labeling and generative end-to-end extraction (including few-shot GPT-3.5). Our results showed that the task of email event extraction is far from being addressed, due to challenges lying in, e.g., extracting non-continuous, shared trigger spans, extracting non-named entity arguments, and modeling the email conversational history. Our work thus suggests more investigations in this domain-specific event extraction task in the future.\footnote{The source code and dataset can be obtained from \url{https://github.com/salokr/Email-Event-Extraction}. | ['Ziyu Yao', 'Joshua Poore', 'Paulo Costa', 'Ali Raz', 'Shou Matsumoto', 'Gaurav Singh', 'Saurabh Srivastava'] | 2023-05-22 | null | null | null | null | ['event-extraction'] | ['natural-language-processing'] | [ 3.13661963e-01 3.78498852e-01 8.79176483e-02 -5.03308654e-01
-1.15054941e+00 -8.10885668e-01 8.71646643e-01 3.47471744e-01
-2.88924247e-01 9.57386672e-01 5.99554002e-01 -2.77868509e-01
-6.29168451e-02 -7.22716510e-01 -4.36708272e-01 -3.61131370e-01
-2.13778093e-02 6.78358376e-01 2.90906698e-01 -8.14654529e-02
5.12132704e-01 1.49583861e-01 -1.49730802e+00 6.64480805e-01
3.14285189e-01 7.48553693e-01 -8.29621702e-02 1.06407595e+00
-5.28190792e-01 1.05457282e+00 -8.95557046e-01 -8.22258532e-01
-1.73462883e-01 -9.68370557e-01 -1.23767078e+00 -1.69860885e-01
-2.09002778e-01 -6.65574297e-02 -4.86133456e-01 5.01375258e-01
8.34406734e-01 3.24954450e-01 5.23634195e-01 -1.50729585e+00
1.19204605e-02 1.17432249e+00 -1.77877396e-01 6.67239547e-01
7.26786971e-01 9.07603875e-02 1.20819020e+00 -7.34709144e-01
1.05931652e+00 9.92885232e-01 6.42300665e-01 4.05890346e-01
-7.79993415e-01 -7.15133011e-01 -7.48114139e-02 1.75528482e-01
-1.10744393e+00 -4.54973310e-01 8.30021262e-01 -4.50751334e-01
1.28649604e+00 4.47466880e-01 5.46061516e-01 1.76246130e+00
4.25716303e-03 8.32044244e-01 1.10877764e+00 -5.50479412e-01
2.24434793e-01 -4.25746990e-03 7.13153660e-01 3.28682512e-01
-2.52004806e-02 -4.49235812e-02 -1.04214513e+00 -7.40675986e-01
3.55278134e-01 -2.98238248e-01 -1.01245612e-01 4.66320485e-01
-1.25006402e+00 7.78137445e-01 -2.57765651e-01 4.23568606e-01
-6.22964680e-01 1.38852715e-01 8.18286657e-01 1.52171120e-01
5.18180668e-01 2.78924942e-01 -7.01571703e-01 -1.04672861e+00
-5.83224297e-01 5.75112283e-01 1.38954461e+00 1.10766828e+00
5.51600754e-01 -5.17100692e-01 -3.17902416e-01 6.98370337e-01
-1.38223274e-02 -1.91133007e-01 2.88622200e-01 -8.06054413e-01
7.72427917e-01 5.99480331e-01 1.57867193e-01 -5.80138087e-01
-4.93378699e-01 1.53397247e-02 -2.87396550e-01 -4.78543520e-01
7.19337225e-01 -6.29015446e-01 -3.76526535e-01 1.59270501e+00
4.92648870e-01 3.81224424e-01 1.58945218e-01 3.22095573e-01
1.17463732e+00 6.52878881e-01 2.61594206e-01 -5.46905041e-01
1.94596493e+00 -6.48690164e-01 -1.05660963e+00 -1.72252670e-01
5.11496842e-01 -1.06139588e+00 9.32497799e-01 1.07272834e-01
-1.11960077e+00 -2.19246998e-01 -5.06377816e-01 9.05541703e-02
-4.76162285e-01 -2.93631852e-01 7.58595049e-01 3.77026856e-01
-3.29615206e-01 5.52796841e-01 -7.17498362e-01 -8.16511571e-01
2.07001731e-01 4.51309383e-02 1.19554877e-01 4.14572150e-01
-1.55610788e+00 5.81301332e-01 5.68048716e-01 -3.53830218e-01
-4.64765519e-01 -6.49334669e-01 -5.13608456e-01 -8.00361782e-02
6.44454241e-01 -4.05236572e-01 1.87172234e+00 -3.14659715e-01
-1.37706232e+00 8.88554692e-01 -4.08238471e-01 -4.15747732e-01
4.14205730e-01 -3.26686472e-01 -3.97701085e-01 1.43171325e-01
1.09010451e-01 1.09451309e-01 4.42772955e-01 -9.80061710e-01
-7.78433263e-01 -2.82160938e-01 1.58043131e-02 -4.69930954e-02
-1.65145144e-01 6.65743113e-01 -3.19573224e-01 -8.83215129e-01
-3.81777853e-01 -1.00470328e+00 -6.62747249e-02 -1.05932748e+00
-5.40144205e-01 -6.82353079e-01 5.85993171e-01 -8.81538391e-01
1.61094654e+00 -1.93308008e+00 -1.30994767e-01 -9.51620638e-02
-3.88376489e-02 -6.76800385e-02 2.15238705e-01 1.25268555e+00
-6.39829934e-02 1.75146595e-01 -3.65465395e-02 -4.02665079e-01
2.85910219e-01 8.25683549e-02 -4.89769787e-01 -8.20043497e-03
1.22616202e-01 1.13633657e+00 -9.78590667e-01 -7.87789285e-01
5.02845310e-02 4.15992260e-01 -1.41716108e-01 4.31083322e-01
-4.87466037e-01 4.84353930e-01 -6.20841742e-01 4.14812863e-01
1.73850343e-01 -4.03443575e-01 3.23147148e-01 -2.34098621e-02
-7.92421624e-02 8.80931914e-01 -1.36789203e+00 1.66366029e+00
-2.65723526e-01 7.87018001e-01 -3.37982297e-01 -6.59035921e-01
6.37755632e-01 7.85490215e-01 5.85846007e-01 -3.06733042e-01
4.55049068e-01 1.30983284e-02 -2.71875024e-01 -6.11160219e-01
7.30333924e-01 1.48056269e-01 -6.49985254e-01 1.02138209e+00
3.09110969e-01 2.93615460e-01 5.34932315e-01 4.20957804e-01
1.44593239e+00 1.91190362e-01 6.62990093e-01 7.23130256e-02
1.84266239e-01 3.43111783e-01 5.37217498e-01 9.71417129e-01
-1.65440470e-01 5.33533335e-01 8.94003093e-01 -3.17791849e-01
-8.87241185e-01 -7.18040466e-01 -5.56769408e-02 1.35696661e+00
-2.45645106e-01 -9.51516807e-01 -9.19327497e-01 -8.64495516e-01
-3.40733141e-01 9.03432071e-01 -4.67358559e-01 3.74062121e-01
-1.08386195e+00 -9.29864824e-01 1.00011480e+00 5.36444843e-01
3.24254543e-01 -1.50351191e+00 -6.77205443e-01 6.73731208e-01
-8.10266078e-01 -1.34892535e+00 -2.90710777e-01 4.35195088e-01
-4.70849484e-01 -1.25521314e+00 -3.79890382e-01 -5.78144133e-01
-4.20340262e-02 -2.13502422e-01 1.37880743e+00 -2.10065678e-01
-5.45463800e-01 4.75799769e-01 -7.14069843e-01 -7.54302025e-01
-5.15818417e-01 2.78177083e-01 -5.14509857e-01 -2.71113306e-01
8.45034003e-01 -6.55369520e-01 -5.82562745e-01 2.48148069e-01
-6.55187786e-01 -2.48912022e-01 2.28004396e-01 5.86859584e-01
2.91060865e-01 -4.74942336e-03 8.30246687e-01 -1.12684548e+00
9.62664306e-01 -6.33563757e-01 4.97059850e-03 3.70744206e-02
-2.05980241e-01 -2.59220958e-01 2.74017781e-01 -4.88359004e-01
-1.42380166e+00 -6.83006495e-02 -3.51217300e-01 3.36764753e-01
-5.95862806e-01 2.36168712e-01 -1.12468004e-03 7.38266706e-01
7.60538757e-01 1.93760887e-01 -3.64711136e-01 -4.87259716e-01
3.00875962e-01 9.44672704e-01 2.96398163e-01 -6.73885405e-01
3.59467059e-01 4.18793589e-01 -5.38083792e-01 -9.02080595e-01
-9.00154293e-01 -6.33447230e-01 -3.54979485e-01 -3.10405999e-01
9.24042821e-01 -7.27667272e-01 -8.49669397e-01 6.02949977e-01
-1.36735833e+00 -5.08328795e-01 -5.08348644e-01 3.79050285e-01
-6.34329021e-01 5.05094111e-01 -1.16978586e+00 -9.29031968e-01
-5.64819098e-01 -5.94349265e-01 1.00622141e+00 2.67979145e-01
-1.03718245e+00 -7.90620685e-01 2.95822740e-01 6.26756012e-01
8.99594929e-03 3.76836538e-01 5.89667499e-01 -1.28786910e+00
-3.33079815e-01 -1.29175931e-01 4.91728224e-02 -1.66519850e-01
-1.07685916e-01 -9.85041484e-02 -1.07376623e+00 4.75220568e-02
-6.06432073e-02 -1.95827946e-01 5.45769155e-01 1.35003611e-01
7.89613485e-01 -2.31195733e-01 -4.83905435e-01 -8.95247161e-02
9.51289237e-01 4.54637885e-01 6.26301527e-01 2.99932986e-01
3.45092893e-01 8.36352885e-01 8.18104327e-01 9.65952575e-01
3.51850837e-01 5.80762923e-01 -1.36659546e-02 3.38305086e-01
-1.50594801e-01 -1.71581775e-01 1.84199348e-01 6.26320302e-01
-8.11006054e-02 -7.40477979e-01 -1.02204442e+00 6.99017823e-01
-2.13469863e+00 -1.26521146e+00 -4.78404194e-01 1.61102521e+00
1.05655992e+00 4.38521057e-01 4.15388465e-01 1.95313141e-01
8.91933560e-01 2.27374524e-01 -1.28454462e-01 -3.35217088e-01
1.68877557e-01 5.39181471e-01 1.01726577e-01 3.45922112e-01
-1.03674293e+00 8.84900033e-01 5.28413916e+00 8.40900302e-01
-4.50810552e-01 3.14220428e-01 2.62120068e-01 -1.57868847e-01
-3.19919176e-02 3.19836289e-01 -1.38174725e+00 7.53241420e-01
1.41392100e+00 -2.25272939e-01 1.98989466e-01 5.56506872e-01
2.12854415e-01 -1.39783934e-01 -1.00662911e+00 7.37475276e-01
-7.78850093e-02 -1.44135094e+00 -3.91566306e-01 1.56955659e-01
3.16514343e-01 -1.56560451e-01 -6.91033185e-01 6.12900198e-01
4.93884295e-01 -5.48587680e-01 4.58681375e-01 3.49662185e-01
2.92256266e-01 -7.82942951e-01 7.55922794e-01 3.35868210e-01
-1.31590593e+00 2.41871342e-01 2.64844090e-01 -1.60828948e-01
7.46547103e-01 8.10775876e-01 -1.10056722e+00 6.96911335e-01
7.70077467e-01 4.38724846e-01 -2.59043634e-01 7.56555259e-01
-2.80083805e-01 9.95210946e-01 -4.27756131e-01 -2.53861517e-01
-4.80731092e-02 1.36587277e-01 7.42117226e-01 1.81960845e+00
1.18498787e-01 4.30002302e-01 1.92351595e-01 4.19581592e-01
-1.26843750e-01 6.83913082e-02 -5.48481584e-01 -3.34159583e-01
6.63092792e-01 1.24701560e+00 -1.23921871e+00 -5.16362906e-01
-2.60355949e-01 9.36029196e-01 1.19881697e-01 2.16276988e-01
-1.18537152e+00 -7.68487513e-01 6.34394646e-01 -2.42006476e-03
3.53832781e-01 -1.52319923e-01 -2.52475709e-01 -9.65481758e-01
-6.46152953e-03 -8.09588432e-01 8.56701255e-01 -5.17343402e-01
-1.43939960e+00 6.03054941e-01 2.69431591e-01 -6.84244633e-01
-5.99566281e-01 -1.50149852e-01 -7.15582490e-01 6.40134335e-01
-1.02999973e+00 -1.04800737e+00 -4.46605414e-01 5.12527823e-01
8.33254755e-01 3.01760703e-01 8.28588009e-01 4.46982205e-01
-5.56269050e-01 4.38538224e-01 -3.99969280e-01 3.25138956e-01
8.77865076e-01 -1.15029466e+00 7.06260562e-01 5.47994733e-01
1.37493595e-01 5.04801393e-01 8.05526495e-01 -8.74463916e-01
-1.21393836e+00 -8.81649911e-01 1.56142759e+00 -8.45883012e-01
7.55240798e-01 -5.80090821e-01 -8.78689706e-01 9.35449958e-01
6.92291915e-01 -5.53171098e-01 1.09435844e+00 2.39650130e-01
3.08530089e-02 5.13763249e-01 -9.80854273e-01 5.88694632e-01
1.30951381e+00 -5.03885686e-01 -9.24488783e-01 5.10084808e-01
8.05609703e-01 -5.17331362e-01 -1.13287318e+00 1.91514254e-01
3.45025182e-01 -7.77532935e-01 9.12048399e-01 -6.54665649e-01
3.44921440e-01 1.38047442e-01 9.75801516e-03 -7.33083665e-01
-1.31470123e-02 -1.08096051e+00 -4.79319334e-01 2.00852823e+00
4.32119817e-01 -5.60459256e-01 7.13547409e-01 3.71728927e-01
-4.82302457e-02 -6.02485895e-01 -7.18982041e-01 -4.32669252e-01
-4.21187341e-01 -8.99430931e-01 5.68296015e-01 9.18845177e-01
1.95135832e-01 8.12738836e-01 -2.65385091e-01 -2.92697474e-02
4.73685950e-01 3.13026816e-01 7.61872649e-01 -1.26861465e+00
-3.76432806e-01 -2.40211248e-01 2.09284678e-01 -7.02322066e-01
1.09593950e-01 -6.64133966e-01 -7.51186833e-02 -1.63862193e+00
2.45312974e-01 -1.31529868e-01 6.48261234e-02 4.21889126e-01
-2.00634062e-01 -5.07676229e-03 -1.05766661e-01 1.27110213e-01
-7.08232522e-01 1.70772716e-01 7.67336309e-01 4.19965714e-01
-4.18085128e-01 1.18772052e-01 -6.03685737e-01 6.73345864e-01
1.16501510e+00 -7.06538498e-01 -6.86151981e-02 3.70649606e-01
2.91768670e-01 3.30565631e-01 3.02672744e-01 -5.91734946e-01
1.62238404e-01 -1.70188516e-01 6.33121356e-02 -9.56004381e-01
3.93406659e-01 -2.83671200e-01 2.83746600e-01 1.91637367e-01
-4.25086975e-01 5.03235543e-03 1.98002294e-01 6.79797590e-01
-1.36132300e-01 -3.92104238e-01 3.64011765e-01 -4.41163301e-01
-7.06690431e-01 -1.31277412e-01 -6.87908709e-01 5.92716396e-01
1.17868996e+00 -3.47375907e-02 -4.51355964e-01 -3.27937484e-01
-1.01135135e+00 2.79935822e-02 9.36707929e-02 5.11438727e-01
1.51887536e-01 -9.31176007e-01 -8.05321753e-01 -3.27625722e-01
1.32899255e-01 -2.26909176e-01 3.53127152e-01 6.38331115e-01
-9.25474018e-02 3.29496056e-01 1.71941116e-01 -2.75011063e-01
-1.54350734e+00 3.72071505e-01 -2.09195495e-01 -6.58918142e-01
-7.85254657e-01 8.16464722e-01 -3.08886141e-01 -3.19182158e-01
4.87220377e-01 -3.21060978e-02 -1.54030696e-01 6.72311485e-01
5.72187304e-01 5.91094732e-01 1.48990184e-01 -1.65331453e-01
-5.00295222e-01 7.89725333e-02 -8.35745633e-02 -3.66297752e-01
1.29507899e+00 -1.61268532e-01 1.41999926e-02 5.14273822e-01
1.08939373e+00 5.25340103e-02 -8.43421876e-01 -2.50263929e-01
4.56569076e-01 -1.05074011e-01 -6.24675691e-01 -7.69740403e-01
-3.06149095e-01 4.00225222e-01 9.71761197e-02 6.76462471e-01
9.04675245e-01 5.06403923e-01 1.13717246e+00 1.90765381e-01
2.37335607e-01 -1.10884452e+00 2.39873260e-01 7.72839785e-01
6.11768603e-01 -1.06080413e+00 -2.74219543e-01 -5.76664686e-01
-6.71178937e-01 9.16482985e-01 4.19313729e-01 9.66081694e-02
5.82369983e-01 4.08398479e-01 -1.40942708e-01 -6.74858212e-01
-9.22895849e-01 -3.44819844e-01 -1.78433821e-01 9.46943909e-02
7.51201689e-01 4.25642217e-03 -8.08232605e-01 7.69720852e-01
-4.62660372e-01 9.60943624e-02 4.47802991e-01 1.52619886e+00
-9.64019373e-02 -1.41503823e+00 -1.82357371e-01 4.32820290e-01
-9.27069664e-01 -8.06077272e-02 -8.11608493e-01 8.57131183e-01
1.38976658e-02 1.21660221e+00 -5.97410500e-02 -2.23467574e-01
5.71144640e-01 7.90980279e-01 3.44714999e-01 -7.08122790e-01
-1.18140841e+00 -1.08350702e-01 7.68775702e-01 -3.79812837e-01
-3.87374938e-01 -9.78037119e-01 -1.36553979e+00 -3.54701400e-01
-1.67430088e-01 5.26300073e-01 4.92716104e-01 7.03111351e-01
5.95163584e-01 7.76433229e-01 2.60927200e-01 -6.47224426e-01
-2.62613744e-01 -1.16261506e+00 -1.55066058e-01 4.61174548e-01
-2.52567232e-01 -3.73379767e-01 -4.72547203e-01 1.84837461e-01] | [9.086450576782227, 9.260741233825684] |
9bd20296-1059-4047-a5f3-3eb44415f461 | graph-combined-coreference-resolution-methods | null | null | https://aclanthology.org/2022.dialdoc-1.8 | https://aclanthology.org/2022.dialdoc-1.8.pdf | Graph-combined Coreference Resolution Methods on Conversational Machine Reading Comprehension with Pre-trained Language Model | Coreference resolution such as for anaphora has been an essential challenge that is commonly found in conversational machine reading comprehension (CMRC). This task aims to determine the referential entity to which a pronoun refers on the basis of contextual information. Existing approaches based on pre-trained language models (PLMs) mainly rely on an end-to-end method, which still has limitations in clarifying referential dependency. In this study, a novel graph-based approach is proposed to integrate the coreference of given text into graph structures (called coreference graphs), which can pinpoint a pronoun’s referential entity. We propose two graph-combined methods, evidence-enhanced and the fusion model, for CMRC to integrate coreference graphs from different levels of the PLM architecture. Evidence-enhanced refers to textual level methods that include an evidence generator (for generating new text to elaborate a pronoun) and enhanced question (for rewriting a pronoun in a question) as PLM input. The fusion model is a structural level method that combines the PLM with a graph neural network. We evaluated these approaches on a CoQA pronoun-containing dataset and the whole CoQA dataset. The result showed that our methods can outperform baseline PLM methods with BERT and RoBERTa. | ['Kazunori Komatani', 'Zhaodong Wang'] | null | null | null | null | dialdoc-acl-2022-5 | ['machine-reading-comprehension'] | ['natural-language-processing'] | [ 5.22699594e-01 8.10644746e-01 -4.29860651e-02 -4.33985084e-01
-1.06437969e+00 -5.43754160e-01 1.07140708e+00 5.75156987e-01
-4.15646762e-01 7.18288302e-01 9.07429039e-01 -3.93029600e-01
-3.27261806e-01 -8.88121247e-01 -6.13459945e-01 -1.74375311e-01
1.71582446e-01 1.14941621e+00 4.73820686e-01 -8.12943101e-01
5.48776090e-01 3.20253551e-01 -1.28221750e+00 7.42924988e-01
1.07630539e+00 2.28012294e-01 3.26941967e-01 6.12649739e-01
-8.09141517e-01 1.03521967e+00 -8.63469958e-01 -8.14715981e-01
-5.09159923e-01 -7.07821667e-01 -1.55352366e+00 -5.02047300e-01
5.68636656e-01 1.43698528e-01 -1.54519692e-01 1.11102843e+00
5.50774455e-01 2.48137414e-01 7.26099849e-01 -1.16672349e+00
-2.97489256e-01 1.10324419e+00 -3.14114749e-01 4.57606941e-01
1.08691788e+00 -2.37998873e-01 1.27346194e+00 -4.85641092e-01
1.05204988e+00 2.06861496e+00 3.30257028e-01 7.84204245e-01
-8.34249079e-01 -2.50805974e-01 1.13902874e-01 8.88461411e-01
-7.81562924e-01 -3.55978161e-01 9.36150312e-01 -1.34790629e-01
1.56601179e+00 4.12468940e-01 1.85778216e-01 1.30687332e+00
5.30228794e-01 7.14192092e-01 7.89217591e-01 -8.30531597e-01
-6.63681626e-02 -5.26616454e-01 6.44647002e-01 6.17173851e-01
1.44392297e-01 -1.17986118e-02 -7.95803308e-01 -2.44744524e-01
2.00087368e-01 -7.74701536e-01 -6.19069517e-01 -9.91903618e-02
-9.38490868e-01 1.02812290e+00 4.90519732e-01 5.55197179e-01
-4.67499554e-01 -1.70208693e-01 5.04113436e-01 2.36407474e-01
-4.09626514e-02 5.34322977e-01 -5.25364699e-03 7.32114911e-02
-5.12706518e-01 4.39658016e-01 1.36480796e+00 9.92580116e-01
4.19480383e-01 -5.36638558e-01 -6.61527157e-01 7.33868301e-01
5.17219961e-01 3.53936881e-01 2.79924333e-01 -1.06731451e+00
1.14740801e+00 1.06410730e+00 -6.79190308e-02 -1.17074764e+00
-7.48032987e-01 -2.36563340e-01 -4.99940991e-01 -1.68216482e-01
2.75014371e-01 7.09090829e-02 -4.00784612e-01 1.96283007e+00
4.32970434e-01 -2.23684553e-02 6.08739913e-01 1.18020260e+00
1.59908664e+00 6.39426410e-01 3.36196065e-01 -3.36076558e-01
1.85729623e+00 -1.11730576e+00 -1.28323197e+00 -3.59354049e-01
5.02803028e-01 -6.38046503e-01 8.59981894e-01 1.98238697e-02
-1.25288785e+00 -2.62738258e-01 -1.23783994e+00 -5.29907882e-01
-2.31404483e-01 -6.15873456e-01 1.58270866e-01 1.07299000e-01
-9.06967640e-01 4.51178402e-01 -3.25852513e-01 -6.82752490e-01
-1.54572457e-01 1.37591109e-01 -5.70997298e-01 5.03929369e-02
-1.77656269e+00 1.49060822e+00 7.32091844e-01 1.70260817e-02
-6.13105953e-01 -1.64531499e-01 -1.12989640e+00 3.18180442e-01
6.47350252e-01 -1.09790969e+00 1.52308333e+00 -7.89702058e-01
-1.36600268e+00 1.02216673e+00 -3.43004286e-01 -7.15735555e-01
1.30592927e-01 -1.86945006e-01 -4.21937019e-01 4.62425590e-01
2.95468867e-01 5.31586528e-01 6.02667034e-01 -1.35053611e+00
-6.66078448e-01 -4.78217572e-01 3.48137259e-01 7.09496975e-01
5.16879678e-01 2.39247486e-01 -3.58758152e-01 -2.59877473e-01
1.21703729e-01 -4.29071456e-01 2.57433414e-01 -9.18372989e-01
-4.69191343e-01 -9.95887280e-01 6.68630302e-01 -7.13846147e-01
1.40341556e+00 -1.53953063e+00 6.87469602e-01 -1.41198590e-01
2.82709301e-01 4.04211730e-01 -4.86269861e-01 8.99463475e-01
-7.09559619e-02 -1.92147978e-02 -2.46815160e-01 -1.57115445e-01
1.44701540e-01 3.89646977e-01 -3.03576767e-01 -2.19107140e-02
3.77080232e-01 1.00734031e+00 -1.12763691e+00 -8.29348147e-01
-3.59063260e-02 2.23728687e-01 -2.71029532e-01 6.44473851e-01
-6.18662238e-01 2.18892589e-01 -2.92268366e-01 3.06353867e-01
6.48880064e-01 -9.15154815e-02 6.67914391e-01 -4.75205243e-01
1.66005790e-01 6.66812658e-01 -1.05121613e+00 1.87362301e+00
-2.27986649e-01 3.07947129e-01 2.21923649e-01 -9.61629450e-01
9.37125325e-01 5.36207199e-01 -4.02206540e-01 -7.46017992e-01
2.43715256e-01 2.25142539e-01 2.83954948e-01 -7.69924521e-01
4.67863142e-01 -1.24412999e-01 -1.72915950e-01 3.79719704e-01
4.27307844e-01 -6.31593689e-02 4.86706853e-01 6.33524716e-01
1.21033394e+00 2.59505153e-01 5.69918990e-01 -2.75210261e-01
1.22531545e+00 2.89513081e-01 5.47730207e-01 6.27499580e-01
-2.04855464e-02 3.09248686e-01 9.56269622e-01 -1.25212967e-01
-4.07639623e-01 -8.69684815e-01 2.92491227e-01 1.05375445e+00
3.71758968e-01 -6.56038404e-01 -1.19187009e+00 -1.00739336e+00
-3.74618530e-01 1.41024685e+00 -3.46699297e-01 -4.78145741e-02
-1.15657842e+00 -1.98648974e-01 6.07983708e-01 3.54101181e-01
6.50608420e-01 -1.60076523e+00 -5.42966545e-01 3.60458404e-01
-9.07713830e-01 -1.15658844e+00 -5.43942273e-01 -1.51419103e-01
-6.18166745e-01 -1.53097928e+00 1.38127552e-02 -9.87108171e-01
3.07674378e-01 2.37324946e-02 1.61786985e+00 4.24668998e-01
2.24975824e-01 5.13143718e-01 -3.44857365e-01 -5.07210433e-01
-8.12852263e-01 1.43080518e-01 -4.86386955e-01 -2.27204621e-01
7.00359821e-01 -5.13001204e-01 -3.24461102e-01 1.03612043e-01
-7.04454660e-01 2.61983305e-01 4.80889171e-01 7.25830436e-01
3.03187251e-01 -7.43794084e-01 6.91585124e-01 -1.11348295e+00
1.32742763e+00 -4.43607420e-01 -4.30199891e-01 6.50028765e-01
-3.86402518e-01 4.34670746e-01 1.49488822e-01 -4.64230962e-02
-1.55620527e+00 -4.91020054e-01 -2.09737644e-01 1.33768305e-01
-4.04458568e-02 7.94051528e-01 -5.19201159e-01 5.43533981e-01
7.72967458e-01 -2.79870421e-01 5.40826470e-02 -3.43424648e-01
6.10317945e-01 5.91515183e-01 1.02330232e+00 -9.46173489e-01
3.55838984e-01 -1.45357132e-01 1.71968192e-01 -4.41728294e-01
-9.57014143e-01 -6.39744222e-01 -7.30009496e-01 -2.90847242e-01
1.13206279e+00 -4.96666908e-01 -8.32582235e-01 -1.70397665e-02
-1.99559855e+00 9.54010803e-03 5.74093908e-02 2.56659508e-01
-6.45784318e-01 6.80824935e-01 -6.15686893e-01 -5.44358790e-01
-8.94308388e-01 -8.77938688e-01 1.13411736e+00 4.02014166e-01
-5.46325862e-01 -1.22371471e+00 3.38829070e-01 8.25965762e-01
2.41742358e-01 3.49253088e-01 1.43142331e+00 -1.18677557e+00
-3.83851767e-01 1.24234483e-01 -3.23510826e-01 -1.01516165e-01
-1.99818119e-01 -4.83983666e-01 -5.80139399e-01 -1.33579165e-01
1.27455026e-01 -2.79096693e-01 5.76243460e-01 -9.45406482e-02
2.03717351e-01 -4.75320846e-01 -4.90667969e-01 1.39502145e-03
1.18447912e+00 1.49602726e-01 8.16528916e-01 3.85662198e-01
4.98313874e-01 1.00382316e+00 7.01702654e-01 -3.75560492e-01
7.82131493e-01 7.14612246e-01 5.63803613e-01 2.43417159e-01
-6.44836426e-01 -3.77495050e-01 2.72364587e-01 1.05520999e+00
-2.71670341e-01 -6.50862515e-01 -1.07276642e+00 3.87148678e-01
-2.29233837e+00 -1.12065506e+00 -6.04776800e-01 1.68370020e+00
1.10456538e+00 9.98437181e-02 -4.81365204e-01 -1.14736624e-01
1.04637480e+00 1.46007165e-01 -1.67176485e-01 -7.20489562e-01
-4.35514331e-01 3.19243461e-01 -2.13712350e-01 1.22237551e+00
-7.55386472e-01 1.16567242e+00 5.10678673e+00 5.24064600e-01
-4.29454237e-01 1.28349274e-01 -3.07379872e-01 6.93185627e-01
-4.03852582e-01 2.85911351e-01 -7.62107849e-01 -1.96693778e-01
1.03844297e+00 -2.23549828e-01 2.23732084e-01 3.17991257e-01
2.66560800e-02 -1.64769128e-01 -1.32688892e+00 7.20639646e-01
5.87011099e-01 -1.43193233e+00 5.73489428e-01 -4.15449142e-01
1.20168544e-01 -3.24732751e-01 -8.50884914e-01 5.84720612e-01
1.78273201e-01 -9.42569613e-01 3.20013344e-01 8.44220817e-01
2.34133571e-01 -4.53926265e-01 1.27027488e+00 5.15726864e-01
-1.02839911e+00 2.57399857e-01 -1.99870989e-01 1.48638012e-02
5.23782134e-01 -3.33457738e-01 -1.06795847e+00 1.11085522e+00
2.43047878e-01 2.20761597e-01 -4.75695908e-01 6.42463207e-01
-1.05851889e+00 5.01116097e-01 1.30553216e-01 -3.33010554e-01
1.53522030e-01 -1.72373012e-01 1.07609713e+00 1.54760981e+00
-1.71271890e-01 4.48460788e-01 -8.54003355e-02 9.86121893e-01
-2.30032951e-01 3.97262394e-01 -3.71791422e-01 3.92582476e-01
6.87373042e-01 1.36223626e+00 -1.55696452e-01 -3.33086073e-01
-4.35146511e-01 8.41381788e-01 6.62394702e-01 1.83527604e-01
-4.71246123e-01 -4.73612815e-01 -1.76234022e-01 -4.48530763e-02
-2.05242455e-01 1.17114112e-01 2.40176186e-01 -8.86547148e-01
-6.04496663e-03 -1.30035162e+00 1.14465952e+00 -1.01453471e+00
-1.36596727e+00 6.57730043e-01 3.54249775e-01 -6.46083355e-01
-6.73004627e-01 -4.92400229e-01 -1.10333586e+00 1.15143490e+00
-1.55996442e+00 -1.28994477e+00 -5.08647084e-01 6.42508090e-01
7.06253648e-01 -2.04190671e-01 1.07997596e+00 -1.69483513e-01
-2.99839616e-01 2.74526209e-01 -9.26139176e-01 2.45972157e-01
6.86461210e-01 -1.51731360e+00 1.91328526e-01 9.65730488e-01
1.40105456e-01 7.60336757e-01 7.55946934e-01 -7.53683269e-01
-1.48018968e+00 -7.22243726e-01 1.75000405e+00 -5.79614162e-01
5.70185304e-01 -1.01930320e-01 -1.44404769e+00 7.66824603e-01
1.06383228e+00 -7.41390288e-01 4.58944917e-01 1.88527375e-01
-3.06549788e-01 1.84939280e-01 -1.07268286e+00 5.51852345e-01
1.28966689e+00 -3.75382930e-01 -2.05590558e+00 2.96304613e-01
9.93105710e-01 -6.44529104e-01 -7.62507677e-01 6.35127902e-01
-2.19762266e-01 -7.83280909e-01 6.33509576e-01 -1.09622359e+00
3.28203142e-01 -3.46152067e-01 -2.26854295e-01 -1.14110374e+00
-1.90162510e-01 -6.51641369e-01 -2.83343673e-01 1.46091831e+00
5.42666674e-01 -4.53598738e-01 9.14750472e-02 3.56600821e-01
-3.70571852e-01 -1.86989725e-01 -1.18030012e+00 -3.51251036e-01
1.57102831e-02 -1.88909531e-01 6.37335777e-01 7.62243867e-01
6.15333319e-01 1.59253383e+00 2.58104771e-01 2.84814805e-01
5.99208713e-01 1.89969882e-01 6.70446634e-01 -1.45949435e+00
5.66963032e-02 -6.12106264e-01 -6.95256889e-02 -8.40396404e-01
7.92768359e-01 -1.31552470e+00 -9.74143520e-02 -2.19715929e+00
3.59049916e-01 2.36952022e-01 -1.79449786e-02 3.09675574e-01
-4.75042582e-01 -6.49556637e-01 1.86308503e-01 1.80037647e-01
-7.73014188e-01 4.02480215e-01 1.24364388e+00 -3.73601258e-01
-1.24376856e-01 -2.50387162e-01 -5.09448946e-01 7.92474508e-01
5.33389866e-01 -2.26801321e-01 -3.96303743e-01 -3.43427330e-01
2.59413213e-01 6.95430636e-01 3.27803135e-01 -5.37277639e-01
7.41289914e-01 -1.98228117e-02 -3.40720862e-01 -7.80562818e-01
9.71774012e-02 -3.76260430e-01 -2.55304962e-01 5.40599048e-01
-4.76705521e-01 3.26044828e-01 1.41297281e-01 4.46398109e-01
-4.95090693e-01 -6.73068047e-01 3.15590352e-01 -2.10412472e-01
-8.03764880e-01 -3.13207239e-01 -1.53126344e-01 5.00976324e-01
4.40896571e-01 3.85767877e-01 -8.83777559e-01 -5.35439491e-01
-6.20230496e-01 5.51938415e-01 -2.67124981e-01 7.26586342e-01
8.11476529e-01 -1.19633996e+00 -1.14048409e+00 -4.76669967e-01
1.35572359e-01 5.31327389e-02 -2.62266584e-02 6.96527958e-01
-4.38028365e-01 6.51454389e-01 -1.31298393e-01 -5.09965003e-01
-1.49464834e+00 7.46135175e-01 5.48730254e-01 -6.06983364e-01
-4.94059920e-01 5.15911162e-01 -4.50994447e-02 -5.96851110e-01
3.53000015e-01 -7.78395236e-02 -1.15260315e+00 2.15291515e-01
5.34949303e-01 2.62420893e-01 1.87653840e-01 -8.33414853e-01
-5.04817128e-01 4.87199187e-01 6.19105771e-02 -2.55220622e-01
9.61053610e-01 -1.15966447e-01 -6.98425233e-01 2.31584772e-01
6.83760107e-01 7.65460357e-02 -4.16620523e-01 -4.59743172e-01
5.87767839e-01 1.86737806e-01 -3.00421506e-01 -1.05689180e+00
-4.13055003e-01 7.90966809e-01 2.23038077e-01 9.02198479e-02
8.21876824e-01 3.64106327e-01 6.05832219e-01 5.84219277e-01
3.74125272e-01 -9.61362064e-01 4.13329788e-02 7.62583971e-01
1.79166555e+00 -1.03579855e+00 -1.69866905e-01 -6.82683229e-01
-6.41099155e-01 1.53986597e+00 8.37529361e-01 2.79884607e-01
7.87263960e-02 -1.55447759e-02 1.83740631e-01 -6.58151805e-01
-9.93935406e-01 -3.18617404e-01 5.83862782e-01 5.36749125e-01
5.74923158e-01 3.93628292e-02 -1.17284560e+00 1.02984881e+00
-3.55171978e-01 -5.51946819e-01 4.03216004e-01 7.92170167e-01
-5.62719941e-01 -1.32108343e+00 -4.49567735e-01 -5.97827435e-02
-1.99341506e-01 -1.84530810e-01 -8.15195262e-01 1.02246833e+00
-1.00934103e-01 1.50171566e+00 -4.70971130e-03 -1.14932209e-01
9.24075603e-01 3.12949270e-01 6.48416400e-01 -5.78476489e-01
-8.17901433e-01 -2.50941694e-01 9.00955141e-01 -6.71563447e-01
-8.72958124e-01 -6.91655695e-01 -1.80609751e+00 -6.45126449e-03
-3.28472316e-01 5.33652723e-01 3.03831637e-01 1.19202840e+00
2.46998832e-01 4.92499560e-01 5.29326126e-02 -2.64155239e-01
-5.79811096e-01 -1.38900697e+00 1.09091140e-01 8.72116029e-01
-1.58760473e-02 -6.24527633e-01 -3.58302027e-01 -1.81563750e-01] | [9.413331031799316, 9.28394603729248] |
77e610e6-8aee-43b5-b549-bd72ec315db2 | triangular-contrastive-learning-on-molecular | 2205.13279 | null | https://arxiv.org/abs/2205.13279v1 | https://arxiv.org/pdf/2205.13279v1.pdf | Triangular Contrastive Learning on Molecular Graphs | Recent contrastive learning methods have shown to be effective in various tasks, learning generalizable representations invariant to data augmentation thereby leading to state of the art performances. Regarding the multifaceted nature of large unlabeled data used in self-supervised learning while majority of real-word downstream tasks use single format of data, a multimodal framework that can train single modality to learn diverse perspectives from other modalities is an important challenge. In this paper, we propose TriCL (Triangular Contrastive Learning), a universal framework for trimodal contrastive learning. TriCL takes advantage of Triangular Area Loss, a novel intermodal contrastive loss that learns the angular geometry of the embedding space through simultaneously contrasting the area of positive and negative triplets. Systematic observation on embedding space in terms of alignment and uniformity showed that Triangular Area Loss can address the line-collapsing problem by discriminating modalities by angle. Our experimental results also demonstrate the outperformance of TriCL on downstream task of molecular property prediction which implies that the advantages of the embedding space indeed benefits the performance on downstream tasks. | ['Sun Kim', 'Yijingxiu Lu', 'Wonseok Shin', 'MinGyu Choi'] | 2022-05-26 | null | null | null | null | ['molecular-property-prediction'] | ['miscellaneous'] | [ 5.15055418e-01 -7.94453919e-02 -5.68380058e-01 -3.82560074e-01
-8.41629565e-01 -9.17123497e-01 9.94603336e-01 3.09846401e-01
-5.74843168e-01 9.08339739e-01 1.99225366e-01 -2.32727453e-01
-2.81729102e-01 -4.32880193e-01 -8.29922318e-01 -1.15159428e+00
-1.42067686e-01 2.65173972e-01 -1.12327732e-01 -5.15795112e-01
2.66422123e-01 5.71431994e-01 -1.62208700e+00 5.41545093e-01
6.28832698e-01 1.11211741e+00 -9.67279002e-02 3.56453598e-01
-8.63051787e-02 2.84611702e-01 -1.18733957e-01 -3.85707647e-01
4.18316156e-01 -3.98970336e-01 -6.42836869e-01 -4.10650522e-01
8.94028366e-01 1.84735835e-01 -2.33124375e-01 7.64632523e-01
7.27961719e-01 -1.90602154e-01 1.26708198e+00 -1.48454750e+00
-6.97968543e-01 2.98440456e-01 -5.17268777e-01 3.41781586e-01
3.61545891e-01 1.97350278e-01 1.58059311e+00 -1.16620529e+00
7.44214892e-01 1.04400814e+00 5.77189445e-01 5.20281792e-01
-1.56796980e+00 -6.29860580e-01 1.20735407e-01 3.83083552e-01
-1.12381220e+00 -3.16895187e-01 8.69409204e-01 -5.63954234e-01
8.30718279e-01 3.09529245e-01 2.98223615e-01 1.22833633e+00
2.47095004e-01 5.40197849e-01 1.59878385e+00 -4.94437873e-01
-9.28023010e-02 1.80344760e-01 1.07562728e-02 7.08523571e-01
7.66124353e-02 2.85824418e-01 -1.00685644e+00 -6.34544268e-02
5.08043647e-01 -2.41404487e-04 -2.04412326e-01 -8.77378106e-01
-1.33815801e+00 8.19362760e-01 6.28177822e-01 2.21378326e-01
2.23872643e-02 -2.06949577e-01 4.73088354e-01 6.64187133e-01
2.36892208e-01 5.11834264e-01 -5.95347464e-01 1.67844296e-01
-5.55013001e-01 -5.43593727e-02 3.63655537e-01 7.07528532e-01
6.25561059e-01 -3.05230856e-01 -3.40731978e-01 6.66744471e-01
2.94921190e-01 4.82934862e-01 4.53662604e-01 -5.00401735e-01
7.10995674e-01 7.49095023e-01 -3.47115993e-01 -6.20595515e-01
-6.93992555e-01 -6.07387125e-01 -9.02410686e-01 3.74587566e-01
7.37493038e-01 1.46814436e-01 -6.83358788e-01 2.15751863e+00
2.55967081e-01 -2.13892564e-01 6.15965575e-02 7.91579664e-01
7.93221176e-01 4.53901678e-01 4.52313662e-01 -1.77622736e-01
1.35170972e+00 -6.91316426e-01 -3.71072322e-01 1.82036255e-02
6.85632765e-01 -6.58850849e-01 1.08797097e+00 3.55687678e-01
-9.09108281e-01 -4.15356666e-01 -1.19646955e+00 -4.34572577e-01
-8.46544981e-01 1.60375552e-03 7.82548308e-01 5.39806902e-01
-8.29773366e-01 5.07821381e-01 -3.42491746e-01 -1.85752600e-01
6.11820757e-01 5.72162271e-01 -9.36120808e-01 -9.05429274e-02
-1.19802141e+00 8.93726587e-01 3.83199543e-01 -7.46325254e-02
-5.55183649e-01 -8.47319365e-01 -8.35236311e-01 1.34126812e-01
5.92196174e-02 -5.73773563e-01 5.12829721e-01 -8.81095648e-01
-1.35774338e+00 1.06728840e+00 2.50932664e-01 -1.86570346e-01
5.26664495e-01 6.48676604e-02 -2.83604950e-01 2.56040096e-01
-1.08376570e-01 9.44643378e-01 9.56548333e-01 -1.01686633e+00
-2.69927382e-01 -7.90843189e-01 -8.12667683e-02 3.91017795e-01
-8.14009309e-01 -7.35332310e-01 1.76461682e-01 -5.05608559e-01
1.62059531e-01 -8.80972147e-01 1.83367923e-01 3.12588632e-01
-3.04622680e-01 -2.42411301e-01 8.41437936e-01 -1.98325291e-01
1.01931560e+00 -2.17836690e+00 5.14888644e-01 2.71084189e-01
3.26559812e-01 1.06490999e-01 -5.74442387e-01 8.21221352e-01
-5.38590729e-01 -8.28513950e-02 -1.76077485e-01 4.05946188e-02
1.47775099e-01 -1.55201510e-01 -2.05458462e-01 7.15466440e-01
4.63339865e-01 1.01997638e+00 -7.67308354e-01 -4.86293972e-01
1.73523962e-01 5.89656413e-01 -4.95508045e-01 1.99777380e-01
-5.08969016e-02 5.82496345e-01 -1.41948611e-01 6.94740057e-01
6.88461602e-01 -2.72313923e-01 2.84682244e-01 -7.44869292e-01
-1.47119477e-01 -1.94286965e-02 -7.47848094e-01 1.84328866e+00
-3.67722541e-01 6.12908721e-01 -1.95317775e-01 -1.32573748e+00
7.93976068e-01 1.43442973e-01 3.32639217e-01 -9.55241621e-01
-8.54075328e-02 3.32378656e-01 1.64136574e-01 -3.51309091e-01
5.64723611e-02 -4.49202150e-01 -1.18407020e-02 2.12525874e-01
4.73818839e-01 -4.18426323e-04 -8.95267501e-02 5.26216887e-02
7.71212399e-01 2.17944682e-01 4.39389974e-01 -3.77653837e-01
7.73894250e-01 -2.76551336e-01 1.17150374e-01 5.76622427e-01
-3.06316972e-01 5.43652713e-01 8.08565497e-01 -2.09663868e-01
-1.10806227e+00 -1.25658655e+00 -6.02956891e-01 1.41369045e+00
1.14191711e-01 -1.34992227e-01 -1.62765816e-01 -7.69736707e-01
1.72700748e-01 2.62160361e-01 -8.16240191e-01 -3.53694648e-01
-2.47553617e-01 -1.09237266e+00 4.81242001e-01 4.38335150e-01
3.12380791e-01 -7.14772522e-01 -1.93103224e-01 -2.38454953e-01
1.94122538e-01 -9.11102712e-01 -3.46212775e-01 6.11128032e-01
-9.77427602e-01 -1.05743539e+00 -7.04450309e-01 -9.23227370e-01
5.14424622e-01 1.34236231e-01 8.88616920e-01 -3.63882869e-01
-3.99425328e-01 4.63320404e-01 -8.83118436e-02 -1.17259182e-01
-1.25795886e-01 2.12131977e-01 -7.77163878e-02 4.00766432e-02
2.94277072e-01 -7.34805465e-01 -8.47300351e-01 1.25581086e-01
-1.03294289e+00 -1.18272208e-01 8.19287241e-01 1.27027619e+00
5.47120571e-01 -6.48884952e-01 8.91229331e-01 -9.24405634e-01
6.10844076e-01 -5.40642798e-01 -2.58109838e-01 4.05778527e-01
-8.71731997e-01 4.91350979e-01 5.46695173e-01 -3.45528841e-01
-7.24424362e-01 -8.24508667e-02 1.54961441e-02 -6.08035512e-02
-4.01804335e-02 4.99688953e-01 -1.68171152e-01 -2.16352984e-01
6.27787471e-01 2.61449099e-01 2.83597678e-01 -2.66260356e-01
5.77780128e-01 3.93642038e-01 2.66445935e-01 -6.34657383e-01
6.15433455e-01 5.09465635e-01 6.59404814e-01 -7.84761429e-01
-7.00565875e-01 -4.17345166e-01 -8.88148844e-01 -1.49479527e-02
8.30760419e-01 -9.19272721e-01 -9.71500933e-01 3.68117094e-01
-7.09407568e-01 1.24866471e-01 -1.01599999e-01 3.92303795e-01
-5.09076416e-01 4.08481628e-01 -4.10748720e-01 -4.68022406e-01
-2.61383772e-01 -1.01425600e+00 1.20476854e+00 3.44916880e-02
8.26496780e-02 -1.20214796e+00 3.50394160e-01 4.05777484e-01
3.78107488e-01 2.97831804e-01 1.46435750e+00 -8.71846437e-01
-3.89505446e-01 -1.28460273e-01 -2.78188378e-01 2.19913572e-01
5.80373667e-02 -2.05032513e-01 -1.21279931e+00 -3.91953796e-01
-4.93677825e-01 -7.25073874e-01 1.16402125e+00 6.40339926e-02
1.03363180e+00 5.30275889e-02 -2.04921201e-01 5.81744373e-01
1.55795968e+00 -1.46536469e-01 4.43572342e-01 3.59592050e-01
7.23191202e-01 7.90186942e-01 4.06093448e-01 2.50696272e-01
1.72374472e-01 7.05178678e-01 5.30957222e-01 -2.26909250e-01
-6.99724481e-02 -2.92428493e-01 3.91502172e-01 6.44411147e-01
1.60963144e-02 -2.42988616e-02 -5.99245429e-01 1.35751456e-01
-1.58211756e+00 -1.07148170e+00 4.82670292e-02 2.30888891e+00
1.01103580e+00 1.76596522e-01 2.67476916e-01 -2.01011933e-02
4.39678520e-01 2.19596446e-01 -6.87993050e-01 -3.84008437e-01
-5.78858316e-01 2.46904761e-01 3.64543557e-01 3.36007595e-01
-1.06496644e+00 6.25079036e-01 6.03175783e+00 1.02044415e+00
-1.39849126e+00 2.23355070e-02 6.46976709e-01 -3.33798230e-02
-5.88200867e-01 -3.24167728e-01 -6.89406633e-01 3.88713539e-01
7.47132540e-01 2.14046955e-01 1.63945377e-01 3.60740066e-01
-1.61089703e-01 5.59707172e-02 -1.47207963e+00 1.12504530e+00
9.18443650e-02 -1.40166450e+00 3.12078893e-01 6.59276620e-02
6.99630141e-01 1.86038949e-02 7.23153591e-01 4.08570886e-01
-5.79522014e-01 -1.28542852e+00 3.49523246e-01 5.10775328e-01
1.00350761e+00 -6.05730653e-01 5.95331728e-01 2.46806834e-02
-9.74897981e-01 -1.34510055e-01 -1.35144055e-01 1.20084442e-01
-2.64493138e-01 2.00844482e-01 -7.40272760e-01 4.63088810e-01
2.89461702e-01 8.23132813e-01 -8.34176183e-01 8.30813825e-01
9.45236012e-02 1.60905272e-01 -7.69432858e-02 5.78001561e-03
2.31458470e-01 -4.56107885e-01 4.77556586e-01 1.23351932e+00
4.71748970e-03 -4.51016933e-01 -9.17510465e-02 6.39090955e-01
-2.09390879e-01 4.02643323e-01 -8.98203671e-01 -3.59969884e-01
3.46498847e-01 1.23979449e+00 -3.21943760e-01 4.82543521e-02
-5.61858892e-01 8.46807301e-01 4.99113172e-01 2.81452656e-01
-6.46877587e-01 -2.87289739e-01 3.84233534e-01 -5.99950477e-02
3.97454768e-01 -5.62674031e-02 -3.87931556e-01 -1.01771879e+00
-1.06012820e-04 -8.39791119e-01 6.07007921e-01 -4.82241690e-01
-1.81192219e+00 5.74930608e-01 -1.43443514e-03 -1.47694182e+00
1.74193144e-01 -1.24795389e+00 -2.81301290e-01 7.17188835e-01
-1.68472028e+00 -1.50725651e+00 2.12722388e-03 6.34404421e-01
1.75524935e-01 -4.69008386e-01 9.44248140e-01 2.54242718e-01
-4.11998481e-01 9.86012459e-01 3.48417759e-01 -2.50914395e-01
1.03673470e+00 -1.43492925e+00 -5.30923724e-01 1.61302164e-01
1.53600171e-01 8.07985604e-01 5.88648736e-01 -7.47370645e-02
-1.56738830e+00 -6.09997272e-01 6.43232644e-01 -5.01292109e-01
8.12037885e-01 -4.24504519e-01 -6.55865550e-01 3.50834429e-01
4.17048067e-01 1.97555572e-01 1.18541777e+00 3.35804760e-01
-1.04270124e+00 -3.25684577e-01 -1.08290219e+00 5.47148883e-01
1.02453434e+00 -8.60831678e-01 -3.85946870e-01 3.73988628e-01
3.99458647e-01 8.70173499e-02 -1.07933772e+00 5.17345548e-01
7.88956940e-01 -1.14943171e+00 1.00942743e+00 -9.95882154e-01
6.62416577e-01 -1.15702495e-01 -3.24994475e-01 -1.08783913e+00
9.11854580e-03 -2.34308481e-01 -4.92747352e-02 1.01631701e+00
7.41965294e-01 -8.21154773e-01 5.33379138e-01 2.64903188e-01
8.55725142e-04 -9.54215765e-01 -1.17162812e+00 -5.52658737e-01
5.54611146e-01 9.12799165e-02 2.88241031e-03 1.09589267e+00
2.48660475e-01 8.70795608e-01 -2.67841101e-01 -2.19424263e-01
6.03662252e-01 3.12355369e-01 3.79026145e-01 -1.26104009e+00
-3.10651928e-01 -6.17137134e-01 -5.76331377e-01 -8.79708707e-01
1.60189167e-01 -1.33039153e+00 -4.82862264e-01 -9.63709593e-01
4.78646308e-01 -2.07109362e-01 -6.85391068e-01 4.29251999e-01
-2.17794962e-02 5.33557892e-01 1.23631962e-01 1.46228224e-01
-6.79449379e-01 7.16780901e-01 1.37924814e+00 -4.07003701e-01
-9.59627982e-03 -3.12708706e-01 -6.89727724e-01 2.77230769e-01
5.45307636e-01 -1.86416298e-01 -5.14593840e-01 -7.98529610e-02
5.62728584e-01 -2.59455532e-01 3.83107722e-01 -5.78822672e-01
1.55894114e-02 6.74199313e-02 5.72475195e-01 -3.31697553e-01
5.77961922e-01 -8.37105453e-01 -3.11577350e-01 3.18390042e-01
-6.98419929e-01 8.07485208e-02 2.96307296e-01 8.18326831e-01
-2.60921538e-01 -7.38079324e-02 7.09974229e-01 1.34176388e-01
-5.97745657e-01 3.85918528e-01 4.97504771e-02 1.70273066e-01
9.47618008e-01 -2.26955488e-01 -5.41415930e-01 -5.45805655e-02
-9.07813668e-01 -7.38703236e-02 2.65609294e-01 3.26621681e-01
4.08122480e-01 -1.48067796e+00 -5.71501255e-01 3.27726275e-01
5.45956612e-01 -6.37686431e-01 3.55014920e-01 1.31851745e+00
-3.17719370e-01 5.73361874e-01 -6.75613046e-01 -7.61966527e-01
-1.28109288e+00 5.62929451e-01 2.06807509e-01 -1.89599648e-01
-2.21451536e-01 8.47399533e-01 4.53436285e-01 -5.66742480e-01
8.99502560e-02 1.02155231e-01 -3.76395643e-01 5.28546154e-01
2.25551352e-01 1.91276625e-01 2.35907748e-01 -6.49675131e-01
-4.90777582e-01 6.57190084e-01 -2.27488220e-01 -7.11378977e-02
1.45261931e+00 -4.65254635e-02 -1.22261956e-01 6.43660069e-01
1.67332113e+00 -5.89355938e-02 -1.23052740e+00 -1.78479508e-01
-2.13597212e-02 -1.47563934e-01 -2.03285858e-01 -9.99981761e-01
-6.24782145e-01 1.20754516e+00 1.03731585e+00 1.67097256e-01
1.00992274e+00 4.53030728e-02 3.48030746e-01 4.27162528e-01
6.75438046e-02 -1.03430271e+00 3.98576200e-01 3.31431061e-01
8.95264685e-01 -1.66055870e+00 3.18824910e-02 -3.13859254e-01
-6.38071537e-01 1.40478599e+00 4.89141703e-01 1.42902032e-01
6.09970808e-01 -2.08191469e-01 -1.24184424e-02 -3.39992195e-01
-6.92937791e-01 -8.64246115e-02 6.93122923e-01 6.13713443e-01
7.45136440e-01 -3.27836797e-02 -6.52139544e-01 2.55052924e-01
1.82359785e-01 -6.46440268e-01 6.20829277e-02 7.35046029e-01
-2.25444645e-01 -1.38606715e+00 -8.49285647e-02 3.77518862e-01
-2.68940419e-01 -2.31824040e-01 -5.07195055e-01 8.82606149e-01
1.37260348e-01 5.10410726e-01 -1.20296329e-01 -2.50447005e-01
1.54016271e-01 2.47504696e-01 9.89469826e-01 -4.01368707e-01
-3.82053643e-01 -3.92421596e-02 -1.60662666e-01 -2.55357265e-01
-6.07721210e-01 -6.49381936e-01 -9.47774827e-01 5.14388941e-02
-2.06407934e-01 -6.42356575e-02 5.64430356e-01 9.57064331e-01
4.52861249e-01 1.79823652e-01 6.83081031e-01 -7.71526277e-01
-8.90579283e-01 -8.04580569e-01 -5.90148389e-01 8.50614727e-01
6.16172254e-01 -9.99301732e-01 -5.33529997e-01 -1.06601991e-01] | [9.665116310119629, 2.8160207271575928] |
00fbc089-b134-4259-b4e9-a9dca0816e21 | using-gaze-for-behavioural-biometrics | null | null | https://www.mdpi.com/1424-8220/23/3/1262 | https://www.mdpi.com/1424-8220/23/3/1262 | Using Gaze for Behavioural Biometrics | A principled approach to the analysis of eye movements for behavioural biometrics is laid down. The approach grounds in foraging theory, which provides a sound basis to capture the uniqueness of individual eye movement behaviour. We propose a composite Ornstein-Uhlenbeck process for quantifying the exploration/exploitation signature characterising the foraging eye behaviour. The relevant parameters of the composite model, inferred from eye-tracking data via Bayesian analysis, are shown to yield a suitable feature set for biometric identification; the latter is eventually accomplished via a classical classification technique. A proof of concept of the method is provided by measuring its identification performance on a publicly available dataset. Data and code for reproducing the analyses are made available. Overall, we argue that the approach offers a fresh view on either the analyses of eye-tracking data and prospective applications in this field. | ['Giuseppe Boccignone', 'Vittorio Cuculo', 'Sathya Bursic', 'Sabrina Patania', 'Alessandro D’Amelio'] | 2023-01-22 | null | null | null | sensors-2023-1 | ['gaze-estimation'] | ['computer-vision'] | [ 3.19500148e-01 -6.77794814e-02 -5.71336411e-02 -2.45027632e-01
4.92186025e-02 -5.16181231e-01 8.00366521e-01 -2.07974672e-01
-8.52947295e-01 4.82961029e-01 -2.02269122e-01 -3.79566669e-01
-8.49678338e-01 -1.09740824e-01 -2.11036205e-01 -1.11887360e+00
-1.92976028e-01 -8.24724436e-02 -9.96390432e-02 -3.49903405e-02
5.56173682e-01 8.30462158e-01 -2.18750453e+00 -4.15957361e-01
2.97000200e-01 1.27084696e+00 1.21116892e-01 8.37683082e-01
2.69358933e-01 2.67051518e-01 -2.35676110e-01 -7.63826668e-01
4.45129544e-01 -5.22347152e-01 -5.83853602e-01 1.16581898e-02
3.06609329e-02 -3.24112773e-01 1.98477715e-01 9.67976093e-01
4.16128635e-01 1.54986465e-02 8.99340510e-01 -1.14753556e+00
-5.60127258e-01 -1.12160770e-02 -4.43058699e-01 4.67213839e-01
5.28647602e-01 3.33043903e-01 1.19797468e+00 -3.92492652e-01
3.89556706e-01 1.13620710e+00 6.38067842e-01 7.50986993e-01
-1.21708155e+00 -4.58343148e-01 -3.80658776e-01 -9.18982252e-02
-1.34968746e+00 -8.41232836e-01 3.33188266e-01 -6.45460665e-01
3.17034215e-01 5.44164002e-01 9.31914568e-01 1.06756532e+00
1.80003554e-01 7.11475492e-01 1.70717669e+00 -4.93093193e-01
3.27679336e-01 2.08827138e-01 3.07667434e-01 5.75719237e-01
6.58454716e-01 1.00706446e+00 -9.58069742e-01 -2.49642149e-01
6.55234218e-01 -1.23564057e-01 1.09052300e-01 -4.63318139e-01
-5.79737127e-01 6.43051326e-01 -2.92631462e-02 -1.45132065e-01
-5.42051494e-01 2.31977049e-02 -1.54055521e-01 4.27476257e-01
5.44890642e-01 7.99853280e-02 -3.46400023e-01 -3.71372849e-01
-9.48539734e-01 3.28727961e-01 8.61618042e-01 6.35766804e-01
4.69305515e-01 -4.75763083e-01 -9.48726162e-02 4.02457535e-01
1.16683495e+00 8.31374407e-01 2.78380394e-01 -9.18630004e-01
-4.80889410e-01 5.48550785e-01 4.00621176e-01 -3.97212476e-01
-6.25608265e-01 1.36292297e-02 -1.36269242e-01 4.56978202e-01
7.97659099e-01 -1.50791243e-01 -5.22248566e-01 1.60190082e+00
6.11180127e-01 -1.39815789e-02 -2.02690855e-01 5.30083120e-01
5.79969227e-01 -2.63243347e-01 1.57268643e-01 -4.56622064e-01
1.62247813e+00 -1.63599793e-02 -6.73096836e-01 1.70505807e-01
2.83460677e-01 -4.70448047e-01 6.51607394e-01 4.01371926e-01
-1.01258600e+00 -9.14897770e-02 -6.29941642e-01 4.01049137e-01
-3.61309588e-01 -6.30241856e-02 7.59650171e-01 1.44877684e+00
-1.15718281e+00 5.24362624e-01 -8.79900813e-01 -7.47200131e-01
4.84898657e-01 4.56495672e-01 -2.41764247e-01 5.59557676e-01
-5.91709495e-01 7.90650129e-01 -5.45424037e-02 5.55650234e-01
-5.34904003e-01 -3.00386488e-01 -6.50337100e-01 -3.83303344e-01
2.37030685e-02 -3.06458890e-01 1.22724092e+00 -7.14380980e-01
-1.85633504e+00 1.38619018e+00 -2.97207385e-01 -4.46878642e-01
6.29370332e-01 1.12625144e-01 -2.48508960e-01 1.73476502e-01
-1.52559176e-01 2.75247127e-01 1.17847157e+00 -9.08417642e-01
-7.78570354e-01 -7.75528669e-01 -3.74455750e-01 -7.89797008e-02
-1.97364345e-01 6.65186584e-01 1.37866423e-01 -4.22271818e-01
-1.64918274e-01 -1.12796867e+00 2.17923492e-01 1.25961915e-01
1.51168719e-01 -2.56766528e-01 1.42758086e-01 -3.79394144e-01
1.41954744e+00 -2.13506389e+00 -2.00006645e-03 2.92778820e-01
1.90767154e-01 1.35377452e-01 4.39979956e-02 5.83527327e-01
3.47468883e-01 2.90364996e-02 -2.72273302e-01 -4.71148252e-01
3.55511874e-01 -4.34280187e-02 -1.01594716e-01 1.14939785e+00
1.51875272e-01 7.88664639e-01 -9.20146525e-01 -1.58408895e-01
2.23125834e-02 3.39161843e-01 -2.76771873e-01 1.38267264e-01
2.71947622e-01 4.18158829e-01 -4.05936092e-01 9.25325394e-01
5.78132331e-01 8.11972767e-02 3.24597553e-04 4.40942824e-01
-4.70288008e-01 2.33983807e-02 -1.07261932e+00 1.18723893e+00
1.19344138e-01 8.06056738e-01 1.98821843e-01 -4.68671143e-01
8.41721416e-01 -4.68688691e-03 2.86715597e-01 -4.35250163e-01
4.28404927e-01 1.47101298e-01 3.18971246e-01 -4.90424454e-01
3.78637373e-01 -3.09191167e-01 2.00675249e-01 7.99994111e-01
1.51607618e-01 2.59161800e-01 8.04852769e-02 -5.71312487e-01
7.44158328e-01 7.24817216e-01 4.13419843e-01 -8.01441431e-01
5.79576969e-01 -3.49515378e-01 1.27734363e-01 7.58855402e-01
-6.99401498e-01 1.84392184e-02 2.71592647e-01 -3.62434983e-01
-4.91781592e-01 -9.46825206e-01 -9.06270802e-01 1.11278522e+00
9.50211212e-02 -1.20029770e-01 -1.00549209e+00 -1.40643150e-01
2.08935931e-01 2.97607183e-01 -1.23609948e+00 -2.06150100e-01
2.86835372e-01 -1.20489883e+00 6.89513683e-01 1.96609274e-02
6.35737553e-02 -1.16879106e+00 -1.44006610e+00 -2.96569198e-01
4.91556287e-01 -5.65843225e-01 -1.37715545e-02 2.10999846e-02
-8.97078156e-01 -1.41977632e+00 -7.41784632e-01 -2.49776188e-02
4.22809929e-01 1.58023447e-01 6.03207052e-01 2.74046093e-01
-4.91073191e-01 8.48439038e-01 -2.16150999e-01 -7.61844873e-01
-3.69456828e-01 -3.63854975e-01 5.97237527e-01 5.45588136e-01
1.28792751e+00 -2.00810954e-01 -7.30800986e-01 4.80382442e-01
-7.26815999e-01 -6.53643847e-01 4.15182918e-01 7.17868447e-01
1.79388553e-01 -2.86335409e-01 1.26357347e-01 -3.07085007e-01
7.62197375e-01 -4.06550020e-01 -9.70773697e-01 2.05620915e-01
-8.21409643e-01 1.68152705e-01 -2.97360510e-01 -3.07681233e-01
-9.67292488e-01 -1.21391721e-01 2.06673041e-01 1.62781492e-01
-4.80810225e-01 2.31132880e-01 3.07971895e-01 -5.87781489e-01
4.36071008e-01 1.76794529e-01 5.37066698e-01 -7.08154917e-01
1.08620659e-01 9.34710205e-01 2.61252582e-01 -4.16639447e-01
5.19416928e-01 8.30200374e-01 3.99592251e-01 -1.29715168e+00
-5.24372518e-01 -6.86657906e-01 -8.23786438e-01 -4.99245107e-01
8.03914547e-01 -4.50625420e-01 -1.68672597e+00 7.93812692e-01
-5.07808447e-01 -3.22821409e-01 -5.73248640e-02 7.57206261e-01
-8.87211978e-01 4.07296479e-01 -1.83008194e-01 -1.91590977e+00
5.15051447e-02 -9.06743824e-01 1.21181631e+00 4.42986816e-01
-2.62731433e-01 -1.18176091e+00 3.51134300e-01 1.78240374e-01
1.45658091e-01 1.01916172e-01 3.91822308e-02 -7.18051493e-01
-4.09961045e-01 -2.44741783e-01 9.68262926e-02 1.78470649e-02
1.24845780e-01 5.34521401e-01 -1.31676769e+00 -3.08589429e-01
2.16311678e-01 -3.75128374e-03 7.30600536e-01 6.53145254e-01
3.35183859e-01 3.75032611e-02 -1.33491307e-01 7.36649036e-01
1.17999434e+00 6.24718666e-02 4.27968860e-01 5.41706800e-01
-3.38961631e-02 1.30512977e+00 4.54428464e-01 7.32530713e-01
2.60071427e-01 6.36921883e-01 4.11004663e-01 4.03992355e-01
2.32005015e-01 -1.30435720e-01 4.20979232e-01 -1.46964742e-02
-5.47420979e-01 4.20688838e-02 -7.86654294e-01 3.21900219e-01
-1.72289813e+00 -1.01704383e+00 -3.25011849e-01 2.79520154e+00
3.55035126e-01 -1.24180533e-01 9.25368726e-01 -6.84847310e-03
5.18634021e-01 -3.39041948e-01 -3.16320956e-01 -3.67254108e-01
2.16157902e-02 1.71807766e-01 9.15388107e-01 4.62918758e-01
-7.93527603e-01 6.00563049e-01 8.03215790e+00 3.80612165e-01
-4.58903491e-01 -2.75677860e-01 5.10224663e-02 -9.27912146e-02
5.98906092e-02 -2.23017652e-02 -1.14624047e+00 7.01785326e-01
1.32786047e+00 -9.15451422e-02 7.17588603e-01 1.49448529e-01
3.36550653e-01 -6.77855551e-01 -1.00575149e+00 8.41009498e-01
-8.40117410e-02 -6.82870269e-01 -3.35394800e-01 8.94013584e-01
1.49383649e-01 -1.27531096e-01 4.53133583e-01 -2.81087667e-01
3.59197855e-01 -7.37291813e-01 6.92044616e-01 9.91931617e-01
5.57667673e-01 -4.72099543e-01 5.13307393e-01 4.02772665e-01
-7.59736240e-01 -2.89940625e-01 -3.37097377e-01 -5.23655891e-01
-1.70215577e-01 -1.38769880e-01 -4.34021056e-01 2.68481284e-01
9.03484285e-01 6.00861609e-01 -8.93187582e-01 1.35019910e+00
9.33103543e-03 4.41191286e-01 -6.35242462e-01 -2.79878229e-01
1.30262718e-01 -7.56694376e-01 7.72606015e-01 8.83942962e-01
2.61362791e-01 -1.97531134e-01 -8.31929028e-01 1.04692411e+00
5.13210893e-01 -8.95848498e-02 -6.92512870e-01 -4.29862291e-01
4.19571251e-01 1.21813166e+00 -8.84028614e-01 2.68742651e-01
-3.84296924e-01 5.85688949e-01 7.82579836e-03 1.93745986e-01
-1.95399731e-01 -7.95752704e-02 1.10468602e+00 -2.03298911e-01
3.33852798e-01 1.29603706e-02 -2.37747788e-01 -8.62000644e-01
-1.69717059e-01 -5.13133764e-01 3.74460608e-01 -3.63282412e-01
-1.25366998e+00 3.11292589e-01 3.66704941e-01 -1.10573924e+00
-3.26150179e-01 -1.18361986e+00 -2.99810827e-01 1.22569120e+00
-1.45238113e+00 -1.00890970e+00 -2.34266028e-01 1.75650656e-01
-2.00198993e-01 -3.19859505e-01 7.07614660e-01 -2.72682071e-01
-8.10964763e-01 5.28552353e-01 3.93536568e-01 -2.02564746e-01
3.29790533e-01 -1.26901972e+00 5.01950741e-01 8.47317696e-01
2.61372805e-01 1.11495459e+00 9.10669744e-01 -4.68430519e-01
-1.49502707e+00 -9.54468474e-02 7.48323381e-01 -1.26397800e+00
1.13518822e+00 -2.52497792e-01 -4.08605725e-01 4.68946993e-01
5.58023117e-02 -2.67902225e-01 1.30882406e+00 1.03927799e-01
1.37566760e-01 1.65604785e-01 -1.07954156e+00 4.98982638e-01
1.08984530e+00 -6.87383890e-01 -8.61597478e-01 -1.77514359e-01
-3.02046388e-01 9.89402533e-02 -8.52905750e-01 -4.30367701e-02
1.32926583e+00 -1.33464122e+00 7.32395411e-01 -8.94792557e-01
-1.89266637e-01 -2.19740883e-01 -5.44360876e-02 -7.38506913e-01
-1.42475650e-01 -1.28453779e+00 -1.98291436e-01 9.63649690e-01
8.78849477e-02 -1.03057313e+00 5.81435323e-01 7.28765726e-01
6.02047086e-01 -4.40873206e-01 -1.03417301e+00 -1.01074958e+00
-1.10730253e-01 -3.94599825e-01 5.34844100e-01 2.00143516e-01
-8.38225409e-02 -2.46698007e-01 -2.36019805e-01 5.18056341e-02
1.28909469e+00 -1.54720783e-01 8.13070297e-01 -1.92044842e+00
-1.60518765e-01 -7.83160865e-01 -6.28585517e-01 -7.20730424e-01
1.24785811e-01 -4.19666886e-01 -8.04444328e-02 -7.15888441e-01
-7.87799582e-02 2.21810844e-02 -4.72430676e-01 -1.63242176e-01
-8.74716416e-02 5.10438502e-01 -2.42678225e-02 4.20924962e-01
-1.92061618e-01 1.25892609e-01 7.61026561e-01 5.98224878e-01
-2.66228557e-01 4.52044666e-01 -8.17511380e-01 6.24881566e-01
4.04937655e-01 -4.44929928e-01 -1.15074597e-01 2.60881364e-01
4.70319360e-01 -2.89016455e-01 8.06239784e-01 -5.06927788e-01
1.93382397e-01 -1.69869304e-01 -3.96640003e-02 -2.55964041e-01
2.88567096e-01 -8.68140876e-01 2.03750938e-01 5.34198523e-01
-2.61610299e-01 -7.71339461e-02 1.43800914e-01 1.03174984e+00
2.44535998e-01 -4.31087345e-01 7.69494951e-01 1.04161173e-01
-6.45701468e-01 2.19031811e-01 -3.53156686e-01 -2.47655183e-01
9.93314981e-01 -7.91244328e-01 -6.90775290e-02 -1.22194782e-01
-7.24875927e-01 7.15767033e-03 8.41378570e-01 1.48320079e-01
2.12337270e-01 -9.40687418e-01 -6.37676597e-01 7.19421506e-01
4.10117179e-01 -9.29543018e-01 9.79639683e-03 1.35855579e+00
-4.96751159e-01 3.40247005e-01 -4.98391211e-01 -5.73635876e-01
-1.49123967e+00 6.53968394e-01 5.46568751e-01 4.85310644e-01
-1.70919374e-01 8.54546070e-01 9.58052352e-02 1.59630273e-02
1.07156329e-01 -1.45466477e-01 -3.91619891e-01 4.51010615e-01
8.38378906e-01 6.53370202e-01 -2.70370334e-01 -1.06682777e+00
-5.81473827e-01 5.79638004e-01 3.58929127e-01 -3.49060774e-01
1.25281966e+00 -7.78293490e-01 -2.92794138e-01 6.46560192e-01
7.42535472e-01 -2.08268180e-01 -1.51743197e+00 -3.57597172e-02
5.31675935e-01 -5.95107257e-01 1.60861209e-01 -4.37175840e-01
-3.21928233e-01 8.09284925e-01 8.53431344e-01 7.67813921e-01
9.07896459e-01 -1.47961438e-01 -8.48477110e-02 3.89451683e-01
3.56019646e-01 -1.05996311e+00 -7.81718254e-01 -1.02915525e-01
5.03265083e-01 -1.29450953e+00 1.38131723e-01 1.20876245e-01
-3.58402103e-01 1.11925757e+00 -1.80787042e-01 5.65575901e-03
8.64775956e-01 -3.86265367e-02 -2.10838635e-02 -4.24170762e-01
-6.08361840e-01 -8.47793221e-01 7.27554679e-01 9.26182866e-01
3.02246004e-01 5.61967082e-02 -5.55748284e-01 4.60490942e-01
-2.38505512e-01 1.97258908e-02 3.62550467e-01 7.60589480e-01
-2.27372229e-01 -9.85753179e-01 -3.39417934e-01 3.20778698e-01
-7.12690711e-01 -5.70365461e-04 -8.35516572e-01 8.34867895e-01
-3.87236804e-01 1.11491823e+00 3.63464430e-02 -1.89106315e-01
3.89512591e-02 2.10373253e-01 7.40584731e-01 -3.51950854e-01
-5.47222078e-01 1.49244210e-02 -1.49774387e-01 -5.08698463e-01
-1.10859418e+00 -1.19969952e+00 -2.89694309e-01 -3.63503993e-01
-4.31822807e-01 9.87446085e-02 7.24333465e-01 9.54886734e-01
8.65791142e-02 -1.93674520e-01 6.46753371e-01 -8.60169351e-01
-7.81344295e-01 -9.00966048e-01 -1.15473461e+00 1.71420708e-01
6.39230430e-01 -1.11068523e+00 -8.34536254e-01 4.43973169e-02] | [10.118263244628906, 1.6666233539581299] |
aafb4116-3d74-45ed-b5c0-4f802f2d4a8f | a-joint-model-for-structure-based-news-genre | null | null | https://aclanthology.org/2021.findings-acl.295 | https://aclanthology.org/2021.findings-acl.295.pdf | A Joint Model for Structure-based News Genre Classification with Application to Text Summarization | null | ['Ruihong Huang', 'Zeyu Dai'] | null | null | null | null | findings-acl-2021-8 | ['genre-classification'] | ['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] | [-7.274194240570068, 3.7999160289764404] |
4071c718-dacb-4c43-96b5-937b58f16303 | divide-and-conquer-large-scale-capacitated | 1912.12667 | null | https://arxiv.org/abs/1912.12667v2 | https://arxiv.org/pdf/1912.12667v2.pdf | Divide-and-Conquer Large Scale Capacitated Arc Routing Problems with Route Cutting Off Decomposition | The capacitated arc routing problem is a very important problem with many practical applications. This paper focuses on the large scale capacitated arc routing problem. Traditional solution optimization approaches usually fail because of their poor scalability. The divide-and-conquer strategy has achieved great success in solving large scale optimization problems by decomposing the original large problem into smaller sub-problems and solving them separately. For arc routing, a commonly used divide-and-conquer strategy is to divide the tasks into subsets, and then solve the sub-problems induced by the task subsets separately. However, the success of a divide-and-conquer strategy relies on a proper task division, which is non-trivial due to the complex interactions between the tasks. This paper proposes a novel problem decomposition operator, named the route cutting off operator, which considers the interactions between the tasks in a sophisticated way. To examine the effectiveness of the route cutting off operator, we integrate it with two state-of-the-art divide-and-conquer algorithms, and compared with the original counterparts on a wide range of benchmark instances. The results show that the route cutting off operator can improve the effectiveness of the decomposition, and lead to significantly better results especially when the problem size is very large and the time budget is very tight. | ['Keqin Jiang', 'Buzhong Zhang', 'Yuzhou Zhang', 'Yi Mei'] | 2019-12-29 | null | null | null | null | ['problem-decomposition'] | ['miscellaneous'] | [ 2.42603257e-01 -7.51259476e-02 -9.85267162e-02 -4.85868081e-02
-6.55763626e-01 -7.30400503e-01 -1.85866449e-02 1.78513959e-01
-2.77200192e-01 9.17173088e-01 -1.99233547e-01 -3.47901672e-01
-7.23889649e-01 -7.43209064e-01 -4.84484911e-01 -9.36784744e-01
-1.06558166e-01 9.03994977e-01 3.25677246e-01 -2.18777150e-01
3.18628460e-01 4.55407470e-01 -1.37875056e+00 9.48280916e-02
1.29525626e+00 9.25620079e-01 5.37605286e-01 1.77736297e-01
-5.39145887e-01 -8.12256988e-03 -6.80577099e-01 -2.60884147e-02
4.70248818e-01 -4.39086020e-01 -1.04332948e+00 4.88410592e-01
-3.24579030e-01 3.40908110e-01 2.90590912e-01 9.22244370e-01
1.65390804e-01 2.29768336e-01 3.94913018e-01 -1.75443375e+00
-1.69320941e-01 4.35330421e-01 -1.02496099e+00 8.28778893e-02
1.69416979e-01 -3.99842143e-01 8.97035897e-01 -2.03141674e-01
4.31063294e-01 1.08195186e+00 2.53761172e-01 1.46341950e-01
-1.43743277e+00 -4.91976768e-01 6.88078523e-01 2.49345288e-01
-1.34846163e+00 1.70833632e-01 5.29984355e-01 -1.61135927e-01
1.02827621e+00 4.86485690e-01 2.76720256e-01 1.97775483e-01
3.50806266e-01 5.16147196e-01 1.13465738e+00 -2.59170175e-01
3.15685928e-01 -2.00741622e-03 2.37038121e-01 1.39525846e-01
4.99790728e-01 -3.19234371e-01 3.71440411e-01 -1.84354529e-01
3.83087695e-01 1.19675234e-01 -3.04612339e-01 -5.08002758e-01
-1.14065218e+00 7.47250319e-01 4.12149340e-01 4.18060005e-01
-3.83654505e-01 -1.20704891e-02 3.89608294e-01 3.68348420e-01
3.62275124e-01 5.81218839e-01 -5.34646988e-01 1.15913182e-01
-6.53409958e-01 4.24932748e-01 8.57570112e-01 9.44416940e-01
7.91316330e-01 -3.10551673e-01 -7.58129433e-02 8.40647101e-01
-1.75879434e-01 1.04249731e-01 -8.76652263e-03 -7.50629485e-01
9.21771705e-01 7.46080160e-01 1.57648042e-01 -1.00675452e+00
-6.40360653e-01 -3.73886764e-01 -7.28292584e-01 1.14906460e-01
5.83727956e-01 -3.00675422e-01 -7.76911318e-01 1.50334668e+00
5.39031148e-01 -2.17186525e-01 3.43460068e-02 9.85976458e-01
4.68648262e-02 9.17999983e-01 -1.40160471e-01 -6.52483523e-01
1.40699482e+00 -1.52530634e+00 -7.88327098e-01 -3.83735478e-01
7.66774893e-01 -7.80383766e-01 5.70228517e-01 5.33592582e-01
-1.02463639e+00 -1.47100344e-01 -8.56216371e-01 3.01247537e-01
-4.33004081e-01 -7.90620372e-02 6.87088251e-01 6.16115749e-01
-8.89866292e-01 2.83134431e-01 -4.58363980e-01 -1.95794478e-01
3.08662113e-02 5.80681443e-01 -5.58550775e-01 -6.00575149e-01
-8.28866661e-01 7.71673203e-01 3.19779187e-01 4.10014749e-01
-4.07134354e-01 -3.09181541e-01 -6.07557118e-01 3.77850235e-01
1.34969401e+00 -3.62533689e-01 9.40653801e-01 -7.60452330e-01
-1.05870354e+00 2.03279048e-01 -2.32405201e-01 1.95531175e-01
2.96632141e-01 3.37295979e-01 -1.66884065e-01 -7.15541393e-02
1.64071500e-01 4.62483838e-02 4.53995645e-01 -1.41083264e+00
-8.95984888e-01 -4.92682248e-01 4.22332585e-01 3.01995665e-01
-1.41365692e-01 1.06271893e-01 -7.21535444e-01 -4.59145248e-01
2.19489634e-01 -1.19216180e+00 -7.78296769e-01 -8.38867128e-01
-5.77746093e-01 -4.31242287e-01 5.87705433e-01 -1.85310870e-01
1.49753368e+00 -2.11945438e+00 6.64895773e-01 2.69628435e-01
2.25916743e-01 1.51359627e-03 -3.84893566e-01 6.87201142e-01
-3.00239205e-01 1.58672914e-01 -3.54956508e-01 -2.45405614e-01
-3.40362266e-02 4.23542857e-01 1.30050898e-01 3.44525933e-01
2.84057278e-02 4.76566702e-01 -6.91205382e-01 -1.99014470e-01
-2.39900798e-01 -1.37877762e-01 -4.91316706e-01 7.16332644e-02
-2.74160564e-01 1.99339390e-01 -7.25194156e-01 5.30107677e-01
8.87294948e-01 4.70869131e-02 4.44326252e-01 1.14949666e-01
-3.90237868e-01 4.76646125e-02 -1.62446642e+00 1.19187939e+00
-4.20135230e-01 3.26363504e-01 6.75543249e-01 -1.52321243e+00
8.50113750e-01 3.35386664e-01 7.09345758e-01 -6.08551204e-01
1.28150135e-01 4.34996754e-01 9.61239263e-02 -4.04756099e-01
4.07160223e-01 -2.36747324e-01 -3.66769165e-01 6.68392003e-01
-7.76427448e-01 -3.69226113e-02 7.80390918e-01 1.37106642e-01
1.07360446e+00 -2.87347108e-01 -1.24250189e-03 -4.93822366e-01
7.25142837e-01 4.00383890e-01 8.94105673e-01 2.58327425e-01
1.41708210e-01 5.10199666e-01 1.29942441e+00 -3.34923208e-01
-5.70207477e-01 -4.70799804e-01 4.66957688e-02 9.54395115e-01
4.51984286e-01 -2.84539491e-01 -8.36023510e-01 -8.61424208e-01
1.61057174e-01 4.49439287e-01 -7.34191656e-01 1.42798051e-01
-6.27101064e-01 -1.11493611e+00 -3.40785570e-02 5.45935452e-01
1.89758539e-01 -8.67609024e-01 -5.30146122e-01 5.17707884e-01
-3.88485491e-01 -1.10659301e+00 -5.12768984e-01 4.63555545e-01
-8.66163254e-01 -1.32534480e+00 -7.24053383e-01 -8.57632101e-01
1.13139713e+00 7.69753516e-01 6.78072870e-01 1.60020888e-01
-2.10042566e-01 -1.60874262e-01 -4.98874366e-01 -1.52732968e-01
1.03579268e-01 3.62265646e-01 -1.07419543e-01 -4.82975021e-02
-4.21029814e-02 -2.97145814e-01 -1.13471277e-01 1.04292583e+00
-1.02812529e+00 -2.60629207e-01 7.05718398e-01 6.50207520e-01
7.34895587e-01 9.57585990e-01 7.72231281e-01 -1.04177713e+00
8.64565849e-01 -4.38235700e-01 -8.36412609e-01 4.92247641e-01
-6.99059784e-01 1.47475228e-01 8.09243202e-01 -3.79435301e-01
-8.16419125e-01 8.55580866e-02 3.23215753e-01 -1.78250134e-01
1.22810729e-01 6.42433286e-01 -4.65344399e-01 -7.35280514e-02
-2.10117489e-01 -9.66090485e-02 -1.27479970e-01 -5.75714111e-01
-3.67028639e-02 4.00760353e-01 5.32237813e-02 -7.47394621e-01
7.24175990e-01 1.92364380e-01 2.51808852e-01 -4.67687458e-01
-4.66617346e-01 -4.47140396e-01 -3.24061930e-01 -6.92277029e-02
7.67858744e-01 -2.72794992e-01 -9.50338960e-01 2.86191285e-01
-1.08574617e+00 -3.16257685e-01 7.11379498e-02 2.72140235e-01
-2.69078404e-01 3.52239132e-01 -2.00005531e-01 -7.04968512e-01
1.13491274e-01 -1.60602415e+00 6.29082620e-01 2.25107282e-01
2.27063045e-01 -7.61812866e-01 -1.42246410e-02 5.25185883e-01
4.78573352e-01 4.75860238e-01 1.22950792e+00 -6.42350316e-01
-6.20320380e-01 -3.17011535e-01 -4.34986234e-01 1.60547439e-02
1.48743063e-01 -1.27292305e-01 -9.98786762e-02 -4.42400843e-01
-4.35518380e-03 6.48582727e-02 6.40185177e-01 2.89924800e-01
9.01407838e-01 -1.00052975e-01 -7.40320086e-01 3.09485435e-01
1.65412390e+00 4.74967062e-01 7.13063538e-01 6.12578034e-01
5.79178035e-01 1.05995500e+00 9.63505507e-01 1.92800045e-01
5.31396568e-01 7.41010547e-01 5.58846831e-01 -2.09238082e-01
4.76804703e-01 5.10169625e-01 2.28397623e-02 6.24646246e-01
-3.31725568e-01 -5.26198685e-01 -8.49158049e-01 4.47270513e-01
-2.09407067e+00 -4.80961293e-01 -4.00382042e-01 2.16730118e+00
4.17563647e-01 2.57219166e-01 2.75658429e-01 4.64538038e-01
7.74856746e-01 -1.96461901e-02 -1.62521884e-01 -8.94727230e-01
2.66938865e-01 -2.37944663e-01 5.57847142e-01 4.87680137e-01
-7.21688688e-01 6.91519737e-01 5.86546516e+00 8.64159942e-01
-9.24626887e-01 -5.17774932e-02 4.10339832e-01 -5.29497080e-02
-1.75521761e-01 1.19132362e-01 -8.48002851e-01 3.26407850e-01
5.06934643e-01 -3.67950380e-01 7.26262033e-01 7.36576498e-01
1.29893869e-01 -5.49205720e-01 -1.12551987e+00 4.43321794e-01
-7.83742070e-02 -8.02207947e-01 -3.70442390e-01 4.40341711e-01
7.17220962e-01 -4.83060032e-01 -2.06241116e-01 2.39767089e-01
1.36794681e-02 -9.63954866e-01 4.41057593e-01 -1.03155531e-01
4.22176570e-01 -1.15982592e+00 9.56710935e-01 3.69314492e-01
-1.56675828e+00 -4.15906698e-01 -4.77520764e-01 -2.61389881e-01
4.91724759e-01 7.06038833e-01 -1.29579693e-01 1.01006961e+00
6.57006860e-01 6.13221787e-02 -1.18030407e-01 1.30213296e+00
-3.62141393e-02 -3.05090137e-02 -3.27046961e-01 3.52156628e-03
5.48603594e-01 -6.86880827e-01 4.62928951e-01 9.31259632e-01
1.30204603e-01 3.13811958e-01 5.16639531e-01 5.85791588e-01
2.54805744e-01 1.64895326e-01 -2.60029852e-01 -8.43006745e-02
2.61106580e-01 1.34437346e+00 -1.34434628e+00 -5.40045798e-02
-4.02253300e-01 4.87112999e-01 3.25619221e-01 5.59662461e-01
-8.85300875e-01 -8.86403263e-01 5.55583537e-01 1.47951469e-01
5.58552265e-01 -3.57473195e-01 -3.40425491e-01 -5.90427339e-01
3.47131819e-01 -8.40246141e-01 4.59846705e-01 -3.13231647e-01
-8.38073373e-01 9.14020360e-01 3.64639997e-01 -9.36700821e-01
1.04357325e-01 -5.03072858e-01 -7.40037084e-01 8.13855290e-01
-1.69031930e+00 -6.75022304e-01 -3.19463164e-01 5.38708866e-01
6.33051813e-01 3.19380134e-01 4.46216941e-01 4.56543684e-01
-1.15689445e+00 3.26056123e-01 2.94378072e-01 -3.41667980e-01
3.62740666e-01 -1.08110356e+00 -2.10606098e-01 7.17359781e-01
-4.91496593e-01 3.55615765e-01 6.98822200e-01 -4.95607942e-01
-1.57393718e+00 -7.59418249e-01 8.79612923e-01 7.48289675e-02
4.48477864e-01 -3.91514778e-01 -8.80865216e-01 5.42171896e-01
1.60936132e-01 -2.20476359e-01 4.04237509e-01 2.73019284e-01
4.35181782e-02 -4.17211503e-01 -1.12476850e+00 2.96467990e-01
6.77495956e-01 3.57357085e-01 -4.97723907e-01 4.88832146e-01
6.34860992e-01 -9.87779945e-02 -6.07796311e-01 3.60435724e-01
2.47854754e-01 -7.51405239e-01 7.49467731e-01 -4.61096436e-01
4.94191498e-01 -3.50350827e-01 2.70233750e-02 -1.49132216e+00
-5.51208138e-01 -4.97052342e-01 5.02997518e-01 1.31476367e+00
7.35607326e-01 -1.08967531e+00 5.49979210e-01 7.66780436e-01
-2.33461514e-01 -1.12783551e+00 -9.46051002e-01 -1.12776041e+00
1.06194047e-02 -1.22749973e-02 8.02590609e-01 9.31529820e-01
2.04088196e-01 2.78474927e-01 -1.16466820e-01 1.92856103e-01
4.59462315e-01 6.10943079e-01 5.98456979e-01 -1.22856891e+00
-2.86287278e-01 -4.56094772e-01 1.07010812e-01 -7.73766458e-01
6.24508485e-02 -6.00369751e-01 2.91968822e-01 -2.11453295e+00
-2.63127834e-02 -9.19800401e-01 -1.99340612e-01 7.15263128e-01
-2.26318836e-01 -8.94149318e-02 4.80268896e-01 1.41446769e-01
-7.10942686e-01 3.25495750e-01 1.42881691e+00 -3.50071974e-02
-4.86461639e-01 2.20099315e-01 -7.99220860e-01 3.06407213e-01
7.69439936e-01 -7.88586617e-01 -4.07002211e-01 -6.84679270e-01
2.07758740e-01 4.04330730e-01 -4.65712279e-01 -6.57088876e-01
3.17326635e-01 -6.82933211e-01 -4.20467436e-01 -4.87385005e-01
1.60473704e-01 -1.23608363e+00 4.21245515e-01 4.73573327e-01
1.02076717e-01 4.70914930e-01 3.83983016e-01 5.91388106e-01
-1.71538398e-01 -3.66843492e-01 4.69870985e-01 -3.61854732e-02
-3.67656529e-01 9.29749310e-02 -4.41050947e-01 -1.50281880e-02
1.71806467e+00 -2.78428823e-01 -4.12954658e-01 8.76348913e-02
-6.43155575e-01 9.92906332e-01 2.22578317e-01 5.18744469e-01
2.51277983e-01 -9.69847918e-01 -5.60295939e-01 1.09376177e-01
-1.22303687e-01 3.14418554e-01 2.46632889e-01 1.32330203e+00
-4.70665157e-01 6.32329881e-01 -3.00852835e-01 -2.94156522e-01
-1.35578585e+00 1.12702608e+00 -6.88130632e-02 -7.82532394e-01
-4.55674291e-01 4.78880376e-01 4.17398661e-01 -9.50784534e-02
2.64914811e-01 -3.47568423e-01 -3.86762291e-01 3.79008263e-01
3.12890679e-01 7.49947608e-01 2.36615613e-02 -2.67448276e-01
-6.52352154e-01 7.40302026e-01 -8.79622698e-02 8.35784450e-02
1.61259770e+00 -1.72764063e-01 -6.78772211e-01 -8.62792879e-02
8.19106042e-01 3.88640934e-03 -7.17976630e-01 -7.25387856e-02
-1.04358494e-02 -7.02644706e-01 -7.76048750e-02 -5.25266349e-01
-1.56308949e+00 5.67260206e-01 -2.28679985e-01 8.11223745e-01
1.43825912e+00 -8.14263616e-03 8.04373145e-01 2.28282139e-01
5.28857589e-01 -1.22478259e+00 -1.69699132e-01 4.70364153e-01
8.40276420e-01 -6.48913801e-01 1.92189738e-01 -1.16316032e+00
-7.79114783e-01 9.61637795e-01 6.36988044e-01 -9.10155252e-02
3.60251427e-01 4.05822635e-01 -2.47099832e-01 -2.69803982e-02
-7.04788208e-01 -2.24068984e-01 -2.15865113e-02 2.37408251e-01
1.40510164e-02 1.85916066e-01 -9.58896339e-01 8.42121899e-01
2.76810199e-01 -1.91457272e-01 5.88467181e-01 1.14456260e+00
-5.03515244e-01 -1.45399034e+00 -7.25762606e-01 3.23222011e-01
-2.65795529e-01 3.31582040e-01 -3.32517564e-01 8.40496004e-01
2.01359585e-01 1.17789114e+00 -1.30891930e-02 -2.04439133e-01
7.22576082e-01 -1.61391690e-01 2.99060553e-01 -5.75126767e-01
-7.11531341e-01 2.85356253e-01 3.64534706e-01 -5.62734544e-01
-2.93218285e-01 -3.99289876e-01 -1.38592410e+00 -3.76637548e-01
-6.03510976e-01 5.89069188e-01 7.06999898e-01 1.03810894e+00
2.84082741e-01 1.02122080e+00 9.21849430e-01 -8.39111865e-01
-5.67836225e-01 -5.40442586e-01 -7.32327700e-01 -6.43406156e-03
8.59518200e-02 -9.13425207e-01 -3.10784549e-01 -5.13466418e-01] | [5.1996893882751465, 2.8969240188598633] |
58bed502-8d6b-420e-9508-269aa2fec9cb | learning-to-substitute-ingredients-in-recipes | 2302.07960 | null | https://arxiv.org/abs/2302.07960v1 | https://arxiv.org/pdf/2302.07960v1.pdf | Learning to Substitute Ingredients in Recipes | Recipe personalization through ingredient substitution has the potential to help people meet their dietary needs and preferences, avoid potential allergens, and ease culinary exploration in everyone's kitchen. To address ingredient substitution, we build a benchmark, composed of a dataset of substitution pairs with standardized splits, evaluation metrics, and baselines. We further introduce Graph-based Ingredient Substitution Module (GISMo), a novel model that leverages the context of a recipe as well as generic ingredient relational information encoded within a graph to rank plausible substitutions. We show through comprehensive experimental validation that GISMo surpasses the best performing baseline by a large margin in terms of mean reciprocal rank. Finally, we highlight the benefits of GISMo by integrating it in an improved image-to-recipe generation pipeline, enabling recipe personalization through user intervention. Quantitative and qualitative results show the efficacy of our proposed system, paving the road towards truly personalized cooking and tasting experiences. | ['Adriana Romero-Soriano', 'Michal Drozdzal', 'Rohit Girdhar', 'Quentin Duval', 'Bahare Fatemi'] | 2023-02-15 | null | null | null | null | ['recipe-generation'] | ['miscellaneous'] | [ 3.86377484e-01 6.17834218e-02 -4.07079279e-01 -3.22469622e-01
-4.36701179e-01 -8.73808563e-01 2.93610573e-01 6.84194326e-01
1.82721525e-01 1.12393811e-01 1.01311886e+00 1.30842909e-01
6.20198585e-02 -1.05315781e+00 -8.94777894e-01 -1.43492669e-01
8.00678879e-02 1.63075905e-02 -1.94259956e-01 -5.73420942e-01
-1.04231387e-01 -1.59454912e-01 -1.36284006e+00 7.23984063e-01
1.17699921e+00 6.37559175e-01 1.44579843e-01 3.79419118e-01
-1.25209801e-02 4.03431803e-01 9.69458297e-02 -7.01235294e-01
3.97495121e-01 -6.71022117e-01 -3.68717462e-01 2.96398276e-03
7.31910586e-01 -2.75426954e-01 7.78611600e-02 8.34716976e-01
5.73743343e-01 4.13952827e-01 4.35283452e-01 -8.51871312e-01
-1.39479184e+00 1.08921838e+00 -1.72163472e-01 -5.19293427e-01
7.02580571e-01 7.27952778e-01 1.26308906e+00 -4.74875093e-01
6.80146635e-01 1.03811586e+00 1.02860951e+00 4.61013794e-01
-1.75246608e+00 -3.09915364e-01 3.34846646e-01 1.50012653e-02
-9.27291095e-01 -2.09760442e-01 8.73957336e-01 -2.94736922e-01
9.23476636e-01 4.13420379e-01 1.30461323e+00 1.03143930e+00
-2.45710105e-01 8.54015112e-01 9.52122986e-01 -2.41471767e-01
2.38728374e-01 -1.89128071e-02 1.44331053e-01 7.33208120e-01
1.88225105e-01 1.81308344e-01 -5.75585067e-01 1.12056211e-02
4.00187403e-01 -3.22621223e-03 -1.67898461e-01 -6.48867130e-01
-1.41650486e+00 7.43708491e-01 8.41406524e-01 -1.69485375e-01
-7.40983427e-01 2.40822792e-01 3.81344110e-01 2.57990602e-02
5.38868308e-01 9.14062202e-01 -3.93784702e-01 4.21991289e-01
-6.08776569e-01 5.10004759e-01 6.51242375e-01 1.04256248e+00
5.98160923e-01 -3.35710108e-01 -6.58033133e-01 9.42511737e-01
1.50785759e-01 5.12907028e-01 -2.42267117e-01 -9.46120083e-01
1.99415341e-01 8.57750773e-01 2.49012664e-01 -1.01249814e+00
-5.50499618e-01 -1.50288343e-01 -5.75176358e-01 -1.48312360e-01
3.67544927e-02 6.32230937e-03 -6.59988403e-01 1.88655174e+00
5.82419276e-01 -6.08932376e-02 -3.71522129e-01 1.02605081e+00
1.18849981e+00 3.96432549e-01 9.46751893e-01 2.79949725e-01
1.40322161e+00 -1.22646749e+00 -5.98413289e-01 1.70678571e-01
6.92238748e-01 -6.93976045e-01 1.48033941e+00 2.53557146e-01
-8.93194735e-01 -5.76207697e-01 -1.05582106e+00 -2.82776505e-01
-5.33143222e-01 1.02254592e-01 1.03627729e+00 9.01737273e-01
-8.58570814e-01 9.21339929e-01 -4.69860107e-01 -5.86820364e-01
3.56910348e-01 5.53079620e-02 -3.03596705e-01 -4.49076682e-01
-1.13202357e+00 6.20369017e-01 4.12887216e-01 -1.24199919e-01
-5.55568099e-01 -1.49443626e+00 -1.29733503e+00 -4.24225722e-03
4.72628176e-01 -1.40651309e+00 9.78588581e-01 -7.81151474e-01
-1.67039716e+00 5.62943399e-01 3.46970320e-01 -1.32701501e-01
2.47157678e-01 -4.21375424e-01 -5.62493265e-01 -3.03372797e-02
1.55893773e-01 1.10042703e+00 2.51649797e-01 -8.94933343e-01
-2.95015752e-01 -2.15027571e-01 3.65671664e-01 4.16222423e-01
-5.83097823e-02 -4.13684398e-01 -2.61548422e-02 -8.47172320e-01
-2.19449252e-01 -8.51171315e-01 -4.82846588e-01 2.41548732e-01
-4.76225615e-01 -1.07170127e-01 6.42484501e-02 -9.19244111e-01
1.19576669e+00 -2.27465940e+00 1.57707393e-01 3.43658142e-02
1.37657030e-02 2.02407483e-02 -5.70331931e-01 8.24171782e-01
-2.60204002e-02 2.11776853e-01 9.66744125e-03 -2.13742957e-01
5.53082943e-01 -1.66589096e-01 2.79641986e-01 2.72597671e-01
1.39260650e-01 1.33525383e+00 -1.44572866e+00 -7.22641870e-02
7.66016662e-01 5.92322886e-01 -9.80690300e-01 -7.50887266e-04
-1.01231635e+00 2.24085793e-01 -1.60774752e-01 5.50144434e-01
6.56788826e-01 -1.82429120e-01 5.94327986e-01 -1.07902873e+00
-8.23981985e-02 3.96919698e-01 -1.17053163e+00 2.35783482e+00
-4.68582034e-01 -2.72955984e-01 -1.48546249e-01 -3.47771704e-01
6.41047835e-01 2.32629031e-02 5.90119898e-01 -8.19810987e-01
-2.30250403e-01 -1.48666829e-01 -6.16263211e-01 -4.42629158e-01
7.93710828e-01 8.36500227e-02 -1.10196069e-01 4.66518581e-01
-9.50842723e-02 -1.89199537e-01 3.47948134e-01 2.05611527e-01
7.47848809e-01 8.10892344e-01 2.81505972e-01 -5.80500841e-01
8.25454518e-02 3.86917144e-01 1.41605720e-01 3.23805392e-01
8.63254443e-02 4.16099995e-01 -2.37679303e-01 -4.53509957e-01
-1.15360498e+00 -1.23979890e+00 3.86712104e-01 1.38112533e+00
1.60916284e-01 -7.38458633e-01 -5.23299873e-01 -5.26832640e-01
2.58900017e-01 1.19197404e+00 -7.54209936e-01 -2.72676766e-01
-4.09060240e-01 -6.99715018e-01 1.58710629e-02 3.33215117e-01
4.87232745e-01 -8.34318340e-01 -2.86684394e-01 2.24236548e-01
-3.39654952e-01 -8.02184522e-01 -1.23982763e+00 -2.44765565e-01
-5.86294353e-01 -1.17486477e+00 -4.57188100e-01 -6.14541829e-01
3.09312105e-01 7.32892752e-01 1.65684295e+00 -6.12296723e-02
-3.38663042e-01 5.58819354e-01 -6.70544744e-01 2.02113334e-02
-5.83534658e-01 2.02622563e-01 -3.03182542e-01 -3.04007381e-01
2.84606457e-01 -4.62231129e-01 -1.43431270e+00 1.44899622e-01
-8.97329748e-01 6.49644494e-01 1.06996685e-01 4.32941705e-01
7.36948609e-01 -2.30377629e-01 5.32781482e-01 -9.39300776e-01
5.46559930e-01 -6.90597117e-01 -2.23149389e-01 4.91171300e-01
-7.53922045e-01 9.06556547e-02 6.73003674e-01 -3.44078988e-01
-9.57201958e-01 3.93301696e-01 -5.19437017e-03 3.25191230e-01
-1.10873543e-01 3.71942431e-01 -3.48157734e-01 2.80644029e-01
8.01500142e-01 -2.25382447e-01 -1.86072901e-01 -7.59896338e-01
1.64606535e+00 2.11408995e-02 6.08428240e-01 -6.47551298e-01
2.46824980e-01 1.98424816e-01 -1.71184361e-01 -4.26884890e-01
-5.40050030e-01 -4.57845271e-01 -2.65452087e-01 -1.21974364e-01
9.56708550e-01 -1.15961099e+00 -9.40281391e-01 1.63256407e-01
-5.40143728e-01 -6.92692876e-01 -5.16986728e-01 3.20593357e-01
-2.64463961e-01 3.70803326e-01 -6.14427686e-01 -3.54753256e-01
-6.42278969e-01 -7.83650994e-01 1.03905022e+00 2.13196591e-01
-4.68630552e-01 -9.90997970e-01 3.27491194e-01 5.66041052e-01
4.89088953e-01 6.96079850e-01 1.15860617e+00 -1.27825439e-01
-6.36673748e-01 2.25423127e-01 -3.06304663e-01 7.21002221e-02
5.00102580e-01 -1.09075539e-01 -6.39942706e-01 6.52566850e-02
-7.50527918e-01 -1.00161314e-01 9.22834396e-01 5.62167704e-01
5.36911309e-01 -1.92467451e-01 -1.56057850e-01 6.44896984e-01
1.54828703e+00 -4.32456702e-01 4.44342941e-01 1.34860948e-01
9.14113164e-01 6.66612923e-01 4.41664100e-01 3.25748891e-01
9.28624451e-01 7.10315764e-01 4.81456667e-01 -4.67888087e-01
-7.71369338e-01 -7.20113873e-01 3.79254997e-01 4.18886095e-01
2.64147192e-01 -2.25522950e-01 -3.45170423e-02 6.99912608e-01
-1.77640820e+00 -9.05075610e-01 -3.36521119e-01 2.06313014e+00
9.71948981e-01 -5.96523464e-01 6.39461160e-01 -5.04916847e-01
5.15997648e-01 1.80783719e-01 -5.56653619e-01 -4.26331758e-01
-1.36654794e-01 3.14520538e-01 5.66989720e-01 3.16029847e-01
-1.14629507e+00 9.17663097e-01 6.64547491e+00 3.54203850e-01
-6.34641171e-01 5.93057880e-03 6.20236695e-01 -1.05076373e-01
-1.00927353e+00 -3.43538821e-01 -2.19431907e-01 2.95052707e-01
7.02632606e-01 2.10607097e-01 1.12252021e+00 6.76909685e-01
3.89293373e-01 5.84438071e-03 -1.36240995e+00 5.63964427e-01
-3.55331674e-02 -1.56762028e+00 -4.80919629e-02 -4.46189970e-01
9.89091575e-01 -1.08200923e-01 5.36747128e-02 2.47968212e-01
9.81248856e-01 -8.62062514e-01 7.35783577e-01 4.43707764e-01
6.27704084e-01 -5.87424934e-01 1.20497175e-01 -2.31064796e-01
-1.61431599e+00 2.98469335e-01 -8.54901522e-02 1.49515286e-01
2.34301373e-01 6.28122807e-01 -7.75584340e-01 9.35488224e-01
4.80294883e-01 9.62753057e-01 -5.42142570e-01 8.83342862e-01
-1.71615660e-01 2.39946038e-01 -4.60424542e-01 7.13043213e-02
-3.76037471e-02 -5.14100373e-01 1.64365426e-01 1.36570632e+00
3.19055736e-01 1.79886997e-01 2.25268230e-01 1.21374285e+00
-1.56784251e-01 6.38593316e-01 -3.23684424e-01 -2.72301704e-01
2.37549737e-01 1.32925451e+00 -5.86379111e-01 -1.87704295e-01
-2.93648988e-01 1.46482038e+00 -8.24830867e-03 1.79140836e-01
-8.49607468e-01 2.32728124e-01 8.31152022e-01 -7.06508234e-02
3.71989131e-01 7.04002380e-02 -3.74972969e-01 -1.07352638e+00
-2.43439272e-01 -1.10602868e+00 5.24333715e-01 -6.99056625e-01
-1.57035136e+00 2.32015364e-02 3.26445103e-02 -6.49485111e-01
2.14880079e-01 -3.59835029e-01 -3.21127743e-01 7.06026614e-01
-1.29186237e+00 -1.73230064e+00 -4.91875559e-01 1.72203258e-01
4.54155326e-01 6.03196621e-01 9.96994734e-01 4.78157192e-01
-4.83014286e-01 4.89115030e-01 -1.75383221e-02 -4.37074244e-01
8.06516528e-01 -1.28398657e+00 9.94211435e-01 5.40561318e-01
6.80399016e-02 6.49652123e-01 7.91392803e-01 -9.03572619e-01
-1.66274881e+00 -1.43513274e+00 5.67909420e-01 -5.13504565e-01
2.07253978e-01 -2.83946663e-01 -3.35158348e-01 4.31830019e-01
2.72757351e-01 -5.67927718e-01 1.03728080e+00 2.89916486e-01
-7.18397021e-01 -5.21087535e-02 -1.39547086e+00 8.80751550e-01
1.56640220e+00 -4.28138226e-01 -3.21664773e-02 5.63066006e-01
1.10896778e+00 -1.46388263e-01 -1.32950735e+00 2.95216382e-01
8.64529788e-01 -9.89387453e-01 1.44761693e+00 -4.02169079e-01
4.86488581e-01 -4.47677612e-01 -4.00705040e-01 -1.67625117e+00
-7.49062836e-01 -5.17366886e-01 1.28034875e-01 1.11480331e+00
3.30719858e-01 -4.00346145e-02 6.15867674e-01 8.88903856e-01
-3.82238805e-01 -4.42471117e-01 -3.74713354e-02 -3.60004395e-01
-2.71756887e-01 -2.16012701e-01 1.36600995e+00 1.02964842e+00
9.38370824e-02 3.20438594e-01 -6.99092746e-01 9.13727880e-02
6.81911826e-01 4.23982292e-01 7.62861729e-01 -5.90228915e-01
-6.33313894e-01 -5.69536150e-01 3.72755527e-01 -8.98998141e-01
-3.89717191e-01 -1.26881444e+00 1.06848873e-01 -2.02135181e+00
3.43830913e-01 -3.95460218e-01 -3.52184862e-01 4.68106270e-01
-5.62603295e-01 4.87265319e-01 4.47198898e-01 -4.12847817e-01
-7.26244807e-01 4.62461323e-01 1.40000296e+00 -2.85738707e-01
-6.32877707e-01 -5.45523703e-01 -1.16645122e+00 1.53239131e-01
7.66935527e-01 -1.24655984e-01 -5.70072651e-01 -4.49229658e-01
6.19810820e-01 -4.35049415e-01 2.81897128e-01 -6.08091712e-01
-3.59726161e-01 -3.02343607e-01 1.72088265e-01 -2.98319817e-01
-3.35015282e-02 -6.73023701e-01 1.11283958e+00 4.85986412e-01
-6.59941018e-01 -1.09200217e-01 2.19320506e-01 5.36437750e-01
7.33823419e-01 2.88912594e-01 3.37537616e-01 -1.95731521e-01
-8.70141089e-01 1.50187254e-01 9.70710814e-02 -2.72987992e-01
7.43123829e-01 8.28726441e-02 -4.26262885e-01 -3.89330834e-02
-6.94926023e-01 3.16013157e-01 9.06688392e-01 5.19579232e-01
2.20919654e-01 -1.68168819e+00 -7.08027124e-01 -1.77075434e-02
5.29037178e-01 -5.71063042e-01 7.02685475e-01 6.56788349e-01
-6.02496624e-01 3.30155343e-02 -2.31727481e-01 -2.19606385e-01
-9.65425968e-01 1.02237225e+00 2.00373441e-01 -2.07631052e-01
-7.46225476e-01 5.91932833e-01 2.97597528e-01 -8.85904312e-01
-1.43838614e-01 -8.04695249e-01 2.66553406e-02 -3.20493758e-01
4.43001300e-01 5.09158313e-01 -1.71412647e-01 -2.22604826e-01
-1.54352114e-01 2.81019032e-01 3.39576274e-01 4.49794084e-01
1.37004471e+00 -3.51703197e-01 3.42297368e-02 5.83542557e-03
7.55420506e-01 1.09709643e-01 -1.35563767e+00 -1.66710213e-01
-2.99995303e-01 -4.60193425e-01 -9.62419361e-02 -1.44245732e+00
-9.42877471e-01 1.21240884e-01 8.42662275e-01 2.30280496e-02
1.15914118e+00 -1.51418865e-01 1.26497734e+00 -2.70836782e-02
2.64193326e-01 -8.54846478e-01 3.21648344e-02 -3.27552736e-01
6.70546055e-01 -1.26454628e+00 1.54682711e-01 -7.29444921e-01
-7.46171951e-01 6.58840179e-01 2.23339647e-01 -1.62547499e-01
4.34482485e-01 -4.16410059e-01 1.26380529e-02 -3.03724974e-01
-4.96310294e-01 -2.27435350e-01 6.59481287e-01 9.44033682e-01
6.77912951e-01 6.45133913e-01 -3.73977512e-01 7.10661709e-01
-6.11297227e-02 3.53229612e-01 -2.05301747e-01 2.43557081e-01
-2.52121717e-01 -1.40909886e+00 1.44509152e-01 2.09458232e-01
6.00173734e-02 -6.42077506e-01 -1.25725225e-01 3.35158378e-01
5.99628866e-01 1.05772305e+00 -4.73530859e-01 -4.87473100e-01
6.09984338e-01 -1.02304637e-01 7.69102395e-01 -5.99624574e-01
-1.22575033e+00 5.88693880e-02 5.28854311e-01 -9.56807733e-01
-5.58665752e-01 -5.02633035e-01 -8.92925203e-01 -8.42022657e-01
-8.10375065e-02 -2.75606215e-01 6.61697865e-01 4.46256995e-01
7.34096467e-01 6.65877819e-01 4.06361878e-01 -7.29732871e-01
-2.45523721e-01 -6.43509448e-01 -3.64718437e-01 1.00240254e+00
-1.78888604e-01 -2.45425224e-01 2.98171908e-01 3.35334748e-01] | [11.51645278930664, 4.532201766967773] |
44c746dd-cb7c-4103-ae30-4a682ce50d25 | simgans-simulator-based-generative | 2006.15353 | null | https://arxiv.org/abs/2006.15353v1 | https://arxiv.org/pdf/2006.15353v1.pdf | SimGANs: Simulator-Based Generative Adversarial Networks for ECG Synthesis to Improve Deep ECG Classification | Generating training examples for supervised tasks is a long sought after goal in AI. We study the problem of heart signal electrocardiogram (ECG) synthesis for improved heartbeat classification. ECG synthesis is challenging: the generation of training examples for such biological-physiological systems is not straightforward, due to their dynamic nature in which the various parts of the system interact in complex ways. However, an understanding of these dynamics has been developed for years in the form of mathematical process simulators. We study how to incorporate this knowledge into the generative process by leveraging a biological simulator for the task of ECG classification. Specifically, we use a system of ordinary differential equations representing heart dynamics, and incorporate this ODE system into the optimization process of a generative adversarial network to create biologically plausible ECG training examples. We perform empirical evaluation and show that heart simulation knowledge during the generation process improves ECG classification. | ['Kira Radinsky', 'Daniel Freedman', 'Tomer Golany'] | 2020-06-27 | null | https://proceedings.icml.cc/static/paper_files/icml/2020/2829-Paper.pdf | https://proceedings.icml.cc/static/paper_files/icml/2020/2829-Paper.pdf | icml-2020-1 | ['ecg-classification', 'heartbeat-classification'] | ['medical', 'medical'] | [ 7.23958731e-01 3.85388404e-01 5.41584671e-01 -4.77022529e-02
-4.01724398e-01 -7.09142983e-01 4.13864344e-01 -7.66793191e-02
-5.92021458e-02 8.61392617e-01 -2.07290143e-01 -3.70264560e-01
4.92099263e-02 -5.36477327e-01 -7.01725602e-01 -7.22930670e-01
-1.98658988e-01 4.42354023e-01 -3.72078151e-01 -1.35868862e-01
-1.25403762e-01 4.62525815e-01 -7.39652097e-01 -1.75174266e-01
8.04491699e-01 4.80903655e-01 -5.63517690e-01 1.37648404e+00
5.03924489e-01 6.14530444e-01 -1.10752583e+00 -1.96832344e-01
3.67952734e-01 -1.41946149e+00 -5.07268906e-01 6.48532063e-02
7.98891336e-02 1.26649827e-01 -2.11771578e-01 6.15232408e-01
9.23200250e-01 5.74730784e-02 9.03709114e-01 -1.25095093e+00
-3.18473011e-01 4.80394125e-01 4.51734848e-02 1.90624610e-01
9.41063315e-02 5.82706213e-01 4.72870231e-01 -3.40212166e-01
5.11896789e-01 8.90037954e-01 8.93689990e-01 8.78314078e-01
-1.85227835e+00 -5.72287500e-01 -5.20009220e-01 -5.65139234e-01
-1.42542231e+00 -2.61584312e-01 7.26004004e-01 -7.01109767e-01
3.74680012e-01 2.14944035e-01 1.14814460e+00 1.23524165e+00
8.25846434e-01 4.72211808e-01 1.03441095e+00 -1.09062746e-01
5.30644655e-01 -4.98387329e-02 -1.03875794e-01 7.43460655e-01
2.27823198e-01 3.44833285e-01 -4.21273023e-01 -3.88087094e-01
1.24651241e+00 -2.11076528e-01 -3.69422913e-01 -9.84544009e-02
-1.14282811e+00 6.16961181e-01 2.45294586e-01 1.31658033e-01
-4.02676314e-01 6.54426098e-01 1.17180899e-01 5.02545178e-01
1.84069023e-01 1.21249783e+00 -2.29500175e-01 -1.43438414e-01
-8.54515851e-01 7.67472625e-01 1.21214426e+00 4.59926903e-01
2.31630757e-01 5.53797185e-01 -2.03383654e-01 4.20162141e-01
9.61719826e-02 4.38612431e-01 3.86775643e-01 -1.26253164e+00
-2.01760888e-01 2.39458248e-01 -1.03470564e-01 -7.15898275e-01
-3.70039403e-01 -9.26825285e-01 -1.13695645e+00 2.90411532e-01
6.81240201e-01 -7.95336425e-01 -1.01507998e+00 1.77823913e+00
2.74396986e-01 6.22065485e-01 2.50609726e-01 7.46209800e-01
3.83814454e-01 6.29776359e-01 3.55457276e-01 -2.93690950e-01
9.78960514e-01 -2.75673538e-01 -5.27908921e-01 -3.41550447e-02
3.71586710e-01 -3.60258877e-01 6.72735929e-01 2.12337539e-01
-1.34032214e+00 -6.73862994e-01 -1.02153945e+00 3.00026417e-01
-2.55601704e-02 -4.23867911e-01 5.75324893e-01 6.71124578e-01
-7.79330194e-01 1.12561524e+00 -1.05178726e+00 -1.03199229e-01
5.35467386e-01 3.96615714e-01 1.58132121e-01 3.48689258e-01
-1.40905571e+00 8.35664988e-01 -4.40312140e-02 5.42210154e-02
-1.22282755e+00 -1.12086916e+00 -7.02174664e-01 5.02876975e-02
1.85553096e-02 -1.44713736e+00 1.24937236e+00 -9.32762444e-01
-1.57513869e+00 5.82816064e-01 8.82954597e-02 -6.70536697e-01
7.23993361e-01 1.98547184e-01 5.98703399e-02 -2.92821284e-02
-3.73095036e-01 6.77012205e-01 9.80055630e-01 -1.17551017e+00
1.53340474e-01 -9.38179716e-02 -2.20184013e-01 1.77206233e-01
2.87039429e-01 -4.54354733e-01 8.67525637e-02 -9.91449237e-01
-9.66846794e-02 -1.35459638e+00 -6.79256737e-01 1.02729887e-01
-2.85452873e-01 5.26511669e-01 3.97011727e-01 -5.47269762e-01
7.96935976e-01 -1.97306061e+00 2.35752434e-01 3.34270269e-01
4.13534611e-01 2.64844775e-01 1.29364148e-01 3.88253540e-01
-2.84381080e-02 4.12256062e-01 -5.80672324e-01 -2.99317360e-01
-1.40502512e-01 2.23901868e-01 -4.28565770e-01 1.91720605e-01
4.85448927e-01 1.30211484e+00 -9.20454860e-01 -4.18616146e-01
-9.67373028e-02 8.06003928e-01 -5.58362126e-01 3.05536360e-01
-3.35315108e-01 1.12293768e+00 -6.56851888e-01 3.02928209e-01
-1.79838151e-01 -2.27589816e-01 1.23174571e-01 5.06328084e-02
4.16638851e-01 -1.16519153e-01 -7.72677302e-01 1.32464039e+00
-3.23742718e-01 5.12338340e-01 -2.71624029e-01 -9.22625244e-01
8.13643277e-01 6.85577452e-01 6.51722312e-01 -3.49320769e-02
4.43814397e-01 -4.24623676e-02 6.61564291e-01 -1.50975481e-01
-1.98310286e-01 -7.83147991e-01 -1.81147188e-01 6.15332365e-01
-1.57745466e-01 -9.63700771e-01 -1.36209846e-01 5.55918738e-02
1.40534651e+00 2.91833043e-01 1.35402679e-01 -7.50196278e-02
1.13681547e-01 6.45829365e-02 5.31052232e-01 8.55738819e-01
-1.91839516e-01 7.89508700e-01 6.10532999e-01 -3.68680626e-01
-1.26873016e+00 -1.08877718e+00 -1.06760532e-01 7.14837238e-02
-2.30169311e-01 -8.21003914e-02 -1.15564692e+00 -1.76905021e-01
5.46186045e-02 5.36349118e-01 -7.74871290e-01 -6.65982723e-01
-5.25465190e-01 -1.05440044e+00 1.14438641e+00 5.49177289e-01
1.39042810e-01 -1.39096689e+00 -1.13387251e+00 6.88192487e-01
1.97140902e-01 -8.29819500e-01 -3.86923522e-01 2.96537519e-01
-1.08910692e+00 -8.19624007e-01 -9.31331575e-01 -4.80985552e-01
6.21792853e-01 -8.49376500e-01 1.17231703e+00 1.74894661e-01
-9.82771516e-01 5.35863876e-01 1.91677362e-01 -9.40021694e-01
-1.22098172e+00 -8.41558352e-02 -7.02386424e-02 -1.11905754e-01
-3.12030286e-01 -8.63063395e-01 -6.54426038e-01 6.61336333e-02
-8.89856577e-01 3.81083965e-01 3.56283575e-01 1.08088052e+00
6.68723762e-01 1.81210798e-03 9.12478566e-01 -1.05964959e+00
1.05786860e+00 -1.98147103e-01 -3.84071529e-01 5.25556467e-02
-5.19080639e-01 1.71992004e-01 6.42241359e-01 -8.02072346e-01
-6.18070722e-01 2.96042979e-01 -1.38873309e-01 -4.82546866e-01
-4.63699289e-02 4.56535637e-01 7.58077055e-02 -2.06961125e-01
8.74252915e-01 3.60046923e-01 3.72082651e-01 7.14774355e-02
2.24671796e-01 -3.10723651e-02 7.20751345e-01 -8.09833050e-01
8.73766243e-01 5.87434284e-02 5.79753876e-01 -8.90642524e-01
-5.01527607e-01 3.89627784e-01 -4.86240625e-01 -3.14324141e-01
7.29126334e-01 -6.01871014e-01 -7.97385931e-01 5.15142083e-01
-8.96438837e-01 -9.23909903e-01 -8.08855176e-01 2.09758580e-01
-9.19043899e-01 -1.31762385e-01 -7.88470745e-01 -7.85738707e-01
-5.62546432e-01 -1.02770007e+00 9.12464917e-01 2.55065262e-01
-6.94974542e-01 -1.07148540e+00 2.59015352e-01 -1.83084998e-02
4.90366280e-01 1.09677887e+00 1.03059065e+00 -2.99793869e-01
-5.35081565e-01 -2.85347730e-01 4.57237720e-01 4.55501944e-01
-1.04143642e-01 5.42818084e-02 -9.59353745e-01 -2.90608332e-02
2.87466854e-01 -8.14764425e-02 3.46996933e-01 5.91414750e-01
9.86256659e-01 -1.80032536e-01 -3.74589592e-01 6.48849428e-01
1.13585210e+00 3.00057441e-01 6.30175829e-01 -4.87281084e-01
7.42326677e-01 4.74382222e-01 9.50604398e-03 2.92300701e-01
1.10828178e-02 3.72662663e-01 -1.06177941e-01 -4.07288313e-01
-1.31272748e-01 -2.45152771e-01 8.48977268e-03 5.22082150e-01
-2.15932056e-01 -2.36323014e-01 -1.12805724e+00 2.42479354e-01
-1.54812467e+00 -1.01043403e+00 1.48204476e-01 1.96732712e+00
1.20709348e+00 2.24086344e-01 1.51748791e-01 4.86578614e-01
3.99946809e-01 -4.67865258e-01 -9.71330583e-01 -2.97206163e-01
1.14070296e-01 7.45192409e-01 8.53252113e-02 4.25965935e-01
-6.47605062e-01 5.08602858e-01 6.92022943e+00 6.36411160e-02
-1.14595187e+00 -3.09148401e-01 1.10426748e+00 1.69173926e-01
5.21519594e-02 -5.92395030e-02 -3.64369959e-01 3.08566272e-01
1.24144244e+00 -5.29418528e-01 5.16534865e-01 3.14773381e-01
3.99227202e-01 1.92786172e-01 -1.39082742e+00 7.80034244e-01
-1.36229098e-01 -1.26551044e+00 -1.10846505e-01 1.83838606e-01
7.19285607e-01 -4.90262151e-01 -1.79258659e-02 2.18665630e-01
3.50351661e-01 -1.46584094e+00 4.74546105e-01 8.66011083e-01
5.85003436e-01 -7.55831063e-01 3.82624984e-01 5.08424044e-01
-6.98949158e-01 2.65456796e-01 9.04226601e-02 2.40616221e-02
3.87394160e-01 6.19432509e-01 -9.99900520e-01 -1.33146271e-02
3.43043581e-02 5.43030202e-01 -4.79968131e-01 1.10477817e+00
-6.55339882e-02 1.16451287e+00 -2.36501157e-01 4.97928448e-02
-3.02659452e-01 -5.28233275e-02 6.44441903e-01 9.23726618e-01
1.98542029e-01 3.69287997e-01 1.06713831e-01 1.36994302e+00
-6.03039376e-02 -2.29582399e-01 -6.56268418e-01 -4.46483910e-01
4.26380374e-02 1.01138389e+00 -8.05537522e-01 -4.33027864e-01
3.66394728e-01 7.64952004e-01 -1.23547472e-01 3.85857999e-01
-9.80925441e-01 -3.22895050e-01 6.77151024e-01 3.40014338e-01
-2.87334174e-01 -2.35454291e-01 -6.50885463e-01 -9.27937210e-01
-4.33849394e-01 -1.17823136e+00 4.41204607e-02 -7.91964293e-01
-1.20582247e+00 5.15024781e-01 -1.46672890e-01 -8.29424441e-01
-4.48566854e-01 -3.70253995e-02 -7.18002737e-01 1.13692653e+00
-8.87435317e-01 -6.40621305e-01 -2.55330771e-01 3.71263355e-01
3.02776873e-01 1.15113087e-01 9.92946982e-01 3.10120881e-02
-4.15398210e-01 5.07857144e-01 -3.27258736e-01 8.10857341e-02
3.24877977e-01 -1.33719575e+00 8.82473886e-01 5.70940256e-01
1.67269260e-02 7.16084719e-01 1.04861927e+00 -7.21960485e-01
-1.43484378e+00 -1.17114508e+00 4.91765380e-01 -7.37724662e-01
3.89450252e-01 -3.35379720e-01 -9.85071540e-01 3.82872224e-01
-1.29010454e-01 6.98384494e-02 6.32939458e-01 -3.81949097e-01
2.38716021e-01 1.83089972e-01 -1.07417345e+00 8.08720410e-01
8.70802224e-01 -4.27084059e-01 -4.99931365e-01 9.54795778e-02
4.26664412e-01 -6.41894519e-01 -1.07235122e+00 4.82351899e-01
6.72104657e-01 -1.33103877e-01 1.09337819e+00 -5.48391044e-01
4.19837862e-01 -4.38500017e-01 5.55094481e-01 -1.61556602e+00
1.57256767e-01 -1.12593031e+00 -2.08923072e-01 8.33638370e-01
3.92794847e-01 -5.40558159e-01 7.91725993e-01 7.07311273e-01
1.37401670e-01 -8.71956468e-01 -3.43752921e-01 -5.17477334e-01
2.70886868e-01 -1.81150973e-01 2.71203518e-01 7.24043787e-01
-1.25446856e-01 7.37487197e-01 -2.51009166e-01 -5.03579825e-02
6.33278728e-01 6.28747686e-04 9.16865349e-01 -1.27836323e+00
-9.02161062e-01 -2.97347665e-01 -4.39366519e-01 -5.45813024e-01
-5.10756224e-02 -7.73683429e-01 4.18512166e-01 -1.18578362e+00
-7.27162883e-02 -6.17643595e-01 1.83449034e-02 2.19280392e-01
-5.21081865e-01 3.61453474e-01 3.76575142e-01 8.34327787e-02
2.16664091e-01 3.22807491e-01 1.47348893e+00 2.36273911e-02
-3.09568405e-01 1.64475814e-01 -5.16474247e-01 5.12549460e-01
8.53806615e-01 -7.71333456e-01 -6.31998658e-01 1.32340223e-01
8.92660096e-02 6.97652578e-01 6.03291929e-01 -1.51846933e+00
-1.08258791e-01 1.19089767e-01 7.49208987e-01 1.30846426e-01
4.14955258e-01 -4.22437012e-01 7.10678041e-01 9.57707763e-01
-6.29917145e-01 2.07105935e-01 2.01920003e-01 5.40090442e-01
-1.15538150e-01 9.57016796e-02 1.00553858e+00 -2.11596444e-01
2.79419839e-01 3.51283461e-01 -6.13386273e-01 4.94478345e-01
8.09751272e-01 -7.00275972e-02 1.34716958e-01 -6.58460259e-01
-1.14471579e+00 -5.40332012e-02 3.33025098e-01 8.90674070e-03
3.80088061e-01 -8.30565512e-01 -8.30650568e-01 2.29228467e-01
-3.53890955e-01 7.21253678e-02 -1.52627930e-01 6.67523146e-01
-6.85666680e-01 -1.43353358e-01 -2.33049661e-01 -6.12989485e-01
-9.99708772e-01 2.51707733e-01 9.72849965e-01 -2.93157905e-01
-6.19699478e-01 4.95759785e-01 3.17791700e-01 -1.15893289e-01
2.62095109e-02 -2.79068828e-01 8.09218436e-02 -4.65952188e-01
1.81648493e-01 -8.40465352e-03 -3.54873121e-01 -1.10409774e-01
7.58763924e-02 3.96250546e-01 5.82486153e-01 -3.41483504e-01
1.27968252e+00 2.91986406e-01 3.21018368e-01 4.83880162e-01
6.56412303e-01 -4.80785966e-01 -1.27382386e+00 2.34251127e-01
-2.16349497e-01 1.62311271e-01 -2.52597392e-01 -8.88048112e-01
-8.87664974e-01 9.60512698e-01 5.14360785e-01 1.30382732e-01
8.96518409e-01 -4.96471167e-01 7.73872077e-01 3.85441720e-01
1.95554927e-01 -5.23047030e-01 2.20554993e-01 2.98545659e-01
8.23851824e-01 -6.43755674e-01 -2.66003478e-02 -3.77448678e-01
-6.56124949e-01 9.59671021e-01 4.40008432e-01 -5.31966805e-01
8.18269074e-01 7.13055611e-01 2.28348985e-01 -1.55102775e-01
-8.07249367e-01 3.69444251e-01 2.45891854e-01 5.47122419e-01
5.74421883e-01 -8.68159533e-02 -3.05265427e-01 4.38525259e-01
-3.14415872e-01 4.13389713e-01 6.54872358e-01 1.00942206e+00
6.91471994e-02 -1.03561401e+00 -1.49829999e-01 3.62962157e-01
-6.93426073e-01 -9.94752571e-02 -3.43780786e-01 5.17056108e-01
1.06806032e-01 4.67654109e-01 -1.78656816e-01 -1.05727665e-01
1.83472872e-01 5.26931822e-01 6.68356299e-01 -7.88474381e-01
-9.33149755e-01 -6.56444067e-03 -2.19620988e-01 -5.22993989e-02
-1.18921371e-02 -7.61201382e-01 -1.37069619e+00 -9.57994610e-02
-7.59183988e-02 1.36477321e-01 7.72087395e-01 7.20612288e-01
3.00187647e-01 1.10114491e+00 3.69559020e-01 -9.16842878e-01
-5.96470296e-01 -6.56713903e-01 -4.18239444e-01 4.35631454e-01
3.16238403e-01 -2.21891135e-01 -2.48383015e-01 7.09021747e-01] | [14.280930519104004, 3.0417656898498535] |
2143a436-0366-4e70-9213-3bfe883b2c57 | imad-image-augmented-multi-modal-dialogue | 2305.10512 | null | https://arxiv.org/abs/2305.10512v1 | https://arxiv.org/pdf/2305.10512v1.pdf | IMAD: IMage-Augmented multi-modal Dialogue | Currently, dialogue systems have achieved high performance in processing text-based communication. However, they have not yet effectively incorporated visual information, which poses a significant challenge. Furthermore, existing models that incorporate images in dialogue generation focus on discussing the image itself. Our proposed approach presents a novel perspective on multi-modal dialogue systems, which interprets the image in the context of the dialogue. By doing so, we aim to expand the capabilities of current dialogue systems and transition them from single modality (text) to multi-modality. However, there is a lack of validated English datasets that contain both images and dialogue contexts for this task. Thus, we propose a two-stage approach to automatically construct a multi-modal dialogue dataset. In the first stage, we utilize text-to-image similarity and sentence similarity to identify which utterances could be replaced with an image. In the second stage, we replace those utterances by selecting a subset of relevant images and filtering them with a visual question answering model. We used this approach, along with additional labeling, to create the IMage Augmented multi-modal Dialogue dataset (IMAD), which can serve as a validated dataset for this task. Furthermore, we propose a baseline model trained on this dataset, which outperforms model trained on the same data without images and BlenderBot. | ['Kuznetsov Denis', 'Frolov Anton', 'Moskvoretskii Viktor'] | 2023-05-17 | null | null | null | null | ['dialogue-generation', 'dialogue-generation'] | ['natural-language-processing', 'speech'] | [ 3.15167069e-01 2.19834596e-01 2.44152322e-01 -4.57376987e-01
-7.06404746e-01 -6.01136208e-01 1.09718251e+00 1.93264019e-02
-5.00692666e-01 7.17602968e-01 3.77717346e-01 -1.48150653e-01
4.95946646e-01 -8.21397007e-01 -3.44983369e-01 -4.80737925e-01
7.33553171e-01 5.88530540e-01 5.36951900e-01 -5.20616114e-01
3.51605237e-01 1.31103605e-01 -1.46343982e+00 9.65164483e-01
8.10089707e-01 7.82981813e-01 4.98027176e-01 6.62048340e-01
-5.91331303e-01 8.87703001e-01 -8.90674829e-01 -5.42086780e-01
-4.23180079e-03 -8.86350691e-01 -1.22138786e+00 5.46949148e-01
2.63571411e-01 -7.96516597e-01 -6.80416748e-02 8.39575708e-01
5.90190291e-01 3.48571062e-01 7.59485960e-01 -1.34752345e+00
-7.27039754e-01 3.70673865e-01 -4.73071992e-01 -1.31680429e-01
9.21640515e-01 2.76400000e-01 7.56750166e-01 -1.12811303e+00
9.35431242e-01 1.64760804e+00 2.10886970e-01 7.54441857e-01
-8.55021060e-01 -2.79165924e-01 6.41874224e-02 1.91730782e-01
-9.70157921e-01 -4.96507019e-01 9.22491550e-01 -2.81923622e-01
8.46300364e-01 3.21981907e-01 7.29554653e-01 1.16982043e+00
-3.13773364e-01 1.16384375e+00 1.37245905e+00 -9.38157618e-01
-9.40428227e-02 5.94192326e-01 -3.24182272e-01 7.29576230e-01
-6.02920592e-01 -3.54934305e-01 -3.47970665e-01 1.18663706e-01
7.21083343e-01 -2.02364385e-01 -2.57564038e-01 -3.35127056e-01
-1.21642327e+00 9.79917407e-01 2.52258986e-01 4.87164736e-01
-1.77346557e-01 -4.17032450e-01 4.83286291e-01 2.07057849e-01
3.72138351e-01 2.84329623e-01 9.63800475e-02 7.63797387e-02
-6.87498510e-01 2.72273511e-01 8.10267985e-01 6.94445252e-01
8.44755769e-01 -3.72993201e-01 -4.80595052e-01 1.20466578e+00
3.87751192e-01 3.04078966e-01 3.95893782e-01 -8.00450265e-01
6.28158271e-01 9.72837269e-01 -3.00024636e-02 -1.01354134e+00
-3.27048302e-01 5.19933581e-01 -6.47522211e-01 5.92403561e-02
5.18648803e-01 -1.83373064e-01 -8.80568087e-01 1.39260638e+00
6.68048143e-01 -3.60463381e-01 4.91650283e-01 1.17640638e+00
1.27215695e+00 7.55680203e-01 1.56869173e-01 -1.47653371e-01
1.66603243e+00 -1.36367261e+00 -8.80208552e-01 -1.14879258e-01
5.55307090e-01 -1.14356399e+00 1.41072702e+00 8.17449614e-02
-1.10631251e+00 -6.28051341e-01 -7.71617055e-01 -3.06661695e-01
-6.33145988e-01 2.57085085e-01 4.29423839e-01 5.00158310e-01
-1.09933913e+00 -1.19973920e-01 -6.76665455e-02 -7.48834193e-01
-1.36253774e-01 -2.10475817e-01 -4.55758512e-01 -1.55948684e-01
-1.42557943e+00 1.23380399e+00 6.07583463e-01 1.83953401e-02
-8.19894969e-01 1.45103391e-02 -9.62939382e-01 -4.02226448e-01
5.53486466e-01 -6.52285695e-01 1.24048293e+00 -1.15593493e+00
-1.57432652e+00 9.82178092e-01 1.26912277e-02 -1.97915778e-01
6.25780523e-01 1.18804209e-01 -3.04037273e-01 6.71844602e-01
1.67658448e-01 1.09073234e+00 7.47634828e-01 -1.84544683e+00
-8.51242244e-01 -3.21476042e-01 8.62043321e-01 8.16746175e-01
-4.32728976e-01 1.01457313e-01 -9.91677225e-01 -4.01193768e-01
-1.08750664e-01 -8.92593563e-01 -1.37257744e-02 1.03353411e-02
-4.31636840e-01 -2.18395755e-01 1.09856546e+00 -7.18865991e-01
9.20499027e-01 -1.96382046e+00 8.28712732e-02 -1.26997709e-01
-5.14079370e-02 3.08098733e-01 -2.45403007e-01 6.49448216e-01
4.35107112e-01 -1.52677298e-02 -1.82905599e-01 -6.05702698e-01
-2.34270006e-01 2.24247545e-01 -7.37079754e-02 -7.51805957e-03
3.10919195e-01 7.36987233e-01 -9.28621471e-01 -1.12376249e+00
5.66488683e-01 2.79215276e-01 -3.15830499e-01 5.20627797e-01
-4.50823337e-01 8.25981379e-01 -4.08986509e-01 5.47235250e-01
6.34926796e-01 1.23036862e-03 2.09500939e-01 -5.92810810e-01
-1.32309869e-01 -1.72267020e-01 -1.06400275e+00 1.82323539e+00
-6.71347260e-01 5.55583715e-01 1.90435991e-01 -9.38297093e-01
8.48368108e-01 3.96253198e-01 3.86269569e-01 -8.47741961e-01
1.60017505e-01 -2.03605816e-02 -2.30295047e-01 -9.94795203e-01
9.61411536e-01 -6.97719976e-02 -1.73150569e-01 6.33471012e-01
1.00070558e-01 -4.22170252e-01 5.61473131e-01 3.44892383e-01
6.27976477e-01 2.60179937e-01 8.86669531e-02 4.28556353e-01
9.21946704e-01 4.32131737e-01 9.78427231e-02 6.47246897e-01
-1.72596782e-01 7.66939223e-01 3.97455782e-01 -2.60876536e-01
-1.13072002e+00 -8.28445137e-01 1.52363926e-01 1.19157648e+00
4.26655799e-01 -1.56187966e-01 -9.11356866e-01 -9.20594752e-01
-3.84304881e-01 5.59857488e-01 -4.99410808e-01 2.09627271e-01
-4.18845832e-01 -5.86745679e-01 7.53678441e-01 -2.95753311e-03
1.06224787e+00 -1.47540748e+00 -5.64810097e-01 6.05176277e-02
-6.80329084e-01 -1.24402344e+00 -4.49255109e-01 -3.20729285e-01
-2.65540451e-01 -1.28067064e+00 -9.29750919e-01 -9.70639527e-01
8.51597607e-01 4.01432484e-01 9.55151498e-01 1.57713786e-01
-2.00547680e-01 5.80121279e-01 -7.89269328e-01 -2.27169126e-01
-8.08955491e-01 -2.48779684e-01 -3.98447245e-01 1.55992478e-01
1.35027587e-01 2.17457950e-01 -7.10140169e-01 4.05062050e-01
-1.20541501e+00 5.88249922e-01 6.77446485e-01 9.77569640e-01
3.12715262e-01 -3.31578851e-01 5.77906132e-01 -9.37524080e-01
1.13377607e+00 -2.16343209e-01 6.41166745e-03 6.70121908e-01
-1.02548771e-01 -2.68301368e-01 4.75923747e-01 -4.44968730e-01
-1.63291419e+00 2.37677604e-01 -3.15457106e-01 -2.03871951e-01
-4.05327350e-01 4.65115160e-01 -1.40008837e-01 2.92636249e-02
4.22771633e-01 2.01265037e-01 2.54635632e-01 -2.00951129e-01
7.48455167e-01 1.12825382e+00 4.77960110e-01 -5.02647281e-01
3.39764088e-01 4.51747209e-01 -4.66884404e-01 -9.41114426e-01
-6.79588854e-01 -5.72088778e-01 -7.75229156e-01 -7.10175991e-01
1.15004790e+00 -7.26419508e-01 -3.93098176e-01 4.09179986e-01
-1.32338572e+00 -2.21043259e-01 7.46832713e-02 2.97052026e-01
-5.27690232e-01 7.76787102e-01 -4.79322106e-01 -1.03467226e+00
-3.34904343e-01 -1.17252505e+00 1.14062333e+00 2.74338990e-01
-4.90698870e-03 -1.00475335e+00 -1.43990114e-01 9.21175480e-01
2.39162102e-01 2.06735149e-01 7.76042461e-01 -7.42283762e-01
-1.85826331e-01 1.38742208e-01 -5.07937253e-01 2.23696128e-01
1.19816862e-01 -6.02102652e-03 -9.06048059e-01 -4.55890261e-02
-2.61637215e-02 -7.96627522e-01 6.17658854e-01 1.29124159e-02
7.83842802e-01 -1.67575523e-01 -1.70006797e-01 4.60407995e-02
1.03024673e+00 3.38043928e-01 5.97723484e-01 3.91795546e-01
6.09836638e-01 1.09093332e+00 1.14781678e+00 3.18927586e-01
8.60635757e-01 8.41770589e-01 3.40469301e-01 -6.29194021e-01
-2.76876688e-01 -2.97130018e-01 2.06513792e-01 7.12666512e-01
9.78525057e-02 -3.71708602e-01 -8.36209178e-01 6.41554296e-01
-1.99771011e+00 -8.84205639e-01 -1.06432155e-01 1.70774937e+00
9.93027031e-01 -1.34801492e-01 2.20974877e-01 -1.57476902e-01
8.47752750e-01 9.71129313e-02 -2.80929208e-01 -5.05918801e-01
-3.18763405e-01 -4.29557055e-01 -1.82573572e-01 3.32387686e-01
-1.13648260e+00 1.08756697e+00 5.93937111e+00 7.59479284e-01
-9.98806655e-01 1.18428737e-01 6.56941712e-01 4.12556857e-01
-3.72044176e-01 2.91691273e-02 -6.05984509e-01 2.17958882e-01
5.84586084e-01 2.07526028e-01 2.02740401e-01 6.25163317e-01
3.35800409e-01 -5.40167093e-01 -8.51265490e-01 8.80796731e-01
6.36002839e-01 -1.20397520e+00 4.01936680e-01 -2.54692644e-01
5.47220647e-01 -6.11908913e-01 -2.83664048e-01 4.16313440e-01
9.38730985e-02 -7.80592322e-01 5.23077488e-01 3.75176430e-01
7.51321375e-01 -4.53424215e-01 8.11056793e-01 4.34929401e-01
-1.10206568e+00 1.63992822e-01 -1.72817186e-01 1.47967413e-01
4.65944767e-01 -2.54630446e-02 -1.34964466e+00 6.33978963e-01
5.76625466e-01 2.05297187e-01 -7.94155896e-01 7.36245513e-01
-8.30165818e-02 -1.42650858e-01 1.24207567e-02 -1.23277657e-01
2.50575334e-01 -2.65474439e-01 1.88253954e-01 1.26056623e+00
1.59122825e-01 1.21769104e-02 5.08913815e-01 7.16328561e-01
2.39486843e-02 5.74596286e-01 -7.60362327e-01 2.36403327e-02
5.38052976e-01 1.51189995e+00 -9.12843645e-01 -5.32063425e-01
-8.74783456e-01 1.31330788e+00 1.60568640e-01 2.25570723e-01
-6.58933818e-01 -3.81006628e-01 -1.50778472e-01 -2.22331777e-01
-1.92309514e-01 1.48043126e-01 8.53178054e-02 -1.06535518e+00
-2.03718632e-01 -1.08001077e+00 4.98537034e-01 -1.14645100e+00
-1.20337129e+00 7.94971108e-01 2.01831058e-01 -1.16930950e+00
-5.26534140e-01 -3.96723151e-01 -4.36787128e-01 7.74243355e-01
-1.43061531e+00 -1.81896484e+00 -6.02797687e-01 8.63143384e-01
9.47278440e-01 -1.20217219e-01 9.29363728e-01 2.57647008e-01
-3.57482493e-01 3.25097144e-01 -1.94473743e-01 3.35086107e-01
1.07583511e+00 -1.25669742e+00 -2.72980910e-02 6.00365102e-01
2.45686062e-02 3.05560887e-01 6.29125416e-01 -5.95753193e-01
-1.20168674e+00 -9.09092426e-01 4.99464720e-01 -2.80891627e-01
4.86272424e-01 -2.50169486e-01 -8.14405322e-01 2.65596300e-01
7.01762676e-01 -5.06477892e-01 6.10775650e-01 -2.95937300e-01
2.02045426e-01 3.17333639e-01 -1.39364445e+00 8.31694126e-01
6.69539988e-01 -6.03266656e-01 -8.16395700e-01 3.76294822e-01
4.84158963e-01 -6.58323109e-01 -9.10408020e-01 3.02191198e-01
3.92401516e-01 -9.62879598e-01 9.75121140e-01 -1.54137626e-01
6.69391334e-01 -2.65958577e-01 -4.65534180e-02 -1.28566456e+00
4.99904603e-01 -2.22792730e-01 2.58997887e-01 1.67853558e+00
3.96494925e-01 -1.92237675e-01 5.35020292e-01 4.93051231e-01
-4.89483029e-02 -5.74685454e-01 -6.09982729e-01 -1.12719834e-01
-1.32767126e-01 3.14406045e-02 3.90864998e-01 9.48246956e-01
1.07199281e-01 6.78971112e-01 -6.19404435e-01 -1.28631592e-01
2.05564335e-01 2.76941866e-01 1.24808943e+00 -7.71903396e-01
6.00190908e-02 -1.78489864e-01 4.30229716e-02 -1.21677744e+00
1.08350061e-01 -4.58570510e-01 2.60118157e-01 -1.98275959e+00
4.56187397e-01 -2.98058003e-01 3.77493441e-01 1.97106600e-01
-4.07933444e-01 5.73611677e-01 3.71457040e-01 3.02869022e-01
-8.62404585e-01 6.48933709e-01 1.63659680e+00 -1.68478325e-01
-2.97906488e-01 -3.10928762e-01 -3.89232159e-01 8.15049887e-01
6.83571041e-01 2.17349634e-01 -5.38647056e-01 -3.34705621e-01
-2.91804701e-01 3.85192573e-01 2.37601638e-01 -5.84573686e-01
2.37239331e-01 -3.18235248e-01 4.01014060e-01 -8.82921636e-01
7.22672403e-01 -7.61288047e-01 -1.65751085e-01 2.06429958e-01
-5.15240967e-01 1.54370308e-01 1.99579191e-03 4.17859763e-01
-5.36148071e-01 -5.65459371e-01 6.52300656e-01 -4.58269536e-01
-1.16003144e+00 -1.10738628e-01 -5.13477802e-01 1.14241958e-01
1.26948428e+00 -2.89149284e-01 -5.70223451e-01 -6.48365617e-01
-6.78084731e-01 5.88470459e-01 6.70449197e-01 6.64613545e-01
9.49959695e-01 -1.24819851e+00 -5.50393820e-01 -3.67533378e-02
3.62872034e-01 -3.51137184e-02 4.32699382e-01 6.81750178e-01
-5.99553227e-01 2.94964731e-01 -5.05217791e-01 -5.84045112e-01
-1.59785616e+00 4.74365234e-01 1.47341758e-01 -1.48557544e-01
-5.00644565e-01 3.97986412e-01 2.12132037e-01 -7.28255391e-01
9.62290689e-02 1.45476148e-01 -8.69561374e-01 4.46357340e-01
4.32716668e-01 -1.83898225e-01 -2.01231390e-01 -1.09901190e+00
-1.14329398e-01 3.57195735e-01 1.22056138e-02 -5.54736137e-01
9.13371146e-01 -7.24673629e-01 -1.73975363e-01 5.45417905e-01
9.05484974e-01 -6.59812093e-02 -1.03987026e+00 -2.24428475e-01
-2.26210639e-01 -5.15143394e-01 -2.92609036e-01 -8.46907616e-01
-7.90803432e-01 8.24534774e-01 5.80816865e-01 5.22247553e-01
1.14099610e+00 4.51709293e-02 6.74647152e-01 2.25760683e-01
1.85256541e-01 -1.33747280e+00 7.62954235e-01 5.54210603e-01
1.10279775e+00 -1.61102438e+00 -1.58159837e-01 -4.56936836e-01
-1.27306402e+00 1.14151835e+00 1.07326293e+00 3.97769332e-01
1.16128035e-01 -1.47166952e-01 4.96999562e-01 -1.51047841e-01
-6.43816352e-01 -5.88432610e-01 1.78417146e-01 7.75069058e-01
5.12232184e-01 -2.13043317e-01 -5.05241215e-01 2.12385610e-01
2.29772888e-02 -2.17318624e-01 6.40138328e-01 9.31687534e-01
-5.05589783e-01 -1.28203082e+00 -5.87797582e-01 3.33598107e-01
-3.26289922e-01 -3.21534649e-02 -7.43356705e-01 8.95716608e-01
3.87943722e-02 1.40867031e+00 -1.09326616e-02 -4.34667021e-01
4.64640766e-01 7.09675476e-02 3.95747006e-01 -6.95492983e-01
-6.56696498e-01 5.23680001e-02 4.89916533e-01 -1.62265331e-01
-8.87272060e-01 -4.08488363e-01 -1.06789029e+00 -1.53443083e-01
-4.66068327e-01 1.61323220e-01 7.95508742e-01 1.06542540e+00
2.61996761e-02 5.09072900e-01 5.36769569e-01 -1.02248406e+00
-1.89749181e-01 -1.22606230e+00 -1.28373250e-01 9.13393199e-01
-1.31375669e-03 -6.61213636e-01 -6.00150861e-02 2.33209461e-01] | [11.007502555847168, 1.3197672367095947] |
8bb709d0-8884-4e54-b851-a7df8d36ecfe | automated-vehicle-highway-merging-motion | 2211.02225 | null | https://arxiv.org/abs/2211.02225v1 | https://arxiv.org/pdf/2211.02225v1.pdf | Automated Vehicle Highway Merging: Motion Planning via Adaptive Interactive Mixed-Integer MPC | A new motion planning framework for automated highway merging is presented in this paper. To plan the merge and predict the motion of the neighboring vehicle, the ego automated vehicle solves a joint optimization of both vehicle costs over a receding horizon. The non-convex nature of feasible regions and lane discipline is handled by introducing integer decision variables resulting in a mixed integer quadratic programming (MIQP) formulation of the model predictive control (MPC) problem. Furthermore, the ego uses an inverse optimal control approach to impute the weights of neighboring vehicle cost by observing the neighbor's recent motion and adapts its solution accordingly. We call this adaptive interactive mixed integer MPC (aiMPC). Simulation results show the effectiveness of the proposed framework. | ['Ardalan Vahidi', 'Viranjan Bhattacharyya'] | 2022-11-04 | null | null | null | null | ['motion-planning'] | ['robots'] | [-1.64159909e-02 8.13433707e-01 -8.12391698e-01 -3.13322872e-01
-7.04897940e-01 -3.68561953e-01 3.28300416e-01 -6.81984052e-02
-4.38573003e-01 1.13245690e+00 -1.77084088e-01 -5.26501119e-01
-5.66237271e-01 -7.88136959e-01 -7.92565465e-01 -8.34438622e-01
-2.78639287e-01 5.20466805e-01 1.56425521e-01 -1.29782381e-02
3.62789065e-01 6.34697795e-01 -1.10951293e+00 -4.29337680e-01
1.13971496e+00 7.73855925e-01 6.31255507e-01 8.02488685e-01
2.51220167e-01 5.20968199e-01 1.73648044e-01 -1.46669075e-02
4.83242631e-01 3.69219720e-01 -6.87284172e-01 7.07427740e-01
-9.15215164e-02 -6.02240562e-01 -3.81677717e-01 9.35617149e-01
-1.07812949e-01 7.15200603e-01 5.00794709e-01 -2.03309870e+00
7.81067610e-02 -3.65963648e-03 -4.90075827e-01 2.94734128e-02
-1.31367937e-01 4.76074547e-01 7.06127346e-01 -5.51110923e-01
8.38759720e-01 1.31726003e+00 5.79984337e-02 1.22359872e-01
-1.15569782e+00 -5.93042932e-03 6.74202323e-01 6.24888003e-01
-1.46717024e+00 -2.98265904e-01 5.23997843e-01 -4.29388374e-01
8.02970648e-01 2.98554868e-01 7.43554294e-01 -1.83064282e-01
6.08932793e-01 7.93481886e-01 6.04962111e-01 -6.34347945e-02
3.07330281e-01 2.44510680e-01 7.72351176e-02 5.75355947e-01
2.14118719e-01 2.71118522e-01 3.86399567e-01 -1.09378742e-02
1.72625676e-01 -9.08657238e-02 -7.75714591e-02 -5.75296938e-01
-1.03034091e+00 9.84138846e-01 2.14615732e-01 -6.19522750e-01
-8.79901409e-01 1.48210034e-01 8.75334144e-02 1.35715917e-01
2.76557326e-01 3.02019045e-02 -2.49095917e-01 1.44253284e-01
-6.56786203e-01 9.30439055e-01 6.98395669e-01 1.49606788e+00
8.46621811e-01 2.21981153e-01 -3.33960801e-01 4.38037992e-01
5.45975149e-01 3.82111758e-01 -4.49688822e-01 -1.94419193e+00
8.56238544e-01 3.75337183e-01 7.47145653e-01 -9.80806351e-01
-2.96264082e-01 -5.21267094e-02 -3.02073032e-01 6.20406866e-01
7.65596405e-02 -4.56910014e-01 -6.71641409e-01 1.36413455e+00
6.26679420e-01 2.55365938e-01 1.65186986e-01 9.01218057e-01
-2.92467356e-01 1.29903400e+00 1.76956117e-01 -7.71902978e-01
7.35996544e-01 -1.20621228e+00 -9.39705253e-01 -3.03121567e-01
6.75062716e-01 -2.96048194e-01 -3.71038467e-02 1.00852706e-01
-1.23648918e+00 -1.94895625e-01 -1.10356176e+00 1.54711053e-01
-2.47746021e-01 -4.85799797e-02 1.05301915e-02 1.73320547e-01
-1.08960617e+00 4.04810905e-01 -5.46589375e-01 2.07887162e-02
1.98371768e-01 7.57369995e-01 -5.24470722e-03 -2.70715266e-01
-8.98965776e-01 1.16692102e+00 5.71075976e-01 3.50588143e-01
-9.88711655e-01 -6.16902649e-01 -8.95415187e-01 -9.28727463e-02
7.97064900e-01 -4.98426408e-01 1.24766326e+00 -5.96181333e-01
-1.52514291e+00 3.84978265e-01 -4.10319000e-01 -5.53227127e-01
6.40887916e-01 2.28162725e-02 -1.73863143e-01 -3.20395194e-02
4.37687159e-01 8.35969388e-01 4.60598588e-01 -1.34629071e+00
-1.27007079e+00 -1.52238056e-01 2.41939887e-01 4.94193763e-01
5.17649710e-01 -5.20876646e-01 -6.39788389e-01 -9.74947140e-02
-1.55864060e-01 -1.34030437e+00 -9.99887943e-01 -1.07188329e-01
-3.44893456e-01 -1.61579072e-01 1.07106328e+00 -7.27599561e-01
1.34176540e+00 -1.81653714e+00 4.35596555e-01 4.49452162e-01
-9.22013372e-02 -1.19951824e-02 -1.36986360e-01 3.92504305e-01
4.74624366e-01 -1.15520567e-01 -2.43742496e-01 -1.28648490e-01
9.24718939e-03 7.26270080e-01 -1.70515195e-01 7.46739686e-01
7.44266659e-02 5.50965071e-01 -6.97231531e-01 -5.11049032e-01
4.30966824e-01 -7.04476684e-02 -6.86914861e-01 2.90611207e-01
-4.05126899e-01 2.76903182e-01 -9.72129643e-01 4.27623242e-01
9.89788175e-01 5.22866547e-01 9.56740752e-02 1.03889957e-01
-1.01435566e+00 -3.16478759e-01 -1.37373781e+00 8.60706627e-01
-3.06670725e-01 5.95520020e-01 7.43711650e-01 -9.54414606e-01
5.88351667e-01 9.10698473e-02 7.81611800e-01 -4.59843963e-01
7.12033361e-02 -6.18851967e-02 -2.91148037e-01 -6.35723948e-01
1.02277946e+00 9.90227163e-02 4.59431075e-02 -5.89966290e-02
-5.98486304e-01 -2.36954898e-01 9.66116264e-02 5.20594716e-02
5.67612469e-01 -6.58268332e-02 4.08859193e-01 -2.75388420e-01
9.11548674e-01 7.44236827e-01 9.16919053e-01 4.17893469e-01
-6.15050197e-01 -2.80441701e-01 5.11338651e-01 -1.55101836e-01
-1.22331238e+00 -6.84783161e-01 8.44345316e-02 6.16771519e-01
6.35160446e-01 4.47665721e-01 -5.55440545e-01 -2.06472442e-01
2.67922014e-01 1.17585528e+00 -2.17538521e-01 1.70335129e-01
-9.64550555e-01 -2.47323051e-01 -3.37455809e-01 2.33013108e-02
3.12090158e-01 -1.80503756e-01 -5.10880351e-01 6.68378532e-01
-3.63219641e-02 -1.21325839e+00 -5.85606217e-01 -3.42563719e-01
-7.07857668e-01 -8.91855896e-01 -3.61172199e-01 -8.14622104e-01
7.04735935e-01 5.72930098e-01 2.29165867e-01 -1.37872949e-01
1.98986024e-01 4.96496797e-01 1.45469040e-01 -4.05727506e-01
-3.85196984e-01 -8.29732940e-02 -3.22650969e-02 2.50517488e-01
-2.89088905e-01 1.00025646e-01 -4.38507199e-01 4.49062586e-01
-4.04034615e-01 2.86720365e-01 2.89037377e-01 5.08205771e-01
1.08653879e+00 4.68306661e-01 7.09316730e-01 -5.09173989e-01
4.85284626e-01 -8.95293534e-01 -1.29780126e+00 1.84448764e-01
-6.61139846e-01 -6.60507306e-02 3.40456635e-01 -1.77569211e-01
-1.22084880e+00 4.46978956e-01 3.45553488e-01 -4.53520894e-01
1.49551407e-01 3.90670717e-01 -5.07659256e-01 -2.67671913e-01
-5.65006435e-01 8.79218951e-02 2.79200524e-01 1.98655486e-01
4.57104594e-01 4.32674587e-01 6.62571251e-01 -2.48571143e-01
8.78177762e-01 4.31737661e-01 4.49844480e-01 -7.01830983e-01
-2.89242715e-01 -5.31837344e-01 -5.84588230e-01 -8.52907836e-01
1.12906969e+00 -9.66102123e-01 -9.42347229e-01 -4.47155647e-02
-1.22317290e+00 -3.62876326e-01 -6.72819167e-02 4.98270929e-01
-1.10477865e+00 2.71666050e-01 -3.93835492e-02 -1.35235894e+00
2.13390619e-01 -1.40498316e+00 5.51143110e-01 2.78008968e-01
1.39009520e-01 -9.52503681e-01 -2.84781028e-02 4.53450471e-01
9.55578610e-02 6.90909028e-01 7.27590919e-01 1.44891798e-01
-1.27182567e+00 -3.74126315e-01 -1.61744565e-01 -8.10758471e-02
-4.26042914e-01 1.97830528e-01 -2.85258800e-01 -3.40311795e-01
-1.25266939e-01 5.59655964e-01 2.35200897e-01 7.42657959e-01
6.49400175e-01 -1.07935071e+00 -9.28572476e-01 2.92874426e-01
1.82892346e+00 8.73376131e-01 4.86326694e-01 7.99017370e-01
4.64674652e-01 1.07791412e+00 1.66475666e+00 4.30072188e-01
1.06706190e+00 1.00937521e+00 1.02136362e+00 2.71199286e-01
7.71822333e-01 6.93148449e-02 3.38961810e-01 1.61139399e-01
5.00886552e-02 -2.66774982e-01 -7.57419467e-01 8.75076890e-01
-2.28293824e+00 -9.78120267e-01 -5.55180371e-01 2.13887930e+00
1.84190497e-01 -9.59363654e-02 -1.12993205e-02 -3.57395522e-02
9.03166771e-01 -5.49582280e-02 -5.02394974e-01 -1.03983724e+00
1.53218716e-01 -7.13008463e-01 1.66393554e+00 1.08318579e+00
-1.14735997e+00 7.30069160e-01 5.86602211e+00 6.84151351e-01
-6.36371672e-01 7.81955495e-02 6.31391466e-01 -1.78854354e-02
-1.25314116e-01 3.38705570e-01 -8.87665749e-01 5.02510250e-01
1.08904016e+00 -8.14427257e-01 6.06805921e-01 8.94991517e-01
1.33401299e+00 -5.39672732e-01 -6.38932228e-01 2.47110516e-01
-5.69012702e-01 -1.57609940e+00 -2.96613872e-01 3.89730155e-01
9.99810457e-01 -1.69512630e-01 -5.24255969e-02 1.97016150e-01
3.75565469e-01 -5.97948134e-01 8.93040180e-01 7.70785987e-01
2.11019441e-01 -1.30077350e+00 5.75854242e-01 7.01637506e-01
-1.19822812e+00 -6.98386252e-01 1.12213986e-02 -9.47903171e-02
1.16257048e+00 -2.13433076e-02 -1.00067151e+00 4.70401436e-01
6.35851314e-03 4.65804487e-01 1.95501924e-01 1.33398390e+00
5.24988845e-02 -1.89432539e-02 -6.49706125e-02 8.64479765e-02
8.01091671e-01 -7.55167484e-01 1.25143480e+00 8.75021577e-01
1.59449160e-01 4.97008026e-01 7.14426398e-01 7.12755740e-01
5.44594169e-01 -6.34349212e-02 -6.90169036e-01 2.65017539e-01
4.15087044e-01 9.51545477e-01 -1.83198258e-01 -2.75637299e-01
-3.71935546e-01 4.29040581e-01 5.04020900e-02 5.95547736e-01
-9.51224923e-01 -9.85979438e-02 9.68743265e-01 7.91644529e-02
2.39001468e-01 -4.85819101e-01 -5.24358273e-01 -2.23867327e-01
-1.07359722e-01 1.76240206e-02 4.42245342e-02 -6.35896802e-01
-3.53500783e-01 6.57920586e-03 3.68923753e-01 -1.40097892e+00
-2.83730268e-01 -2.55338222e-01 -9.19209242e-01 8.23791802e-01
-1.87197793e+00 -9.50577080e-01 1.40121460e-01 1.74080342e-01
9.37158227e-01 4.48774397e-02 -3.95840518e-02 2.76172966e-01
-8.58316481e-01 -1.38673499e-01 2.27975860e-01 -6.74588442e-01
-1.19782940e-01 -9.58347917e-01 -2.08105311e-01 9.75431263e-01
-1.45049798e+00 -1.81021467e-01 1.22731066e+00 -7.76178002e-01
-1.87176156e+00 -1.68098128e+00 9.63760257e-01 2.80686677e-01
7.73411274e-01 5.48805296e-01 -6.72451019e-01 7.41385520e-01
1.18339591e-01 -3.77315342e-01 -1.35910660e-01 -9.23406780e-01
7.19128728e-01 -1.15980953e-01 -1.40742612e+00 6.92310333e-01
5.09686291e-01 1.51860118e-01 -2.22347770e-02 3.31497937e-01
9.65608895e-01 -3.52238387e-01 -7.68590987e-01 5.01247168e-01
3.08135569e-01 -1.72081888e-01 7.51996279e-01 -5.05180538e-01
-1.47205353e-01 -4.74900544e-01 -3.15883875e-01 -1.17151797e+00
-4.19645190e-01 -6.85733020e-01 -6.86383471e-02 9.90886211e-01
4.21431035e-01 -6.64636433e-01 8.32779765e-01 1.18263018e+00
-6.45719767e-01 -8.80781412e-01 -1.40146327e+00 -7.60902882e-01
-4.37680259e-02 -4.85274732e-01 2.07176208e-01 4.42690611e-01
-6.37977496e-02 -1.97418436e-01 -5.38784504e-01 8.86418521e-01
9.45585251e-01 -5.27344048e-02 8.05551589e-01 -5.79017460e-01
8.81366879e-02 -2.83078849e-01 -3.09089303e-01 -1.04690742e+00
6.57121658e-01 -6.31943524e-01 4.82209086e-01 -1.52980900e+00
-1.65614709e-01 -3.90003890e-01 3.30427915e-01 -7.28547573e-02
1.27741113e-01 -5.96801519e-01 3.84990782e-01 -3.73434462e-02
-5.01939237e-01 6.62241340e-01 1.28251696e+00 -4.26199853e-01
-5.50553083e-01 6.12012327e-01 -5.60803190e-02 4.18479919e-01
8.84372711e-01 -3.64567876e-01 -4.67100799e-01 -3.19997102e-01
-3.01551044e-01 1.05385947e+00 3.23234081e-01 -6.39568150e-01
5.35658956e-01 -1.19888878e+00 -6.90095246e-01 -1.19740677e+00
5.07025659e-01 -1.21347916e+00 5.89199245e-01 7.01252222e-01
-1.90709338e-01 8.05109590e-02 9.95150432e-02 1.00393367e+00
-2.54113257e-01 -2.60128319e-01 9.35793638e-01 3.06202501e-01
-1.16011333e+00 5.66545129e-01 -1.11738348e+00 -5.27849853e-01
1.96652174e+00 -5.20631433e-01 4.39554937e-02 -4.04219031e-01
-7.35419154e-01 1.36865258e+00 1.99114427e-01 4.06750739e-01
5.23174167e-01 -1.11451590e+00 -6.14634871e-01 -2.80525863e-01
-2.79848397e-01 -7.13214055e-02 5.73950112e-01 1.11405277e+00
-5.10037482e-01 8.48177373e-01 -1.83110848e-01 -4.05139357e-01
-1.09405231e+00 6.92849755e-01 4.03723270e-01 -2.58807003e-01
-2.04422176e-01 6.78662285e-02 7.86880627e-02 -2.60761082e-01
-5.87241538e-02 3.41151282e-02 -5.03331423e-01 9.30099487e-02
1.53439850e-01 1.25166750e+00 -5.29221952e-01 -1.13159239e+00
-8.17080364e-02 3.75350386e-01 3.13102573e-01 -2.97820866e-01
1.22537637e+00 -1.00976062e+00 -3.22215557e-02 9.93920490e-02
1.22611606e+00 -5.12387872e-01 -1.75635076e+00 1.52937993e-01
1.90437898e-01 -4.07294691e-01 3.46910596e-01 -1.78635508e-01
-9.47820425e-01 5.95463403e-02 3.95207912e-01 -1.16384283e-01
7.68444180e-01 -4.55599874e-01 7.15306878e-01 8.78310874e-02
5.30430019e-01 -1.48766184e+00 -8.29316556e-01 6.91193998e-01
8.48006606e-01 -9.80040014e-01 6.93950355e-02 -8.03755164e-01
-9.03492272e-01 9.22666371e-01 6.66084468e-01 -2.71170139e-01
7.77829409e-01 1.34294942e-01 -5.09026051e-01 3.16947639e-01
-1.13962317e+00 -2.22364068e-01 -1.57307312e-02 3.53260815e-01
-5.02009153e-01 5.35548866e-01 -7.97395349e-01 1.06115356e-01
2.99490005e-01 6.30885586e-02 8.30895126e-01 9.80315924e-01
-9.27984416e-01 -6.82117820e-01 -5.44362307e-01 2.45885968e-01
4.63856943e-02 6.52787089e-01 3.32129151e-01 7.24705756e-01
1.35663599e-01 1.23204625e+00 1.07527003e-01 -4.56829034e-02
4.43574995e-01 -3.07329327e-01 -2.70017385e-01 -3.34678531e-01
8.02483261e-02 -4.70456146e-02 4.16882783e-01 -5.60640872e-01
-2.97676116e-01 -8.69600832e-01 -1.45280266e+00 -3.42247784e-01
-1.60146311e-01 1.01713508e-01 7.55100727e-01 1.06146634e+00
2.65524566e-01 2.71768957e-01 1.18838811e+00 -1.07392728e+00
-3.94331664e-01 -2.99719602e-01 -3.37329298e-01 -5.11144578e-01
4.86841232e-01 -6.52502716e-01 -1.52566403e-01 -1.64114773e-01] | [5.449989318847656, 1.7033543586730957] |
72c0a88b-be8f-4e76-9028-28192a00ea6c | is-cross-modal-information-retrieval-possible | 2304.11095 | null | https://arxiv.org/abs/2304.11095v1 | https://arxiv.org/pdf/2304.11095v1.pdf | Is Cross-modal Information Retrieval Possible without Training? | Encoded representations from a pretrained deep learning model (e.g., BERT text embeddings, penultimate CNN layer activations of an image) convey a rich set of features beneficial for information retrieval. Embeddings for a particular modality of data occupy a high-dimensional space of its own, but it can be semantically aligned to another by a simple mapping without training a deep neural net. In this paper, we take a simple mapping computed from the least squares and singular value decomposition (SVD) for a solution to the Procrustes problem to serve a means to cross-modal information retrieval. That is, given information in one modality such as text, the mapping helps us locate a semantically equivalent data item in another modality such as image. Using off-the-shelf pretrained deep learning models, we have experimented the aforementioned simple cross-modal mappings in tasks of text-to-image and image-to-text retrieval. Despite simplicity, our mappings perform reasonably well reaching the highest accuracy of 77% on recall@10, which is comparable to those requiring costly neural net training and fine-tuning. We have improved the simple mappings by contrastive learning on the pretrained models. Contrastive learning can be thought as properly biasing the pretrained encoders to enhance the cross-modal mapping quality. We have further improved the performance by multilayer perceptron with gating (gMLP), a simple neural architecture. | ['Youngjune L. Gwon', 'Seongho Joe', 'Hyunjae Lee', 'Hyunjin Choi'] | 2023-04-20 | null | null | null | null | ['cross-modal-information-retrieval'] | ['miscellaneous'] | [ 2.23881364e-01 -1.02975243e-03 -2.56276667e-01 -4.59588796e-01
-1.05822861e+00 -5.30134559e-01 1.01443124e+00 2.51378149e-01
-7.38466203e-01 2.18765005e-01 2.86798477e-01 -2.24248856e-01
-1.79704264e-01 -7.47469723e-01 -1.06866062e+00 -4.76833194e-01
1.37182131e-01 5.47830224e-01 -4.56604622e-02 -3.06499481e-01
7.24435076e-02 3.97247642e-01 -1.49390161e+00 7.96179116e-01
3.09535295e-01 1.36429346e+00 2.75938183e-01 5.43733001e-01
-3.21704626e-01 5.28576076e-01 -4.50347900e-01 -6.20989263e-01
3.18234414e-01 -1.09382167e-01 -8.81233811e-01 -1.73347548e-01
7.77537346e-01 -5.11375129e-01 -7.34880388e-01 9.27888036e-01
3.53173733e-01 -1.82211399e-02 1.00086868e+00 -1.08402586e+00
-1.37279844e+00 3.10302585e-01 -5.50862014e-01 5.69158532e-02
4.03583348e-01 -1.22661807e-01 1.28018486e+00 -1.27213323e+00
5.90261579e-01 1.40305412e+00 4.89483416e-01 4.60643739e-01
-1.18850136e+00 -3.95683974e-01 -1.04773059e-01 1.25340939e-01
-1.33551216e+00 -2.83414811e-01 5.53384423e-01 -3.78738791e-01
9.38342392e-01 2.36738324e-01 4.12752301e-01 1.17798007e+00
4.28305149e-01 1.01131094e+00 8.16483200e-01 -5.19055605e-01
-2.18375057e-01 2.71657258e-01 -1.62005007e-01 6.15808666e-01
2.05308162e-02 9.92231257e-03 -3.65149945e-01 -5.12960367e-02
6.81067348e-01 4.61408436e-01 -1.97266117e-01 -4.50694472e-01
-1.48787010e+00 8.99097443e-01 1.16505134e+00 3.60881329e-01
-2.76831418e-01 4.29373324e-01 5.49073577e-01 5.20410359e-01
3.01505953e-01 7.69037604e-01 -4.44098860e-01 2.78681278e-01
-8.66757393e-01 -5.62722608e-02 5.26179612e-01 8.71110857e-01
8.80898535e-01 -3.54245573e-01 -2.84742028e-01 9.59995389e-01
4.36242670e-01 5.07547796e-01 7.97110796e-01 -5.38303494e-01
5.49646378e-01 6.04435205e-01 -5.31390458e-02 -1.07145977e+00
-3.50463480e-01 -3.21185440e-01 -8.99813175e-01 -9.68733430e-02
2.93377489e-01 2.53878623e-01 -1.12862098e+00 1.67390800e+00
-1.73329085e-01 -2.26617590e-01 1.40515760e-01 1.10729790e+00
9.00260508e-01 8.54956269e-01 7.01830983e-02 2.80305028e-01
1.59612548e+00 -1.16031969e+00 -6.24241114e-01 -3.85277092e-01
6.66181922e-01 -8.18258166e-01 1.26239049e+00 -3.87446247e-02
-1.06596017e+00 -7.09190190e-01 -1.18183935e+00 -5.38842261e-01
-8.96337867e-01 4.72034961e-02 5.05380690e-01 1.52482763e-01
-1.27977252e+00 4.49112535e-01 -5.26416361e-01 -5.35720289e-01
3.58478963e-01 3.36394936e-01 -7.55555332e-01 -2.67107189e-01
-1.45793569e+00 1.06351900e+00 3.96552503e-01 -1.04696929e-01
-7.35253513e-01 -6.41751111e-01 -9.11593258e-01 3.03986520e-01
9.39176455e-02 -8.67852509e-01 1.05551624e+00 -1.14720380e+00
-1.04421961e+00 1.30106604e+00 -2.97496263e-02 -4.74062800e-01
2.57993132e-01 -3.03042859e-01 -4.98779714e-01 4.69423264e-01
1.84088379e-01 1.27660835e+00 1.12426984e+00 -1.26421332e+00
-3.15249085e-01 -4.66170520e-01 3.00787896e-01 2.13915929e-01
-7.35460281e-01 -1.94722153e-02 -6.39913261e-01 -7.68168628e-01
1.87687650e-01 -8.09696913e-01 1.91602841e-01 3.92264605e-01
-1.49526596e-01 -2.29669139e-01 6.52127445e-01 -6.77480161e-01
1.07887816e+00 -2.33549023e+00 1.29773110e-01 -2.20443495e-02
1.52874976e-01 8.10344219e-02 -5.72616041e-01 6.67541564e-01
-3.18913907e-01 2.43259370e-01 1.57011636e-02 -3.97328109e-01
1.98677525e-01 2.66159922e-01 -5.39321840e-01 3.57162863e-01
4.17326003e-01 1.44302297e+00 -8.03681195e-01 -4.12439764e-01
8.93585086e-02 5.40320575e-01 -4.34072614e-01 2.80404925e-01
-1.39933854e-01 -1.67463198e-01 -8.99232104e-02 4.51115072e-01
4.78979975e-01 -5.49669445e-01 -1.74770713e-01 -5.25021195e-01
4.91997868e-01 2.26552546e-01 -7.45550990e-01 2.16250777e+00
-7.26933300e-01 8.28324556e-01 -2.93654591e-01 -1.13917172e+00
7.06361234e-01 2.63560653e-01 4.31864649e-01 -1.17687488e+00
7.75093213e-02 2.45576322e-01 -2.28641599e-01 -4.58119273e-01
8.55304897e-01 -1.96558371e-01 -2.61745185e-01 6.14990056e-01
4.50929552e-01 1.04736269e-01 -1.75216928e-01 2.59745449e-01
7.95077562e-01 -1.07837543e-01 6.89387843e-02 -1.16010271e-01
2.56162852e-01 -1.52511984e-01 -3.30643743e-01 8.64901423e-01
1.56111360e-01 9.29410160e-01 3.33710372e-01 -5.69395602e-01
-1.23656023e+00 -1.11548972e+00 -3.66623670e-01 1.46176958e+00
1.84249803e-01 -3.47528726e-01 -4.17704165e-01 -6.35984957e-01
1.11760937e-01 3.02586257e-01 -9.26316023e-01 -4.09230828e-01
-2.33172417e-01 -4.88393575e-01 5.92898369e-01 5.76906443e-01
3.68398577e-01 -9.68222082e-01 -3.14222276e-01 -2.48498619e-02
-2.78834105e-02 -1.05296624e+00 -6.03720129e-01 4.44512188e-01
-7.03993320e-01 -8.50204825e-01 -1.00336123e+00 -1.12907243e+00
7.18811035e-01 3.56247813e-01 1.23074925e+00 1.07807375e-01
-1.91886961e-01 6.81689560e-01 -2.93546468e-01 -1.43057704e-01
-1.83354557e-01 8.03054944e-02 -1.27945635e-02 7.97443092e-02
4.65213507e-01 -1.25791132e-01 -8.21675062e-01 2.34874442e-01
-1.40439320e+00 -7.29608536e-02 7.30697155e-01 1.27783704e+00
6.08983815e-01 -2.07757592e-01 3.29363614e-01 -6.63660407e-01
8.16549838e-01 -5.51540315e-01 -1.85868993e-01 4.34035033e-01
-4.44294959e-01 3.65936995e-01 4.55700994e-01 -4.67590451e-01
-6.00785077e-01 -1.32235810e-01 -9.80835557e-02 -7.19191432e-01
1.58421472e-02 7.51119435e-01 1.05754279e-01 1.12674914e-01
7.99551725e-01 3.00265819e-01 1.57991782e-01 -4.68270510e-01
6.54417098e-01 7.55440295e-01 4.86661345e-01 -4.65808362e-01
4.74506140e-01 4.51534331e-01 -1.39853328e-01 -3.40721786e-01
-1.01880836e+00 -4.19752628e-01 -6.20228469e-01 1.52358189e-01
9.98928547e-01 -1.08733404e+00 -5.67800760e-01 6.90926686e-02
-1.15500355e+00 -1.03791341e-01 -2.65312642e-01 3.64895254e-01
-4.86386955e-01 8.20238665e-02 -6.01542771e-01 -1.28731146e-01
-2.72128403e-01 -9.56337035e-01 1.44196141e+00 -5.42449728e-02
-5.94295971e-02 -1.24908626e+00 -6.58299923e-02 1.36371762e-01
4.94744182e-01 -1.28137365e-01 1.04416084e+00 -9.55714524e-01
-3.66645694e-01 -4.53029066e-01 -5.77871025e-01 3.60413462e-01
4.90287542e-02 -3.34425002e-01 -1.08086562e+00 -6.28753245e-01
-2.73220725e-02 -5.69239378e-01 1.12934422e+00 1.70906231e-01
1.40636611e+00 -3.54878515e-01 -4.26028669e-01 5.69416583e-01
1.53833759e+00 -6.05542101e-02 6.73898458e-01 6.04130447e-01
6.83698475e-01 5.22020936e-01 3.66663486e-01 8.29652771e-02
3.94216120e-01 6.29545152e-01 5.33371747e-01 -3.18795949e-01
-2.46237770e-01 -5.10428548e-01 3.76616716e-02 6.32920027e-01
5.05126774e-01 -3.58182311e-01 -8.02713811e-01 6.04281604e-01
-1.73758793e+00 -7.60285258e-01 4.02908683e-01 1.99112308e+00
7.85884500e-01 -3.79378647e-02 -3.43528032e-01 -2.01968357e-01
5.51125646e-01 2.70951927e-01 -4.12975639e-01 -4.22222018e-01
-1.64912030e-01 1.30031571e-01 3.89391869e-01 3.10172200e-01
-1.17824376e+00 7.52140284e-01 6.46392441e+00 8.66094530e-01
-1.41025925e+00 9.97463837e-02 6.03970587e-01 -3.33486833e-02
-4.81940657e-01 -3.26414466e-01 -5.35520732e-01 4.29233551e-01
9.49740887e-01 1.46614924e-01 4.73554432e-01 6.74285114e-01
-4.34793591e-01 1.79854512e-01 -1.61479938e+00 1.34072244e+00
2.61761367e-01 -1.50385845e+00 5.50258577e-01 -5.96780293e-02
7.38188863e-01 2.14170948e-01 4.98513728e-01 4.58363742e-01
1.13896132e-02 -1.25173688e+00 5.09651184e-01 6.25629008e-01
1.03016841e+00 -4.57781166e-01 7.25221038e-01 1.63568065e-01
-8.01757157e-01 -5.42894080e-02 -5.59774518e-01 2.78970629e-01
-1.26147568e-01 1.57066718e-01 -6.78013444e-01 4.09747422e-01
6.15576923e-01 7.84553051e-01 -7.60896444e-01 7.34795153e-01
1.19561926e-01 -5.05077094e-02 -2.43451938e-01 1.45003766e-01
5.24123013e-01 2.78703630e-01 1.34676665e-01 1.22625148e+00
3.26578617e-01 -2.12769300e-01 -7.86751434e-02 7.99593329e-01
-5.01313984e-01 4.53041829e-02 -8.28013420e-01 -1.79388344e-01
2.81203598e-01 1.22628224e+00 -3.65421653e-01 -4.48823571e-01
-6.94614232e-01 1.17720008e+00 4.21020627e-01 6.04729772e-01
-6.85006618e-01 -4.10833567e-01 5.82465112e-01 1.04476534e-01
3.91053647e-01 3.79430875e-02 -2.56503671e-01 -1.24229038e+00
1.26688451e-01 -6.65649354e-01 5.52326381e-01 -1.16588950e+00
-1.60880041e+00 7.72931099e-01 -2.29269281e-01 -1.35491765e+00
-3.41776162e-01 -8.80640924e-01 -1.63059518e-01 9.65712786e-01
-1.57278395e+00 -1.25291312e+00 -1.35733455e-01 8.46481383e-01
3.08141649e-01 -3.02028775e-01 1.03097427e+00 3.81850272e-01
-7.87057634e-03 8.69427085e-01 3.28993291e-01 3.43423039e-01
1.05215144e+00 -1.15640306e+00 2.45038882e-01 4.38077807e-01
5.94917536e-01 8.87543559e-01 2.78903663e-01 -1.83278292e-01
-1.63378537e+00 -1.04323339e+00 8.07837129e-01 -5.12093604e-01
9.71808493e-01 -4.63052094e-01 -9.97803271e-01 8.16006005e-01
5.31959713e-01 2.95064300e-01 6.99973464e-01 5.47782481e-02
-7.42569804e-01 -1.57515004e-01 -9.56005871e-01 6.30204797e-01
6.93028629e-01 -1.21505260e+00 -8.14270377e-01 4.91202384e-01
1.01840460e+00 -4.65547681e-01 -1.07377458e+00 1.89780459e-01
6.73325598e-01 -6.61512494e-01 1.20239282e+00 -9.87291217e-01
8.62958014e-01 -6.84073195e-02 -5.05595982e-01 -1.31418848e+00
-3.74460578e-01 -1.17260553e-01 -1.17277175e-01 8.91156495e-01
4.63689178e-01 -5.16163707e-01 4.27891403e-01 5.99537313e-01
-8.77579972e-02 -8.88318419e-01 -7.80436814e-01 -6.50076389e-01
2.76790082e-01 -3.93740356e-01 6.11596048e-01 1.00701308e+00
1.88816749e-02 5.27724445e-01 -2.46691972e-01 6.16764389e-02
1.94343135e-01 6.14870973e-02 5.98629832e-01 -9.85213637e-01
-2.07507521e-01 -4.34312582e-01 -4.79000479e-01 -1.50756621e+00
2.01774031e-01 -1.39322805e+00 -4.38054390e-02 -1.63734925e+00
3.96654308e-01 -3.68165076e-01 -6.72926307e-01 5.86360276e-01
-5.90078123e-02 5.93139112e-01 3.67767811e-01 4.09267843e-01
-7.15067625e-01 5.21458805e-01 1.35132134e+00 -5.68994582e-01
2.25420982e-01 -4.98804957e-01 -9.13602352e-01 3.31524283e-01
4.47946668e-01 -4.35387075e-01 -2.10827515e-01 -1.01082933e+00
5.38007855e-01 2.13781238e-01 5.35371184e-01 -6.07292652e-01
4.27535504e-01 4.26725745e-01 7.16656148e-01 -4.78903562e-01
5.37231326e-01 -1.20307803e+00 -9.56409276e-02 2.98451573e-01
-7.37885296e-01 2.00623930e-01 3.88214737e-01 5.72886288e-01
-7.16179550e-01 -2.88490385e-01 3.81710589e-01 -6.17962144e-02
-7.78748870e-01 3.70373577e-01 -1.30442888e-01 -4.28040214e-02
6.42476499e-01 -2.01189220e-01 -4.74301636e-01 -4.49250013e-01
-8.70607734e-01 1.74573943e-01 3.98955077e-01 8.61175478e-01
7.25372553e-01 -1.69929552e+00 -4.01102006e-01 3.70482206e-01
5.79549313e-01 -2.52618194e-01 2.56003737e-02 5.48700452e-01
-3.51408899e-01 7.71484554e-01 -4.42677677e-01 -9.56533790e-01
-8.79914224e-01 9.19847667e-01 3.12409848e-01 -3.10831487e-01
-3.50576192e-01 7.72298634e-01 5.04567802e-01 -5.19267142e-01
2.04297781e-01 -2.73274779e-01 -1.55963060e-02 2.44658113e-01
6.08997107e-01 -1.61725625e-01 2.61358470e-01 -5.47322690e-01
-3.13729584e-01 6.05303288e-01 -3.43479693e-01 -6.20924979e-02
1.26840329e+00 -1.52065665e-01 -2.72786707e-01 3.64857554e-01
2.03393650e+00 -5.02891302e-01 -1.01970994e+00 -5.85791469e-01
-1.08228825e-01 -4.16365385e-01 1.35761917e-01 -7.96927989e-01
-1.06069767e+00 1.11961579e+00 6.78905845e-01 2.70772398e-01
1.13715160e+00 2.44653940e-01 7.28689134e-01 7.16224134e-01
1.74052462e-01 -7.97267139e-01 5.04872262e-01 5.42840004e-01
1.14606404e+00 -1.59658742e+00 -1.75350040e-01 3.27675700e-01
-6.12789869e-01 1.25358272e+00 4.61133987e-01 -3.35854709e-01
8.13199878e-01 -4.96046618e-02 -5.86641859e-03 -4.18990016e-01
-7.49567091e-01 -3.40111367e-02 8.11261058e-01 2.71126717e-01
3.64575535e-01 -5.27904071e-02 2.45669615e-02 3.65458161e-01
-4.28108452e-03 -2.92780638e-01 -1.27605405e-02 7.19638944e-01
-2.50711292e-01 -8.14393759e-01 -2.98058897e-01 6.04412436e-01
-5.19753337e-01 -3.77417594e-01 -2.43520901e-01 8.16013694e-01
-2.31689662e-01 4.70498115e-01 4.72879827e-01 -3.77613634e-01
1.70130089e-01 1.22074530e-01 5.62063873e-01 -5.61305881e-01
-5.10088682e-01 -1.50853515e-01 -3.96227300e-01 -5.00458062e-01
-3.78866822e-01 -1.26114473e-01 -9.75767553e-01 -1.09893851e-01
-1.72825038e-01 -1.74362987e-01 7.41966724e-01 9.37774658e-01
4.22572196e-01 4.87673312e-01 5.39125144e-01 -8.68337035e-01
-4.27303404e-01 -9.88228977e-01 -4.00483370e-01 7.66749144e-01
4.44375008e-01 -5.48441648e-01 -1.22130625e-01 4.89761457e-02] | [10.629119873046875, 1.7162545919418335] |
b48b83e8-e93a-4490-a0cc-43f77df9f1aa | a-priori-compression-of-convolutional-neural | 2304.04964 | null | https://arxiv.org/abs/2304.04964v2 | https://arxiv.org/pdf/2304.04964v2.pdf | A priori compression of convolutional neural networks for wave simulators | Convolutional neural networks are now seeing widespread use in a variety of fields, including image classification, facial and object recognition, medical imaging analysis, and many more. In addition, there are applications such as physics-informed simulators in which accurate forecasts in real time with a minimal lag are required. The present neural network designs include millions of parameters, which makes it difficult to install such complex models on devices that have limited memory. Compression techniques might be able to resolve these issues by decreasing the size of CNN models that are created by reducing the number of parameters that contribute to the complexity of the models. We propose a compressed tensor format of convolutional layer, a priori, before the training of the neural network. 3-way kernels or 2-way kernels in convolutional layers are replaced by one-way fiters. The overfitting phenomena will be reduced also. The time needed to make predictions or time required for training using the original Convolutional Neural Networks model would be cut significantly if there were fewer parameters to deal with. In this paper we present a method of a priori compressing convolutional neural networks for finite element (FE) predictions of physical data. Afterwards we validate our a priori compressed models on physical data from a FE model solving a 2D wave equation. We show that the proposed convolutinal compression technique achieves equivalent performance as classical convolutional layers with fewer trainable parameters and lower memory footprint. | ['David Ryckelynck', 'Fabien Casenave', 'Nissrine Akkari', 'Hamza Boukraichi'] | 2023-04-11 | null | null | null | null | ['object-recognition'] | ['computer-vision'] | [ 1.50116652e-01 1.05887137e-01 3.37524354e-01 -4.79233563e-01
1.80138499e-01 -5.96283898e-02 8.72352496e-02 -6.11237995e-02
-5.44701338e-01 5.58194101e-01 -3.37023020e-01 -6.02495253e-01
-4.59693968e-01 -1.04098535e+00 -1.07203019e+00 -6.12376690e-01
-2.47018650e-01 3.02873969e-01 1.81552470e-01 -2.65376002e-01
2.77534612e-02 8.60525727e-01 -1.64536715e+00 3.11198026e-01
4.94048506e-01 1.46106815e+00 2.32475027e-01 6.28102958e-01
1.20195471e-01 7.13402927e-01 -2.01255456e-01 -2.18705431e-01
3.03355902e-01 1.46659806e-01 -8.38427842e-01 -2.11035818e-01
1.51367798e-01 -6.93031311e-01 -6.15807772e-01 7.81495214e-01
2.61614352e-01 1.92835718e-01 4.33564752e-01 -7.62413979e-01
-1.85947698e-02 4.94782716e-01 -1.59018338e-01 4.92157452e-02
-4.54858184e-01 1.62271172e-01 2.52817899e-01 -5.61044574e-01
1.68793216e-01 1.05337608e+00 1.05964994e+00 3.55675012e-01
-9.91958380e-01 -5.83218753e-01 -5.24980545e-01 9.52468961e-02
-1.27727401e+00 -3.66296738e-01 7.70970523e-01 -3.69289905e-01
1.25912488e+00 4.05944943e-01 7.22245574e-01 4.54680175e-01
5.65908909e-01 1.82291325e-02 7.68783510e-01 -3.01917493e-01
2.64923096e-01 9.20513831e-03 2.14608863e-01 8.56554806e-01
4.37578022e-01 2.18729809e-01 -1.22566402e-01 -1.56544805e-01
9.04852629e-01 7.44870827e-02 -2.53346860e-01 4.27951552e-02
-8.46486270e-01 9.04090762e-01 4.63842362e-01 2.62183398e-01
-5.57505429e-01 6.24308884e-01 6.09049380e-01 1.39104515e-01
4.58032459e-01 4.34786379e-01 -7.50062346e-01 -5.20631187e-02
-1.25880027e+00 3.63806337e-01 7.63571143e-01 6.54857695e-01
6.35712028e-01 5.06911397e-01 4.93922979e-01 8.02390873e-01
5.77238277e-02 4.82520074e-01 6.81338370e-01 -9.64788675e-01
1.17766753e-01 5.32615006e-01 -2.44408622e-01 -1.29265010e+00
-6.46130919e-01 -6.40331805e-01 -1.41056323e+00 1.31750599e-01
7.35765975e-03 -2.44717747e-01 -7.93341398e-01 1.27697623e+00
2.53490090e-01 2.93598086e-01 8.85256827e-02 8.86546135e-01
6.69001281e-01 8.52175593e-01 -2.81209260e-01 4.99994457e-02
1.37432480e+00 -6.65652633e-01 -3.80728692e-01 -4.56972383e-02
9.81922984e-01 -8.46402287e-01 6.35168672e-01 5.36878526e-01
-1.04877698e+00 -7.99294829e-01 -1.23225832e+00 -5.69132203e-03
-3.37399691e-01 2.85504788e-01 9.91740882e-01 6.05601490e-01
-9.17079329e-01 1.39416218e+00 -1.07527423e+00 1.51526332e-01
2.91282833e-01 8.65945458e-01 -2.77133763e-01 1.49547800e-01
-1.26930153e+00 8.21625471e-01 5.20798683e-01 4.51261491e-01
-5.92029929e-01 -7.63116241e-01 -6.44728839e-01 5.03829896e-01
-1.25131935e-01 -5.25387883e-01 1.02304697e+00 -8.45817864e-01
-1.38686144e+00 1.35264710e-01 1.79932743e-01 -7.81474829e-01
8.05211514e-02 -4.87046391e-02 -3.66004348e-01 7.13893548e-02
-6.88395619e-01 5.72795987e-01 7.79690385e-01 -6.64380193e-01
-7.93647952e-03 1.05773814e-01 1.06803402e-01 -3.49512160e-01
-4.86546457e-01 -2.43739471e-01 1.29322782e-01 -5.14268577e-01
2.19137445e-01 -9.98729110e-01 -4.80132341e-01 1.55436784e-01
-1.23035096e-01 1.01494595e-01 1.03659308e+00 -7.46836960e-01
8.84041488e-01 -1.94074214e+00 -1.53793588e-01 3.27544212e-01
3.37332338e-01 6.27692223e-01 1.67006254e-02 1.83819100e-01
-2.54818469e-01 1.56459492e-02 -1.91802397e-01 -1.49230510e-01
-4.14169908e-01 4.68514860e-01 -3.34364057e-01 4.29185897e-01
1.72928035e-01 5.10582328e-01 -1.79774716e-01 -3.21952999e-01
8.51375163e-02 7.94695973e-01 -9.31280851e-01 4.40443903e-02
-1.64904743e-01 1.46056682e-01 -4.23714578e-01 4.56947535e-02
9.48800683e-01 -3.93257588e-01 -8.19308385e-02 -5.27042925e-01
-7.98127651e-02 3.95966917e-01 -1.12522125e+00 1.21623230e+00
-6.19838595e-01 7.04424500e-01 6.44713268e-02 -1.45624399e+00
9.51080084e-01 4.33101326e-01 5.54229975e-01 -5.07420778e-01
5.31023562e-01 1.77435040e-01 2.74018288e-01 -4.26356643e-01
6.54371500e-01 -1.39205933e-01 1.53239295e-01 1.80335522e-01
5.00510447e-02 -2.12858722e-01 -1.19652055e-01 -4.51546209e-03
9.56286013e-01 -1.10283323e-01 -2.35875338e-01 -3.45470309e-01
3.94705892e-01 1.60314545e-01 4.41521257e-01 2.45457694e-01
3.62899125e-01 2.58606553e-01 2.79674143e-01 -9.56949294e-01
-1.55191934e+00 -2.50432938e-01 -3.20311546e-01 3.19383651e-01
-3.84560168e-01 -4.43281740e-01 -8.25952053e-01 2.09615707e-01
-2.65854836e-01 4.12307143e-01 -2.73236811e-01 -3.34486663e-01
-8.12558711e-01 -6.41917944e-01 7.29460120e-01 5.07391751e-01
6.70502245e-01 -7.12375820e-01 -9.72847641e-01 2.61272043e-01
2.74148345e-01 -1.19711232e+00 3.05093497e-01 3.84079307e-01
-1.50153244e+00 -7.52866745e-01 -2.95140803e-01 -7.48079836e-01
6.77084148e-01 -4.49268855e-02 7.08631635e-01 5.05413890e-01
-1.76853120e-01 -2.73533940e-01 -1.16119236e-01 -3.20864022e-01
-5.08930922e-01 7.99686834e-02 3.86753976e-01 -3.25093806e-01
-2.22278871e-02 -7.54130065e-01 -4.89088595e-01 5.54401204e-02
-1.09570384e+00 6.67047977e-01 4.78867173e-01 9.68034804e-01
3.60005915e-01 3.13811094e-01 1.49658203e-01 -6.99534416e-01
4.17178988e-01 -2.66327709e-01 -9.17171717e-01 -1.07258223e-01
-5.71514726e-01 4.57585841e-01 1.09292388e+00 -5.85998416e-01
-7.31523991e-01 9.23378766e-02 -3.74479860e-01 -8.38407874e-01
6.82275146e-02 8.30416203e-01 5.11782527e-01 -6.15866065e-01
4.92911965e-01 1.92437228e-03 2.09083334e-01 -6.83394432e-01
-1.49989009e-01 6.32631838e-01 2.82127202e-01 -4.96704310e-01
5.68941891e-01 1.15722820e-01 6.79848671e-01 -9.86452520e-01
-2.54440635e-01 4.41282056e-02 -3.42631638e-01 -2.31144726e-01
4.74288613e-01 -5.77517629e-01 -9.92290974e-01 4.89811540e-01
-1.49315274e+00 -1.70855403e-01 -1.56797692e-01 7.72890866e-01
-2.73569793e-01 2.53413647e-01 -9.41209197e-01 -4.24915165e-01
-5.49680889e-01 -1.30784011e+00 5.09894490e-01 1.52559787e-01
3.68029885e-02 -7.44858325e-01 -1.99511543e-01 7.96026662e-02
9.67595398e-01 6.24849424e-02 1.08685613e+00 -3.92851323e-01
-5.71821332e-01 -3.73766094e-01 -1.67476892e-01 8.13325405e-01
-2.78825343e-01 -4.27858122e-02 -9.29129779e-01 -3.59423488e-01
3.23438823e-01 -2.21460313e-01 6.35341406e-01 4.46260333e-01
1.65251791e+00 -6.88889265e-01 -2.32350558e-01 9.64549839e-01
1.52962470e+00 1.36425659e-01 7.03236580e-01 -2.34531462e-01
6.19728088e-01 5.37268281e-01 1.78815629e-02 3.96639347e-01
-2.61522323e-01 4.62238580e-01 4.29244548e-01 -1.27206489e-01
1.00626215e-01 1.89604312e-01 1.83614027e-02 1.53958237e+00
-4.29906189e-01 1.17678856e-02 -1.00286710e+00 2.85577595e-01
-1.33860230e+00 -7.05673039e-01 -2.57165670e-01 2.02538323e+00
5.56528330e-01 1.73751161e-01 -5.83643019e-01 4.09496129e-01
2.28928342e-01 -2.56073326e-01 -3.18350464e-01 -8.67881536e-01
2.28670672e-01 6.67230129e-01 8.49977911e-01 3.87965411e-01
-7.92306662e-01 5.35315096e-01 6.02822304e+00 8.51787508e-01
-1.74301386e+00 1.08638309e-01 8.48239541e-01 -1.87286600e-01
6.35209829e-02 1.80967852e-01 -6.69055045e-01 4.21292514e-01
1.60104167e+00 2.12226585e-01 4.68576282e-01 9.43152487e-01
2.65900820e-01 -8.29645246e-02 -9.42576289e-01 9.82605219e-01
-3.59829336e-01 -1.82149589e+00 1.65164787e-02 1.30647942e-01
5.01355052e-01 1.43319756e-01 -1.56120971e-01 9.84536186e-02
-2.44234383e-01 -1.20579422e+00 4.72228885e-01 7.30178714e-01
9.52579439e-01 -8.62676322e-01 9.11115229e-01 5.20587623e-01
-9.34414625e-01 1.10317126e-01 -9.31627154e-01 -4.92225438e-01
-3.06969695e-02 8.71490359e-01 -7.42074728e-01 3.50860775e-01
7.44802117e-01 1.39274240e-01 -3.81215364e-01 9.23107088e-01
7.56072462e-01 8.46114874e-01 -7.64846802e-01 -2.05235273e-01
2.74448514e-01 -1.32533550e-01 2.23488405e-01 8.17247927e-01
4.79750037e-01 3.49538624e-01 -2.48185664e-01 6.65159643e-01
7.21342303e-03 -1.28997520e-01 -5.22901893e-01 -1.40760392e-01
6.26503229e-02 1.12889099e+00 -6.83574498e-01 -3.72003376e-01
-2.79082924e-01 5.43544471e-01 3.30492780e-02 2.43879706e-02
-8.39460671e-01 -5.02261698e-01 4.75045085e-01 4.97534811e-01
2.67406911e-01 -4.52729374e-01 -2.40034580e-01 -9.20473337e-01
-1.20789044e-01 -5.17701507e-01 -2.81940669e-01 -8.71341646e-01
-7.16092288e-01 8.54260504e-01 5.70128597e-02 -1.19165015e+00
-3.14340055e-01 -1.00818992e+00 -3.46355051e-01 8.32721710e-01
-1.21065581e+00 -8.57129991e-01 -1.77509785e-01 2.86781758e-01
1.66496187e-01 -1.03198759e-01 1.00308919e+00 6.03243351e-01
-3.75304163e-01 4.50108975e-01 1.42220005e-01 2.47027874e-01
1.17216915e-01 -1.73514336e-01 2.32930660e-01 6.38836265e-01
-4.32264954e-01 6.63355589e-01 7.68549383e-01 -5.38406849e-01
-1.48121572e+00 -1.04660738e+00 8.01092684e-01 4.63379204e-01
3.48400295e-01 -2.10104927e-01 -1.04096651e+00 5.34799576e-01
-6.58669025e-02 2.99709350e-01 3.29176098e-01 -1.65265322e-01
1.33290589e-02 -3.47559333e-01 -1.15470457e+00 2.41672844e-01
4.28517759e-01 -2.91329592e-01 -1.64951921e-01 3.95777166e-01
8.80309820e-01 -5.35985172e-01 -1.03123057e+00 5.70525885e-01
5.76079071e-01 -7.46238947e-01 9.54265416e-01 -6.19129121e-01
8.10402572e-01 -1.37178302e-02 -1.60417348e-01 -1.03504193e+00
-2.31709629e-01 -4.22877610e-01 -1.35539949e-01 5.65599859e-01
2.54945338e-01 -5.90871632e-01 8.12769413e-01 7.25573003e-01
-3.71068865e-01 -1.28000236e+00 -9.87898350e-01 -6.00987554e-01
1.92152321e-01 -5.89297652e-01 8.27585876e-01 8.37863326e-01
-2.93494791e-01 4.39800881e-02 -4.25605774e-01 1.75062314e-01
3.20601106e-01 -2.03790635e-01 4.05044824e-01 -1.32039082e+00
-5.59809506e-01 -6.41918108e-02 -6.54512465e-01 -9.95786786e-01
-1.92918330e-01 -7.12373257e-01 -1.79455251e-01 -9.97643948e-01
-6.53225034e-02 -1.01929700e+00 5.34918234e-02 5.35120487e-01
6.75352871e-01 2.05747858e-01 4.92046066e-02 2.24198222e-01
1.42934844e-01 5.35189867e-01 1.12880552e+00 -1.81932393e-02
1.95255637e-01 -1.68435395e-01 3.55903581e-02 8.60373914e-01
8.91384184e-01 -6.11701071e-01 -3.50143701e-01 -6.43223822e-01
4.66317981e-01 4.12368685e-01 6.88257873e-01 -1.50963736e+00
4.25005764e-01 6.34611174e-02 5.83538771e-01 -5.40331304e-01
5.95214725e-01 -1.04031599e+00 5.97930729e-01 6.91031635e-01
-1.52939290e-01 3.56547862e-01 5.33852100e-01 8.89782514e-03
-1.34662896e-01 -6.16091967e-01 8.46439421e-01 -5.31373508e-02
-2.59648085e-01 3.43103617e-01 -4.46952313e-01 -7.14157164e-01
5.45347154e-01 -1.91510916e-01 -3.60058025e-02 -1.71655416e-01
-6.05854094e-01 -2.89035082e-01 1.14752702e-01 3.85635197e-02
7.69223869e-01 -1.13947439e+00 -3.91460180e-01 5.50417721e-01
-6.84558511e-01 2.93842196e-01 6.85711861e-01 6.59232020e-01
-1.16742635e+00 6.95130527e-01 -3.51769179e-01 -5.01318693e-01
-1.02484775e+00 4.49854076e-01 6.12389505e-01 -3.98745596e-01
-6.18048966e-01 6.08765960e-01 -2.12562978e-01 -4.31911319e-01
-6.71818256e-02 -7.31389999e-01 2.44124979e-02 -5.48689246e-01
4.47630763e-01 4.22927022e-01 5.53017974e-01 -5.33718169e-01
-7.45610595e-02 5.63759446e-01 1.19762998e-02 2.50357121e-01
1.67395377e+00 5.04288733e-01 -3.42125237e-01 3.59898806e-02
1.43611681e+00 -5.04441082e-01 -1.02923334e+00 -4.31970991e-02
-5.29547513e-01 -2.12396011e-02 6.20780051e-01 -3.96270543e-01
-1.48127294e+00 1.14920962e+00 7.51965940e-01 2.69779321e-02
1.25385320e+00 -3.72007012e-01 1.13003409e+00 7.81048656e-01
3.94746929e-01 -9.80851889e-01 -3.43742579e-01 7.38184810e-01
6.99546576e-01 -6.57255173e-01 2.73456454e-01 -4.20599669e-01
5.21620959e-02 1.50285959e+00 3.87480915e-01 -4.75949049e-01
1.19900560e+00 7.41970956e-01 -5.84151864e-01 -3.64569217e-01
-7.99869180e-01 5.42421401e-01 2.87173599e-01 6.44209608e-02
3.79557133e-01 4.99592870e-02 -1.93149224e-01 7.59560645e-01
-5.76966524e-01 2.62301117e-01 6.01804316e-01 7.93384612e-01
-4.43018377e-01 -8.58282685e-01 -2.99263716e-01 8.06846499e-01
-4.22362953e-01 -1.42697334e-01 4.56848145e-01 5.37853599e-01
3.28582555e-01 4.81090993e-01 3.48243892e-01 -7.67216206e-01
-3.64067517e-02 -1.04775848e-02 4.31286305e-01 -1.18952930e-01
-5.41936040e-01 -1.79175600e-01 7.29896054e-02 -5.68262517e-01
-3.09473544e-01 -1.11636229e-01 -1.32070816e+00 -1.04723215e+00
-5.20609856e-01 1.02132052e-01 1.08258438e+00 1.02300763e+00
4.70934451e-01 7.25047290e-01 2.71201968e-01 -1.08355153e+00
-6.11348689e-01 -9.71445084e-01 -3.30284953e-01 -4.15156521e-02
2.62303680e-01 -6.08905077e-01 -1.09210774e-01 1.23287588e-01] | [8.527362823486328, 2.9968721866607666] |
bd932d10-82cd-426e-ae3a-772dccd82f22 | deep-reformulated-laplacian-tone-mapping | 2102.00348 | null | https://arxiv.org/abs/2102.00348v1 | https://arxiv.org/pdf/2102.00348v1.pdf | Deep Reformulated Laplacian Tone Mapping | Wide dynamic range (WDR) images contain more scene details and contrast when compared to common images. However, it requires tone mapping to process the pixel values in order to display properly. The details of WDR images can diminish during the tone mapping process. In this work, we address the problem by combining a novel reformulated Laplacian pyramid and deep learning. The reformulated Laplacian pyramid always decompose a WDR image into two frequency bands where the low-frequency band is global feature-oriented, and the high-frequency band is local feature-oriented. The reformulation preserves the local features in its original resolution and condenses the global features into a low-resolution image. The generated frequency bands are reconstructed and fine-tuned to output the final tone mapped image that can display on the screen with minimum detail and contrast loss. The experimental results demonstrate that the proposed method outperforms state-of-the-art WDR image tone mapping methods. The code is made publicly available at https://github.com/linmc86/Deep-Reformulated-Laplacian-Tone-Mapping. | ['Orly Yadid-Pecht', 'Svetlana Yanushkevich', 'Mengchen Lin', 'Ziyi Liu', 'Jie Yang'] | 2021-01-31 | null | null | null | null | ['tone-mapping'] | ['computer-vision'] | [ 4.23974395e-01 -5.05086660e-01 -1.42638907e-01 -9.67956409e-02
-6.46315336e-01 -2.98419446e-01 1.43638954e-01 -4.89975661e-01
-1.38160259e-01 6.14008844e-01 3.32274526e-01 4.94509339e-02
-1.27732992e-01 -1.19241405e+00 -5.80591977e-01 -7.53178120e-01
1.66898802e-01 -4.31554586e-01 5.45419872e-01 -3.36361736e-01
2.29214162e-01 3.38173062e-01 -1.61307859e+00 6.29805088e-01
9.23001230e-01 1.02278411e+00 6.58488512e-01 5.92309833e-01
-2.48460528e-02 8.25619519e-01 -4.33299035e-01 1.90721020e-01
5.08302927e-01 -4.43650514e-01 -4.93427992e-01 7.94304013e-02
4.64331686e-01 -3.83720368e-01 -6.08726382e-01 1.53367507e+00
5.64150751e-01 1.07087970e-01 2.20405221e-01 -7.55216360e-01
-1.40766549e+00 2.12994814e-01 -1.17603207e+00 4.74562883e-01
3.07956845e-01 1.59924515e-02 5.82308948e-01 -9.92800474e-01
4.93800998e-01 1.34842408e+00 3.27170670e-01 2.33596787e-01
-1.43337107e+00 -9.75593805e-01 2.22243480e-02 2.97957897e-01
-1.59324801e+00 -1.93876684e-01 1.08678913e+00 -1.49331734e-01
6.47876441e-01 4.01352614e-01 5.03053367e-01 5.08524358e-01
5.18123031e-01 1.84546441e-01 1.64532244e+00 -4.11147892e-01
-2.54615575e-01 -8.69826507e-03 -1.60176337e-01 6.25747204e-01
1.79151744e-01 2.77003139e-01 -5.75103641e-01 3.71462792e-01
1.26897264e+00 1.88113287e-01 -7.66515255e-01 -1.24023214e-01
-1.21077907e+00 4.40770000e-01 6.29675150e-01 5.96692324e-01
-1.70104027e-01 8.26875418e-02 1.86981726e-02 5.07690251e-01
5.05761147e-01 5.59656918e-02 2.35906970e-02 1.73082918e-01
-7.91983485e-01 -2.00545847e-01 3.14577371e-02 6.05051041e-01
1.08578086e+00 1.47717297e-01 -1.41536996e-01 1.33643198e+00
4.17063832e-02 6.10435188e-01 6.34166181e-01 -9.46444154e-01
4.38556105e-01 2.99869388e-01 6.44694418e-02 -1.16304481e+00
-1.03437111e-01 -3.26773465e-01 -1.14269722e+00 5.88207006e-01
-2.48042531e-02 -1.23707354e-01 -7.52495468e-01 1.41270256e+00
7.76129812e-02 2.73233503e-01 -1.18198156e-01 1.32394493e+00
9.72781479e-01 1.35236537e+00 -1.80557549e-01 -1.10485248e-01
1.30876470e+00 -7.74130225e-01 -1.06056738e+00 -1.41488954e-01
-2.78832287e-01 -1.14887357e+00 1.63136744e+00 3.56951326e-01
-1.23688626e+00 -1.16552448e+00 -1.39858222e+00 -4.32949156e-01
-2.26752296e-01 1.95376545e-01 9.54981819e-02 4.20876861e-01
-1.11172771e+00 3.64583284e-01 -3.35925639e-01 4.57723923e-02
2.07360446e-01 -4.46134154e-03 -2.91131139e-01 -2.31565520e-01
-1.34490168e+00 7.02051520e-01 3.21773380e-01 -6.17870837e-02
-4.06929314e-01 -8.14005494e-01 -7.21663117e-01 2.08721668e-01
1.12522982e-01 -2.51236111e-01 9.67524052e-01 -8.09777975e-01
-1.57328212e+00 8.22072089e-01 -2.14808458e-03 1.50817245e-01
4.25534457e-01 4.54259291e-02 -7.69303799e-01 2.14333341e-01
4.19619121e-02 4.74532813e-01 9.67359722e-01 -1.39746678e+00
-8.72534096e-01 -1.01202674e-01 -4.98611964e-02 5.24101496e-01
-3.16595614e-01 -1.60769433e-01 -6.63117409e-01 -8.83219600e-01
1.93416446e-01 -4.09522772e-01 2.39044756e-01 1.08224452e-01
-1.35339618e-01 2.87040979e-01 1.28918600e+00 -6.30470216e-01
1.42936802e+00 -2.45044255e+00 -1.27715528e-01 7.95260891e-02
1.10498577e-01 1.33264720e-01 -1.32801130e-01 2.63164312e-01
-4.51601744e-01 -1.01322822e-01 -6.02566153e-02 2.29321435e-01
-3.53950053e-01 -4.40825939e-01 -4.53884304e-01 5.20864964e-01
3.12745348e-02 6.64203227e-01 -6.64920151e-01 -3.79317552e-01
7.02277660e-01 8.80006313e-01 -2.16657430e-01 3.65962125e-02
2.16161385e-01 3.26507509e-01 -2.49194741e-01 5.17763436e-01
1.17377198e+00 -2.51726627e-01 -1.01287618e-01 -8.04647028e-01
-5.00851572e-01 -1.10407434e-01 -1.28316152e+00 1.53164423e+00
-6.85042381e-01 9.62449551e-01 -1.66165666e-03 -5.35239995e-01
1.34040821e+00 -1.17463037e-01 2.40648299e-01 -1.28047454e+00
3.08029130e-02 1.43955708e-01 -2.66506761e-01 -3.81711602e-01
5.35371244e-01 -3.13360423e-01 -7.50368237e-02 3.51118118e-01
-3.64660233e-01 -2.42768064e-01 7.31263980e-02 -2.34466985e-01
4.76752222e-01 4.10747081e-02 2.99962729e-01 -3.05411875e-01
7.72698402e-01 -1.49618953e-01 3.30979645e-01 3.93446565e-01
-6.21898696e-02 9.43977594e-01 1.99702755e-01 -4.88520145e-01
-1.08622921e+00 -1.35470688e+00 -3.93848807e-01 1.09508908e+00
8.19047630e-01 -9.17691886e-02 -6.52301192e-01 8.90585557e-02
-2.91908950e-01 5.57984591e-01 -5.07105947e-01 -3.19345236e-01
-5.85628033e-01 -4.39345598e-01 1.18470997e-01 1.53056562e-01
1.20510960e+00 -1.03865087e+00 -7.39720941e-01 -3.38565279e-03
-4.73898679e-01 -8.01283777e-01 -1.00560844e+00 -8.86675641e-02
-4.13592041e-01 -7.00633645e-01 -9.65926111e-01 -1.25875509e+00
4.53120112e-01 7.53496408e-01 8.03863585e-01 -2.06894711e-01
-5.76659024e-01 -7.38546997e-02 -3.23602170e-01 1.80061057e-01
-1.33085281e-01 -3.19401503e-01 -2.63506889e-01 2.16255739e-01
2.36676365e-01 -6.49284840e-01 -9.24568295e-01 3.96557480e-01
-1.05673325e+00 4.12152737e-01 8.18167448e-01 7.73395956e-01
1.09092474e+00 6.70209885e-01 4.16255802e-01 -6.38384998e-01
8.19363713e-01 6.09976687e-02 -7.04888225e-01 2.48896152e-01
-1.20468281e-01 -2.81414241e-01 7.60903478e-01 -6.83038831e-01
-1.48509550e+00 -2.21497431e-01 9.49942470e-02 -4.28414881e-01
3.49471569e-02 1.04968302e-01 -1.68375909e-01 -2.53682911e-01
6.96208119e-01 6.60742104e-01 -2.47713804e-01 -4.66324002e-01
3.45350921e-01 8.33264410e-01 8.75746369e-01 -3.11058700e-01
9.35727775e-01 5.41358173e-01 -3.04999202e-01 -8.42696786e-01
-5.92328250e-01 -1.81964189e-02 -2.33205661e-01 -4.68582809e-01
9.19919372e-01 -1.06180024e+00 -3.98447514e-01 4.26023394e-01
-6.22713983e-01 -4.51385289e-01 -3.22887897e-01 3.54071856e-01
-5.65121531e-01 2.09776461e-01 -7.19070554e-01 -2.52015173e-01
-3.50301117e-01 -1.11427689e+00 9.41535473e-01 6.51315808e-01
2.55043864e-01 -5.73814869e-01 8.66570696e-02 -5.24782203e-02
7.15809405e-01 1.07221350e-01 8.35957587e-01 5.84148824e-01
-7.08108366e-01 1.54716805e-01 -7.00370014e-01 1.46181494e-01
6.61192060e-01 -2.14563236e-01 -8.89055789e-01 -3.93582225e-01
6.30697757e-02 -2.13456959e-01 9.14520085e-01 5.67713141e-01
1.42455363e+00 -7.58932307e-02 4.82944287e-02 9.04664874e-01
1.86440074e+00 3.35079223e-01 9.09237921e-01 4.83768731e-01
3.86447549e-01 1.92026600e-01 7.20427096e-01 1.94144756e-01
-2.30328087e-03 8.29280138e-01 -5.73538281e-02 -7.71139026e-01
-6.77776396e-01 -3.60646784e-01 1.80374652e-01 4.73495960e-01
3.97991464e-02 4.02502976e-02 -5.37418842e-01 2.29287609e-01
-1.43160367e+00 -1.17608809e+00 2.21874699e-01 2.05964470e+00
1.13035595e+00 8.10210705e-02 -3.18419844e-01 -8.11916497e-03
1.12020111e+00 5.30538201e-01 -6.21714950e-01 -3.88129473e-01
-3.67660791e-01 1.67589217e-01 4.19548124e-01 7.22461939e-01
-1.02638733e+00 7.82101274e-01 5.35079956e+00 1.26268840e+00
-1.54245317e+00 6.21195994e-02 8.44555676e-01 -2.53126711e-01
-2.66070813e-01 -4.98554349e-01 -4.55388486e-01 5.03485978e-01
3.17682117e-01 -5.07657230e-01 6.98884428e-01 6.13536000e-01
4.14758116e-01 -1.96972743e-01 -4.76012319e-01 1.37192583e+00
-6.28646910e-02 -1.24373496e+00 9.13600996e-02 -1.98258176e-01
9.85242605e-01 -3.15335810e-01 7.62727678e-01 -7.99112394e-03
1.94219835e-02 -1.01119030e+00 7.43281186e-01 5.63069761e-01
1.52013719e+00 -1.01852143e+00 3.11259449e-01 -1.80475451e-02
-1.71410251e+00 -3.60976458e-02 -7.46386766e-01 2.32633114e-01
-5.77308871e-02 6.26735985e-01 -4.05452214e-02 4.48185921e-01
1.24867928e+00 7.00461805e-01 -4.69779372e-01 7.07310617e-01
8.93416256e-02 9.55144465e-02 -7.98913836e-03 4.37938839e-01
3.36326770e-02 -4.82343584e-01 3.26097101e-01 1.22827172e+00
5.06472886e-01 2.98775285e-01 1.26749218e-01 8.99211228e-01
-4.83363457e-02 -1.21686226e-02 -5.11451483e-01 3.66309196e-01
5.70674062e-01 1.35341394e+00 -6.69689417e-01 -2.43089586e-01
-6.84104800e-01 1.15790284e+00 5.72865037e-03 4.71768677e-01
-9.23204005e-01 -9.21611965e-01 5.13513863e-01 2.50715613e-01
3.86557668e-01 8.33584461e-03 -2.73734033e-01 -1.03964007e+00
7.77326897e-02 -9.44801927e-01 1.96910664e-01 -1.19085085e+00
-1.14798486e+00 8.23592603e-01 -1.03216700e-01 -1.69171429e+00
2.95774281e-01 -3.05273920e-01 -4.76100653e-01 1.24860835e+00
-1.88549304e+00 -9.34725404e-01 -7.43790030e-01 9.42506611e-01
6.32592022e-01 6.04888983e-02 4.25196856e-01 4.86887544e-01
-2.19856918e-01 4.06169772e-01 2.11736202e-01 -3.38838808e-02
9.10215378e-01 -9.68771279e-01 4.26299945e-02 9.65676844e-01
-2.85257250e-01 4.24250066e-01 5.29500127e-01 -7.01042056e-01
-1.05321944e+00 -1.27516198e+00 3.45218629e-01 3.30192327e-01
3.69466901e-01 -3.08839560e-01 -1.16178143e+00 3.34836721e-01
3.72625977e-01 1.62188262e-01 4.87622291e-01 -5.69864631e-01
-4.36208546e-01 -5.82536876e-01 -1.24109840e+00 6.52229249e-01
7.28856087e-01 -7.05166996e-01 -4.45752650e-01 -7.00531900e-02
8.05552721e-01 -5.53596795e-01 -8.49545836e-01 3.19287717e-01
5.34918129e-01 -1.38774502e+00 1.11191297e+00 4.55446690e-01
4.51905131e-01 -9.68880355e-01 -2.70952672e-01 -1.16007340e+00
-8.09995055e-01 -3.41340184e-01 3.18259597e-01 1.08758771e+00
3.18859696e-01 -6.42008364e-01 2.37245753e-01 -8.15319046e-02
7.29418620e-02 -5.79471409e-01 -7.48852015e-01 -5.66506505e-01
-1.47503778e-01 1.21770985e-01 5.14649630e-01 8.63801301e-01
-1.66053861e-01 6.47999793e-02 -6.40332937e-01 3.35719287e-01
6.91908598e-01 6.27187073e-01 2.39662409e-01 -6.67632043e-01
-1.06020704e-01 -3.52090478e-01 -1.68939605e-01 -9.01444316e-01
-3.15964818e-01 -5.11411428e-01 -9.04115289e-02 -1.57074082e+00
3.56396437e-01 -8.60970020e-02 -2.78364420e-01 3.34955722e-01
-1.27155200e-01 9.46187139e-01 2.53873646e-01 2.81509638e-01
-2.72494972e-01 4.44364309e-01 1.73272371e+00 -1.20783798e-01
-4.30419296e-01 -5.21883011e-01 -7.80908644e-01 5.48048079e-01
1.08385324e+00 1.69006158e-02 -4.93157059e-01 -4.45513189e-01
-2.51312494e-01 9.71223116e-02 1.28905550e-01 -1.12920237e+00
2.47229666e-01 -2.15798080e-01 9.88961339e-01 -6.61302567e-01
3.62791419e-01 -8.26245129e-01 4.80126083e-01 5.63438892e-01
-4.61799324e-01 -2.97950841e-02 2.09471777e-01 3.20711046e-01
-4.02476668e-01 3.48579437e-01 1.52953780e+00 -9.36162621e-02
-9.72977996e-01 2.65862435e-01 -2.45005116e-01 -1.23713866e-01
1.03437090e+00 -4.36147273e-01 -5.92424750e-01 -4.46996689e-01
-4.42219585e-01 -2.49598905e-01 6.56368494e-01 5.65201581e-01
8.81875575e-01 -1.55843961e+00 -7.34470963e-01 4.99000311e-01
-1.14202872e-01 -3.25070590e-01 8.92010927e-01 3.67041081e-01
-7.39001095e-01 2.46712208e-01 -8.28320622e-01 -4.15242732e-01
-1.24619758e+00 6.50480688e-01 5.44625521e-01 -3.09076309e-02
-1.04841232e+00 4.83521342e-01 6.75138354e-01 6.77984133e-02
1.81292016e-02 -2.26340577e-01 -2.48699218e-01 -2.45173782e-01
9.70559299e-01 2.64840275e-01 -2.45109379e-01 -7.22988069e-01
-8.22372213e-02 1.44240046e+00 -1.22988135e-01 -9.30629745e-02
1.33952725e+00 -7.10333765e-01 -2.82305777e-01 4.12113667e-01
1.37659419e+00 3.19275379e-01 -1.37693751e+00 -4.54099149e-01
-7.80318856e-01 -9.93526876e-01 3.93651217e-01 -8.41989636e-01
-1.22204006e+00 8.25026631e-01 1.10627043e+00 3.11844975e-01
1.90527570e+00 -1.11855865e-01 6.32544637e-01 -3.13975252e-02
2.45459467e-01 -9.71193194e-01 3.53227586e-01 1.71206415e-01
1.07723510e+00 -8.80840719e-01 5.33310138e-02 -5.45394838e-01
-5.89229405e-01 1.13488007e+00 8.01534414e-01 -4.14725214e-01
5.48634470e-01 4.31293607e-01 2.66557455e-01 9.04954076e-02
-5.35887718e-01 -2.25568160e-01 3.71181995e-01 7.76918828e-01
2.91623116e-01 -1.54349282e-01 -6.61557242e-02 3.57483387e-01
-3.18381846e-01 -3.48184966e-02 6.56101406e-01 4.48404104e-01
-7.64155865e-01 -6.12367213e-01 -6.10886991e-01 1.92685440e-01
-5.12410879e-01 -1.61186874e-01 6.06879853e-02 6.75568461e-01
2.08885506e-01 8.87782454e-01 2.14996785e-01 -4.05227482e-01
5.33433378e-01 -4.28872466e-01 4.00874615e-01 -2.97094852e-01
-4.10680845e-02 5.83989918e-01 -3.62503052e-01 -6.55013561e-01
-1.87225088e-01 -6.09761663e-03 -1.17100191e+00 -3.54004055e-01
-4.69725244e-02 -1.43217921e-01 1.95139170e-01 5.57561517e-02
1.70165226e-01 8.55192363e-01 8.02290738e-01 -8.94944012e-01
-2.05330979e-02 -7.41363168e-01 -9.23190832e-01 3.25966477e-01
7.45371938e-01 -3.87411565e-01 -2.87140697e-01 3.22257757e-01] | [10.963947296142578, -2.210054874420166] |
a71cf2d5-8215-4991-809a-46f95f09ae81 | temporal-label-smoothing-for-early-prediction | 2208.13764 | null | https://arxiv.org/abs/2208.13764v2 | https://arxiv.org/pdf/2208.13764v2.pdf | Temporal Label Smoothing for Early Event Prediction | Models that can predict the occurrence of events ahead of time with low false-alarm rates are critical to the acceptance of decision support systems in the medical community. This challenging task is typically treated as a simple binary classification, ignoring temporal dependencies between samples, whereas we propose to exploit this structure. We first introduce a common theoretical framework unifying dynamic survival analysis and early event prediction. Following an analysis of objectives from both fields, we propose Temporal Label Smoothing (TLS), a simpler, yet best-performing method that preserves prediction monotonicity over time. By focusing the objective on areas with a stronger predictive signal, TLS improves performance over all baselines on two large-scale benchmark tasks. Gains are particularly notable along clinically relevant measures, such as event recall at low false-alarm rates. TLS reduces the number of missed events by up to a factor of two over previously used approaches in early event prediction. | ['Rita Kuznetsova', 'Gunnar Rätsch', 'Alizée Pace', 'Hugo Yèche'] | 2022-08-29 | null | null | null | null | ['respiratory-failure', 'circulatory-failure', 'survival-analysis'] | ['medical', 'medical', 'miscellaneous'] | [ 5.38224220e-01 1.56441107e-01 -6.50247276e-01 -5.68860054e-01
-9.79221523e-01 -2.59658933e-01 8.30448568e-01 1.03609359e+00
-6.32697105e-01 8.01402450e-01 4.50412363e-01 -4.66232508e-01
-5.50118864e-01 -4.61014956e-01 -2.45227516e-01 -5.93470514e-01
-5.74310899e-01 4.08956766e-01 5.34228802e-01 1.53809831e-01
1.69964686e-01 3.44980538e-01 -1.16944098e+00 6.39728487e-01
5.72817206e-01 9.40521240e-01 -4.35033083e-01 6.25862241e-01
3.94633144e-01 1.03156304e+00 -1.92541167e-01 -4.57437664e-01
-1.19706824e-01 -3.62296343e-01 -9.16002274e-01 -4.39799607e-01
-9.79384631e-02 -6.89301491e-02 -1.93750635e-01 4.40581352e-01
4.49095547e-01 5.36221676e-02 8.15704167e-01 -1.16537845e+00
1.09543391e-01 5.76706886e-01 -3.56213838e-01 5.43816090e-01
1.81977242e-01 -1.22109815e-01 1.03306139e+00 -6.03447497e-01
6.53269470e-01 1.04781151e+00 1.17257118e+00 4.00825053e-01
-1.66755879e+00 -4.70945776e-01 1.90598279e-01 -1.93198174e-02
-1.16442251e+00 -5.84736705e-01 1.23491161e-01 -6.29565954e-01
9.85460937e-01 6.03358686e-01 2.26846948e-01 9.99741435e-01
8.08152497e-01 6.42834544e-01 9.47761416e-01 -3.13939244e-01
3.32635880e-01 -1.05817869e-01 7.00460494e-01 3.42984796e-01
2.26464570e-01 5.79057038e-01 -5.61412573e-01 -6.93126261e-01
3.17364752e-01 4.26471889e-01 -4.23126370e-02 -8.61487761e-02
-1.30394590e+00 8.14820647e-01 5.56912050e-02 1.77415609e-01
-4.21539634e-01 -7.92863294e-02 7.39167273e-01 3.09494913e-01
8.68719876e-01 3.21209848e-01 -6.56859875e-01 -1.41038105e-01
-1.14772141e+00 1.77949518e-01 6.74806297e-01 4.24507529e-01
-3.91092226e-02 -4.27569360e-01 -7.70523131e-01 5.75910568e-01
-4.26050052e-02 1.64722675e-03 3.37056726e-01 -6.16795719e-01
-4.08610776e-02 5.03768146e-01 2.75488943e-01 -6.57093585e-01
-9.75678265e-01 -9.03740525e-01 -7.88147390e-01 5.17314374e-02
6.06641948e-01 -1.00158714e-01 -8.50446701e-01 1.78582561e+00
3.22666407e-01 3.24468523e-01 -9.15752873e-02 3.23535532e-01
2.19043687e-01 3.80460471e-01 7.28940308e-01 -7.85844922e-01
1.41754830e+00 -4.91775960e-01 -5.47967553e-01 -2.54562676e-01
1.08284521e+00 -4.86527860e-01 4.99931514e-01 4.48978692e-01
-8.77473652e-01 -1.32142492e-02 -6.62195444e-01 4.63901274e-02
-7.10354447e-02 -1.17373459e-01 8.00899625e-01 4.58268076e-01
-8.40240836e-01 1.05778158e+00 -1.32238615e+00 -3.82582188e-01
5.31412780e-01 3.29685926e-01 -2.68457144e-01 -3.23622348e-03
-1.02205050e+00 9.38586533e-01 4.11035091e-01 -3.23241174e-01
-7.19792187e-01 -1.21521902e+00 -5.20019054e-01 1.39471933e-01
3.09957027e-01 -8.24860811e-01 1.33774316e+00 -7.82981396e-01
-8.09291303e-01 7.59699643e-01 -4.66811329e-01 -1.07225597e+00
7.26467133e-01 -2.96779364e-01 -5.60400903e-01 -1.88812509e-01
1.45414531e-01 2.93061197e-01 4.38328922e-01 -7.02641785e-01
-1.08247995e+00 -3.76694560e-01 -4.97060806e-01 -1.33678287e-01
-3.32643986e-01 2.64521480e-01 4.10222858e-02 -8.80471647e-01
-1.12551853e-01 -7.48480320e-01 -7.20541358e-01 -5.95054217e-02
-2.14567959e-01 -3.70546639e-01 3.99324596e-01 -6.94987297e-01
1.68884480e+00 -2.13059163e+00 -4.04838949e-01 9.53740403e-02
4.63267595e-01 5.40718660e-02 2.93953031e-01 3.08078378e-01
-5.13002396e-01 -3.76579612e-02 -2.39866868e-01 -3.91546071e-01
-4.18377161e-01 2.13476196e-02 -4.06855613e-01 5.01540780e-01
3.20320427e-01 9.00279045e-01 -1.09008420e+00 -5.02606809e-01
2.58170068e-02 3.08650732e-01 -5.13550818e-01 -8.07727352e-02
9.85003188e-02 4.85931844e-01 -1.87981904e-01 4.23288226e-01
-1.98604967e-02 -7.07012653e-01 2.95874268e-01 1.93510011e-01
-4.88171689e-02 4.46163833e-01 -8.56110096e-01 1.26051486e+00
-1.93007678e-01 3.29171568e-01 -5.99002123e-01 -1.02526367e+00
5.16343474e-01 5.41291356e-01 9.25721407e-01 -4.48932260e-01
-6.62012696e-02 2.46147409e-01 -6.09395392e-02 -1.43348962e-01
1.85751706e-01 -6.49024904e-01 -1.46974862e-01 2.41908386e-01
-2.74454892e-01 8.70912373e-01 1.08590081e-01 2.79325604e-01
1.53655827e+00 -3.23431373e-01 9.41442668e-01 -3.63125920e-01
6.82888553e-02 1.66380435e-01 9.08231318e-01 1.07416117e+00
-2.87618995e-01 2.64671922e-01 6.44184530e-01 -6.01424158e-01
-7.47671723e-01 -1.14030194e+00 -7.38825738e-01 1.21560073e+00
-3.08323801e-01 -5.15166044e-01 -9.11397710e-02 -8.96496236e-01
1.05827205e-01 8.62671733e-01 -8.72830272e-01 -3.38445932e-01
-4.25867856e-01 -1.50528824e+00 4.83220428e-01 6.96993589e-01
-2.82284230e-01 -7.15162039e-01 -7.22367644e-01 5.48659444e-01
3.86687070e-02 -8.05387676e-01 -1.68672979e-01 4.65258896e-01
-1.24716067e+00 -1.11562765e+00 -7.87619531e-01 -4.98184949e-01
3.62402231e-01 -2.90058225e-01 1.38591576e+00 -2.94388682e-02
-4.08367872e-01 2.01960251e-01 -3.00040483e-01 -6.00423396e-01
-4.91010189e-01 6.34399289e-03 3.66056450e-02 2.15331316e-02
3.69238853e-01 -3.34875852e-01 -7.93879032e-01 6.88066408e-02
-6.63607001e-01 -9.78471804e-03 5.23293674e-01 1.00291598e+00
6.08780801e-01 -1.10449120e-01 1.04953492e+00 -1.28392267e+00
3.07509094e-01 -8.05949867e-01 -2.42469475e-01 1.97466806e-01
-1.15429950e+00 9.18206573e-02 4.50888544e-01 -4.24658269e-01
-9.58512247e-01 9.21725258e-02 1.05395362e-01 2.30516940e-01
-2.14736477e-01 4.76644367e-01 5.61373591e-01 1.46005124e-01
7.94938564e-01 6.66691586e-02 -3.28346975e-02 -4.47370052e-01
-1.44699663e-01 3.37286592e-01 4.09680068e-01 -1.44719973e-01
1.68493912e-01 6.56189740e-01 4.45680052e-01 -4.80397493e-01
-1.02750754e+00 -6.64584816e-01 -3.93626601e-01 6.06461093e-02
5.79375625e-01 -7.61177421e-01 -6.96657419e-01 1.33998469e-01
-7.38302529e-01 -3.09652567e-01 -4.80233818e-01 5.27303696e-01
-3.34555864e-01 2.19700590e-01 -6.66641235e-01 -8.62033546e-01
-3.73475373e-01 -5.44786990e-01 1.03457177e+00 -1.27201408e-01
-6.87349379e-01 -1.28668332e+00 3.15178096e-01 -1.86852291e-01
3.45349640e-01 5.59274495e-01 1.06526911e+00 -1.13182700e+00
2.97946781e-02 -4.74380672e-01 -1.41203746e-01 -6.67318329e-02
4.75549810e-02 -1.71644643e-01 -8.15353572e-01 -4.36987132e-01
-2.06010312e-01 1.35493010e-01 1.29042113e+00 7.96894372e-01
1.20563376e+00 -2.12347388e-01 -1.05674148e+00 2.87556589e-01
1.16772401e+00 4.02072847e-01 3.76488060e-01 1.39916703e-01
1.23176806e-01 6.26098573e-01 7.21203089e-01 6.06272817e-01
3.14447761e-01 6.21225536e-01 -4.62603271e-02 -3.34313899e-01
3.56528051e-02 -6.17935508e-03 1.47458598e-01 1.94366887e-01
-3.05102635e-02 -5.27998917e-02 -1.09963715e+00 8.13440144e-01
-1.98209834e+00 -9.03979599e-01 -2.95439363e-01 2.69243574e+00
1.02321482e+00 6.63399935e-01 4.51774359e-01 3.32350463e-01
4.99189436e-01 -2.17885405e-01 -4.03819531e-01 -1.33168712e-01
-4.64070821e-03 2.25248039e-01 6.38674736e-01 4.31264728e-01
-1.35570633e+00 4.66592997e-01 7.30868292e+00 7.64054954e-01
-1.04876888e+00 2.00097188e-01 1.26821220e+00 -3.98185223e-01
1.42053738e-01 1.43310532e-01 -7.84734666e-01 4.89698052e-01
1.51595056e+00 -2.95918196e-01 -2.66620964e-01 7.05602765e-01
5.18498361e-01 -1.60844147e-01 -1.49405658e+00 4.98495370e-01
-3.30042511e-01 -1.47703207e+00 -2.38183320e-01 -4.22238223e-02
6.63090169e-01 -8.30464587e-02 -6.33209422e-02 2.81609595e-01
4.26328301e-01 -9.27885115e-01 5.29189229e-01 8.49563003e-01
8.33589315e-01 -6.72876358e-01 8.36706340e-01 2.95760721e-01
-8.92948747e-01 -3.30126494e-01 2.86594659e-01 -1.57412499e-01
4.59096938e-01 1.12001657e+00 -1.02423406e+00 4.94009882e-01
4.91181672e-01 9.69928205e-01 -6.79651618e-01 1.44451666e+00
2.69670576e-01 1.11886632e+00 -3.55742186e-01 3.28020304e-01
-8.01332891e-02 5.51880777e-01 5.68110883e-01 1.55684805e+00
1.44170612e-01 2.72254765e-01 3.63889962e-01 7.17080235e-02
2.86915332e-01 1.75645962e-01 -4.47918624e-01 1.87668353e-01
3.26508015e-01 8.32301259e-01 -9.72039282e-01 -5.97263992e-01
-2.47278064e-01 6.10762596e-01 3.30800191e-02 8.66955519e-02
-7.34229684e-01 -1.61275461e-01 2.82552958e-01 4.96853113e-01
-1.10967271e-01 1.67157382e-01 -8.24783504e-01 -7.91111112e-01
-3.20747942e-01 -5.28523445e-01 1.15342677e+00 -8.60721990e-02
-1.30366957e+00 3.29831094e-01 4.54259217e-02 -1.26298189e+00
-4.11925107e-01 -4.00277346e-01 -4.84238684e-01 5.24726808e-01
-1.30949271e+00 -7.72962332e-01 1.26316160e-01 3.16444367e-01
4.56911862e-01 2.85563976e-01 8.78828049e-01 3.52154016e-01
-5.39369464e-01 6.45317614e-01 2.01158240e-01 -1.51545614e-01
9.71840203e-01 -1.21407402e+00 3.23803961e-01 8.27113271e-01
-9.80844721e-02 4.64233637e-01 7.79276609e-01 -9.01817918e-01
-4.61764395e-01 -1.21447015e+00 1.39818263e+00 -6.12049580e-01
6.47701681e-01 -1.11276098e-02 -1.00520205e+00 7.72497475e-01
-4.74140942e-01 1.11802761e-02 8.77321661e-01 5.91528654e-01
-1.69973433e-01 -9.46655273e-02 -1.04653943e+00 4.07204270e-01
8.98156762e-01 -1.45246476e-01 -5.03296971e-01 5.39437354e-01
5.24585783e-01 -1.42553568e-01 -1.12704217e+00 8.08879912e-01
6.71536982e-01 -7.88680255e-01 1.05971515e+00 -1.17436409e+00
3.78499299e-01 1.58775300e-01 1.71255261e-01 -8.69882464e-01
-6.98176086e-01 -7.10111797e-01 -1.79667890e-01 8.79507601e-01
6.03256524e-01 -7.51879156e-01 7.63844669e-01 6.70063376e-01
1.37713542e-02 -8.77321661e-01 -1.09449589e+00 -8.21645260e-01
8.70971456e-02 -3.80351037e-01 -1.52877150e-02 1.01890171e+00
2.12535873e-01 3.55411887e-01 -4.14609760e-01 1.35598660e-01
6.56497538e-01 -1.53699471e-02 1.64733037e-01 -1.72788131e+00
-2.62300432e-01 -3.95360142e-01 -5.26296377e-01 -5.41903198e-01
-2.82868713e-01 -9.70267713e-01 -1.25023410e-01 -1.22470140e+00
5.17879188e-01 -4.60130215e-01 -7.59304881e-01 7.03645527e-01
-5.56521297e-01 2.12371409e-01 -3.05041134e-01 2.43697345e-01
-7.24328041e-01 1.25127599e-01 4.81667727e-01 2.63153225e-01
-3.57561111e-01 3.60592276e-01 -6.28071785e-01 8.56486440e-01
6.24726176e-01 -9.42753673e-01 -1.87802032e-01 1.41525075e-01
1.39350682e-01 4.73616719e-01 6.12286866e-01 -9.46867883e-01
1.07448742e-01 -3.51796716e-01 3.99476588e-01 -4.88203794e-01
-5.20770531e-03 -4.99819726e-01 2.89848328e-01 9.87666249e-01
-7.15855777e-01 -2.58603115e-02 2.33469352e-01 1.00735188e+00
1.59128904e-02 1.72955096e-01 9.68119979e-01 2.86683887e-01
-4.98672903e-01 1.58635095e-01 -3.45566958e-01 7.13175759e-02
1.24672472e+00 1.15431435e-02 -2.93763310e-01 -1.35715261e-01
-1.16519320e+00 2.18228683e-01 3.99240069e-02 2.32196480e-01
1.86086580e-01 -1.00718427e+00 -9.12446082e-01 -5.69756217e-02
2.27562740e-01 -4.66857791e-01 2.46900007e-01 1.36539686e+00
-7.89718404e-02 6.46545231e-01 2.55728394e-01 -6.55998409e-01
-1.54519272e+00 6.36863351e-01 2.51215696e-01 -7.99542964e-01
-6.92397654e-01 7.45355844e-01 3.79496843e-01 2.16494173e-01
2.74087489e-01 -3.24681252e-01 -1.14791237e-01 1.13880739e-01
7.24604309e-01 6.52187467e-01 2.13334799e-01 -2.34209448e-01
-5.94618499e-01 -1.05898730e-01 -4.60056573e-01 1.18782401e-01
1.44083798e+00 1.23801650e-02 8.77600089e-02 7.20419943e-01
9.14651513e-01 -2.77911037e-01 -1.15131235e+00 -3.63642931e-01
6.18833065e-01 -2.01810926e-01 1.37610853e-01 -1.13701367e+00
-3.80631506e-01 5.32810330e-01 6.84917390e-01 2.61921018e-01
1.13355970e+00 1.06530592e-01 6.16517782e-01 -1.73557959e-02
1.74363688e-01 -7.46374309e-01 -1.66635424e-01 2.41708323e-01
5.15047193e-01 -1.18984640e+00 1.04002573e-01 -4.52314734e-01
-5.19923210e-01 9.09292281e-01 -8.30433816e-02 -2.10279264e-02
7.92052567e-01 2.35677630e-01 -3.85797292e-01 -1.22909307e-01
-1.28850520e+00 1.07545726e-01 4.71083969e-01 2.23183677e-01
5.82000792e-01 1.99590638e-01 -6.63534403e-01 8.79839122e-01
2.17543244e-01 3.84623259e-01 1.79963574e-01 8.49647343e-01
-4.83614147e-01 -9.67796922e-01 -1.61709934e-01 1.03469503e+00
-9.86536503e-01 -2.63429493e-01 -2.45017335e-02 5.44013262e-01
-1.89303532e-01 9.23701882e-01 2.13213250e-01 -1.25289604e-01
2.04247892e-01 3.98046196e-01 -3.53927873e-02 -5.97419560e-01
-5.88907361e-01 1.09877206e-01 3.35396707e-01 -6.82630599e-01
-2.49038905e-01 -1.05075836e+00 -1.15469158e+00 -1.31961793e-01
-2.18123868e-01 2.90568713e-02 9.18223485e-02 1.06204128e+00
5.63068151e-01 7.26899683e-01 4.62187707e-01 -2.27322400e-01
-6.86409950e-01 -6.40519202e-01 -4.67089325e-01 4.25411910e-01
5.20465970e-01 -6.60674274e-01 -3.62124503e-01 2.10372090e-01] | [7.844993591308594, 5.716399669647217] |
e7af91e7-75e8-4df0-a0eb-21178364dc5f | learning-transferable-deep-models-for-land | 1807.05713 | null | https://arxiv.org/abs/1807.05713v3 | https://arxiv.org/pdf/1807.05713v3.pdf | Land-Cover Classification with High-Resolution Remote Sensing Images Using Transferable Deep Models | In recent years, large amount of high spatial-resolution remote sensing (HRRS) images are available for land-cover mapping. However, due to the complex information brought by the increased spatial resolution and the data disturbances caused by different conditions of image acquisition, it is often difficult to find an efficient method for achieving accurate land-cover classification with high-resolution and heterogeneous remote sensing images. In this paper, we propose a scheme to apply deep model obtained from labeled land-cover dataset to classify unlabeled HRRS images. The main idea is to rely on deep neural networks for presenting the contextual information contained in different types of land-covers and propose a pseudo-labeling and sample selection scheme for improving the transferability of deep models. More precisely, a deep Convolutional Neural Networks is first pre-trained with a well-annotated land-cover dataset, referred to as the source data. Then, given a target image with no labels, the pre-trained CNN model is utilized to classify the image in a patch-wise manner. The patches with high confidence are assigned with pseudo-labels and employed as the queries to retrieve related samples from the source data. The pseudo-labels confirmed with the retrieved results are regarded as supervised information for fine-tuning the pre-trained deep model. To obtain a pixel-wise land-cover classification with the target image, we rely on the fine-tuned CNN and develop a hybrid classification by combining patch-wise classification and hierarchical segmentation. In addition, we create a large-scale land-cover dataset containing 150 Gaofen-2 satellite images for CNN pre-training. Experiments on multi-source HRRS images show encouraging results and demonstrate the applicability of the proposed scheme to land-cover classification. | ['Huanfeng Shen', 'Qikai Lu', 'Gui-Song Xia', 'Xin-Yi Tong', 'Shucheng You', 'Shengyang Li', 'Liangpei Zhang'] | 2018-07-16 | null | null | null | null | ['segmentation-of-remote-sensing-imagery', 'remote-sensing-image-classification'] | ['miscellaneous', 'miscellaneous'] | [ 4.81969804e-01 -3.08864057e-01 -2.65930623e-01 -6.40756428e-01
-9.05512989e-01 -3.64378959e-01 1.29806533e-01 4.49278653e-02
-4.21669543e-01 9.83069003e-01 -2.71687061e-01 -4.36253458e-01
-1.96681485e-01 -1.58584881e+00 -7.01456666e-01 -8.37798536e-01
-4.45937738e-02 3.19538534e-01 1.00657381e-01 -3.93742293e-01
-1.70772582e-01 7.88131773e-01 -1.89667666e+00 1.63194492e-01
1.24775684e+00 1.03139007e+00 7.72224307e-01 1.66322827e-01
-2.87649930e-01 3.51477176e-01 -3.61276537e-01 4.20607060e-01
3.72206837e-01 -4.36816365e-01 -8.23544741e-01 5.67948341e-01
2.05485180e-01 -5.69959342e-01 -1.96342939e-03 1.29525948e+00
2.03824878e-01 1.65400013e-01 4.70096916e-01 -5.68406999e-01
-3.38422865e-01 4.86277491e-01 -7.35039055e-01 9.86966044e-02
-3.91315520e-01 -3.30296420e-02 8.99842203e-01 -5.81031382e-01
3.97933871e-01 1.13245475e+00 4.72815007e-01 2.37514600e-01
-1.07123137e+00 -7.49981582e-01 4.30084497e-01 4.21185046e-03
-1.79304385e+00 -2.01791838e-01 6.57467306e-01 -3.07327569e-01
3.75901669e-01 4.10007715e-01 8.24333072e-01 1.31629050e-01
-1.77582175e-01 7.01074958e-01 1.22449076e+00 -9.28334668e-02
4.65213358e-02 3.40387374e-02 1.92689702e-01 3.66345644e-01
3.92014951e-01 1.61154047e-01 4.48042005e-01 1.07617304e-01
8.53559136e-01 5.56591392e-01 -2.58690655e-01 -4.71436940e-02
-8.59287679e-01 6.44228876e-01 1.25381374e+00 6.38279796e-01
-8.09202850e-01 -9.01257917e-02 8.62109363e-02 2.21958697e-01
8.13480377e-01 1.96041420e-01 -5.20456731e-01 7.68857718e-01
-1.25674450e+00 1.03468038e-01 1.74529254e-01 5.88464618e-01
1.49627197e+00 -7.88145289e-02 1.10941492e-02 1.03299141e+00
4.19162631e-01 7.06789374e-01 1.47727296e-01 -3.51810068e-01
5.05912542e-01 8.70817125e-01 2.50017256e-01 -1.06710041e+00
-2.53567964e-01 -9.34175372e-01 -1.13267648e+00 2.42851302e-01
-1.02884822e-01 8.78364444e-02 -1.37976956e+00 1.27170575e+00
2.93871313e-01 -8.04184899e-02 3.18358183e-01 1.11426437e+00
7.43037224e-01 1.11132014e+00 3.60116631e-01 -1.01323873e-01
1.15122831e+00 -5.28446853e-01 -3.76831442e-01 -3.62269670e-01
4.77607459e-01 -1.56201109e-01 8.26205730e-01 -1.40694723e-01
-1.97206974e-01 -8.39385211e-01 -1.17034543e+00 4.05860245e-01
-5.93247354e-01 5.63734531e-01 3.76008660e-01 2.00174451e-01
-7.74944663e-01 5.04861712e-01 -6.19360626e-01 -3.49823475e-01
9.06306624e-01 1.46763757e-01 -2.58701533e-01 -2.82367319e-01
-1.31443977e+00 4.59421158e-01 9.47592139e-01 7.35565186e-01
-1.13610005e+00 -2.53076136e-01 -7.80047655e-01 1.61787838e-01
2.00735673e-01 1.71577316e-02 5.16808867e-01 -1.49724340e+00
-8.85951042e-01 1.09653139e+00 1.28410190e-01 -1.67581916e-01
2.55726755e-01 1.66394055e-01 -4.81392115e-01 8.92473459e-02
6.54133677e-01 6.25257373e-01 6.39136493e-01 -1.60405493e+00
-1.12036932e+00 -4.15609956e-01 1.77978307e-01 4.82858568e-01
-9.00983885e-02 -2.12650761e-01 -1.65434420e-01 -4.05458480e-01
6.14752710e-01 -8.18561375e-01 -4.43140507e-01 -2.66197115e-01
-1.57066628e-01 7.85790682e-02 6.65954173e-01 -6.99096620e-01
1.07450998e+00 -2.31448221e+00 -2.89956629e-01 4.24771488e-01
-1.32996356e-02 5.01501441e-01 -4.02764112e-01 -2.21699513e-02
-2.06777230e-01 3.47888410e-01 -8.61433864e-01 3.89427930e-01
-4.00283843e-01 3.57262731e-01 -3.81340206e-01 3.86713713e-01
5.00738144e-01 7.63758957e-01 -8.51952851e-01 -5.37683010e-01
3.93197596e-01 2.73586839e-01 1.44470543e-01 3.97021383e-01
-4.85560238e-01 7.03917265e-01 -9.36860919e-01 8.20789516e-01
1.12139308e+00 -9.61342528e-02 -3.45701799e-02 -2.26795137e-01
-3.71124804e-01 -4.31165136e-02 -1.13055587e+00 1.03446198e+00
-4.57566768e-01 3.53265345e-01 7.12006167e-02 -1.25752223e+00
1.22827995e+00 3.67269218e-02 2.55478442e-01 -6.89760447e-01
-5.74005879e-02 5.07835805e-01 -2.28742257e-01 -6.61785781e-01
4.46301848e-01 -2.29037702e-01 6.31747097e-02 1.53931424e-01
-4.21413660e-01 -4.52652015e-02 2.68313214e-02 -4.86289352e-01
3.71564925e-01 2.69879818e-01 4.51888330e-02 -2.91914135e-01
7.64481306e-01 6.44882441e-01 3.99988949e-01 5.82167983e-01
-1.43133759e-01 6.38020396e-01 -8.88071209e-02 -1.00967073e+00
-9.82369125e-01 -4.86909270e-01 -6.34410441e-01 1.13400722e+00
3.59480500e-01 5.45712411e-01 -4.39217091e-01 -6.43110096e-01
-1.24343019e-02 3.37106079e-01 -6.25274122e-01 2.19840094e-01
-3.37335497e-01 -1.33087504e+00 4.48331714e-01 3.53652388e-01
1.39375389e+00 -1.41478157e+00 -4.11502451e-01 4.09590691e-01
-2.87100852e-01 -6.27382576e-01 4.10860717e-01 8.39071348e-02
-1.03933263e+00 -9.06121075e-01 -8.46210480e-01 -7.90435731e-01
8.68429482e-01 6.25450969e-01 8.72746646e-01 3.53805840e-01
-3.58980112e-02 -4.93758321e-01 -5.16195893e-01 1.96786784e-02
-4.13820118e-01 4.40293819e-01 -5.90498269e-01 4.19505656e-01
3.45112741e-01 -3.50488573e-01 -5.25000155e-01 4.18398321e-01
-1.23266923e+00 1.71111345e-01 1.09278584e+00 7.09216237e-01
1.03314507e+00 6.06492937e-01 7.75200725e-01 -7.71814704e-01
-5.50685339e-02 -6.87862575e-01 -7.93897092e-01 3.10080409e-01
-3.96461964e-01 -1.45598993e-01 3.83497775e-01 -3.85232158e-02
-1.04944074e+00 3.54142964e-01 -3.41194183e-01 -5.32154664e-02
-7.51420736e-01 1.11989558e+00 -3.85393888e-01 7.68760517e-02
7.87202895e-01 5.98816693e-01 -2.57625669e-01 -4.27856714e-01
9.25208926e-02 1.28094757e+00 2.84079015e-01 -1.22883052e-01
1.00970948e+00 5.39141178e-01 -2.91390598e-01 -8.35185766e-01
-1.10918391e+00 -5.31840086e-01 -7.41227329e-01 -5.16604818e-02
9.36262369e-01 -1.30352676e+00 1.48626193e-01 6.06551051e-01
-7.24293530e-01 -4.10364151e-01 -3.83434370e-02 4.26970243e-01
-1.07791699e-01 -5.59542589e-02 -3.47937159e-02 -8.69540155e-01
-4.42273527e-01 -9.42657411e-01 1.06954408e+00 4.14490819e-01
5.64585507e-01 -5.51113725e-01 -1.58547178e-01 1.08970806e-01
5.71387470e-01 3.05421978e-01 7.41190016e-01 -3.62083733e-01
-8.67333055e-01 -3.96409303e-01 -7.98447967e-01 5.55501699e-01
6.02508307e-01 -1.99994743e-01 -9.45380569e-01 -2.32311860e-01
-2.51159221e-01 -2.88255483e-01 1.27916002e+00 6.02431715e-01
1.20197332e+00 -1.90402701e-01 -4.41211164e-01 7.52430618e-01
1.93159902e+00 1.86982632e-01 6.93258584e-01 6.37078404e-01
8.74450743e-01 7.12551713e-01 1.09357846e+00 9.09571722e-02
1.22925952e-01 1.85592651e-01 7.17471123e-01 -6.04052663e-01
3.32290858e-01 -1.33308157e-01 -1.87865645e-01 1.05231732e-01
-3.13624829e-01 4.28462774e-02 -8.96156430e-01 7.18149424e-01
-1.80199754e+00 -1.02276623e+00 -4.96669710e-02 2.02994466e+00
9.09277618e-01 -1.30243510e-01 -1.46713927e-01 1.36943325e-01
1.09778559e+00 4.73642677e-01 -7.27002084e-01 5.27319372e-01
-3.61907989e-01 1.02416068e-01 8.84853125e-01 4.25088525e-01
-1.65355849e+00 1.29480541e+00 4.86287975e+00 1.05336535e+00
-1.42682385e+00 -3.92993130e-02 8.05439770e-01 6.05890989e-01
-3.09688777e-01 -1.37356758e-01 -8.44717741e-01 1.54959857e-01
5.50035834e-01 1.49521530e-01 3.00280213e-01 8.41776073e-01
3.63297135e-01 -1.07287653e-01 -3.01533818e-01 6.33949757e-01
-3.84235710e-01 -1.20838678e+00 2.35826567e-01 4.73147817e-02
9.28679347e-01 4.07745898e-01 -2.77881145e-01 2.45986238e-01
2.93897539e-01 -9.15781260e-01 4.33782548e-01 4.51084167e-01
1.15277255e+00 -6.71243191e-01 1.07844651e+00 4.12201375e-01
-1.55521691e+00 -1.39199376e-01 -8.60505104e-01 1.33099407e-01
-4.50661689e-01 4.53617513e-01 -4.11929399e-01 9.37903404e-01
7.87297428e-01 1.14995348e+00 -4.77963567e-01 1.01286769e+00
-3.18450898e-01 7.67554283e-01 -2.93379754e-01 2.60693073e-01
4.28145736e-01 -3.55333060e-01 7.78963193e-02 9.13178444e-01
3.30029458e-01 3.98832291e-01 4.41659927e-01 1.00665772e+00
-9.17080324e-03 2.17799276e-01 -5.18459141e-01 -1.97045133e-03
4.14943457e-01 1.49155462e+00 -8.65041494e-01 -4.81009334e-01
-8.70126858e-02 5.72136879e-01 6.13366812e-02 6.23593390e-01
-4.59099352e-01 -4.22956526e-01 3.19915235e-01 1.26938686e-01
2.86969572e-01 -7.82505870e-02 -1.02129161e-01 -1.06179070e+00
-1.10268809e-01 -6.43368065e-01 2.34347522e-01 -7.85635173e-01
-1.10155725e+00 9.58346844e-01 4.84888367e-02 -1.64668119e+00
1.30810291e-01 -3.38422835e-01 -3.74338239e-01 1.39074898e+00
-2.25545478e+00 -1.36368155e+00 -1.00829518e+00 4.33092326e-01
3.74648958e-01 -5.60882874e-02 8.95220995e-01 2.23027393e-01
-5.24355233e-01 -1.69762522e-01 2.31586874e-01 4.46618199e-01
1.44873857e-01 -9.45936084e-01 1.51574507e-01 9.75526750e-01
-3.26437622e-01 1.01724967e-01 2.10001349e-01 -6.86993897e-01
-7.84149528e-01 -1.98201513e+00 4.67387825e-01 4.28454757e-01
3.41739118e-01 2.26642281e-01 -1.17714584e+00 5.98613083e-01
-4.72574919e-01 2.74324059e-01 3.42755675e-01 -2.68368185e-01
3.84562127e-02 -8.48053336e-01 -1.21940172e+00 4.04341109e-02
7.19201863e-01 -5.30463636e-01 -5.14647126e-01 4.05213803e-01
5.30392230e-01 -1.29592657e-01 -5.63576818e-01 7.32323468e-01
2.93644190e-01 -5.46411037e-01 6.93724990e-01 -1.94447681e-01
6.64761484e-01 -7.13470459e-01 -4.12846297e-01 -1.30878925e+00
-4.97586012e-01 3.76829594e-01 1.06580460e+00 1.05599427e+00
3.55478108e-01 -7.18262672e-01 5.96479297e-01 1.45774394e-01
-1.61375508e-01 -3.21253031e-01 -6.07225180e-01 -5.85407317e-01
2.26669945e-02 -1.91284686e-01 1.06830907e+00 1.01238394e+00
-8.75348806e-01 -6.07460774e-02 6.76506162e-02 7.36471474e-01
4.97334838e-01 6.55114472e-01 6.64200068e-01 -1.44887376e+00
2.92111307e-01 -2.61013240e-01 -2.07852662e-01 -9.81513143e-01
1.94556624e-01 -9.73653376e-01 3.07045609e-01 -1.82162249e+00
1.47165820e-01 -1.17540717e+00 -5.63686252e-01 9.39042985e-01
-4.42167252e-01 5.34094214e-01 -2.66628563e-01 8.37685287e-01
-2.44718432e-01 7.42795885e-01 1.15545130e+00 -5.00631273e-01
-5.30904412e-01 3.48699301e-01 -5.13241887e-01 2.83920735e-01
8.79530072e-01 -5.02560735e-01 -4.34512720e-02 -5.03671825e-01
7.97532201e-02 7.79252313e-03 5.48360288e-01 -1.10854828e+00
-2.94145167e-01 -3.66065681e-01 5.19221485e-01 -1.10468149e+00
-2.93539226e-01 -1.08639753e+00 4.11197424e-01 4.67937618e-01
-3.39829087e-01 -7.46302426e-01 2.98594952e-01 3.02395493e-01
-5.05467653e-01 -1.92338705e-01 8.84731948e-01 -3.47984016e-01
-1.06429350e+00 7.49825180e-01 -2.65446186e-01 -5.14062762e-01
7.93315709e-01 -1.79634005e-01 -9.78070199e-02 -2.51081325e-02
-7.37837493e-01 3.75773311e-01 4.01600182e-01 2.69425631e-01
4.32760119e-01 -1.29803431e+00 -1.08097720e+00 5.01356781e-01
5.55149913e-01 4.90805000e-01 4.30189639e-01 2.37465441e-01
-8.96204710e-01 1.47340730e-01 -4.28098738e-01 -8.29259992e-01
-8.10574949e-01 2.85182446e-01 8.99403036e-01 -1.21164463e-01
-3.59021872e-01 3.29810590e-01 1.71684071e-01 -7.10853815e-01
-3.67511928e-01 -2.94024199e-01 -6.87605202e-01 2.17370495e-01
4.29469913e-01 -2.54813343e-01 4.02333848e-02 -9.73162889e-01
-2.68082768e-01 7.79942155e-01 3.51662606e-01 1.04399800e-01
1.64487565e+00 -2.84012377e-01 -4.93706673e-01 3.41751307e-01
9.62902248e-01 -4.99493808e-01 -1.32283795e+00 -6.32235646e-01
-1.64222866e-01 -5.51297605e-01 3.92646134e-01 -8.15285742e-01
-1.53492415e+00 7.52474546e-01 9.73366976e-01 3.54267836e-01
1.34418595e+00 -4.46327329e-02 3.55384797e-01 4.94583130e-01
4.20114160e-01 -8.02596986e-01 -7.46647298e-01 3.06822479e-01
8.12004507e-01 -1.62301087e+00 -4.04518433e-02 -2.08244815e-01
-2.41138473e-01 1.05981338e+00 4.73295450e-01 -2.63794422e-01
7.41035938e-01 -3.17568600e-01 3.22149336e-01 -9.98555869e-02
6.69750124e-02 -8.83649886e-01 2.22111493e-01 4.44018811e-01
6.91620186e-02 4.88697916e-01 -8.92134234e-02 3.25599849e-01
1.23184361e-01 3.59831095e-01 1.03614815e-01 6.86596096e-01
-9.78698790e-01 -7.46674478e-01 -5.58812797e-01 5.74598968e-01
-8.79317075e-02 -1.29569009e-01 1.62856743e-01 5.88032305e-01
3.02173644e-01 9.19435441e-01 1.99925378e-01 -4.36683863e-01
1.74023420e-01 -3.63922596e-01 -9.67570394e-02 -8.80751789e-01
-3.14392090e-01 1.08095117e-01 -1.56873882e-01 -1.59129485e-01
-9.77791131e-01 -2.70737320e-01 -1.19760680e+00 -6.87457994e-02
-3.22776794e-01 2.96661764e-01 5.86980104e-01 1.15447164e+00
2.72146910e-01 3.34022224e-01 1.08036923e+00 -1.09524834e+00
-3.03917646e-01 -1.22427988e+00 -8.61912429e-01 2.84143448e-01
5.82557321e-01 -3.48808676e-01 -3.37527633e-01 -1.33924693e-01] | [9.608356475830078, -1.42849862575531] |
8f685749-1cfc-4788-9732-f42342ddb86c | textbugger-generating-adversarial-text | 1812.05271 | null | http://arxiv.org/abs/1812.05271v1 | http://arxiv.org/pdf/1812.05271v1.pdf | TextBugger: Generating Adversarial Text Against Real-world Applications | Deep Learning-based Text Understanding (DLTU) is the backbone technique
behind various applications, including question answering, machine translation,
and text classification. Despite its tremendous popularity, the security
vulnerabilities of DLTU are still largely unknown, which is highly concerning
given its increasing use in security-sensitive applications such as sentiment
analysis and toxic content detection. In this paper, we show that DLTU is
inherently vulnerable to adversarial text attacks, in which maliciously crafted
texts trigger target DLTU systems and services to misbehave. Specifically, we
present TextBugger, a general attack framework for generating adversarial
texts. In contrast to prior works, TextBugger differs in significant ways: (i)
effective -- it outperforms state-of-the-art attacks in terms of attack success
rate; (ii) evasive -- it preserves the utility of benign text, with 94.9\% of
the adversarial text correctly recognized by human readers; and (iii) efficient
-- it generates adversarial text with computational complexity sub-linear to
the text length. We empirically evaluate TextBugger on a set of real-world DLTU
systems and services used for sentiment analysis and toxic content detection,
demonstrating its effectiveness, evasiveness, and efficiency. For instance,
TextBugger achieves 100\% success rate on the IMDB dataset based on Amazon AWS
Comprehend within 4.61 seconds and preserves 97\% semantic similarity. We
further discuss possible defense mechanisms to mitigate such attack and the
adversary's potential countermeasures, which leads to promising directions for
further research. | ['Ting Wang', 'Bo Li', 'Shouling Ji', 'Jinfeng Li', 'Tianyu Du'] | 2018-12-13 | null | null | null | null | ['adversarial-text'] | ['adversarial'] | [ 3.20135683e-01 -7.61870891e-02 1.54066026e-01 -1.21289052e-01
-8.63476872e-01 -1.31002629e+00 7.56446719e-01 4.12030250e-01
-3.58121485e-01 3.17444474e-01 2.20509887e-01 -8.53283703e-01
3.30870539e-01 -1.09126103e+00 -7.14954555e-01 -5.75497270e-01
1.92797959e-01 4.43662763e-01 2.69125775e-02 -6.50083482e-01
2.23831668e-01 3.12108636e-01 -1.00469291e+00 5.36892712e-01
6.51455045e-01 9.19939280e-01 -5.16691804e-01 9.52490151e-01
-2.08645239e-01 8.76361609e-01 -1.07004070e+00 -1.17167866e+00
4.10601288e-01 -3.59239459e-01 -9.06726778e-01 -4.23714161e-01
1.62167966e-01 -6.13077283e-01 -5.84010661e-01 1.28039646e+00
6.32028639e-01 -1.13721944e-01 4.50668246e-01 -1.37308228e+00
-7.77886391e-01 9.59148943e-01 -5.91871500e-01 1.50931388e-01
3.46414447e-01 2.74892986e-01 1.09316778e+00 -6.11274183e-01
2.70749837e-01 1.34299660e+00 3.56183618e-01 1.04082131e+00
-7.92176068e-01 -1.17973793e+00 1.94667414e-01 -1.19157828e-01
-9.08632040e-01 -2.43987992e-01 5.74433208e-01 -1.65018111e-01
7.88097262e-01 5.88986158e-01 -3.23538631e-02 1.74193060e+00
5.46045959e-01 1.11250329e+00 9.66823816e-01 -2.53206402e-01
3.79409492e-01 1.71828747e-01 6.40763193e-02 4.88892823e-01
2.22460151e-01 9.40793678e-02 -4.90847647e-01 -5.81240237e-01
-7.31518269e-02 -1.67896807e-01 -8.90022889e-02 3.14869702e-01
-7.30428815e-01 1.32934308e+00 1.35258809e-01 1.48825750e-01
4.51816171e-02 1.39532328e-01 9.81534600e-01 4.17953432e-01
6.22484863e-01 6.19027495e-01 -5.20686805e-01 -1.67325616e-01
-4.26798254e-01 4.58662242e-01 1.16003847e+00 7.83891439e-01
-3.65037918e-02 4.12193865e-01 -1.40577834e-02 5.21186948e-01
8.11622515e-02 9.40086067e-01 6.23555779e-01 -2.40194082e-01
6.91044927e-01 3.36575627e-01 -6.50419518e-02 -1.29158342e+00
-4.41076122e-02 -2.69260347e-01 -7.05771744e-01 1.33508220e-02
3.92584294e-01 -5.45432568e-01 -6.28745139e-01 1.64071810e+00
1.36143506e-01 -1.76124573e-01 1.67248473e-01 5.86706400e-01
6.10543430e-01 7.81520128e-01 1.46434546e-01 1.82619825e-01
1.49384975e+00 -6.13847613e-01 -4.86934662e-01 -5.27945399e-01
8.07665050e-01 -8.19119871e-01 1.38766146e+00 5.18093705e-01
-7.46398747e-01 9.11302716e-02 -1.15888393e+00 2.87558764e-01
-8.39718461e-01 -6.60046160e-01 5.68777978e-01 1.45724225e+00
-5.83205640e-01 1.80505216e-01 -3.98247749e-01 -8.80196914e-02
6.25643611e-01 3.51293117e-01 -1.99700907e-01 1.80783099e-03
-1.81088448e+00 5.29594719e-01 9.54194516e-02 -2.80266613e-01
-1.02350819e+00 -6.16839707e-01 -7.43692219e-01 1.13679767e-01
3.92836303e-01 -5.16671717e-01 1.43841076e+00 -7.79659033e-01
-1.49016523e+00 8.24679911e-01 1.90062270e-01 -8.36988747e-01
7.28004813e-01 -4.94883120e-01 -5.92860162e-01 1.43692672e-01
2.09418647e-02 -4.66940477e-02 1.15382576e+00 -1.08533430e+00
-3.75282496e-01 -4.34676290e-01 1.58356085e-01 1.04331985e-01
-9.38034177e-01 3.70522022e-01 -1.07812010e-01 -1.09558618e+00
-4.79735017e-01 -8.03059340e-01 -1.43006757e-01 -2.49830842e-01
-8.25061738e-01 -1.91569012e-02 1.36518538e+00 -4.42171156e-01
1.18878567e+00 -1.94263732e+00 -3.72313857e-01 3.40829790e-01
4.49234724e-01 7.33235478e-01 -2.28024006e-01 6.87643707e-01
-3.29630896e-02 6.89788997e-01 -8.44623521e-02 -1.71630695e-01
3.35540116e-01 -1.34349108e-01 -1.15083265e+00 4.14778292e-01
-1.62754059e-01 1.11546040e+00 -8.32874715e-01 3.55690867e-02
2.16013715e-01 1.42875642e-01 -4.26481366e-01 6.61863834e-02
-5.50407112e-01 -8.57356861e-02 -8.02431941e-01 6.19887948e-01
4.95170593e-01 3.14035229e-02 -5.42063117e-02 1.31881297e-01
5.97230911e-01 2.61223286e-01 -6.01544499e-01 8.57866824e-01
-4.83018994e-01 7.95114696e-01 9.34949368e-02 -7.00838268e-01
7.30862021e-01 1.49723306e-01 1.17835917e-01 -6.56604111e-01
6.65015340e-01 8.56298059e-02 -2.33049721e-01 -2.66254157e-01
4.85072672e-01 -2.10462406e-01 -5.99014103e-01 1.02122736e+00
-4.13591981e-01 -2.50914335e-01 -2.01043516e-01 7.24978626e-01
1.31124246e+00 -6.11420453e-01 3.25271301e-02 -3.31700593e-02
5.87190211e-01 -2.12805897e-01 -9.63565558e-02 1.12609243e+00
-2.99562842e-01 9.26846862e-02 6.71491206e-01 -3.97910684e-01
-9.44172442e-01 -8.48260224e-01 2.57727504e-01 1.38610220e+00
2.50771135e-01 -5.91104329e-01 -1.27096450e+00 -1.28882885e+00
4.19919565e-02 1.14351940e+00 -7.50315011e-01 -6.00172222e-01
-4.15619940e-01 -7.80524552e-01 1.36275053e+00 3.96395922e-01
5.41081011e-01 -1.05536580e+00 -1.25202149e-01 1.13879241e-01
-1.81252331e-01 -1.29811871e+00 -8.22172403e-01 -1.74109548e-01
-2.73879349e-01 -1.02405226e+00 -1.44869298e-01 -3.25896293e-01
3.88968825e-01 3.38128865e-01 9.38561916e-01 1.04520589e-01
-2.71896452e-01 3.97034913e-01 -6.55595481e-01 -7.96753407e-01
-1.06862104e+00 1.61210969e-01 1.09467380e-01 2.18723738e-03
6.56379759e-01 -2.51713783e-01 -4.83142465e-01 3.42146933e-01
-1.45035052e+00 -4.59872305e-01 2.69480407e-01 8.44585836e-01
1.06661720e-02 3.98576707e-01 8.09398115e-01 -1.37293041e+00
1.18950367e+00 -5.87408781e-01 -4.55656916e-01 2.88714189e-02
-6.41213119e-01 -2.29456022e-01 1.25550938e+00 -6.35695457e-01
-8.50242138e-01 -5.90426803e-01 -6.77127957e-01 -2.12782398e-01
-3.41752656e-02 2.55137056e-01 -4.35448349e-01 -1.07712567e-01
9.99984384e-01 4.28139925e-01 -9.03547928e-02 6.44731224e-02
5.75909019e-01 8.47088873e-01 3.26742530e-01 -6.92419410e-01
1.28320062e+00 5.14500082e-01 -4.10497487e-01 -6.96584702e-01
-8.57954979e-01 -9.73682327e-05 2.54502714e-01 -1.08086625e-02
7.38852203e-01 -4.63899583e-01 -1.11604869e+00 8.64883959e-01
-1.01385665e+00 -2.71575898e-01 9.22864676e-02 -2.00664237e-01
-1.29862145e-01 6.78978324e-01 -9.92984712e-01 -7.23735929e-01
-1.19627440e+00 -1.10409796e+00 8.56140018e-01 -1.03195317e-01
-3.16416174e-01 -1.17909038e+00 -2.10910469e-01 7.81529427e-01
3.77156317e-01 2.98213631e-01 1.16445112e+00 -1.45192134e+00
8.22823793e-02 -7.98963487e-01 -2.26907339e-02 5.79005659e-01
-6.96536079e-02 -2.03127548e-01 -1.03505373e+00 -5.02524316e-01
2.93277591e-01 -5.48530757e-01 4.03332263e-01 -2.59504408e-01
1.27629638e+00 -8.04520309e-01 -1.02264732e-01 3.36745024e-01
1.01521719e+00 2.28453219e-01 5.50150812e-01 4.35568720e-01
6.75494075e-01 6.28419876e-01 4.81836379e-01 5.96991658e-01
-1.21402189e-01 2.41808072e-01 8.42398584e-01 3.71465981e-02
4.58576351e-01 -6.04851067e-01 7.10143328e-01 3.67558628e-01
8.67874682e-01 -9.36550736e-01 -8.45730603e-01 1.68950140e-01
-1.52950108e+00 -9.99477983e-01 -2.18012426e-02 1.89928520e+00
8.32838655e-01 5.94438255e-01 3.02268844e-02 3.57547045e-01
7.08996773e-01 3.55622411e-01 -8.66397798e-01 -8.58956039e-01
-2.27208525e-01 5.18661261e-01 6.97396636e-01 4.01651829e-01
-1.15557110e+00 1.31016290e+00 5.85935450e+00 1.23769176e+00
-1.05006957e+00 2.09518597e-01 8.30241740e-01 -1.03572913e-01
-3.92785251e-01 -4.01925117e-01 -6.48356616e-01 6.13581240e-01
9.73797739e-01 -5.75504541e-01 4.15182501e-01 9.97557163e-01
-1.59680303e-02 4.81143236e-01 -7.91526258e-01 5.97469866e-01
3.44801009e-01 -1.24627709e+00 5.04338205e-01 2.80411150e-02
4.39160019e-01 -8.93444940e-02 4.52950001e-01 4.60991472e-01
8.18201721e-01 -1.14046824e+00 7.43222713e-01 -4.79318589e-01
6.74016058e-01 -1.25956893e+00 7.78492033e-01 4.38952893e-01
-7.00081408e-01 -1.00799099e-01 -1.62505299e-01 1.62989780e-01
-1.12209320e-01 4.83229190e-01 -7.88195133e-01 3.57771695e-01
6.07576013e-01 1.22023121e-01 -4.79082167e-01 2.49848841e-03
-3.00238699e-01 1.02621353e+00 -1.87576890e-01 -4.61015075e-01
5.30380607e-01 1.85725152e-01 6.74318373e-01 1.25003564e+00
-6.72541186e-02 1.36218160e-01 1.19902872e-01 6.46045923e-01
-5.60126781e-01 2.20448807e-01 -7.25529015e-01 -4.83374506e-01
6.29633129e-01 1.12158585e+00 -6.20225906e-01 -2.03867212e-01
-1.67417601e-01 1.33631992e+00 -1.48169011e-01 2.32443750e-01
-9.31677639e-01 -7.93749511e-01 7.68660188e-01 -1.04906105e-01
-4.84968461e-02 6.25059307e-02 -4.70470160e-01 -1.18073452e+00
-1.09709315e-01 -1.68811166e+00 5.20298362e-01 -4.81003135e-01
-1.63527954e+00 9.10118282e-01 -4.01429504e-01 -7.98382163e-01
-2.29154795e-01 -7.41244018e-01 -6.29142880e-01 7.53274381e-01
-8.48571658e-01 -1.27335238e+00 9.93615016e-02 8.72453868e-01
7.36936450e-01 -5.07609069e-01 9.70743656e-01 9.47172865e-02
-7.05754519e-01 1.28249657e+00 1.37523696e-01 6.57681286e-01
6.08342469e-01 -1.08091712e+00 1.24008489e+00 1.10520768e+00
1.23160154e-01 6.00784540e-01 9.56358850e-01 -6.29262745e-01
-1.65922093e+00 -1.24752831e+00 5.32644868e-01 -7.34253228e-01
1.17426014e+00 -8.51530015e-01 -8.23863447e-01 5.99609077e-01
8.05987865e-02 -4.05324101e-01 9.56150949e-01 -3.33961248e-01
-9.87664521e-01 1.20444871e-01 -1.44559085e+00 1.16783738e+00
5.38496196e-01 -8.17267895e-01 -2.76874810e-01 5.76361954e-01
9.68906939e-01 -4.03478742e-01 -4.17029858e-01 -2.99149975e-02
4.70231414e-01 -8.72925341e-01 8.02547514e-01 -8.36575925e-01
5.22342503e-01 1.49030343e-01 -2.06063300e-01 -1.01461256e+00
2.22486809e-01 -1.07351029e+00 -2.67028540e-01 1.22825813e+00
2.70293713e-01 -9.19614434e-01 9.06205833e-01 5.92457235e-01
2.16368765e-01 -7.67112851e-01 -5.75961351e-01 -5.47663391e-01
6.54376924e-01 -8.03443551e-01 6.05438530e-01 1.11849737e+00
8.37299228e-03 3.97092372e-01 -4.92110729e-01 2.31647626e-01
5.66142678e-01 -3.55044216e-01 8.65308225e-01 -7.13654697e-01
-3.87659431e-01 -5.55093825e-01 -9.26824659e-02 -9.66450512e-01
5.56769967e-01 -7.56284058e-01 -2.12291464e-01 -8.23965073e-01
-1.53348699e-01 -6.69244751e-02 5.94550818e-02 5.83257020e-01
-3.30329001e-01 3.40315133e-01 2.15409026e-01 -5.32027632e-02
-2.74504781e-01 3.87723148e-01 6.88594699e-01 -4.45848733e-01
1.00216970e-01 1.82653472e-01 -1.23232865e+00 7.23287165e-01
1.11451602e+00 -5.16964853e-01 -6.24306381e-01 -1.61289707e-01
4.69909012e-01 -2.62915850e-01 3.87570076e-02 -3.69035989e-01
-5.93894795e-02 -8.19370598e-02 2.76273452e-02 -3.03595692e-01
2.22605076e-02 -5.99811375e-01 -5.65851927e-01 6.28012478e-01
-5.90571105e-01 1.07369728e-01 3.99112046e-01 6.59326792e-01
1.66584864e-01 -2.51928329e-01 7.79335976e-01 -3.78904715e-02
-2.67849863e-01 4.09525603e-01 -8.84490311e-01 4.35936958e-01
9.80231941e-01 8.12473372e-02 -7.47714281e-01 -8.73056650e-01
-1.07541546e-01 1.27068564e-01 2.90192217e-01 7.07601070e-01
8.12586665e-01 -8.36562634e-01 -7.47483075e-01 1.07735373e-01
1.76324248e-01 -5.25085747e-01 1.61003083e-01 -1.07662929e-02
-4.97301489e-01 2.03472987e-01 3.08843702e-01 -1.87264085e-02
-1.38816643e+00 7.68520236e-01 3.82054567e-01 -3.56169343e-01
-5.23225546e-01 8.81719887e-01 4.28439289e-01 -2.95416057e-01
2.54887164e-01 3.17406744e-01 1.62449896e-01 -3.39368641e-01
8.90912652e-01 2.60676265e-01 2.08997518e-01 -3.71667236e-01
-2.15410039e-01 9.88832768e-03 -6.84629500e-01 -1.52962983e-01
6.67911589e-01 1.03408262e-01 -8.36040825e-02 -1.68867812e-01
1.21105754e+00 4.23928440e-01 -6.22959852e-01 -3.37325066e-01
-5.92361093e-02 -3.77114594e-01 -5.24569079e-02 -1.10169291e+00
-9.25843239e-01 8.66853416e-01 2.45244816e-01 6.18499279e-01
9.88779902e-01 -3.10381770e-01 1.55650723e+00 5.53301036e-01
2.72028357e-01 -7.50774980e-01 4.42205250e-01 5.57091296e-01
7.31878698e-01 -9.21463370e-01 -2.38611341e-01 -3.02294075e-01
-8.44340563e-01 8.97711456e-01 6.68015718e-01 6.24726973e-02
5.00541151e-01 5.38988054e-01 3.62957478e-01 -4.84673120e-02
-8.01744819e-01 3.88432980e-01 -4.14835215e-02 5.15893519e-01
1.06867217e-02 -3.86686809e-02 2.54502892e-02 9.21916604e-01
-5.57874501e-01 -6.58020675e-01 6.24350846e-01 9.02885079e-01
-3.14603627e-01 -1.15935683e+00 -4.82788026e-01 4.48495895e-01
-1.16183043e+00 -4.10147399e-01 -9.21880484e-01 5.65946579e-01
-4.16726798e-01 1.51036501e+00 -4.80621189e-01 -8.95748913e-01
2.95073003e-01 -5.61082587e-02 -9.84507203e-02 -5.14349163e-01
-1.15200019e+00 -2.28515387e-01 1.54772848e-01 -3.55548769e-01
4.75580931e-01 -2.07835630e-01 -1.28131616e+00 -1.01002824e+00
-3.73149097e-01 2.92430043e-01 6.96323633e-01 9.74147916e-01
3.95731330e-01 1.33903742e-01 1.06028676e+00 -3.13364387e-01
-1.05090499e+00 -6.47316217e-01 -4.11375344e-01 6.49450779e-01
1.46012127e-01 -4.21407297e-02 -7.60116339e-01 -1.18249327e-01] | [6.006949424743652, 8.04949951171875] |
17f46da6-d3aa-4d4d-b5e4-818d37825221 | adaptive-statistical-learning-with-bayesian | 1911.00765 | null | https://arxiv.org/abs/1911.00765v1 | https://arxiv.org/pdf/1911.00765v1.pdf | Adaptive Statistical Learning with Bayesian Differential Privacy | In statistical learning, a dataset is often partitioned into two parts: the training set and the holdout (i.e., testing) set. For instance, the training set is used to learn a predictor, and then the holdout set is used for estimating the accuracy of the predictor on the true distribution. However, often in practice, the holdout dataset is reused and the estimates tested on the holdout dataset are chosen adaptively based on the results of prior estimates, leading to that the predictor may become dependent of the holdout set. Hence, overfitting may occur, and the learned models may not generalize well to the unseen datasets. Prior studies have established connections between the stability of a learning algorithm and its ability to generalize, but the traditional generalization is not robust to adaptive composition. Recently, Dwork et al. in NIPS, STOC, and Science 2015 show that the holdout dataset from i.i.d. data samples can be reused in adaptive statistical learning, if the estimates are perturbed and coordinated using techniques developed for differential privacy, which is a widely used notion to quantify privacy. Yet, the results of Dwork et al. are applicable to only the case of i.i.d. samples. In contrast, correlations between data samples exist because of various behavioral, social, and genetic relationships between users. Our results in adaptive statistical learning generalize the results of Dwork et al. for i.i.d. data samples to arbitrarily correlated data. Specifically, we show that the holdout dataset from correlated samples can be reused in adaptive statistical learning, if the estimates are perturbed and coordinated using techniques developed for Bayesian differential privacy, which is a privacy notion recently introduced by Yang et al. in SIGMOD 2015 to broaden the application scenarios of differential privacy when data records are correlated. | ['Jun Zhao'] | 2019-11-02 | null | null | null | null | ['holdout-set'] | ['computer-vision'] | [ 2.19764307e-01 -6.91726059e-02 -3.72390240e-01 -3.91017854e-01
-7.26258993e-01 -8.28993917e-01 8.52240175e-02 4.37645346e-01
-6.04618013e-01 1.01250052e+00 -2.32498080e-01 -1.08040042e-01
-2.72227764e-01 -9.36906815e-01 -9.51822937e-01 -1.03697693e+00
-2.70488411e-01 3.78505677e-01 1.14104394e-02 2.20778316e-01
1.51210688e-02 4.27408487e-01 -1.40136302e+00 5.86806387e-02
7.01733053e-01 8.22109222e-01 -4.07551497e-01 5.13176978e-01
1.92405269e-01 3.62110063e-02 -6.20202243e-01 -3.01302373e-01
5.91590405e-01 -4.51319724e-01 -2.93284506e-01 -8.75328183e-02
4.53788221e-01 -1.15487687e-01 -2.68006712e-01 1.29965568e+00
3.65865916e-01 1.41454950e-01 4.77557153e-01 -1.77679574e+00
-2.67964423e-01 6.68762803e-01 -5.75279355e-01 -1.86286598e-01
1.86729968e-01 -1.70190662e-01 9.22695458e-01 -4.35854316e-01
3.75576735e-01 9.20559227e-01 7.05016017e-01 4.99926746e-01
-1.57862401e+00 -1.21519649e+00 1.77338958e-01 -1.02476515e-02
-1.48033345e+00 -2.03269452e-01 3.85859162e-01 -3.15887243e-01
4.19201851e-02 5.75756550e-01 5.94497979e-01 1.06618190e+00
1.59538999e-01 9.45640683e-01 9.53146875e-01 -1.54929072e-01
7.02833354e-01 4.86932218e-01 3.95213783e-01 3.20471495e-01
7.77930915e-01 2.62727797e-01 -4.81871694e-01 -7.82048821e-01
3.01950097e-01 5.46741188e-01 -4.59482819e-01 -6.99218273e-01
-5.28464556e-01 7.48715937e-01 3.72605026e-02 4.17515524e-02
-1.85154364e-01 -1.57114059e-01 3.24303120e-01 6.39425755e-01
3.82677436e-01 4.29914892e-01 -6.06898427e-01 2.92211175e-02
-9.33381975e-01 5.48164010e-01 1.15738237e+00 1.06183994e+00
9.78220463e-01 -5.41227937e-01 -1.25854328e-01 5.06609559e-01
3.69394273e-02 4.18514848e-01 3.97069305e-01 -7.50703216e-01
3.58732969e-01 2.92571217e-01 1.12254344e-01 -8.96872342e-01
-1.84875533e-01 -2.79501200e-01 -8.71998549e-01 4.33730707e-02
8.22522104e-01 -3.68092686e-01 -4.00040507e-01 2.21213961e+00
3.94112766e-01 2.12892383e-01 -8.12652484e-02 5.62153518e-01
-3.17281708e-02 3.40535998e-01 -4.26014215e-02 -3.99423391e-01
1.01136982e+00 -2.34763950e-01 -5.08175731e-01 6.84567764e-02
6.29827023e-01 -2.15222642e-01 9.43173528e-01 7.27656126e-01
-8.43416154e-01 -2.62088269e-01 -1.03325343e+00 2.41333231e-01
-3.18336517e-01 -3.23129714e-01 3.60000700e-01 8.91955197e-01
-7.84106314e-01 6.44623816e-01 -6.83526397e-01 -4.81767744e-01
7.35967398e-01 3.68584722e-01 -4.51790065e-01 -3.62508982e-01
-1.11409676e+00 2.44699940e-01 3.33571285e-01 -3.36259097e-01
-6.02376223e-01 -8.20251942e-01 -3.56744498e-01 1.20771170e-01
5.27182937e-01 -5.10727048e-01 8.39889050e-01 -1.01941824e+00
-9.81938422e-01 5.90790570e-01 -1.71932429e-01 -6.19214952e-01
9.63020205e-01 -1.45515904e-01 -1.63677320e-01 -2.22845733e-01
-2.58170608e-02 -1.42295420e-01 7.21205831e-01 -1.13900578e+00
-6.68404162e-01 -8.54933262e-01 -1.55328378e-01 -2.99578279e-01
-5.32762468e-01 -4.99895871e-01 -3.38860393e-01 -5.46177447e-01
4.57545929e-02 -1.17837894e+00 -1.89827427e-01 3.03668082e-01
-5.02345145e-01 1.42353833e-01 8.89984965e-01 -3.55660975e-01
1.27076125e+00 -2.55271101e+00 -1.17033638e-01 7.93774962e-01
2.30616540e-01 1.61707681e-02 -5.56398788e-03 3.65629226e-01
-3.82413641e-02 3.14018309e-01 -4.47777539e-01 -3.31793606e-01
-1.38914302e-01 2.91032314e-01 -4.11306232e-01 7.67564178e-01
-2.88610071e-01 2.63848424e-01 -6.89814687e-01 -8.04927573e-02
-2.59639174e-01 1.83617890e-01 -8.19605589e-01 2.39057839e-01
-6.47620410e-02 4.74836200e-01 -5.99033237e-01 3.07779998e-01
9.75426614e-01 -7.12000728e-02 5.08602619e-01 1.23711362e-01
1.82822019e-01 -1.29567116e-01 -1.40403545e+00 1.03524315e+00
-1.05948322e-01 4.25225437e-01 3.52198593e-02 -1.02268040e+00
8.42475951e-01 3.20198148e-01 5.16370833e-01 -7.48799220e-02
-5.73097281e-02 1.52108163e-01 -5.47990426e-02 -2.44348079e-01
3.08988214e-01 -1.54788956e-01 -1.45884454e-01 6.29194319e-01
-2.88802296e-01 1.06155805e-01 -1.37367621e-01 3.78979594e-02
1.18946815e+00 -2.46326342e-01 4.91423070e-01 -1.20619677e-01
3.11684668e-01 -3.65821928e-01 9.02256966e-01 1.16539979e+00
-1.99550405e-01 6.05315506e-01 8.12579274e-01 -1.35103658e-01
-9.94913578e-01 -1.21827793e+00 -5.30994654e-01 9.27091181e-01
-3.86523567e-02 -3.46667290e-01 -7.40300179e-01 -9.92651105e-01
5.83403885e-01 7.07910657e-01 -8.32508206e-01 -3.76354277e-01
-5.75348176e-02 -7.15062380e-01 5.66026807e-01 2.19097748e-01
3.23972851e-01 -4.00721818e-01 -4.04701531e-01 1.44428715e-01
2.85063118e-01 -7.34289765e-01 -6.40862286e-01 1.56756312e-01
-9.19677079e-01 -1.40367401e+00 -3.11447680e-01 -9.11491513e-02
8.62895250e-01 1.30170465e-01 5.86148918e-01 7.62892216e-02
5.89283556e-03 4.59684372e-01 -1.46603763e-01 -5.52214563e-01
-3.34499687e-01 -1.76760316e-01 3.10491264e-01 3.40165228e-01
5.43865740e-01 -7.02745557e-01 -3.45194012e-01 3.57381344e-01
-1.13655293e+00 -5.88262081e-01 3.52140009e-01 9.02987957e-01
7.18159497e-01 2.63118178e-01 5.43309808e-01 -1.52429569e+00
4.01724815e-01 -9.69413280e-01 -6.50429666e-01 4.14066464e-01
-8.60278845e-01 1.18837357e-01 7.78864145e-01 -8.37732971e-01
-6.01099372e-01 -1.58406179e-02 4.01477247e-01 -5.98936677e-01
-3.16980444e-02 3.53676289e-01 -5.75308204e-01 1.32400975e-01
5.04486859e-01 1.74079001e-01 2.36801356e-01 -4.53748196e-01
-2.94489469e-02 8.00820112e-01 2.02727765e-01 -7.35538483e-01
7.52783775e-01 3.52600843e-01 1.74901322e-01 -7.24983752e-01
-6.48965120e-01 -1.83783978e-01 -4.20432836e-01 2.24810764e-01
3.27160299e-01 -5.91005206e-01 -7.17285097e-01 4.18005019e-01
-5.64217448e-01 -8.04325342e-02 -5.58259964e-01 6.16147161e-01
-4.54371542e-01 4.14760292e-01 -6.74344599e-02 -1.08277488e+00
1.98392738e-02 -1.00984240e+00 5.95474839e-01 7.04773292e-02
-4.88639802e-01 -9.98914480e-01 -1.25739440e-01 1.36852324e-01
5.24817258e-02 3.67959499e-01 1.07487345e+00 -1.40914285e+00
-4.40306693e-01 -6.53873801e-01 2.98294038e-01 4.56005156e-01
2.75932640e-01 4.68054563e-02 -8.87564063e-01 -7.70901203e-01
9.06210691e-02 -1.57669738e-01 6.90449953e-01 3.34222734e-01
1.63044214e+00 -7.72811949e-01 -4.47937429e-01 6.49298251e-01
1.21073437e+00 2.76200205e-01 3.87946069e-01 1.55997977e-01
2.27224424e-01 5.71790755e-01 5.09806395e-01 7.40334749e-01
8.70607123e-02 4.36678082e-01 7.50564113e-02 3.14297646e-01
7.36246049e-01 -5.80567360e-01 4.22705591e-01 2.91082561e-02
4.03800637e-01 -1.62355244e-01 -5.76929748e-01 2.22459018e-01
-1.85039461e+00 -9.85618532e-01 1.44333228e-01 3.09051085e+00
1.21053982e+00 -9.41741839e-02 3.20822239e-01 7.73664713e-02
7.29204297e-01 -1.61687568e-01 -1.07678127e+00 -1.71978831e-01
-1.11717910e-01 1.48463979e-01 8.81819665e-01 2.64987260e-01
-8.82993042e-01 3.24716330e-01 5.61557674e+00 6.97680771e-01
-9.89718258e-01 -1.82600655e-02 7.97520518e-01 -4.06820774e-01
-5.12831986e-01 2.01840892e-01 -6.30769134e-01 6.91142201e-01
9.15693939e-01 -6.49556994e-01 4.82064098e-01 9.36491609e-01
-6.00414909e-02 -2.08819509e-01 -1.60694313e+00 8.95232975e-01
-1.26213431e-01 -6.94590449e-01 -3.71201992e-01 4.98263091e-01
7.58539021e-01 -2.65221447e-01 3.67139041e-01 2.79110730e-01
5.37949085e-01 -8.63070488e-01 3.56186926e-01 7.30133832e-01
5.59844971e-01 -9.15275693e-01 7.09701598e-01 7.15051293e-01
-5.09677470e-01 -2.15350106e-01 -2.92313337e-01 1.41802087e-01
-4.50306416e-01 9.32297170e-01 -7.32936502e-01 2.99559742e-01
7.34124005e-01 5.52786231e-01 -3.49688530e-01 1.01497328e+00
9.31499451e-02 7.92691827e-01 -5.30996203e-01 -2.44238460e-03
-3.40868205e-01 -4.70073313e-01 6.52798951e-01 7.57107437e-01
4.32887435e-01 -9.54557732e-02 2.26831466e-01 6.87266707e-01
-2.51106232e-01 1.15241736e-01 -6.79086387e-01 -2.68115737e-02
8.50514650e-01 7.61108041e-01 3.23556699e-02 -1.95103958e-01
-3.53537709e-01 7.29230344e-01 2.46744767e-01 6.11837983e-01
-5.32597125e-01 -1.60631329e-01 9.89760041e-01 2.17354223e-01
1.68533638e-01 3.84322219e-02 -2.92489648e-01 -1.17574847e+00
7.63933584e-02 -9.14530039e-01 8.40920925e-01 1.27461634e-03
-1.75257838e+00 7.66913872e-03 1.93034708e-01 -1.21329069e+00
-1.90468267e-01 -2.15962902e-01 -3.49738777e-01 9.53176200e-01
-9.40358341e-01 -4.81371433e-01 1.69294849e-02 5.64309239e-01
-2.12910280e-01 -2.05816522e-01 7.99671948e-01 -6.63168635e-03
-5.83812714e-01 1.12221086e+00 8.08295012e-01 1.20438382e-01
9.17965591e-01 -1.12252271e+00 -1.64112628e-01 7.34544098e-01
4.31004837e-02 8.63636136e-01 6.28617823e-01 -6.26188874e-01
-1.39631796e+00 -1.15832651e+00 4.80029792e-01 -2.81569690e-01
4.69616413e-01 -4.97141927e-01 -1.36103141e+00 9.41134095e-01
-4.83320087e-01 1.43490985e-01 1.05392909e+00 2.30091572e-01
-6.65641010e-01 -4.67588961e-01 -1.90377462e+00 5.02173543e-01
7.05930591e-01 -5.35353541e-01 -1.03616342e-01 5.61880209e-02
4.44790274e-01 -2.64316052e-01 -9.81397629e-01 1.52401656e-01
8.57567608e-01 -9.28188324e-01 5.85722148e-01 -8.88357699e-01
-7.39605501e-02 -1.02222003e-01 -4.39502984e-01 -1.04245067e+00
8.72431472e-02 -6.54973984e-01 -2.72466063e-01 1.31565583e+00
4.42610770e-01 -1.19733012e+00 7.91102529e-01 1.19817042e+00
6.34447515e-01 -6.22397065e-01 -1.14032578e+00 -8.86741161e-01
2.31441557e-01 -4.43131715e-01 8.04196477e-01 1.08581662e+00
-1.43916205e-01 -3.37359041e-01 -4.32611525e-01 3.42344284e-01
8.42875004e-01 -1.34926178e-02 1.07513475e+00 -1.32680154e+00
-5.81049085e-01 -1.23915717e-01 -3.84204954e-01 -8.43094707e-01
2.66971588e-01 -8.62988234e-01 -1.52107909e-01 -5.45877814e-01
3.34825307e-01 -8.39935958e-01 -5.09537220e-01 4.91210312e-01
-3.35395992e-01 -2.97048062e-01 2.38685042e-01 2.90119529e-01
-4.87204008e-02 3.00717473e-01 8.44513357e-01 8.50777850e-02
-4.57876444e-01 7.39844143e-01 -8.62640858e-01 2.82726347e-01
8.72247517e-01 -7.48929560e-01 -4.78868037e-01 2.85139829e-01
1.31005030e-02 2.05227397e-02 3.27890605e-01 -7.76838481e-01
2.77115494e-01 -2.90858746e-01 3.96028519e-01 -1.79592028e-01
1.96620841e-02 -1.21655381e+00 3.68718177e-01 5.06172299e-01
-6.88052416e-01 -2.72596627e-01 9.19312835e-02 1.12575614e+00
1.07376307e-01 -2.63739377e-01 8.61664057e-01 2.30280116e-01
8.98934752e-02 7.57492661e-01 -4.25597802e-02 3.87030602e-01
1.16041744e+00 -3.30071777e-01 -8.73928517e-02 -5.72403967e-01
-6.29263043e-01 2.70279735e-01 7.13338733e-01 1.54213145e-01
4.76148337e-01 -1.01944065e+00 -6.11998856e-01 5.42685747e-01
1.94839641e-01 1.09261975e-01 1.30789921e-01 8.06370974e-01
2.81862736e-01 -1.74054861e-01 1.89707100e-01 -4.80012268e-01
-1.31531549e+00 6.43175483e-01 1.09621160e-01 -2.72795092e-02
-3.59299779e-01 5.20779431e-01 2.05866292e-01 -3.72498184e-01
3.40793580e-01 -1.98777333e-01 4.37135667e-01 1.84404805e-01
6.02148116e-01 3.10398072e-01 -2.19256163e-01 -1.21645033e-01
-1.95726991e-01 2.28547230e-01 -3.37714076e-01 -2.10281938e-01
1.41345263e+00 -1.77897677e-01 -1.94905013e-01 8.49284708e-01
1.50599635e+00 5.29876649e-01 -1.18612576e+00 -7.35294044e-01
7.83692002e-02 -7.92623401e-01 -3.01252007e-01 -5.36129892e-01
-1.05165684e+00 5.19845426e-01 5.86378694e-01 2.02178970e-01
1.31703115e+00 -1.99264735e-01 4.93787169e-01 4.05743837e-01
6.09753072e-01 -9.88220572e-01 -5.14697909e-01 2.44888604e-01
6.48491621e-01 -1.00888610e+00 1.66806251e-01 -3.36105898e-02
-6.87230647e-01 1.10190904e+00 3.30684870e-01 8.23659673e-02
9.21126068e-01 3.05200338e-01 -3.06110144e-01 3.82694423e-01
-6.96208239e-01 4.36360270e-01 3.83311324e-02 5.23071051e-01
8.13525543e-02 1.84178859e-01 -2.22326607e-01 9.47091758e-01
-2.42133260e-01 -1.44661916e-02 4.83264983e-01 8.91447544e-01
-7.47948140e-02 -1.36350310e+00 -4.00933653e-01 6.81065917e-01
-3.72525752e-01 1.31381795e-01 -4.72487956e-01 8.15413654e-01
3.16135027e-02 6.45848930e-01 2.43177369e-01 -4.33620840e-01
3.41458768e-01 3.23127747e-01 1.92439154e-01 -5.02387822e-01
-5.52263021e-01 -4.59162652e-01 -2.99538642e-01 -4.32245046e-01
-9.63256955e-02 -1.10930252e+00 -1.07255542e+00 -5.41795611e-01
-2.04873487e-01 4.66725260e-01 3.98642361e-01 7.04825580e-01
4.58279043e-01 -8.83464292e-02 1.09637713e+00 -2.08273027e-02
-1.06316876e+00 -6.74019217e-01 -1.04869759e+00 4.92054075e-01
6.95120454e-01 -2.57003605e-01 -7.33800352e-01 -2.22542748e-01] | [6.084505558013916, 6.670045852661133] |
1dcfd664-43b9-417e-b190-981b1b740a7b | prevention-and-resolution-of-conflicts-in | 2106.12113 | null | https://arxiv.org/abs/2106.12113v3 | https://arxiv.org/pdf/2106.12113v3.pdf | Conflict Avoidance in Social Navigation -- a Survey | A major goal in robotics is to enable intelligent mobile robots to operate smoothly in shared human-robot environments. One of the most fundamental capabilities in service of this goal is competent navigation in this ``social" context. As a result, there has been a recent surge of research on social navigation; and especially as it relates to the handling of conflicts between agents during social navigation. These developments introduce a variety of models and algorithms, however as this research area is inherently interdisciplinary, many of the relevant papers are not comparable and there is no shared standard vocabulary. This survey aims to bridge this gap by introducing such a common language, using it to survey existing work, and highlighting open problems. It starts by defining the boundaries of this survey to a limited, yet highly common type of social navigation - conflict avoidance. Within this proposed scope, this survey introduces a detailed taxonomy of the conflict avoidance components. This survey then maps existing work into this taxonomy, while discussing papers using its framing. Finally, this paper proposes some future research directions and open problems that are currently on the frontier of social navigation to aid ongoing and future research. | ['Peter Stone', 'Justin Hart', 'Xuesu Xiao', 'Reuth Mirsky'] | 2021-06-23 | null | null | null | null | ['social-navigation'] | ['robots'] | [ 1.19893327e-01 5.97707391e-01 -4.91156757e-01 -3.51912141e-01
-5.05932532e-02 -5.39261460e-01 8.44895840e-01 -2.41617918e-01
-7.10291505e-01 9.88968968e-01 4.46084663e-02 -4.34747726e-01
-6.04839802e-01 -7.07005739e-01 -2.36204192e-01 -6.95018232e-01
-1.94413483e-01 5.76250553e-01 4.47297364e-01 -1.10966945e+00
5.85073233e-01 3.72758120e-01 -2.01491952e+00 -5.44990778e-01
9.25185263e-01 3.63545805e-01 7.21969485e-01 5.83081722e-01
-4.39040102e-02 6.77342415e-01 -4.15898085e-01 -2.21493125e-01
2.15714604e-01 -4.26668942e-01 -1.29438913e+00 -5.64357191e-02
-1.73676923e-01 1.22554526e-01 -7.97049403e-02 1.13224697e+00
3.92481655e-01 6.02139592e-01 6.61180019e-01 -1.84934878e+00
-2.47789115e-01 3.60448807e-01 3.10788304e-02 -9.65974480e-02
7.85037577e-01 -3.86919230e-01 8.76955271e-01 -5.73475003e-01
8.75087559e-01 1.17155266e+00 6.77794576e-01 7.08366990e-01
-4.40920204e-01 -2.78919369e-01 4.76866394e-01 4.95355874e-01
-1.05112314e+00 -4.72368717e-01 3.19990605e-01 -2.22552106e-01
1.16466391e+00 2.56056815e-01 8.09479713e-01 8.03270400e-01
4.23864335e-01 6.44367933e-01 6.74213886e-01 -6.14688158e-01
3.81188005e-01 -8.12071748e-03 3.08963835e-01 4.47139561e-01
5.47929764e-01 7.12345019e-02 -3.20354640e-01 -3.43480706e-02
4.21111614e-01 -2.41630435e-01 8.73058513e-02 -1.05790138e+00
-1.25187147e+00 8.35908830e-01 3.42426360e-01 5.08603394e-01
-1.94608137e-01 -1.03739714e-02 3.07150960e-01 4.79470223e-01
-2.37282336e-01 5.71123183e-01 3.27603742e-02 -4.93525624e-01
1.65833943e-02 5.75425506e-01 1.31367481e+00 1.00559390e+00
7.39616692e-01 -1.88615724e-01 9.06717122e-01 9.57757413e-01
7.32905269e-01 4.21738088e-01 3.31801891e-01 -1.50520098e+00
2.41957635e-01 2.77705759e-01 4.87549812e-01 -1.16289282e+00
-9.38039541e-01 8.95719677e-02 -3.61567110e-01 8.14845383e-01
3.37683797e-01 -2.98612267e-01 -1.96117863e-01 1.62842774e+00
3.68367881e-01 -5.20144105e-01 6.15649879e-01 7.67711163e-01
7.31204450e-01 3.89721781e-01 -8.24558660e-02 -2.34073669e-01
1.18753314e+00 -1.53728819e+00 -9.87951636e-01 -6.53174698e-01
8.46836746e-01 -7.66110301e-01 5.14068663e-01 1.20224357e-01
-8.68779182e-01 7.32724220e-02 -1.27332067e+00 -7.65865073e-02
-8.35258901e-01 -4.46919620e-01 9.33624625e-01 5.73084831e-01
-1.37512636e+00 3.77784491e-01 -6.22952759e-01 -1.38964736e+00
-2.56769955e-01 4.88818049e-01 -3.93012434e-01 -3.43078934e-02
-1.27795541e+00 1.73651421e+00 1.86339945e-01 6.82410821e-02
-2.14960396e-01 3.25019926e-01 -8.57352078e-01 -7.63406038e-01
6.17279947e-01 -8.20526659e-01 1.64338827e+00 -6.18012488e-01
-1.83722794e+00 9.11214232e-01 -7.97457024e-02 -4.10904825e-01
5.54327786e-01 -2.64515549e-01 -3.12338024e-01 -1.91461146e-01
8.06151032e-01 7.22291231e-01 -1.04374774e-01 -1.40673816e+00
-1.25577545e+00 -3.21570069e-01 4.85848427e-01 9.91628826e-01
2.71268003e-02 2.65421450e-01 -2.99990892e-01 -2.73161441e-01
2.95752913e-01 -1.50444210e+00 -9.18037236e-01 -2.15509921e-01
5.29220379e-05 -5.35790920e-01 6.27398372e-01 1.88479438e-01
9.81138766e-01 -1.80924118e+00 5.02726614e-01 -4.36987504e-02
1.06578358e-02 -7.15671033e-02 -1.47975264e-02 1.00788057e+00
3.38219881e-01 -2.52190590e-01 -4.10427988e-01 -2.64721185e-01
2.18007579e-01 7.26999402e-01 -2.10562810e-01 5.61842322e-01
-6.37450755e-01 6.33745551e-01 -1.49759114e+00 -3.15860271e-01
5.71301699e-01 2.14678824e-01 -3.64213169e-01 -3.27802271e-01
2.08211258e-01 3.99972260e-01 -4.10149306e-01 4.77629662e-01
2.74903715e-01 2.54116625e-01 3.89777392e-01 6.73203170e-01
-6.39605224e-01 1.97401464e-01 -1.27851689e+00 1.73397470e+00
-1.94192082e-01 7.42252409e-01 3.82177383e-01 -1.03860319e+00
6.90826416e-01 1.07550018e-01 6.95330620e-01 -3.84066373e-01
2.35013083e-01 6.88794851e-01 2.08145473e-02 -5.66552162e-01
1.04626775e+00 2.31699497e-01 -4.69784826e-01 6.17765009e-01
-5.35824120e-01 -4.67637181e-01 3.30718189e-01 5.27520711e-03
1.07405269e+00 2.20381409e-01 8.54559600e-01 -5.47280431e-01
7.45439231e-01 3.94734055e-01 2.39445582e-01 1.05154181e+00
-1.01661551e+00 -4.28751949e-03 5.79420887e-02 -3.80420625e-01
-4.86316830e-01 -8.97747397e-01 -2.09304597e-02 1.43120909e+00
1.02360511e+00 -4.07468587e-01 -6.82063341e-01 -4.94424462e-01
-2.52048492e-01 6.01675212e-01 -5.93941391e-01 2.40441844e-01
-6.80633962e-01 -6.22651219e-01 3.55593234e-01 1.80059716e-01
4.60817188e-01 -1.35394752e+00 -1.14776218e+00 6.22286573e-02
-4.83575076e-01 -1.11703825e+00 4.34280694e-01 2.55729586e-01
-6.26913905e-01 -1.46948540e+00 -4.90923643e-01 -1.49427354e+00
6.05360091e-01 1.16476858e+00 7.49294996e-01 7.24553615e-02
1.97012216e-01 7.53142357e-01 -8.10980558e-01 -4.39766943e-01
-2.69577742e-01 2.52649903e-01 3.82188678e-01 -8.05996299e-01
4.71884251e-01 -5.56928098e-01 -1.96566880e-01 7.62541831e-01
-3.96869242e-01 -1.80232406e-01 3.39770794e-01 6.35678589e-01
1.61594432e-02 8.19354951e-02 7.29993761e-01 -3.57194960e-01
8.56151104e-01 -8.04316521e-01 -1.84334219e-01 -6.55853152e-02
-5.44072092e-01 -4.77026522e-01 1.97472628e-02 1.06030665e-01
-1.00546098e+00 -2.06346273e-01 -7.17667416e-02 7.84720242e-01
-3.13462138e-01 4.51078862e-01 3.65447090e-03 -3.51406723e-01
6.74611926e-01 -8.29039738e-02 4.13340360e-01 4.58578914e-02
5.58153331e-01 9.39066708e-01 3.11862171e-01 -3.86392176e-01
3.80090952e-01 7.06312180e-01 1.07297756e-01 -1.16556072e+00
-3.74914557e-01 -5.84672570e-01 -6.76320374e-01 -5.44047713e-01
7.42149413e-01 -4.65269327e-01 -5.71636260e-01 6.58839524e-01
-1.00576365e+00 -6.34012520e-01 -3.01952451e-01 5.73423386e-01
-1.10023797e+00 5.75410247e-01 -4.68996435e-01 -1.11963177e+00
2.20835418e-01 -1.49152017e+00 5.64630747e-01 2.74302721e-01
-5.16385674e-01 -9.61076021e-01 2.87806392e-01 4.72303122e-01
4.26374376e-01 -2.54587550e-03 2.71503031e-01 -4.71568525e-01
-3.96457285e-01 -1.21254392e-01 1.24167763e-02 -3.23533475e-01
2.90940881e-01 -3.00450832e-01 -3.08499277e-01 -8.03390890e-02
4.75475937e-03 -5.34994677e-02 2.10065693e-01 2.48332456e-01
-2.52750933e-01 1.71887308e-01 -7.59057760e-01 -7.00271875e-02
9.44748580e-01 9.37833428e-01 6.47133827e-01 1.36920655e+00
2.15532571e-01 1.47781634e+00 1.36049998e+00 -3.61932628e-02
1.16878200e+00 9.29276645e-01 7.04231381e-01 4.91854101e-01
2.47825086e-01 4.84324366e-01 5.03385067e-01 9.30644691e-01
-4.46151942e-01 -2.66224623e-01 -1.15708196e+00 4.72378612e-01
-2.48109221e+00 -9.66747940e-01 -3.67338300e-01 1.69875491e+00
1.40219536e-02 -2.32095256e-01 2.05058321e-01 1.87062949e-01
9.99083281e-01 9.93792042e-02 -3.08173299e-01 -4.53748792e-01
-9.28826034e-02 -7.89697289e-01 4.07528132e-01 9.64517713e-01
-1.22125924e+00 1.31470454e+00 7.15667009e+00 1.79984748e-01
-4.97750700e-01 1.51249033e-03 -2.72098720e-01 3.26205850e-01
2.09045276e-01 3.09614465e-02 -6.31058753e-01 8.84272233e-02
4.56999570e-01 -2.31169298e-01 4.41044033e-01 1.10940588e+00
2.22857937e-01 -6.45935059e-01 -6.62312984e-01 8.45185637e-01
3.81546170e-01 -7.42388248e-01 -3.58341813e-01 3.20763767e-01
5.02421737e-01 2.85855472e-01 -1.02788627e-01 3.82349849e-01
7.37678468e-01 -7.19443381e-01 8.34543407e-01 2.00168356e-01
-5.59306331e-02 -8.31986904e-01 9.82060134e-01 4.82313365e-01
-1.10183764e+00 -4.05586809e-01 -2.38059208e-01 -8.15722466e-01
8.03585112e-01 -2.35916935e-02 -5.12274802e-01 6.49437904e-01
1.00185680e+00 8.59384775e-01 -2.75202040e-02 1.23448622e+00
-3.27287138e-01 -6.17375672e-01 -1.93799809e-01 -6.96362317e-01
4.67282653e-01 -6.05402768e-01 9.31733727e-01 8.91739786e-01
2.11484745e-01 1.44423053e-01 3.44053537e-01 -2.28373274e-01
8.44762623e-01 2.69015789e-01 -1.16552067e+00 3.47263694e-01
6.11351550e-01 8.36463332e-01 -1.19138074e+00 8.12323242e-02
-4.48580533e-01 9.11545932e-01 1.25467539e-01 1.03580058e-01
-6.48484945e-01 -6.96208477e-01 1.02710199e+00 -3.79903823e-01
-1.98755950e-01 -6.86720967e-01 -3.99679840e-01 -7.04637170e-01
-2.20970318e-01 -6.21704221e-01 1.86121076e-01 -6.70342386e-01
-8.17411959e-01 5.21435440e-01 2.63702810e-01 -1.31463945e+00
-5.75903654e-01 -6.10955417e-01 -2.09770482e-02 1.97898090e-01
-1.47347414e+00 -9.58659530e-01 -3.40483189e-01 1.02832869e-01
7.05783069e-01 -2.98237622e-01 1.00604081e+00 1.01312444e-01
-4.62749034e-01 9.62782949e-02 6.91444799e-02 -3.88022840e-01
6.88533127e-01 -8.99375558e-01 4.98033166e-01 7.54352212e-01
-6.31036520e-01 9.81710076e-01 9.72960234e-01 -8.20363641e-01
-1.35258818e+00 -6.34255230e-01 1.09851706e+00 -6.30404115e-01
5.99423170e-01 1.75537113e-02 -2.11834341e-01 9.39437985e-01
2.77004123e-01 -7.34424531e-01 8.00118327e-01 2.18698695e-01
2.59041846e-01 4.39094990e-01 -1.25039852e+00 1.38880646e+00
1.64106202e+00 -6.51471615e-02 -9.79367018e-01 1.78552970e-01
4.14382726e-01 -3.85610491e-01 -3.40618581e-01 4.24860567e-01
9.17795599e-01 -1.15338933e+00 9.38954592e-01 -1.08086310e-01
-1.51355043e-01 -5.52931786e-01 -5.43053865e-01 -1.19445169e+00
-2.84312278e-01 -6.56579971e-01 2.62798011e-01 7.22256362e-01
2.07405537e-01 -1.19336092e+00 9.54202592e-01 4.09986287e-01
-4.54991519e-01 -4.73479688e-01 -1.05887556e+00 -7.87311614e-01
2.89318636e-02 -6.35273039e-01 2.94642746e-01 1.00904417e+00
7.56124318e-01 4.76575643e-01 -2.21724451e-01 -1.31156966e-01
5.27454793e-01 -3.95140141e-01 1.16697514e+00 -1.34047222e+00
6.25279963e-01 -9.07126844e-01 -7.89361596e-01 -1.21950626e+00
4.14913476e-01 -5.97830534e-01 5.35140395e-01 -2.20716834e+00
-4.77492630e-01 -6.30199432e-01 1.90473527e-01 6.52543828e-02
4.58160907e-01 2.01957718e-01 2.54210085e-01 2.97459036e-01
-1.01951778e+00 3.91108274e-01 1.11311436e+00 3.07622761e-01
-4.77190673e-01 2.06310898e-01 -9.49558794e-01 1.30536151e+00
9.62209105e-01 -6.47983328e-02 -5.76977313e-01 -3.31058830e-01
6.25464976e-01 -1.44581363e-01 -1.67426728e-02 -1.19190562e+00
8.48105729e-01 -5.38463950e-01 -7.39874959e-01 -6.13854527e-01
5.55681169e-01 -9.80105102e-01 -3.83279845e-02 6.63054287e-01
3.37096537e-03 3.13802451e-01 -1.93992093e-01 6.71771109e-01
2.12967675e-02 -5.96480966e-01 5.91437161e-01 -2.64113843e-01
-1.45106483e+00 -5.30765772e-01 -1.33099937e+00 -2.40987822e-01
1.79962385e+00 -7.86066413e-01 -2.70155311e-01 -5.84015548e-01
-7.00489342e-01 6.40720487e-01 8.03031564e-01 7.39020228e-01
3.26006532e-01 -1.09527552e+00 -4.30827327e-02 -2.91654199e-01
2.27974549e-01 4.64795753e-02 2.11599633e-01 6.93776727e-01
-7.97780216e-01 7.05666780e-01 -4.58964646e-01 -4.05044079e-01
-1.17316031e+00 2.71338642e-01 4.05991018e-01 2.57335693e-01
-5.44979930e-01 6.97475970e-01 -1.05913982e-01 -1.05648530e+00
5.81195831e-01 2.14844406e-01 -9.47835565e-01 8.63375664e-02
4.43090409e-01 9.74055350e-01 -3.09511840e-01 -1.29345930e+00
-7.22542703e-01 8.56653094e-01 4.37785715e-01 -4.92354065e-01
1.13640213e+00 -1.01360774e+00 -5.20360231e-01 6.81344688e-01
5.08851230e-01 -2.25312281e-02 -6.60237014e-01 1.16671093e-01
3.48981202e-01 -7.71021470e-02 -4.87690777e-01 -3.68774474e-01
-2.17191666e-01 2.46469811e-01 3.91333640e-01 4.62270170e-01
5.26656449e-01 6.57650642e-03 4.10643011e-01 1.01170158e+00
1.20651972e+00 -1.38492882e+00 -3.38299051e-02 1.41474688e+00
9.43324089e-01 -1.17859781e+00 -4.73576523e-02 -8.56663764e-01
-5.27795732e-01 9.60760176e-01 6.64641380e-01 -1.34558260e-01
8.01423311e-01 6.44818097e-02 5.02294540e-01 -2.32339099e-01
-1.56641737e-01 -4.41986978e-01 -5.72250247e-01 1.17631185e+00
-1.79259609e-02 -4.62297127e-02 -6.93248272e-01 2.59567052e-01
-8.47341895e-01 -2.04426438e-01 8.19258630e-01 1.73769069e+00
-9.83425319e-01 -1.41515434e+00 -4.94308561e-01 2.71985363e-02
2.56751403e-02 5.59931099e-01 -5.44966519e-01 9.88553405e-01
2.16175944e-01 1.61669862e+00 -1.38003618e-01 -4.65378344e-01
5.05232573e-01 -4.16271567e-01 2.50609994e-01 -6.54273570e-01
-2.61504084e-01 -6.70917571e-01 4.80458945e-01 -4.94377911e-01
-1.05674362e+00 -1.00241017e+00 -1.35101390e+00 -3.85140181e-01
-3.03423882e-01 5.22139609e-01 9.11230087e-01 1.04438305e+00
3.81747819e-02 2.60959536e-01 2.23763093e-01 -1.39418042e+00
-1.89219657e-02 -4.76885706e-01 -3.30735445e-01 -2.82855779e-01
2.27426872e-01 -1.20829153e+00 -4.37816739e-01 -4.83119696e-01] | [4.829054355621338, 1.0058289766311646] |
3a326d36-7219-4cb7-bb20-4da83115abe1 | chatgpt-as-your-personal-data-scientist | 2305.13657 | null | https://arxiv.org/abs/2305.13657v1 | https://arxiv.org/pdf/2305.13657v1.pdf | ChatGPT as your Personal Data Scientist | The rise of big data has amplified the need for efficient, user-friendly automated machine learning (AutoML) tools. However, the intricacy of understanding domain-specific data and defining prediction tasks necessitates human intervention making the process time-consuming while preventing full automation. Instead, envision an intelligent agent capable of assisting users in conducting AutoML tasks through intuitive, natural conversations without requiring in-depth knowledge of the underlying machine learning (ML) processes. This agent's key challenge is to accurately comprehend the user's prediction goals and, consequently, formulate precise ML tasks, adjust data sets and model parameters accordingly, and articulate results effectively. In this paper, we take a pioneering step towards this ambitious goal by introducing a ChatGPT-based conversational data-science framework to act as a "personal data scientist". Precisely, we utilize Large Language Models (ChatGPT) to build a natural interface between the users and the ML models (Scikit-Learn), which in turn, allows us to approach this ambitious problem with a realistic solution. Our model pivots around four dialogue states: Data Visualization, Task Formulation, Prediction Engineering, and Result Summary and Recommendation. Each state marks a unique conversation phase, impacting the overall user-system interaction. Multiple LLM instances, serving as "micro-agents", ensure a cohesive conversation flow, granting us granular control over the conversation's progression. In summary, we developed an end-to-end system that not only proves the viability of the novel concept of conversational data science but also underscores the potency of LLMs in solving complex tasks. Interestingly, its development spotlighted several critical weaknesses in the current LLMs (ChatGPT) and highlighted substantial opportunities for improvement. | ['Shubhra Kanti Karmaker Santu', 'Alex Knipper', 'Md Mahadi Hassan'] | 2023-05-23 | null | null | null | null | ['data-visualization', 'automl', 'data-visualization'] | ['methodology', 'methodology', 'miscellaneous'] | [-1.39221072e-01 4.07871723e-01 5.42027876e-02 -4.61306334e-01
-4.93360817e-01 -7.30089068e-01 9.30432379e-01 3.08679700e-01
-1.36169598e-01 5.24783432e-01 4.93749887e-01 -6.90152287e-01
-1.37166560e-01 -5.42022765e-01 -1.25165984e-01 -1.89863339e-01
1.27672046e-01 8.88916194e-01 -2.90772110e-01 -4.45868224e-01
3.81448388e-01 2.56332934e-01 -1.51439857e+00 5.71867108e-01
9.87709939e-01 4.12113339e-01 2.95671999e-01 7.56969750e-01
-6.52469933e-01 1.12365687e+00 -6.72824919e-01 -4.91881996e-01
-1.47656605e-01 -2.91005164e-01 -1.05677903e+00 -1.08546438e-02
-5.36064953e-02 2.65013613e-02 3.80328983e-01 4.48794335e-01
2.16678128e-01 4.81609590e-02 3.13249677e-01 -1.26910210e+00
-2.69971997e-01 4.79668975e-01 4.56600413e-02 -2.40754962e-01
6.65633857e-01 7.16106772e-01 9.30914581e-01 -6.38407052e-01
5.22754490e-01 1.39882493e+00 6.40557706e-01 6.46992445e-01
-1.25064528e+00 -2.51355171e-01 2.24591374e-01 -8.63643065e-02
-8.14601302e-01 -5.90693831e-01 4.62095737e-01 -9.64737713e-01
9.90971863e-01 7.73252726e-01 9.15634811e-01 1.11133516e+00
-1.62750259e-02 7.62195230e-01 1.02593243e+00 -4.16930825e-01
4.29659337e-01 6.29535735e-01 4.45731223e-01 4.93891507e-01
-1.60629943e-01 -2.01326534e-01 -7.96421349e-01 -2.94555426e-01
3.75182420e-01 -1.48962659e-03 -2.11728409e-01 -2.66959816e-01
-1.14883924e+00 6.14611268e-01 -1.24724954e-01 4.61150378e-01
-3.98961514e-01 -4.10681963e-01 4.15881276e-01 5.49425602e-01
5.39521217e-01 8.51679504e-01 -5.56986034e-01 -8.83788288e-01
-3.66662323e-01 3.32208097e-01 1.51224661e+00 6.75685585e-01
6.38217211e-01 -4.95853901e-01 -1.96267441e-01 8.82446468e-01
3.15936387e-01 -5.18727340e-02 2.76595086e-01 -9.71739709e-01
4.39291596e-01 1.16599905e+00 3.84064227e-01 -8.82126212e-01
-5.64389944e-01 -3.00866187e-01 -5.75792551e-01 2.43727416e-01
7.55894840e-01 -3.34998965e-01 -1.83132157e-01 1.52032268e+00
4.65510875e-01 -4.10211861e-01 1.79857165e-01 8.14821601e-01
6.55954540e-01 6.11482143e-01 3.92538339e-01 -1.34292632e-01
1.50358236e+00 -8.51028502e-01 -6.37999237e-01 -3.67952973e-01
1.01039863e+00 -7.06853688e-01 1.75982141e+00 5.12374818e-01
-9.86876488e-01 -4.24455047e-01 -7.39309013e-01 -2.07527116e-01
-4.77895826e-01 -2.37620801e-01 8.13851714e-01 5.58517396e-01
-9.21007872e-01 4.18498307e-01 -7.15765178e-01 -5.88491857e-01
1.04067564e-01 1.28503576e-01 -2.35517055e-01 7.82167315e-02
-1.04885471e+00 1.14117360e+00 4.13758680e-02 1.61716148e-01
-3.32900763e-01 -9.59700108e-01 -5.67161798e-01 6.98905960e-02
4.40011054e-01 -7.58598208e-01 1.61289716e+00 -6.61470950e-01
-1.63363636e+00 7.81054020e-01 -8.20548534e-02 -1.74837798e-01
7.88983226e-01 -2.56628841e-01 -6.36519641e-02 -5.22305906e-01
-2.19587147e-01 3.13461512e-01 1.94041401e-01 -1.06971300e+00
-5.41720569e-01 -2.75038987e-01 2.33480424e-01 3.78096938e-01
-2.92667150e-01 1.21445268e-01 -3.64329726e-01 -4.20972817e-02
-5.08581102e-01 -6.53866112e-01 -3.46373260e-01 -2.22226545e-01
-2.64584869e-01 -5.89971006e-01 7.10562706e-01 -4.01230782e-01
1.45891523e+00 -2.15483379e+00 6.96260715e-03 7.87321664e-03
6.51848972e-01 4.45611179e-01 7.55542666e-02 9.69754279e-01
3.11858296e-01 2.15573564e-01 7.20561445e-02 -6.96248531e-01
3.06270152e-01 1.05032571e-01 -9.60985571e-02 -1.16237842e-01
6.17949590e-02 9.48507071e-01 -9.62845802e-01 -2.22961739e-01
4.86481100e-01 4.29912001e-01 -6.82021260e-01 7.00427413e-01
-6.57354116e-01 6.74929380e-01 -5.08556366e-01 1.43191770e-01
2.05882326e-01 -5.90598404e-01 5.73388159e-01 -1.87463255e-03
-6.20548368e-01 6.63473248e-01 -1.03516388e+00 1.45222771e+00
-7.33405292e-01 6.89695835e-01 4.29259062e-01 -7.08538055e-01
1.08461142e+00 2.39552900e-01 4.12047416e-01 -4.96476442e-01
-1.99380934e-01 -6.95827454e-02 -1.27601445e-01 -9.47684348e-01
3.35477889e-01 -2.00368669e-02 -5.75617850e-02 1.00893342e+00
-2.79898703e-01 -2.20255136e-01 -1.62819214e-02 2.13882327e-01
9.41990137e-01 6.51309825e-03 3.41254592e-01 -1.43092304e-01
3.28651816e-01 2.14773625e-01 2.34729916e-01 7.63402939e-01
7.54950475e-03 8.45190957e-02 7.98559844e-01 -7.76760459e-01
-7.11009741e-01 -6.61479712e-01 2.29448855e-01 1.36864007e+00
-4.15458202e-01 -8.34561348e-01 -7.66212404e-01 -4.73831415e-01
-1.83981713e-02 9.25672650e-01 -3.41294587e-01 1.05350316e-01
-3.96029651e-01 -5.57107329e-01 2.34883159e-01 -1.78064540e-01
4.05716181e-01 -1.14052737e+00 -6.89819455e-01 2.84843564e-01
-4.58378106e-01 -9.51898396e-01 -2.31095240e-01 -1.06726371e-01
-4.55502242e-01 -9.66082156e-01 -1.77558064e-02 -3.03929031e-01
2.93194443e-01 2.25315705e-01 1.45197809e+00 5.62400334e-02
-2.43703932e-01 4.26582813e-01 -1.79509893e-01 -6.88064754e-01
-9.26101923e-01 2.33544752e-01 -1.77399993e-01 -9.37605351e-02
6.44622862e-01 -7.16026068e-01 -5.13397634e-01 3.17415535e-01
-5.20704925e-01 7.16363728e-01 4.18668628e-01 3.47319812e-01
-2.24942580e-01 -4.78982985e-01 7.62779355e-01 -1.15269971e+00
1.34282386e+00 -4.86768186e-01 -3.20128769e-01 4.18093026e-01
-8.24921131e-01 3.06908693e-02 6.94135785e-01 -2.12455451e-01
-1.03261185e+00 -1.71196252e-01 -7.79147074e-02 2.41723016e-01
-2.49619588e-01 6.81985497e-01 -2.60377169e-01 4.84257817e-01
6.57682359e-01 -3.66196446e-02 5.67062318e-01 -6.94763005e-01
5.79857349e-01 1.08869183e+00 3.52024257e-01 -6.96792841e-01
3.40029657e-01 -2.16082875e-02 -7.28447855e-01 -8.01990151e-01
-6.14951491e-01 -2.99899012e-01 -5.13072371e-01 -4.07764196e-01
6.90659523e-01 -6.55663610e-01 -1.61911654e+00 1.93142176e-01
-1.23618734e+00 -6.55647576e-01 -2.63633072e-01 7.16006681e-02
-3.78010154e-01 1.62410900e-01 -3.19269478e-01 -1.07164717e+00
-5.13530016e-01 -1.03531671e+00 6.21998429e-01 1.57514155e-01
-1.06525517e+00 -1.20226836e+00 1.00663990e-01 9.59022641e-01
7.74220526e-01 9.38975960e-02 1.07982874e+00 -8.43956232e-01
-4.92210776e-01 -2.53450125e-01 -1.24186352e-01 6.04891442e-02
2.33989730e-01 1.21817678e-01 -1.26777744e+00 8.56815577e-02
-1.62296332e-02 -2.71198899e-01 -1.55684790e-02 -1.53197616e-01
9.08186376e-01 -5.77214360e-01 -1.64741471e-01 4.98345047e-02
6.26012146e-01 9.55188796e-02 1.46430030e-01 7.02394322e-02
6.01079166e-01 1.10380578e+00 4.80968118e-01 6.54430032e-01
8.63145351e-01 7.48853803e-01 2.03974564e-02 -1.90314397e-01
2.05183581e-01 -3.73060733e-01 2.47603819e-01 7.25616276e-01
4.94738929e-02 -1.24075681e-01 -1.17707229e+00 8.06046799e-02
-2.20721436e+00 -8.15081716e-01 -2.37654105e-01 2.08657575e+00
1.08565319e+00 2.26646692e-01 2.68520236e-01 -2.47280881e-01
1.98301375e-01 2.01941133e-01 -4.39668477e-01 -7.84716845e-01
5.06241620e-01 -1.84378400e-01 -3.43292534e-01 9.16939139e-01
-5.61144471e-01 8.69371772e-01 5.79351187e+00 4.25847352e-01
-1.09352803e+00 -1.00911796e-01 6.00461662e-01 -1.40484452e-01
-4.57602233e-01 -4.73700352e-02 -5.91277421e-01 3.99228394e-01
1.16710103e+00 -4.23043430e-01 6.91809654e-01 8.27396989e-01
9.81145561e-01 -2.56145615e-02 -1.51083493e+00 8.87864769e-01
-4.26342875e-01 -1.72033417e+00 -1.18026525e-01 9.57718417e-02
-7.94458389e-02 -4.56444221e-03 -2.25757435e-01 5.31863451e-01
6.80192053e-01 -1.18425500e+00 3.77142459e-01 7.54886031e-01
3.82289141e-01 -4.48668599e-01 3.41654658e-01 9.74400818e-01
-6.35873973e-01 -6.84980825e-02 1.84232265e-01 -6.91563666e-01
1.39163882e-01 5.30654728e-01 -1.50593519e+00 1.15948185e-01
3.10348511e-01 5.43892324e-01 -2.13715479e-01 4.58608568e-01
2.24324480e-01 4.66226071e-01 -1.14392221e-01 -3.19922656e-01
1.02164403e-01 -3.15207779e-01 4.17933017e-01 1.57370043e+00
-2.04696923e-01 2.42421150e-01 4.33721185e-01 9.73712385e-01
1.90178216e-01 1.78179368e-01 -6.55907929e-01 -4.52739030e-01
6.63559675e-01 1.41672206e+00 -1.60862342e-01 -1.82919294e-01
-3.01638305e-01 3.79878551e-01 5.30326307e-01 1.82247564e-01
-5.98936379e-02 3.95395234e-02 1.13300455e+00 1.37304366e-01
-5.38896739e-01 -4.13454175e-01 -7.91827977e-01 -1.05950987e+00
-2.29627546e-02 -1.52729535e+00 2.39865720e-01 -5.75328410e-01
-1.26296544e+00 3.69803369e-01 -2.79102266e-01 -8.59652698e-01
-6.15023136e-01 -2.41401866e-01 -7.83613741e-01 1.14505601e+00
-9.63667214e-01 -1.09306765e+00 -5.91186285e-01 3.03979993e-01
7.10792661e-01 5.15072718e-02 1.19323802e+00 5.34220152e-02
-6.45991981e-01 2.56858647e-01 -1.38634697e-01 -1.49429187e-01
6.83800876e-01 -1.13250697e+00 5.41863263e-01 2.18741015e-01
-1.50511682e-01 9.67474580e-01 9.92838144e-01 -4.69495952e-01
-1.66167808e+00 -5.94269753e-01 1.25252056e+00 -1.05597913e+00
8.15655112e-01 -8.72454286e-01 -1.04935575e+00 5.37175953e-01
2.08268896e-01 -6.96063876e-01 1.04365611e+00 7.55089045e-01
-1.42503843e-01 -6.39970601e-02 -8.99994552e-01 8.18311632e-01
8.11578393e-01 -5.54607809e-01 -4.69373494e-01 5.50456464e-01
7.49831140e-01 -2.59870172e-01 -8.19141805e-01 4.75942008e-02
5.77091455e-01 -1.11319959e+00 7.56326139e-01 -9.97015357e-01
3.90248746e-01 1.17915958e-01 2.36048952e-01 -1.26381743e+00
4.54574227e-02 -1.29449964e+00 6.83667138e-03 1.29771721e+00
6.13371789e-01 -6.51575089e-01 6.32758856e-01 1.67764676e+00
-1.91801459e-01 -7.95603573e-01 -3.84518862e-01 -1.13477170e-01
-1.74120829e-01 -8.59480321e-01 7.16892123e-01 1.10481119e+00
7.87362933e-01 6.28139198e-01 -2.68456399e-01 -9.57760140e-02
3.44637215e-01 4.67063114e-02 1.41069019e+00 -1.55361938e+00
-3.76524538e-01 -5.92530966e-01 2.59903848e-01 -1.07060456e+00
-2.39237711e-01 -7.42653668e-01 -2.26808369e-01 -1.62580037e+00
2.79489742e-03 -6.97358549e-01 1.93353564e-01 5.99597096e-01
-1.23825803e-01 -5.40105939e-01 3.55925769e-01 3.58219892e-01
-6.38156772e-01 1.00931779e-01 1.18606818e+00 1.84390754e-01
-7.56371617e-01 2.48744518e-01 -1.17315400e+00 6.71063066e-01
6.55342579e-01 6.71552122e-02 -6.08969748e-01 -2.67902434e-01
4.38499451e-01 7.07503334e-02 3.19481671e-01 -6.10613346e-01
4.35855240e-01 -5.98656595e-01 -1.36353239e-01 -2.11603507e-01
2.82794833e-01 -5.53093970e-01 1.61786720e-01 2.51106024e-01
-7.78232157e-01 -1.52003020e-01 1.29300371e-01 2.16583669e-01
7.45390877e-02 1.57137528e-01 4.21974391e-01 -1.87553316e-01
-4.40552473e-01 -1.21871509e-01 -5.90116739e-01 -1.12423636e-02
9.25297201e-01 -1.34807527e-01 -5.38952947e-01 -7.34677494e-01
-8.60990226e-01 6.43311381e-01 2.95074552e-01 6.35917723e-01
1.64326265e-01 -5.80564082e-01 -6.72132969e-01 4.42072541e-01
1.59570366e-01 1.41538396e-01 1.43027440e-01 7.55451679e-01
-2.63434529e-01 4.47791010e-01 7.31936991e-02 -4.61518764e-01
-1.33511686e+00 2.49760136e-01 3.61498326e-01 -3.68849099e-01
-6.82395875e-01 5.97180784e-01 4.62216437e-02 -8.20437014e-01
4.33107227e-01 -2.02347919e-01 -3.40799630e-01 3.11408073e-01
7.80110300e-01 2.52377629e-01 8.23511183e-02 2.84836531e-01
-1.03209756e-01 -1.61120743e-01 -3.17733347e-01 -9.31974575e-02
1.57606196e+00 -3.25715452e-01 -1.53383076e-01 8.18978727e-01
7.21050620e-01 -3.97277549e-02 -1.19721544e+00 -3.11263353e-01
3.07538152e-01 -2.78863221e-01 -3.76294643e-01 -1.38455117e+00
-1.28237367e-01 9.69093025e-01 4.31002155e-02 7.65498400e-01
5.24176240e-01 9.51271653e-02 4.71207589e-01 5.08974373e-01
1.78789631e-01 -1.00926161e+00 -1.88050547e-03 6.55889034e-01
1.08028293e+00 -1.31000173e+00 -3.29926789e-01 -2.05602929e-01
-9.68389273e-01 1.26561546e+00 6.90646112e-01 7.21734822e-01
2.78845817e-01 4.58610237e-01 5.42504311e-01 -4.30152714e-01
-1.39735138e+00 1.85653180e-01 -1.38340229e-02 5.21198332e-01
8.60136151e-01 1.80850357e-01 -1.56877950e-01 6.25437617e-01
-4.89063501e-01 3.16638887e-01 2.91723639e-01 7.35761583e-01
-5.88674486e-01 -1.16070211e+00 -2.72235364e-01 2.90693849e-01
1.53979063e-01 -1.62859261e-02 -7.70152152e-01 6.80498004e-01
-1.78904384e-01 1.26024568e+00 -2.20263693e-02 -4.23547477e-01
4.93262500e-01 4.45965499e-01 -2.26246491e-01 -7.82299042e-01
-1.01627696e+00 -2.08958745e-01 6.51636899e-01 -5.15342057e-01
1.11588813e-01 -4.77868170e-01 -1.06482208e+00 -8.30926895e-01
2.58799255e-01 5.57657063e-01 8.27485085e-01 1.10392904e+00
8.90698195e-01 1.75136939e-01 6.12312675e-01 -6.22529387e-01
-6.88957870e-01 -1.06639695e+00 1.67883039e-01 3.35897505e-01
2.35765740e-01 -1.10112526e-01 -2.43864641e-01 -6.48777187e-02] | [12.59626293182373, 7.842702865600586] |
32457427-6acf-419a-9144-4f8dae229297 | prophetnet-x-large-scale-pre-training-models | 2104.08006 | null | https://arxiv.org/abs/2104.08006v2 | https://arxiv.org/pdf/2104.08006v2.pdf | ProphetNet-X: Large-Scale Pre-training Models for English, Chinese, Multi-lingual, Dialog, and Code Generation | Now, the pre-training technique is ubiquitous in natural language processing field. ProphetNet is a pre-training based natural language generation method which shows powerful performance on English text summarization and question generation tasks. In this paper, we extend ProphetNet into other domains and languages, and present the ProphetNet family pre-training models, named ProphetNet-X, where X can be English, Chinese, Multi-lingual, and so on. We pre-train a cross-lingual generation model ProphetNet-Multi, a Chinese generation model ProphetNet-Zh, two open-domain dialog generation models ProphetNet-Dialog-En and ProphetNet-Dialog-Zh. And also, we provide a PLG (Programming Language Generation) model ProphetNet-Code to show the generation performance besides NLG (Natural Language Generation) tasks. In our experiments, ProphetNet-X models achieve new state-of-the-art performance on 10 benchmarks. All the models of ProphetNet-X share the same model structure, which allows users to easily switch between different models. We make the code and models publicly available, and we will keep updating more pre-training models and finetuning scripts. | ['Nan Duan', 'Houqiang Li', 'Ruofei Zhang', 'Jiusheng Chen', 'Daxin Jiang', 'Biao Cheng', 'Bartuer Zhou', 'Bolun Yao', 'Can Xu', 'Yu Yan', 'Yeyun Gong', 'Weizhen Qi'] | 2021-04-16 | null | https://aclanthology.org/2021.acl-demo.28 | https://aclanthology.org/2021.acl-demo.28.pdf | acl-2021-5 | ['open-domain-dialog'] | ['natural-language-processing'] | [-1.89115494e-01 5.96898496e-01 -2.78056189e-02 -2.59872049e-01
-9.44380283e-01 -5.64273894e-01 1.01442266e+00 1.36321068e-01
-1.95654362e-01 1.36544418e+00 8.76942873e-01 -4.55700636e-01
4.16493893e-01 -1.09499264e+00 -2.94360310e-01 -2.31722206e-01
3.29388767e-01 9.78752017e-01 2.03948557e-01 -8.17394137e-01
2.02805534e-01 -4.37855303e-01 -9.33630645e-01 4.95663851e-01
1.47514820e+00 3.38781297e-01 3.33583772e-01 1.02789330e+00
-5.06099463e-01 9.70872462e-01 -9.06246901e-01 -6.41047060e-01
-1.77276239e-01 -8.95333767e-01 -1.32182598e+00 -4.26824272e-01
7.07467180e-03 -4.12974000e-01 -5.39318137e-02 8.39688122e-01
9.87883091e-01 1.32173255e-01 7.39768147e-01 -1.35660243e+00
-1.18325043e+00 1.30377579e+00 -9.73893106e-02 -2.94625491e-01
8.47413301e-01 5.47541559e-01 9.79790688e-01 -6.14304185e-01
9.52355921e-01 1.58854210e+00 4.94433254e-01 9.56301093e-01
-7.34144747e-01 -6.69285119e-01 -5.74304238e-02 -1.56665266e-01
-1.09747517e+00 -2.68326193e-01 5.08008897e-01 -2.74411172e-01
1.08836067e+00 1.55171186e-01 3.74582529e-01 1.21227300e+00
5.80242872e-01 9.59894478e-01 8.91147971e-01 -3.94287765e-01
1.55030921e-01 1.71697456e-02 1.76190108e-01 6.32073760e-01
1.39546189e-02 -9.44571942e-02 -3.97529691e-01 -3.27286810e-01
4.79727894e-01 -6.27979875e-01 -1.99509963e-01 6.60967648e-01
-1.45400774e+00 1.09970272e+00 2.61621714e-01 4.33611691e-01
-1.61591023e-01 -1.67412609e-02 6.36105180e-01 4.15892929e-01
5.48789859e-01 8.83774102e-01 -3.79138529e-01 -3.00404847e-01
-6.99057639e-01 8.01893890e-01 1.19414151e+00 1.64698017e+00
7.12555289e-01 2.15674102e-01 -8.74761462e-01 8.46461892e-01
3.49208303e-02 3.51518720e-01 1.24197853e+00 -7.38450408e-01
7.04301655e-01 6.71059132e-01 -1.96433952e-03 -4.69277650e-01
-3.47395331e-01 4.50190157e-02 -1.11617041e+00 -5.46456397e-01
-3.19824181e-02 -1.05834794e+00 -8.25207591e-01 1.58874536e+00
1.12281121e-01 -3.25647056e-01 6.37233675e-01 5.71726322e-01
1.59500146e+00 1.29486156e+00 1.98926687e-01 -2.73447633e-01
1.43523133e+00 -1.42966187e+00 -6.93228006e-01 -1.42353371e-01
5.46300650e-01 -9.08169389e-01 1.08328581e+00 1.44674554e-01
-1.23227894e+00 -7.46152759e-01 -7.58493483e-01 -4.23632592e-01
-5.55824339e-01 1.26691148e-01 4.58772749e-01 1.09182321e-01
-1.15032113e+00 4.44804907e-01 -4.82096940e-01 -5.46803832e-01
-1.81452975e-01 -1.27857730e-01 -1.92830395e-02 1.29200116e-01
-1.67662871e+00 8.19073915e-01 1.11764228e+00 -2.85653859e-01
-7.84890234e-01 -5.27589321e-01 -1.03475165e+00 -1.58065953e-03
1.95463553e-01 -1.28171098e+00 1.81052911e+00 -4.53195810e-01
-1.76133311e+00 7.01228321e-01 -5.89992516e-02 -5.26869595e-01
4.51412082e-01 -1.53587461e-01 -5.68570733e-01 -3.78726423e-02
3.98058742e-01 8.98086250e-01 4.92395788e-01 -8.66475999e-01
-4.63269591e-01 3.35934252e-01 1.77797273e-01 3.43050808e-01
4.55672517e-02 6.91187009e-02 -1.01339392e-01 -7.33767450e-01
-6.42986655e-01 -6.04199350e-01 -3.97251964e-01 -7.20476091e-01
-8.77367318e-01 -8.71826530e-01 6.17280304e-01 -7.02447534e-01
1.35989952e+00 -1.49111271e+00 -1.39241695e-01 -5.59698045e-01
9.85358804e-02 6.05425835e-01 -4.60120171e-01 1.14481497e+00
2.51891136e-01 2.74853855e-01 -2.88717270e-01 -2.85533428e-01
3.81516367e-01 1.72629774e-01 -5.72253346e-01 -5.22110999e-01
1.76311076e-01 1.40171921e+00 -1.20913148e+00 -8.09854448e-01
-4.15285341e-02 -2.28931695e-01 -7.77265251e-01 7.01618254e-01
-1.07241762e+00 3.18106443e-01 -7.67062962e-01 3.66739631e-01
5.64837873e-01 -2.24921882e-01 6.24854863e-02 2.77967393e-01
-3.04464072e-01 6.99301362e-01 -6.89998448e-01 1.99599648e+00
-5.37007093e-01 2.58231968e-01 -2.04687312e-01 -4.23521250e-01
1.20273697e+00 6.07968688e-01 -2.43334919e-02 -3.55742574e-01
2.48662382e-01 2.35793799e-01 -1.32806763e-01 -6.60243988e-01
1.30704510e+00 -2.00947031e-01 -5.71356714e-01 9.67261076e-01
5.04107893e-01 -7.35706329e-01 7.27980733e-01 5.61521173e-01
6.28712296e-01 3.06806583e-02 7.59456098e-01 -4.36422169e-01
6.06533945e-01 2.19073877e-01 2.96843916e-01 7.83018410e-01
1.31387889e-01 5.79540610e-01 4.85840529e-01 2.63418369e-02
-6.18599772e-01 -1.01247156e+00 2.62074947e-01 1.37603080e+00
7.02959150e-02 -1.00375950e+00 -9.60095286e-01 -6.25384688e-01
-3.23379040e-01 1.30605721e+00 -1.73144430e-01 -1.97726771e-01
-7.56181479e-01 -6.95111930e-01 8.90092015e-01 3.21129650e-01
1.07971907e+00 -1.81532931e+00 -1.13913417e-01 5.89140058e-01
-6.42693818e-01 -9.39348161e-01 -8.81001115e-01 -4.78122085e-01
-4.91580099e-01 -7.30846405e-01 -7.69190311e-01 -1.34162009e+00
3.54181468e-01 -2.69157976e-01 1.65603292e+00 2.06437632e-01
1.24114372e-01 1.55931413e-01 -7.70829380e-01 -5.73837876e-01
-1.01247799e+00 7.85331547e-01 -3.08630705e-01 -5.63304305e-01
1.51187554e-01 -5.26226044e-01 -2.15158790e-01 -1.03750918e-02
-9.01215076e-01 8.21668267e-01 3.50904286e-01 8.35953414e-01
2.54814327e-01 -3.31109554e-01 1.07294381e+00 -1.14700866e+00
1.65800512e+00 -5.50586104e-01 -3.84010732e-01 4.56137806e-01
-4.61864591e-01 1.88400492e-01 1.04562402e+00 -2.06041023e-01
-1.43633580e+00 -5.78635216e-01 -6.58528805e-01 3.75671834e-01
-5.83498441e-02 1.06376410e+00 -2.42870778e-01 6.84615910e-01
8.11875045e-01 4.57517773e-01 -4.27706331e-01 -4.91957366e-01
8.92849147e-01 9.01864767e-01 8.53870511e-01 -7.85601437e-01
7.65177965e-01 -4.44156826e-01 -7.86728680e-01 -7.08073556e-01
-7.30130374e-01 -6.96942508e-02 -3.81584138e-01 1.30432695e-01
1.16731691e+00 -9.01295185e-01 -2.97173142e-01 5.29300749e-01
-1.75681305e+00 -6.36650980e-01 -3.91947806e-01 3.89882438e-02
-5.40120721e-01 2.01000497e-01 -9.02264476e-01 -4.63330299e-01
-1.23816597e+00 -6.88571870e-01 1.18792021e+00 6.53406024e-01
-2.06642300e-01 -1.34414148e+00 5.06714702e-01 1.34488449e-01
5.92786729e-01 2.57348716e-01 8.68017972e-01 -8.41215014e-01
-3.89381140e-01 2.64825404e-01 -5.23284860e-02 2.27567330e-01
1.71755940e-01 -1.85091153e-01 -4.30180699e-01 -1.28160954e-01
-1.19985826e-01 -6.23941064e-01 7.69363225e-01 1.44152537e-01
8.21588218e-01 -9.47134137e-01 -1.75037891e-01 3.62689704e-01
1.01402438e+00 4.15099189e-02 6.11554921e-01 -7.69668519e-02
5.60024500e-01 3.67210001e-01 6.35912478e-01 3.48474473e-01
1.00849676e+00 4.00408298e-01 -1.55893475e-01 -8.24546739e-02
-1.39174387e-01 -8.24775398e-01 5.29223502e-01 1.43448365e+00
1.40634477e-01 -7.48214662e-01 -6.77630901e-01 5.64467251e-01
-1.98395765e+00 -1.08694494e+00 -2.82531917e-01 1.46336663e+00
1.44941580e+00 -1.05521716e-01 1.01359554e-01 -5.59028208e-01
7.43466973e-01 3.40090901e-01 -3.46393079e-01 -8.49195838e-01
-1.23004250e-01 2.62817323e-01 -8.94388035e-02 8.57794464e-01
-7.77699888e-01 1.51615071e+00 6.14141560e+00 9.36860025e-01
-1.02603161e+00 6.14333265e-02 3.76979232e-01 4.21939284e-01
-4.75503564e-01 1.10941827e-01 -1.04224384e+00 4.71760035e-01
7.82697499e-01 -1.35308230e+00 2.29600832e-01 8.53625476e-01
1.81239456e-01 1.14327423e-01 -1.09248042e+00 7.32431471e-01
2.59232342e-01 -1.48183095e+00 4.65252787e-01 -4.31071758e-01
9.63037789e-01 1.36465922e-01 -6.00686133e-01 1.06303120e+00
1.16132522e+00 -1.02091336e+00 4.98291701e-01 2.46510282e-01
8.60585332e-01 -5.03571689e-01 7.54903495e-01 8.86856735e-01
-1.17053437e+00 2.82382816e-01 -7.17360497e-01 -2.18774915e-01
9.25559342e-01 4.91882443e-01 -1.22366488e+00 1.23199415e+00
-6.43109232e-02 7.38454938e-01 -6.55700326e-01 7.34404624e-01
-7.72881150e-01 6.69796884e-01 1.45731032e-01 -4.36491162e-01
3.21072191e-01 -1.91693991e-01 4.74174052e-01 1.48034811e+00
4.35607165e-01 3.27236563e-01 5.43325067e-01 1.04422188e+00
-3.19274902e-01 2.60185331e-01 -3.63122523e-01 -8.43869597e-02
3.28404367e-01 1.36487651e+00 -8.24667811e-02 -9.20425713e-01
-5.20112962e-02 8.74579966e-01 3.01714186e-02 2.46371984e-01
-6.89796686e-01 -7.36192167e-01 1.36725813e-01 -5.73543794e-02
-1.47709996e-01 -2.21285731e-01 1.04817174e-01 -1.43889797e+00
-4.12432164e-01 -1.14014006e+00 5.87787509e-01 -1.18988419e+00
-1.59664583e+00 1.12557638e+00 3.82488608e-01 -1.04289687e+00
-1.21782112e+00 -1.58328563e-01 -1.13480163e+00 1.03949916e+00
-1.36263430e+00 -1.40514290e+00 -4.16596562e-01 5.50965786e-01
9.74207819e-01 -4.57948267e-01 9.09865081e-01 2.63133198e-02
-1.88749135e-01 5.37210047e-01 -1.49634987e-01 3.52181464e-01
8.56191754e-01 -1.39520025e+00 1.04293072e+00 6.99371219e-01
-2.18522593e-01 6.83353603e-01 6.82102025e-01 -7.73590386e-01
-8.61062884e-01 -1.51192093e+00 1.47157335e+00 -2.34313175e-01
5.88683486e-01 -4.10446137e-01 -7.29499757e-01 8.33465338e-01
1.18650413e+00 -9.06029642e-01 4.20163482e-01 -2.39362791e-01
1.60768405e-01 3.94221604e-01 -9.98518288e-01 8.63610446e-01
9.44016159e-01 -2.23620936e-01 -1.02065206e+00 7.20252931e-01
1.44043505e+00 -6.57718837e-01 -6.36981964e-01 1.12130865e-01
1.23722948e-01 -7.81558573e-01 3.54846746e-01 -5.67740083e-01
8.42970967e-01 -2.21721530e-01 2.87138313e-01 -1.80856574e+00
-1.23648494e-01 -1.22641146e+00 2.16373116e-01 1.75546539e+00
5.95453560e-01 -8.32212925e-01 3.73017192e-01 3.16421121e-01
-5.68895042e-01 -3.92157823e-01 -5.45232296e-01 -7.04864800e-01
6.20443106e-01 1.58112999e-02 1.04710567e+00 9.00202870e-01
3.85421962e-01 1.25104141e+00 -4.59000826e-01 -5.66675127e-01
-2.20310297e-02 3.82456660e-01 1.22000110e+00 -8.81071925e-01
-5.81873596e-01 -4.08253670e-01 4.33840632e-01 -1.58488655e+00
2.51081854e-01 -1.19056606e+00 2.21195444e-01 -2.00046396e+00
-1.65880620e-02 -1.17325358e-01 5.50201297e-01 5.96817374e-01
-4.98508185e-01 -2.64309973e-01 3.16051990e-01 -9.04065557e-03
-5.82080901e-01 1.04509795e+00 1.69716120e+00 -2.44765148e-01
-5.47925651e-01 -7.68786743e-02 -1.17386067e+00 4.09058988e-01
9.36405480e-01 -2.79177099e-01 -6.67455733e-01 -6.98242307e-01
-7.81674162e-02 2.52202272e-01 -1.46548957e-01 -8.89004171e-01
3.27165097e-01 -2.51886815e-01 -1.76396668e-01 -7.06923664e-01
-9.82159078e-02 2.96251446e-01 -6.74136728e-02 4.34497565e-01
-6.17176652e-01 3.69922817e-01 1.78409457e-01 -1.83728501e-01
-3.92177671e-01 -5.22056341e-01 4.18329954e-01 -4.01440829e-01
-4.67775196e-01 3.80072206e-01 -4.53597248e-01 6.78917646e-01
5.35699487e-01 3.23770165e-01 -8.05274725e-01 -9.07266378e-01
-2.42610708e-01 7.78516054e-01 2.50271022e-01 6.00854814e-01
4.34348851e-01 -1.38995099e+00 -1.26541603e+00 -9.85168070e-02
6.34207278e-02 2.34846860e-01 1.33190468e-01 2.72290379e-01
-8.56791675e-01 5.85822105e-01 -7.32531548e-02 -5.92389815e-02
-9.25407946e-01 3.33102435e-01 1.54924408e-01 -9.89233255e-01
-3.99816632e-01 5.66814065e-01 3.51898819e-01 -9.05843318e-01
-3.14789504e-01 -4.21764493e-01 -2.89605916e-01 -2.62583256e-01
7.16385365e-01 2.20856220e-01 -2.49332860e-01 -2.35803992e-01
9.19763520e-02 2.91474648e-02 -1.34805992e-01 -3.99597764e-01
9.44312096e-01 9.29462016e-02 -4.57296401e-01 1.31881297e-01
8.16687346e-01 -2.59740613e-02 -5.16414285e-01 -1.90460622e-01
-1.32353082e-01 2.63511866e-01 -6.60174429e-01 -1.13242304e+00
-5.40509343e-01 6.73367202e-01 -1.90178245e-01 2.92609066e-01
1.04133189e+00 -4.73356573e-03 1.27730906e+00 4.84248936e-01
3.81994367e-01 -1.03196287e+00 1.59659043e-01 1.20803344e+00
1.43032146e+00 -8.13618541e-01 1.06150005e-02 -2.53715396e-01
-9.48783636e-01 8.63158047e-01 9.38694119e-01 -1.05359510e-01
4.72306490e-01 -3.78493741e-02 1.35117024e-01 2.78687235e-02
-1.20326209e+00 -9.17022228e-02 1.05193608e-01 6.64183140e-01
7.97851384e-01 2.66518801e-01 -8.79115403e-01 1.11598027e+00
-1.25256324e+00 1.94478035e-01 7.72418916e-01 7.28686869e-01
-4.56606090e-01 -1.40800023e+00 -7.03316182e-02 3.78671199e-01
-2.86219455e-02 -4.46606070e-01 -7.69483030e-01 8.02949786e-01
2.52072215e-02 1.24311745e+00 -2.20887259e-01 -3.45459133e-01
3.00690770e-01 3.70265208e-02 1.32403284e-01 -1.14993763e+00
-9.97702301e-01 -2.21657112e-01 4.81946170e-01 1.69184133e-02
-8.36728662e-02 -1.87318280e-01 -1.31672025e+00 -5.47325671e-01
-1.66367203e-01 7.33173966e-01 9.08846110e-02 8.31978381e-01
4.11576807e-01 4.57182944e-01 3.33062202e-01 -5.81116915e-01
-5.79411209e-01 -1.58419859e+00 -3.19020897e-01 4.24226373e-02
-6.63039610e-02 -3.22924927e-02 7.46639147e-02 2.92837620e-01] | [11.8883056640625, 9.054323196411133] |
2b99184f-dcc6-4a11-89df-45c2a071006c | learning-joint-2d-3d-diffusion-models-for | 2305.12347 | null | https://arxiv.org/abs/2305.12347v2 | https://arxiv.org/pdf/2305.12347v2.pdf | Learning Joint 2D & 3D Diffusion Models for Complete Molecule Generation | Designing new molecules is essential for drug discovery and material science. Recently, deep generative models that aim to model molecule distribution have made promising progress in narrowing down the chemical research space and generating high-fidelity molecules. However, current generative models only focus on modeling either 2D bonding graphs or 3D geometries, which are two complementary descriptors for molecules. The lack of ability to jointly model both limits the improvement of generation quality and further downstream applications. In this paper, we propose a new joint 2D and 3D diffusion model (JODO) that generates complete molecules with atom types, formal charges, bond information, and 3D coordinates. To capture the correlation between molecular graphs and geometries in the diffusion process, we develop a Diffusion Graph Transformer to parameterize the data prediction model that recovers the original data from noisy data. The Diffusion Graph Transformer interacts node and edge representations based on our relational attention mechanism, while simultaneously propagating and updating scalar features and geometric vectors. Our model can also be extended for inverse molecular design targeting single or multiple quantum properties. In our comprehensive evaluation pipeline for unconditional joint generation, the results of the experiment show that JODO remarkably outperforms the baselines on the QM9 and GEOM-Drugs datasets. Furthermore, our model excels in few-step fast sampling, as well as in inverse molecule design and molecular graph generation. Our code is provided in https://github.com/GRAPH-0/JODO. | ['Weifeng Lv', 'Bowen Du', 'Leilei Sun', 'Han Huang'] | 2023-05-21 | null | null | null | null | ['drug-discovery', '3d-molecule-generation'] | ['medical', 'medical'] | [ 4.19259667e-02 -7.94224069e-02 -2.55605400e-01 -7.85943270e-02
-7.30780721e-01 -6.60348594e-01 7.45631516e-01 3.27564895e-01
1.04708388e-01 1.10962653e+00 3.12224746e-01 -4.93921191e-01
2.26293262e-02 -1.30037248e+00 -8.73696148e-01 -9.65769529e-01
5.08563034e-02 6.02484345e-01 -2.87456840e-01 -2.96018362e-01
3.48947912e-01 6.42385125e-01 -8.61254096e-01 1.46687806e-01
1.14167452e+00 4.39775139e-01 2.76773393e-01 4.79895145e-01
1.52676016e-01 6.70330822e-01 -4.17662144e-01 -5.28383732e-01
-8.01758021e-02 -7.24459112e-01 -5.75461924e-01 -5.73054850e-01
1.49214476e-01 -2.79779911e-01 -6.85344994e-01 8.88358712e-01
1.17990780e+00 2.98658818e-01 9.94836152e-01 -7.83883572e-01
-1.50141644e+00 6.91941321e-01 -2.06727266e-01 -8.39002058e-02
6.27939522e-01 6.50752664e-01 1.04618144e+00 -1.05919147e+00
9.34274137e-01 1.18076956e+00 4.83857572e-01 6.90520406e-01
-1.53850770e+00 -7.96662331e-01 -5.63552883e-03 1.40064806e-01
-1.65188384e+00 -3.34550261e-01 8.14205468e-01 -4.87153679e-01
1.44005704e+00 -1.56333856e-02 7.25208521e-01 1.48776531e+00
6.73718393e-01 3.99888724e-01 5.27459264e-01 1.41363770e-01
4.83610749e-01 -3.70595932e-01 -2.02933982e-01 7.74398565e-01
3.40892911e-01 2.71228105e-01 -5.85224032e-01 -5.05115807e-01
7.49309659e-01 2.57714689e-01 -3.27933341e-01 -2.45220438e-01
-1.06216526e+00 1.25782859e+00 8.64705741e-01 -3.72628495e-02
-6.46937430e-01 3.40213895e-01 -9.20163468e-02 -1.78818345e-01
4.71822739e-01 8.35711420e-01 -1.60497665e-01 -4.90296744e-02
-5.29330552e-01 6.19440973e-01 7.89133608e-01 1.13137329e+00
6.28989756e-01 -8.23238119e-02 -3.68650764e-01 4.29339498e-01
3.23296696e-01 4.20367181e-01 4.98864353e-02 -4.45342898e-01
3.57318580e-01 4.41448510e-01 -1.09005749e-01 -9.33673859e-01
-4.56447542e-01 -4.94463950e-01 -1.06421733e+00 -3.17468941e-01
-2.15469569e-01 -1.81956217e-01 -1.14313781e+00 1.73409343e+00
4.53222543e-01 5.80740273e-02 3.32018286e-02 6.63716316e-01
1.40570915e+00 8.71594846e-01 3.94649506e-01 -9.59090292e-02
8.79525661e-01 -7.78424978e-01 -6.00737453e-01 2.46069655e-01
9.60047364e-01 -6.69008017e-01 5.42115927e-01 2.90412664e-01
-1.13383377e+00 -4.61290598e-01 -8.52387786e-01 -4.67936784e-01
-5.21638572e-01 -1.45285740e-01 1.18036270e+00 3.75866055e-01
-7.38783717e-01 1.15012646e+00 -8.83199275e-01 2.28153631e-01
8.09821069e-01 6.52897596e-01 -3.35939288e-01 -3.63052875e-01
-1.45769382e+00 7.66979694e-01 2.26574749e-01 1.36665270e-01
-1.13619983e+00 -1.04335093e+00 -8.38938475e-01 -2.70203259e-02
3.60609405e-02 -1.44451416e+00 7.13593781e-01 -4.82216254e-02
-1.57142329e+00 8.43810514e-02 -2.39652961e-01 -2.57531464e-01
1.92877248e-01 1.41262189e-01 -1.33457184e-01 -1.41772911e-01
2.43775398e-02 9.31208551e-01 2.93267280e-01 -8.42912257e-01
1.44989356e-01 -3.93801361e-01 -5.39608346e-03 3.26283574e-01
1.47774115e-01 -6.90324843e-01 -2.44375065e-01 -6.90047681e-01
-1.22064598e-01 -9.10397768e-01 -6.76381826e-01 -3.94213676e-01
-8.76803160e-01 -3.39749336e-01 1.71097130e-01 -4.51420218e-01
1.30421007e+00 -1.64601576e+00 8.51748347e-01 2.80845314e-01
4.21601772e-01 1.52066415e-02 -4.38755691e-01 9.25827384e-01
-3.80692363e-01 4.59541649e-01 -2.77527303e-01 -3.01323891e-01
4.00878638e-02 -6.16108477e-02 -2.10947260e-01 2.07528591e-01
3.76026332e-01 1.38946903e+00 -1.07911754e+00 6.57790005e-02
-4.60680313e-02 1.19434118e+00 -1.15688133e+00 1.25120476e-01
-5.20932436e-01 6.51274979e-01 -7.53661931e-01 6.78162754e-01
7.86460102e-01 -5.44710577e-01 2.08310679e-01 -5.04038274e-01
1.25203252e-01 6.84877515e-01 -7.15118527e-01 2.06714630e+00
-1.21861570e-01 -9.15916339e-02 -7.75737643e-01 -5.32741666e-01
8.22986305e-01 2.16520771e-01 5.55967450e-01 -6.57301128e-01
-8.05762336e-02 7.50584602e-02 2.05411449e-01 -1.07673638e-01
4.49662894e-01 -1.87670752e-01 1.95830181e-01 2.55916685e-01
2.15762854e-01 -6.15017474e-01 1.81692436e-01 4.38767314e-01
9.93160784e-01 3.90694141e-01 2.23557010e-01 6.24526106e-02
8.84701014e-02 -3.36900093e-02 3.37752044e-01 4.76689965e-01
4.23773438e-01 6.28862858e-01 5.80307782e-01 -2.26580113e-01
-1.01196420e+00 -9.76257980e-01 -1.75718218e-01 5.54565191e-01
7.14491762e-04 -9.83283162e-01 -5.85476398e-01 -5.91215134e-01
1.90851778e-01 7.83898413e-01 -8.10132265e-01 -7.29138851e-01
-1.87131554e-01 -1.35537684e+00 3.45669985e-01 2.39303097e-01
-9.82809514e-02 -9.16585386e-01 5.49430311e-01 4.78911728e-01
1.82562426e-01 -5.23511648e-01 -7.64231384e-01 2.18408093e-01
-7.09280431e-01 -9.78006661e-01 -6.08774245e-01 -4.34166551e-01
6.53065920e-01 1.09776855e-01 1.02703226e+00 -6.22625723e-02
-4.88789141e-01 -2.14255601e-01 -1.89333916e-01 -2.99793899e-01
-4.46930617e-01 3.93743366e-01 -1.44815817e-01 -4.59257990e-01
3.07435572e-01 -7.94488251e-01 -1.11570466e+00 -1.40519679e-01
-9.15393829e-01 2.50330538e-01 4.60051119e-01 8.98536384e-01
9.33761537e-01 -1.65054396e-01 6.04214013e-01 -1.01381743e+00
8.64858329e-01 -7.78597713e-01 -3.95119697e-01 -6.85741156e-02
-7.43924618e-01 4.44979221e-01 6.64867401e-01 -2.32884407e-01
-6.90315545e-01 5.62181957e-02 -6.47755861e-01 -1.62188962e-01
1.07707240e-01 8.71448398e-01 -4.36070740e-01 -1.52083531e-01
6.50615335e-01 2.93611526e-01 -2.01730803e-01 -4.33428943e-01
6.54768646e-01 2.62366027e-01 -1.40071228e-01 -6.82932734e-01
5.23239195e-01 1.58868670e-01 4.47863817e-01 -6.16029859e-01
-6.05009079e-01 4.66257073e-02 -4.79364544e-01 5.70979297e-01
7.11768508e-01 -1.03490210e+00 -9.83870804e-01 2.23194391e-01
-1.27790439e+00 -3.65187019e-01 -2.33633980e-01 4.90669787e-01
-4.35452998e-01 4.24692303e-01 -7.77668297e-01 -4.07695651e-01
-5.93459487e-01 -1.65759778e+00 1.29912007e+00 6.13125600e-02
-1.09139115e-01 -1.18357885e+00 3.86788934e-01 1.18045531e-01
3.32394898e-01 3.84958535e-01 1.08312738e+00 -4.32041466e-01
-9.50501323e-01 -1.30927980e-01 2.73210220e-02 -1.67847723e-02
3.56029510e-01 6.82649463e-02 -7.75937200e-01 -1.97273359e-01
-5.73443115e-01 -2.78695881e-01 1.16109657e+00 6.13422513e-01
1.17736685e+00 -2.25953728e-01 -6.02236986e-01 9.89559233e-01
1.12243760e+00 3.92574817e-01 7.17108548e-01 -3.84886175e-01
1.24437392e+00 1.84535787e-01 1.46554083e-01 5.69976568e-01
5.13949394e-01 6.76182032e-01 4.69308704e-01 -2.30372921e-01
-2.54483134e-01 -8.66636276e-01 3.20959419e-01 6.99926853e-01
-3.61512959e-01 -6.34340823e-01 -6.36630774e-01 2.49875407e-03
-1.61959445e+00 -1.18525743e+00 -1.77754208e-01 2.03432894e+00
1.19374084e+00 -1.58846512e-01 -1.68400425e-02 -5.15983999e-01
3.13494533e-01 7.28563815e-02 -1.03719604e+00 -8.55348706e-02
-2.36638695e-01 7.71582246e-01 4.00761366e-01 8.26519668e-01
-7.76192784e-01 1.17058504e+00 6.13914204e+00 1.07145774e+00
-1.00688481e+00 -2.10234880e-01 8.86245906e-01 -1.42182708e-01
-9.38709140e-01 1.05151884e-01 -1.09090590e+00 5.49951494e-01
8.69959533e-01 -2.20125541e-01 5.48808575e-01 5.81339419e-01
3.62658620e-01 3.11689824e-01 -1.23207200e+00 9.29726005e-01
-1.20057344e-01 -2.14291024e+00 7.29081154e-01 4.62200582e-01
8.89429450e-01 6.29790202e-02 2.95819730e-01 8.02742764e-02
4.86186177e-01 -1.57908320e+00 3.04281116e-01 8.03311706e-01
8.40670645e-01 -1.01557302e+00 3.94220114e-01 3.54411006e-02
-1.01664877e+00 5.10838509e-01 -3.83570105e-01 2.57943630e-01
1.27716586e-01 5.85822344e-01 -8.89907241e-01 8.45346153e-01
-6.20082580e-02 1.25750661e+00 -3.50338846e-01 8.32196236e-01
-2.64221638e-01 3.45641464e-01 -1.63522452e-01 -2.25187957e-01
9.90591571e-02 -6.99692845e-01 2.80753791e-01 9.46950257e-01
4.69323814e-01 1.43248647e-01 8.41012970e-03 1.46010566e+00
-4.57625031e-01 2.77509093e-02 -7.53007710e-01 -4.86057878e-01
5.95163822e-01 9.64330494e-01 -2.20328182e-01 -1.72930554e-01
-5.34635447e-02 1.05063510e+00 3.13356400e-01 6.56209052e-01
-1.05656540e+00 -3.90505344e-01 8.70872378e-01 1.82225592e-02
2.76212096e-01 -4.37679589e-01 -3.29670347e-02 -1.15822911e+00
-4.69099522e-01 -8.24328363e-01 3.47314291e-02 -7.40372002e-01
-1.39880419e+00 5.76704502e-01 -2.92151034e-01 -6.08322620e-01
3.16136815e-02 -6.79622293e-01 -4.47405219e-01 1.26563680e+00
-1.41509271e+00 -9.98424232e-01 2.41572969e-02 2.93895692e-01
1.20990634e-01 4.01323773e-02 1.06444311e+00 4.21534419e-01
-9.05227900e-01 4.56864834e-01 1.84350565e-01 -3.63143831e-01
7.22969890e-01 -1.19941270e+00 8.63571703e-01 2.34897122e-01
-8.03771988e-03 1.11656189e+00 4.87125605e-01 -9.40079629e-01
-1.78911471e+00 -1.26315820e+00 6.42566383e-01 -7.16142893e-01
4.87313509e-01 -4.72125709e-01 -7.79766679e-01 2.84058452e-01
-9.42399502e-02 -1.47270039e-01 1.07352865e+00 8.71987548e-03
-2.37039790e-01 3.99046779e-01 -8.90250623e-01 5.99094748e-01
1.51997650e+00 -5.84509969e-01 2.14526072e-01 9.46321130e-01
1.06556416e+00 -6.02130473e-01 -1.43238139e+00 2.52920419e-01
2.68207669e-01 -6.46293342e-01 1.28332984e+00 -1.03220832e+00
6.51598632e-01 -2.98925608e-01 -9.22259316e-03 -1.54899514e+00
-6.15346909e-01 -9.53194618e-01 -3.99811983e-01 8.02686810e-01
8.39669585e-01 -5.98870456e-01 7.60991335e-01 5.26197851e-01
-2.69157052e-01 -1.01075459e+00 -6.93978012e-01 -3.19147915e-01
3.53384644e-01 -2.36558154e-01 9.55843449e-01 7.83026218e-01
-1.71783075e-01 9.15733635e-01 -2.94184476e-01 -3.20005231e-02
3.65732193e-01 2.52155334e-01 6.63278997e-01 -9.27188396e-01
-5.46900392e-01 -6.22335136e-01 -1.50350943e-01 -1.31927240e+00
1.39436182e-02 -1.34451783e+00 -5.73392332e-01 -1.86706543e+00
3.86827618e-01 -4.40926105e-01 -1.10470634e-02 2.97461599e-01
-3.88392091e-01 5.22092134e-02 -1.60138950e-01 7.26538822e-02
-2.82771379e-01 1.24441481e+00 1.84831524e+00 -3.35810483e-01
-3.43959451e-01 -1.37333006e-01 -9.60299790e-01 3.14327963e-02
6.42807364e-01 -3.75158370e-01 -4.73001808e-01 -1.82427660e-01
6.93199873e-01 -1.73995271e-02 2.78910279e-01 -5.20437598e-01
1.65005445e-01 -2.51049489e-01 5.54276347e-01 -4.68573332e-01
5.02698898e-01 -1.59468681e-01 5.05830586e-01 4.29801285e-01
-2.60372013e-01 -3.45694013e-02 1.57918241e-02 8.07074904e-01
5.59488013e-02 7.64688477e-02 5.53613722e-01 -1.15384750e-01
6.94341138e-02 1.21588612e+00 7.03191236e-02 -1.98582888e-01
7.82800496e-01 2.36499533e-02 -4.73509997e-01 -2.84583807e-01
-8.83090496e-01 8.99576247e-02 6.10320389e-01 1.98367804e-01
7.77856469e-01 -1.43012190e+00 -6.35543585e-01 1.65859506e-01
8.20968580e-03 3.61370802e-01 3.93756300e-01 6.65818334e-01
-5.96616566e-01 6.70943975e-01 2.79377192e-01 -1.60768747e-01
-7.45507836e-01 7.43721724e-01 5.06931365e-01 -2.80780792e-01
-1.83657721e-01 9.14455950e-01 4.46056396e-01 -3.54904503e-01
-1.19868718e-01 -3.11891079e-01 6.81990758e-02 -5.47093153e-02
3.37825269e-01 8.61010998e-02 5.47506399e-02 -4.18988049e-01
-2.45251313e-01 4.84197229e-01 -2.88469017e-01 4.67011482e-01
1.61538327e+00 4.24716175e-01 -1.05008945e-01 -8.71986896e-02
1.28483701e+00 2.41082255e-02 -1.20515943e+00 1.26736492e-01
-5.62678456e-01 -1.38987303e-01 1.47148237e-01 -7.37289369e-01
-8.33680928e-01 9.43290889e-01 2.21437514e-01 -1.78078189e-01
5.13912916e-01 1.04576731e-02 8.55214715e-01 4.10883635e-01
2.07966462e-01 -5.56132078e-01 8.97885412e-02 4.70121920e-01
9.94759500e-01 -1.06625950e+00 3.02898496e-01 -4.60861325e-01
-5.38599789e-01 9.83497024e-01 3.65622014e-01 -5.74152842e-02
6.59768581e-01 -1.08795464e-01 -7.20407069e-01 -5.46026170e-01
-9.15401936e-01 -1.25605538e-01 5.31575322e-01 7.63097286e-01
1.03118896e+00 1.73932835e-01 -1.33061171e-01 5.76925874e-01
-1.78312629e-01 -6.37715161e-02 2.09401697e-01 6.35678351e-01
-1.93647951e-01 -1.62176096e+00 1.48744926e-01 2.60574043e-01
-1.07977562e-01 -8.25123012e-01 -7.43809223e-01 4.70245689e-01
1.46960393e-01 7.16532588e-01 -2.64591664e-01 -3.87281865e-01
3.30408692e-01 -2.73099601e-01 7.79905260e-01 -8.38766813e-01
-3.96345139e-01 6.58052191e-02 -9.88013763e-03 -4.33947474e-01
-1.45908788e-01 -2.72402197e-01 -1.31990159e+00 -6.66294456e-01
-5.35394609e-01 5.42270303e-01 5.19239902e-01 4.22961354e-01
1.14571548e+00 8.06850255e-01 6.02002025e-01 -9.32889640e-01
-4.09084737e-01 -8.52694154e-01 -3.21884513e-01 2.56128043e-01
1.56010628e-01 -6.52629256e-01 6.97500706e-02 -2.09778652e-01] | [5.017292022705078, 5.758340835571289] |
4dee732d-8151-48ea-98ab-0e55c00da884 | active-dictionary-learning-in-sparse | 1409.5763 | null | http://arxiv.org/abs/1409.5763v2 | http://arxiv.org/pdf/1409.5763v2.pdf | Active Dictionary Learning in Sparse Representation Based Classification | Sparse representation, which uses dictionary atoms to reconstruct input
vectors, has been studied intensively in recent years. A proper dictionary is a
key for the success of sparse representation. In this paper, an active
dictionary learning (ADL) method is introduced, in which classification error
and reconstruction error are considered as the active learning criteria in
selection of the atoms for dictionary construction. The learned dictionaries
are caculated in sparse representation based classification (SRC). The
classification accuracy and reconstruction error are used to evaluate the
proposed dictionary learning method. The performance of the proposed dictionary
learning method is compared with other methods, including unsupervised
dictionary learning and whole-training-data dictionary. The experimental
results based on the UCI data sets and face data set demonstrate the efficiency
of the proposed method. | ['Jin Xu', 'Hong Man', 'Haibo He'] | 2014-09-19 | null | null | null | null | ['sparse-representation-based-classification'] | ['computer-vision'] | [ 1.86630502e-01 -3.21120322e-01 -6.64290905e-01 -2.63261765e-01
-5.19625306e-01 -1.23901302e-02 3.30793381e-01 2.86453307e-01
-3.15863132e-01 8.16044927e-01 3.99499416e-01 2.24613741e-01
-1.11430302e-01 -8.61127496e-01 4.43862006e-02 -1.00421262e+00
5.80784939e-02 2.27836311e-01 -1.10597670e-01 -1.63228244e-01
4.32430923e-01 5.27068138e-01 -1.60787761e+00 2.77048182e-02
9.36093986e-01 1.02566671e+00 3.11464787e-01 -7.98680261e-02
-3.33062172e-01 1.03983331e+00 -5.43955207e-01 3.01905014e-02
1.44418895e-01 -6.08241916e-01 -1.19005844e-01 2.42882118e-01
5.63749229e-04 1.54653668e-01 -4.44741726e-01 1.11710846e+00
6.57453537e-01 4.44557995e-01 6.83365166e-01 -8.08678031e-01
-5.39380431e-01 5.37004948e-01 -5.33131063e-01 7.18500197e-01
3.77225310e-01 -5.16134083e-01 8.41443419e-01 -1.61104906e+00
4.29382235e-01 1.08897138e+00 6.02175057e-01 1.86378181e-01
-8.47784460e-01 -9.44128871e-01 -1.12731829e-02 5.93782902e-01
-1.85957277e+00 -7.33481646e-01 1.22150886e+00 -3.46334606e-01
7.33994842e-01 -6.00093044e-03 8.09063792e-01 5.85002184e-01
2.77921706e-01 8.22722673e-01 1.00097346e+00 -9.46097255e-01
4.23228383e-01 1.41951621e-01 2.62073904e-01 8.77900958e-01
5.70592284e-01 4.32135433e-01 -7.04791784e-01 -4.00748819e-01
9.67444301e-01 3.66458148e-01 -2.94854611e-01 -1.54126570e-01
-8.90708447e-01 1.30261195e+00 1.11698257e-02 6.00140989e-01
-5.63633859e-01 -2.86225647e-01 4.29466337e-01 4.25793678e-01
4.88771796e-01 -2.86045410e-02 2.50052810e-01 1.20872848e-01
-1.04475760e+00 -2.03423485e-01 5.90389252e-01 8.03016663e-01
5.38231969e-01 8.48981380e-01 9.99117270e-02 1.29362011e+00
6.26656592e-01 7.05925465e-01 8.99380267e-01 -3.68494868e-01
4.31475341e-02 7.59984910e-01 -2.17127219e-01 -1.62626779e+00
-7.46480301e-02 -3.51194113e-01 -1.06762910e+00 -8.92273411e-02
-3.55759501e-01 1.56679098e-02 -8.78967702e-01 1.07068610e+00
2.91101575e-01 6.21015787e-01 1.26580939e-01 8.28139246e-01
1.29724658e+00 9.44337428e-01 5.42340390e-02 -8.63226831e-01
7.36215591e-01 -6.38373911e-01 -1.05057752e+00 9.43805799e-02
2.86757112e-01 -1.02222848e+00 5.14613807e-01 6.83652222e-01
-7.37422705e-01 -9.02887940e-01 -1.14099407e+00 3.50886554e-01
3.85487899e-02 3.08346152e-01 9.74217176e-01 4.33617532e-01
-4.97277707e-01 4.23736498e-02 -6.95749581e-01 -2.86794484e-01
4.37875032e-01 4.66919154e-01 -3.77597451e-01 -2.63481081e-01
-1.22440863e+00 6.83885396e-01 6.90213263e-01 -8.13088268e-02
-1.16358030e+00 -3.04859400e-01 -9.87088084e-01 5.60736051e-03
-3.11160218e-02 -7.74911838e-04 6.83065832e-01 -1.05325687e+00
-1.26815474e+00 6.84316039e-01 -1.64604649e-01 -3.09922993e-01
-4.04801220e-01 2.41456464e-01 -8.59036744e-01 2.25018591e-01
-7.50629324e-03 3.81332412e-02 1.20850360e+00 -1.11510026e+00
-5.73264599e-01 -3.30297679e-01 -2.77903974e-01 4.03493613e-01
-4.01502192e-01 -1.22032709e-01 -3.24389428e-01 -1.25837708e+00
6.01769269e-01 -5.58406770e-01 -3.34560692e-01 -1.25919446e-01
3.31154615e-01 -2.90745080e-01 1.04885018e+00 -3.81641686e-01
1.52473724e+00 -2.43539476e+00 3.64519060e-01 7.24143207e-01
-1.05170929e-03 2.42971495e-01 -1.68288201e-01 4.04418707e-01
-2.82884032e-01 -6.14058673e-01 -2.99651057e-01 -2.03456104e-01
-6.11573994e-01 5.76706052e-01 -3.77260566e-01 6.49163902e-01
-2.52916157e-01 2.73689896e-01 -8.62708032e-01 -7.99462378e-01
3.43015343e-01 6.33520901e-01 -3.55981201e-01 4.78813618e-01
4.29143727e-01 4.66179997e-01 -3.58766288e-01 1.10988724e+00
2.99602717e-01 -3.17511782e-02 1.94136560e-01 -2.95419782e-01
6.87616989e-02 -8.97393525e-02 -1.59515429e+00 1.63079417e+00
-4.41731930e-01 4.30569053e-01 -7.25917965e-02 -1.42598283e+00
1.46467924e+00 6.12122416e-01 7.74422586e-01 -7.66724646e-01
5.06821871e-01 3.89644414e-01 -3.14280353e-02 -5.04783154e-01
2.15925634e-01 -3.78133118e-01 2.00291395e-01 2.38985851e-01
2.94593006e-01 6.90491870e-03 4.32793766e-01 1.64581053e-02
6.44225180e-01 -4.49739128e-01 8.50057244e-01 -3.81549984e-01
9.50329363e-01 2.03842953e-01 8.80991459e-01 3.68901402e-01
4.95402217e-02 3.43693078e-01 -3.18140954e-01 -5.02567232e-01
-7.79711246e-01 -6.52026892e-01 -4.37872201e-01 8.26982558e-01
2.98872918e-01 -4.92636979e-01 -4.42413501e-02 -6.21313214e-01
2.80248336e-02 4.87175196e-01 -4.41282988e-01 -2.97140688e-01
-6.36502922e-01 -8.48904788e-01 1.82116836e-01 3.28698486e-01
5.96748233e-01 -1.13068187e+00 -1.50381312e-01 3.88629913e-01
6.39172122e-02 -5.99529982e-01 -1.89546913e-01 1.52974382e-01
-1.13933587e+00 -1.08472729e+00 -5.99181771e-01 -1.38030910e+00
9.95229542e-01 5.87051034e-01 7.96804845e-01 4.91558760e-01
-2.83075213e-01 3.98697108e-01 -8.96274686e-01 -2.87808567e-01
-2.55142510e-01 -2.35888585e-01 3.45736176e-01 2.02969521e-01
4.99668777e-01 -6.86978340e-01 -2.83326000e-01 2.29402915e-01
-8.24777305e-01 -3.82326096e-01 6.60768092e-01 1.23488212e+00
1.27270365e+00 5.59504926e-01 7.23703384e-01 -1.08891034e+00
7.10572839e-01 -7.44729400e-01 -7.61008918e-01 5.67678846e-02
-7.99201548e-01 -3.70682329e-01 5.56599498e-01 -6.15643680e-01
-8.67225766e-01 2.18300417e-01 -2.42225796e-01 -5.85413933e-01
2.38372803e-01 1.03087986e+00 -1.65267140e-01 -5.50606668e-01
6.55826092e-01 6.51682913e-01 -1.08015360e-02 -5.83690405e-01
-4.00989875e-02 8.28435123e-01 1.89691797e-01 -4.52914745e-01
4.96368110e-01 2.27045164e-01 -8.86477828e-02 -9.50557292e-01
-7.87420034e-01 -7.99352467e-01 -5.12712777e-01 -1.44708231e-01
2.16009811e-01 -1.29821408e+00 -6.01111762e-02 9.73454565e-02
-7.08530962e-01 3.08341563e-01 -5.38370132e-01 8.46424460e-01
-3.51149559e-01 5.31174183e-01 -1.83376610e-01 -7.98120141e-01
-5.61131954e-01 -1.11881888e+00 5.98844826e-01 2.13605925e-01
-1.31555378e-01 -1.22223461e+00 3.00799966e-01 1.59781769e-01
3.14850360e-01 2.59459168e-02 7.98327267e-01 -6.28490329e-01
-1.97637305e-01 -3.91988337e-01 2.39534333e-01 4.39500183e-01
5.16304910e-01 -4.45799291e-01 -4.83152032e-01 -6.49643302e-01
4.99166995e-01 -2.84379005e-01 6.42741144e-01 2.03596860e-01
1.11615753e+00 -4.39180940e-01 -3.12920958e-01 5.58848321e-01
1.85792816e+00 7.44705379e-01 8.21398377e-01 -3.02090030e-02
5.85390389e-01 -1.66844770e-01 9.75965917e-01 7.89221287e-01
-2.12281063e-01 4.03219581e-01 -4.14822362e-02 -1.80161908e-01
-2.11312056e-01 -6.71429113e-02 2.72829056e-01 1.63953090e+00
-9.78037436e-03 -2.57545169e-02 -8.32423031e-01 5.34844160e-01
-1.69410801e+00 -1.23913825e+00 1.51540041e-01 2.02945256e+00
1.03492451e+00 5.46688922e-02 -1.69750556e-01 8.43402147e-01
5.53100884e-01 1.66232169e-01 -4.67286348e-01 -1.78384855e-01
-2.22253472e-01 6.59943819e-01 1.01409934e-01 6.81138396e-01
-1.12798226e+00 8.45958769e-01 7.06257296e+00 1.21411824e+00
-1.42007649e+00 3.90324116e-01 1.64372534e-01 1.59984037e-01
-5.77554032e-02 -3.52107314e-03 -9.00632560e-01 5.14441192e-01
4.35727358e-01 -2.45092809e-01 4.21796143e-01 1.17225659e+00
2.39461303e-01 -9.60391760e-02 -6.01644039e-01 1.80376053e+00
6.32058799e-01 -1.32216716e+00 5.79636455e-01 -2.63957322e-01
9.24215555e-01 -2.82544822e-01 -5.27986065e-02 2.98015535e-01
5.77468760e-02 -1.01469934e+00 3.31046402e-01 5.46875775e-01
8.32054257e-01 -9.05282378e-01 8.26350570e-01 3.11559707e-01
-1.41694593e+00 -1.79977685e-01 -6.93741322e-01 1.14525435e-02
-3.23862284e-02 8.37821066e-01 -5.17312407e-01 2.93979764e-01
1.95891008e-01 1.36878622e+00 -1.75384015e-01 1.27561080e+00
-2.80883431e-01 1.03622079e+00 -2.59076804e-01 2.22280473e-01
-1.48438618e-01 -4.79554057e-01 6.91168010e-01 1.17190039e+00
2.89257318e-01 6.76755905e-01 4.81446713e-01 3.56311351e-01
-8.45372677e-02 5.84561288e-01 -7.02892005e-01 -2.02926978e-01
8.81422341e-01 1.03615940e+00 -5.12991905e-01 -3.25015366e-01
-5.13830781e-01 6.49634004e-01 -3.22066322e-02 8.56884271e-02
-5.15658498e-01 -3.95531952e-01 2.32706636e-01 9.14564729e-02
4.03181016e-01 -4.54789639e-01 -1.46331817e-01 -1.34194803e+00
-5.14365852e-01 -1.15215373e+00 5.96939504e-01 -3.02707583e-01
-1.31714249e+00 5.79339445e-01 -1.71734858e-02 -1.60895395e+00
1.20435290e-01 -1.79524824e-01 -5.56102037e-01 6.72703743e-01
-1.57980633e+00 -7.24672019e-01 -4.24724579e-01 1.01331937e+00
8.50126982e-01 -1.17085803e+00 9.71129954e-01 7.90430546e-01
-5.35933137e-01 6.70424402e-01 3.60639691e-01 1.85780495e-01
2.23368853e-01 -6.12123847e-01 -5.69364250e-01 8.26400459e-01
5.97234845e-01 8.42616975e-01 4.19892848e-01 -6.59309804e-01
-1.49979198e+00 -8.68898690e-01 8.44885111e-01 2.47099593e-01
3.41247141e-01 1.35718331e-01 -7.22371995e-01 3.56197387e-01
-1.08697139e-01 3.15567672e-01 1.25819457e+00 -5.63786365e-02
-8.35512429e-02 -3.15051943e-01 -1.25233924e+00 -7.85740837e-02
7.21161008e-01 -4.65968043e-01 -9.15382266e-01 2.68795311e-01
1.43283769e-01 -3.48233551e-01 -1.07369030e+00 5.09574354e-01
3.31296146e-01 -4.39424694e-01 1.02115154e+00 -2.81476498e-01
-1.95884239e-02 -3.93225044e-01 -5.71413040e-01 -1.18591452e+00
-7.23888934e-01 3.30512971e-02 -4.07215357e-01 9.73053992e-01
-1.54696349e-02 -2.65984029e-01 7.05019057e-01 -3.28840435e-01
3.06845736e-02 -8.89274478e-01 -1.00623667e+00 -6.32310450e-01
-5.46578467e-01 1.68812312e-02 4.03184414e-01 1.37977564e+00
-4.93200794e-02 5.16427696e-01 -5.02462983e-01 -2.27484167e-01
7.16125369e-01 2.23791644e-01 3.48539799e-01 -1.42153394e+00
-2.20245257e-01 5.33421943e-03 -6.59892619e-01 -9.02834833e-01
1.93277940e-01 -1.02026689e+00 -3.64137977e-01 -1.45534539e+00
-8.50857422e-03 -9.76512969e-01 -6.56359196e-01 5.90994000e-01
-7.25681782e-02 3.25137645e-01 -7.99708888e-02 8.91009450e-01
-3.64248157e-01 9.25069213e-01 1.06430268e+00 -4.47248101e-01
-2.69570589e-01 -2.53173828e-01 -5.23344278e-01 5.19805193e-01
6.91207111e-01 -7.81138599e-01 -7.77370393e-01 -2.53841728e-01
-8.26365203e-02 1.57318711e-01 -2.06346884e-01 -1.07971668e+00
1.60213664e-01 -2.99186915e-01 4.94374663e-01 -5.85077584e-01
4.94237244e-01 -1.21005917e+00 3.74744803e-01 5.14553726e-01
-3.37321490e-01 -1.13065079e-01 -6.96531609e-02 6.31458163e-01
-7.47579277e-01 -5.14990747e-01 1.10978305e+00 -1.72834784e-01
-9.92021263e-01 5.29716432e-01 -2.06624791e-01 -1.33797750e-01
8.37325454e-01 -4.88902092e-01 5.29138744e-01 -5.30404299e-02
-9.85649586e-01 -3.13379496e-01 9.07897800e-02 2.80007511e-01
1.27559102e+00 -1.74195921e+00 -8.58739853e-01 6.53655589e-01
3.00914645e-01 -3.08536559e-01 -2.98341960e-01 5.00587285e-01
-5.29612780e-01 2.14725748e-01 -2.41918147e-01 -3.86127949e-01
-1.37340748e+00 4.35396671e-01 1.36104286e-01 7.93689117e-02
-5.38721085e-01 7.49631882e-01 -4.12388653e-01 1.54543296e-01
4.35959756e-01 3.48041624e-01 -8.27261388e-01 1.92046717e-01
7.11975396e-01 2.24459141e-01 1.28467575e-01 -9.71903026e-01
-2.25855961e-01 5.82947850e-01 -1.09815285e-01 3.21907848e-01
1.48517954e+00 -2.14773184e-03 -4.27338332e-01 4.07649547e-01
1.03170884e+00 1.78682208e-01 -3.01486671e-01 -6.67977512e-01
-2.67209351e-01 -8.71056020e-01 5.78393519e-01 -4.26865906e-01
-1.50088859e+00 5.38801134e-01 1.06229746e+00 -1.34999037e-01
1.13243604e+00 -4.75605100e-01 8.60721052e-01 3.46582204e-01
8.77853215e-01 -1.00883603e+00 9.67730582e-02 3.65498126e-01
9.56856966e-01 -1.48157239e+00 4.98872578e-01 -4.58949804e-01
-3.74915332e-01 1.08689022e+00 5.39824486e-01 -5.27800381e-01
1.14170003e+00 1.06239714e-01 -6.62683770e-02 -2.00997636e-01
-4.59833622e-01 -1.67482644e-01 2.85816550e-01 4.85467494e-01
4.44482923e-01 -1.97911680e-01 -9.17182505e-01 5.98408103e-01
1.73980981e-01 1.38124749e-01 3.67829204e-02 1.20033324e+00
-7.97850668e-01 -1.26157391e+00 -6.13887072e-01 5.22191048e-01
-1.60200521e-01 -8.60542953e-02 -2.00723872e-01 3.35500926e-01
4.36683893e-01 1.12779200e+00 -1.25998631e-01 -3.98488373e-01
1.45121366e-01 -3.83091927e-01 5.23269057e-01 -9.40218329e-01
-3.87991935e-01 1.06528342e-01 -2.34370247e-01 -1.46331802e-01
-7.46891797e-01 -4.51646686e-01 -1.37314582e+00 -1.97613448e-01
-6.94810271e-01 7.07375348e-01 3.87194425e-01 6.97582185e-01
1.85431883e-01 1.16782002e-02 1.07242525e+00 -4.32488441e-01
-2.59979129e-01 -6.99615300e-01 -8.11428428e-01 3.65884960e-01
-4.23120968e-02 -1.11565781e+00 -4.72905457e-01 2.93616772e-01] | [12.396055221557617, 0.40833550691604614] |
017e0e0e-a125-463e-81b4-706c1a8dbded | embedding-graphs-on-grassmann-manifold | 2205.15068 | null | https://arxiv.org/abs/2205.15068v1 | https://arxiv.org/pdf/2205.15068v1.pdf | Embedding Graphs on Grassmann Manifold | Learning efficient graph representation is the key to favorably addressing downstream tasks on graphs, such as node or graph property prediction. Given the non-Euclidean structural property of graphs, preserving the original graph data's similarity relationship in the embedded space needs specific tools and a similarity metric. This paper develops a new graph representation learning scheme, namely EGG, which embeds approximated second-order graph characteristics into a Grassmann manifold. The proposed strategy leverages graph convolutions to learn hidden representations of the corresponding subspace of the graph, which is then mapped to a Grassmann point of a low dimensional manifold through truncated singular value decomposition (SVD). The established graph embedding approximates denoised correlationship of node attributes, as implemented in the form of a symmetric matrix space for Euclidean calculation. The effectiveness of EGG is demonstrated using both clustering and classification tasks at the node level and graph level. It outperforms baseline models on various benchmarks. | ['Junbin Gao', 'Ming Li', 'Yu Guang Wang', 'Xuebin Zheng', 'Bingxin Zhou'] | 2022-05-30 | null | null | null | null | ['graph-property-prediction'] | ['graphs'] | [-6.49789795e-02 6.97501361e-01 -2.05972925e-01 -1.19994283e-01
-1.93783656e-01 -6.70095921e-01 6.23984039e-01 2.88659990e-01
1.73068747e-01 1.21676520e-01 3.32409918e-01 -3.19499791e-01
-4.62588131e-01 -9.80175555e-01 -6.86897635e-01 -8.68557930e-01
-6.13835931e-01 1.91491574e-01 -3.72745216e-01 9.65345930e-03
-1.17374733e-01 7.63514459e-01 -9.13213193e-01 -1.67143479e-01
3.71380836e-01 6.85024261e-01 -1.48180544e-01 9.26791012e-01
1.75984517e-01 6.60201490e-01 -1.24424547e-01 -4.25422966e-01
3.61299038e-01 -1.18689358e-01 -8.23192656e-01 2.19040334e-01
3.90232801e-01 6.10281937e-02 -1.14751697e+00 1.20745361e+00
1.36408135e-01 2.54527599e-01 7.55039215e-01 -1.67834568e+00
-1.14722514e+00 6.17786825e-01 -4.34145451e-01 -5.62490448e-02
2.56391883e-01 -1.78760722e-01 1.57781076e+00 -8.37410986e-01
6.90799415e-01 1.24330544e+00 8.96963596e-01 1.97753921e-01
-1.61469042e+00 -2.23970920e-01 -6.27663285e-02 4.25944813e-02
-1.68487096e+00 -1.13066554e-01 9.81077611e-01 -5.17645180e-01
9.70081091e-01 3.67949694e-01 6.85183704e-01 7.62707651e-01
3.27829421e-01 1.91467181e-01 5.11238575e-01 -1.17227681e-01
1.19518340e-01 -2.67580032e-01 2.22905368e-01 1.14350367e+00
5.34478366e-01 5.59399240e-02 -3.14006418e-01 -2.18709648e-01
7.94173121e-01 2.97757655e-01 -2.95730054e-01 -9.27548409e-01
-1.22338605e+00 1.00848746e+00 1.14915276e+00 4.20773149e-01
-4.47544992e-01 5.24040580e-01 1.47358090e-01 5.29350996e-01
4.83733952e-01 4.35527503e-01 -2.40307692e-02 3.79856884e-01
-5.54229677e-01 -3.05902332e-01 1.00938070e+00 1.07025146e+00
1.09558523e+00 6.83783442e-02 -1.18157029e-01 2.17823669e-01
6.10593855e-01 3.24814051e-01 9.43135470e-02 -6.85877800e-01
2.01070860e-01 1.10083854e+00 -6.34962320e-01 -1.83244169e+00
-5.52648246e-01 -7.18255401e-01 -1.35681903e+00 -9.05658305e-02
4.36667539e-03 2.34498695e-01 -6.37130916e-01 1.65378571e+00
4.11170244e-01 8.48641872e-01 1.15840659e-01 7.33589232e-01
9.79870796e-01 5.70293427e-01 -2.52841204e-01 2.42776364e-01
1.06281006e+00 -6.30084038e-01 -4.90406483e-01 2.77994089e-02
9.59257185e-01 -1.96504101e-01 7.89476752e-01 -2.52148986e-01
-6.88511312e-01 -2.91020274e-01 -1.25949490e+00 -2.47873709e-01
-5.18390357e-01 1.39676023e-03 7.24622369e-01 5.13080776e-01
-1.53915668e+00 1.01599622e+00 -8.18107247e-01 -6.15757406e-01
3.30155164e-01 3.10291946e-01 -9.78128076e-01 -1.19543098e-01
-1.01775491e+00 4.00497407e-01 1.81010202e-01 1.90208524e-01
-6.77173734e-01 -8.36988628e-01 -1.33370519e+00 4.33568567e-01
-6.21838123e-02 -7.78476238e-01 2.64820337e-01 -4.20958459e-01
-1.03955126e+00 9.09273684e-01 2.85957068e-01 -4.98202801e-01
-2.13261098e-02 1.08054139e-01 -4.99887884e-01 2.30335712e-01
9.97819230e-02 1.30393922e-01 1.35372138e+00 -1.01978159e+00
1.81182191e-01 -5.44179380e-01 1.55255660e-01 1.18801735e-01
-6.04490519e-01 -5.46165347e-01 -3.67471576e-01 -7.26296902e-01
6.52003884e-01 -9.71269786e-01 -3.60499203e-01 -4.99742012e-03
-5.07880807e-01 -1.66527405e-01 8.85577857e-01 -7.63325751e-01
1.37759471e+00 -2.34477687e+00 6.21795833e-01 7.62094617e-01
1.08949566e+00 -1.79679781e-01 -4.86198187e-01 8.54464650e-01
-6.85243607e-01 5.84614128e-02 -2.42823102e-02 -1.86542839e-01
1.44421652e-01 1.37007192e-01 -2.05777064e-01 1.04357708e+00
2.16136411e-01 1.30225885e+00 -1.16009021e+00 -1.04908802e-01
1.98869064e-01 7.18267739e-01 -5.71883917e-01 3.39200139e-01
2.03470662e-01 1.02670468e-01 -3.77787203e-01 2.50560075e-01
7.15834260e-01 -5.78376591e-01 3.94967258e-01 -6.95594192e-01
5.41782379e-01 -1.12159634e-02 -1.12306142e+00 1.90688515e+00
-2.83926159e-01 4.92453665e-01 2.32942507e-01 -1.27496302e+00
1.00275338e+00 -3.38276220e-03 6.02934539e-01 5.11307344e-02
-5.45698125e-03 -3.39739859e-01 -3.04216146e-02 2.30043996e-02
1.46440655e-01 3.75535846e-01 6.10197112e-02 5.33721149e-01
2.32280463e-01 5.16132303e-02 -1.95237458e-01 9.55582738e-01
1.81861401e+00 -2.55082518e-01 3.39611620e-01 -6.04428053e-01
5.08045912e-01 -4.77619946e-01 2.04090611e-03 3.54326129e-01
1.07748315e-01 3.20426702e-01 7.74405718e-01 -5.38838267e-01
-7.35563993e-01 -1.41553497e+00 1.78689390e-01 8.26808751e-01
4.35356377e-03 -9.86961186e-01 -6.14886045e-01 -8.06451261e-01
3.91355455e-01 3.35357368e-01 -8.28312755e-01 -8.48048329e-01
-1.86048657e-01 -4.10839528e-01 4.18311685e-01 3.27129900e-01
1.65675163e-01 -4.44653213e-01 2.89938658e-01 -1.93110481e-02
3.52572262e-01 -1.07872939e+00 -6.73792481e-01 1.02432504e-01
-8.36588383e-01 -1.33012760e+00 -9.82187688e-02 -1.05197346e+00
1.04922533e+00 6.38098359e-01 9.68635261e-01 4.50703740e-01
-3.15583348e-01 9.91587996e-01 -2.45693222e-01 3.74486208e-01
-3.99740994e-01 1.78652778e-01 1.47393063e-01 4.58951980e-01
2.52936661e-01 -1.04182029e+00 -5.09403229e-01 -9.46103930e-02
-7.56360173e-01 -1.06471553e-01 7.18391716e-01 7.92731106e-01
3.72919559e-01 1.90050259e-01 2.28896141e-01 -7.61349022e-01
7.52579629e-01 -7.13107765e-01 -4.04558867e-01 2.45605275e-01
-7.08694637e-01 3.92870247e-01 6.13476098e-01 3.43553163e-02
-9.44040641e-02 5.33400737e-02 4.22267884e-01 -8.75808537e-01
3.33361030e-01 7.42629945e-01 -3.36538076e-01 -4.05720949e-01
6.39873564e-01 2.96274424e-02 3.99766475e-01 -3.85888278e-01
9.78720248e-01 3.05878907e-01 4.36928093e-01 -2.41879031e-01
1.40666354e+00 4.61504251e-01 6.81981921e-01 -1.00107563e+00
-4.60622996e-01 -5.83780646e-01 -1.07462728e+00 -1.04549691e-01
8.03856254e-01 -8.73843014e-01 -6.96775615e-01 7.11910077e-04
-8.84420753e-01 1.30953295e-02 -2.94230133e-01 4.12811816e-01
-3.81775737e-01 7.83415973e-01 -7.80407488e-01 -1.81114569e-01
-3.97367537e-01 -5.48625588e-01 1.13408208e+00 -3.76646698e-01
-6.90347552e-02 -1.44532084e+00 2.94943511e-01 -2.64604506e-03
3.27697098e-02 5.04514277e-01 1.16233754e+00 -5.63388586e-01
-6.76332712e-01 -6.87232673e-01 -4.52160388e-01 3.09735119e-01
2.57697523e-01 -1.12986952e-01 -7.47670472e-01 -8.71849239e-01
-2.78163999e-01 2.24498764e-01 8.51445258e-01 2.43727803e-01
9.15412128e-01 -3.99129480e-01 -5.16393185e-01 9.99413967e-01
1.38560069e+00 -7.21165061e-01 3.99054140e-01 -2.30571643e-01
1.41570246e+00 5.43459058e-01 -7.14913430e-03 2.35244423e-01
3.61794382e-01 2.36531451e-01 6.93710387e-01 -7.26815611e-02
-2.00003788e-01 -5.75274467e-01 2.89285511e-01 1.17136919e+00
1.71826966e-02 1.00875705e-01 -8.48902881e-01 3.34883571e-01
-1.93271220e+00 -7.56045341e-01 -2.36095235e-01 2.08540845e+00
-1.99424345e-02 -2.39148289e-01 -9.25111845e-02 2.70907823e-02
9.47975218e-01 5.09558022e-01 -4.07417744e-01 -2.78814644e-01
-4.27323058e-02 1.42500460e-01 6.60141826e-01 5.79126000e-01
-1.14093471e+00 8.55685055e-01 5.87811565e+00 3.13548118e-01
-6.70961618e-01 -6.49827644e-02 2.74515152e-01 4.56778944e-01
-4.03631896e-01 1.93946645e-01 1.16852120e-01 -3.36603783e-02
1.09101903e+00 -5.68857670e-01 9.53778148e-01 8.27920973e-01
-1.58408970e-01 8.14753473e-01 -1.51385093e+00 1.17621672e+00
1.94171086e-01 -1.45221853e+00 2.65651792e-01 5.07262945e-01
5.10672629e-01 6.12919852e-02 1.99291170e-01 2.97052562e-01
4.66789663e-01 -1.47545421e+00 -8.35408196e-02 4.53199327e-01
8.71693254e-01 -6.53999805e-01 4.51049924e-01 -1.71012431e-01
-1.71613753e+00 2.19179243e-01 -6.19981468e-01 -5.90101294e-02
-3.79143566e-01 5.25279164e-01 -1.15188169e+00 9.43803489e-01
5.14562905e-01 1.52557337e+00 -6.80425823e-01 4.20674890e-01
-2.41224244e-01 4.69922721e-01 -8.72953385e-02 3.16143543e-01
3.58404189e-01 -8.50261807e-01 7.39718735e-01 8.61998856e-01
3.21570188e-01 -3.33644822e-02 9.73381326e-02 1.01023006e+00
-4.42425787e-01 1.11818269e-01 -1.20751417e+00 -6.61307812e-01
3.69056910e-01 1.80391371e+00 -6.59442842e-01 4.83584516e-02
-5.79837739e-01 1.18998754e+00 7.07734466e-01 5.72858989e-01
-5.58255315e-01 -4.49045569e-01 1.07491887e+00 4.40830477e-02
2.96665698e-01 -3.67115945e-01 1.57171175e-01 -1.21119261e+00
-6.87013268e-02 -6.22041404e-01 5.99440634e-01 -4.59357589e-01
-1.40139937e+00 4.97378170e-01 -3.88738006e-01 -1.11570799e+00
-6.13161884e-02 -8.99376035e-01 -8.81920576e-01 7.01911330e-01
-8.77054334e-01 -1.48223495e+00 -4.88525510e-01 9.22954440e-01
-4.39780146e-01 -3.54748756e-01 1.08509457e+00 4.92062606e-02
-5.12831271e-01 5.11601627e-01 3.44327718e-01 5.48122644e-01
1.62362799e-01 -1.61140454e+00 8.38280857e-01 7.75589824e-01
5.45175076e-01 6.56256735e-01 4.27722335e-01 -6.30159795e-01
-2.20982575e+00 -1.65479755e+00 5.27529836e-01 -5.33392012e-01
1.21172726e+00 -6.15685344e-01 -8.44847798e-01 1.16870773e+00
-1.13054328e-01 6.58834159e-01 7.56229758e-01 2.05036312e-01
-7.30925202e-01 -1.02341369e-01 -9.86259997e-01 6.57285273e-01
1.45993257e+00 -1.15247571e+00 -1.59343198e-01 5.50017834e-01
9.23263133e-01 4.09049615e-02 -1.45473230e+00 1.21421732e-01
3.62589180e-01 -5.06885529e-01 1.16163754e+00 -1.19736505e+00
3.88666317e-02 -2.76803285e-01 -3.89028072e-01 -1.62458074e+00
-9.82077003e-01 -9.52411115e-01 -4.71506417e-01 9.63452458e-01
1.01740085e-01 -5.87700665e-01 1.11337185e+00 3.36271703e-01
2.05110852e-02 -6.87484682e-01 -9.36967015e-01 -7.07885802e-01
-7.68811479e-02 -1.09471805e-01 7.90575981e-01 1.30592227e+00
1.47712454e-01 8.09995890e-01 -1.06792636e-01 7.29776680e-01
9.39945340e-01 1.50163889e-01 8.37926149e-01 -1.40392065e+00
-1.44134775e-01 -3.14967155e-01 -1.56513953e+00 -6.77000284e-01
6.60387874e-01 -1.87100470e+00 -6.24043882e-01 -1.53186071e+00
-1.29803836e-01 4.69657928e-02 -4.42038417e-01 2.58751120e-02
1.04117714e-01 1.89398583e-02 1.62935015e-02 -7.06446692e-02
-4.98321325e-01 7.38141000e-01 8.24656844e-01 -2.70541221e-01
4.82468214e-03 -2.10160196e-01 -6.91109300e-01 4.18133587e-01
5.68178177e-01 -1.40494972e-01 -6.22558415e-01 -2.58325160e-01
2.26489037e-01 -1.91068187e-01 4.52466697e-01 -8.84301364e-01
2.91939706e-01 3.12302649e-01 1.24294236e-01 -2.57681102e-01
2.67282754e-01 -1.10883498e+00 4.21935916e-01 5.33388555e-01
-1.61372438e-01 2.50910193e-01 -2.52450734e-01 1.19150436e+00
2.58741640e-02 3.67756397e-01 5.66995680e-01 2.14249536e-01
-6.36491597e-01 7.91747630e-01 3.51554900e-01 -6.03588484e-02
1.24609280e+00 -1.84657320e-01 -1.40077844e-01 -5.37859499e-01
-9.33211684e-01 1.21693760e-01 4.92451489e-01 3.94545496e-01
9.59426463e-01 -1.74791181e+00 -7.70399928e-01 5.11402249e-01
2.46823654e-01 -2.68690884e-01 8.55667219e-02 8.21768403e-01
-5.58922529e-01 2.31724203e-01 -1.12993689e-02 -6.45525515e-01
-1.24351227e+00 9.51524854e-01 3.11154842e-01 -2.64157623e-01
-1.03356481e+00 6.68964386e-01 2.67140865e-01 -7.42596388e-01
1.18512079e-01 -2.65988737e-01 -3.90478596e-02 -7.34190717e-02
2.00483888e-01 4.53961730e-01 8.01477209e-02 -9.96776819e-01
-4.11334395e-01 4.64485139e-01 8.40667263e-02 4.20813769e-01
1.43692231e+00 -1.16542235e-01 -6.67865455e-01 2.41956636e-01
1.83126819e+00 -1.74100056e-01 -9.77680385e-01 -4.40723777e-01
3.26986581e-01 -4.74984258e-01 3.22024375e-01 1.45150602e-01
-1.22502959e+00 6.57251179e-01 2.13719860e-01 4.21951801e-01
6.63997769e-01 3.16996455e-01 4.75915253e-01 6.62680626e-01
1.76370636e-01 -4.55672354e-01 1.94011256e-01 2.31747359e-01
1.10598099e+00 -9.34808433e-01 1.86672121e-01 -8.02005291e-01
-2.81023353e-01 1.05372024e+00 2.33277574e-01 -7.12941647e-01
1.25582671e+00 -2.42522523e-01 -3.19365650e-01 -6.99790061e-01
-6.22640789e-01 -7.43957460e-02 6.39367759e-01 9.59650278e-01
2.20924065e-01 3.19631100e-01 3.01541150e-01 4.13497508e-01
-3.46879214e-01 -8.95445943e-01 3.81138951e-01 4.15181845e-01
-1.35851130e-01 -6.38188541e-01 7.25936750e-03 7.80614972e-01
1.13163263e-01 -1.83443263e-01 -7.04672575e-01 6.33143127e-01
-5.34698129e-01 8.60986233e-01 7.35080093e-02 -7.51617789e-01
9.34763923e-02 -2.05417484e-01 3.15648615e-01 -8.88709903e-01
-1.66224062e-01 -2.65691489e-01 -1.88256517e-01 -9.92485166e-01
-1.70726478e-01 -4.28733945e-01 -1.15704739e+00 -5.73056161e-01
3.21405642e-02 2.81954348e-01 3.71494889e-01 3.49998504e-01
7.80445099e-01 3.30594778e-01 8.59288454e-01 -7.95579255e-01
-7.68793225e-01 -6.87299192e-01 -1.26966047e+00 7.43306637e-01
3.48241180e-01 -4.95681107e-01 -6.77598655e-01 -3.28905374e-01] | [7.091531753540039, 5.96779203414917] |
05a39c8c-4198-48ec-88d7-f690e18daaea | exact-how-to-train-your-accuracy | 2205.09615 | null | https://arxiv.org/abs/2205.09615v3 | https://arxiv.org/pdf/2205.09615v3.pdf | EXACT: How to Train Your Accuracy | Classification tasks are usually evaluated in terms of accuracy. However, accuracy is discontinuous and cannot be directly optimized using gradient ascent. Popular methods minimize cross-entropy, hinge loss, or other surrogate losses, which can lead to suboptimal results. In this paper, we propose a new optimization framework by introducing stochasticity to a model's output and optimizing expected accuracy, i.e. accuracy of the stochastic model. Extensive experiments on linear models and deep image classification show that the proposed optimization method is a powerful alternative to widely used classification losses. | ['Sergey Kolesnikov', 'Stanislav Dereka', 'Ivan Karpukhin'] | 2022-05-19 | null | null | null | null | ['classification'] | ['methodology'] | [-3.24465960e-01 -1.20810293e-01 -3.87961000e-01 -7.55803525e-01
-1.01401508e+00 -2.13309154e-01 8.58585313e-02 2.19862312e-01
-6.19167209e-01 1.25084877e+00 -4.72905010e-01 -4.56413515e-02
-1.39916807e-01 -6.24200463e-01 -6.85631573e-01 -7.53776670e-01
2.44268894e-01 3.16579461e-01 -2.03134447e-01 2.60981679e-01
2.95649111e-01 5.09303689e-01 -1.16514480e+00 -8.24095011e-02
1.07096314e+00 1.59392178e+00 -1.95335954e-01 4.17185307e-01
1.50529593e-02 9.02928770e-01 -6.84136987e-01 -7.10913956e-01
1.36305273e-01 -3.44155014e-01 -6.60648286e-01 -8.77791569e-02
1.72655545e-02 -2.78836370e-01 -1.63165733e-01 1.15359604e+00
3.36672932e-01 1.99918777e-01 7.14229405e-01 -1.53403509e+00
-7.07917571e-01 2.45881736e-01 -2.58322477e-01 -3.00409142e-02
5.42707369e-02 -8.68942887e-02 1.29534805e+00 -1.00896263e+00
2.69879997e-01 1.04634988e+00 1.01126575e+00 4.25025970e-01
-1.35489762e+00 -4.43449438e-01 2.34236449e-01 9.63791609e-02
-1.58226228e+00 -2.54961550e-01 7.91729331e-01 -3.68829250e-01
5.40167212e-01 3.27214509e-01 4.38675612e-01 7.78111637e-01
4.10176814e-01 1.11798227e+00 1.22888899e+00 -2.10713401e-01
2.34731093e-01 7.41412401e-01 1.50917754e-01 9.66046095e-01
1.38322249e-01 8.98356810e-02 -2.39138037e-01 -1.94274485e-01
4.12647128e-01 1.22618757e-01 -3.03372651e-01 -4.90650624e-01
-6.58668458e-01 1.07837784e+00 6.24577284e-01 1.56268943e-03
-4.08413589e-01 1.66547135e-01 2.70307779e-01 3.62479329e-01
8.93077910e-01 3.46398026e-01 -3.37001920e-01 -6.77759647e-02
-9.03402328e-01 2.27832973e-01 8.22943449e-01 6.14035189e-01
5.58630407e-01 4.44171466e-02 -2.23222509e-01 9.47377980e-01
3.30446303e-01 3.28980029e-01 3.23784679e-01 -7.81064510e-01
3.13640177e-01 5.04164696e-01 2.54759699e-01 -9.82659638e-01
-3.81535321e-01 -7.57192075e-01 -8.80947590e-01 3.01917851e-01
3.10127467e-01 -1.59747079e-01 -5.87701440e-01 1.73395514e+00
1.83966607e-01 -2.65562356e-01 1.21170841e-02 9.13027644e-01
4.76086438e-01 4.98977721e-01 9.10376012e-02 -3.15841734e-01
5.05496621e-01 -1.01785183e+00 -8.01721811e-01 1.03535086e-01
5.01636028e-01 -5.71889639e-01 1.20591509e+00 4.24030870e-01
-1.20713723e+00 -2.67722368e-01 -1.03832614e+00 2.11485382e-03
-2.48692960e-01 3.33066434e-01 4.95864093e-01 5.96398175e-01
-8.07074964e-01 8.50902498e-01 -9.59984183e-01 3.49580199e-01
6.84087157e-01 4.43858862e-01 -1.81982949e-01 2.10016444e-01
-1.04644716e+00 1.07834601e+00 2.66526908e-01 3.81648690e-01
-6.86888099e-01 -6.08349562e-01 -7.67564118e-01 2.82899197e-02
9.39862281e-02 -6.24297798e-01 1.21072555e+00 -8.72100949e-01
-1.72627211e+00 8.62746239e-01 -6.79403841e-02 -6.87066615e-01
1.00226092e+00 -3.62062126e-01 -1.09852433e-01 -2.18148500e-01
-2.12408781e-01 3.78647596e-01 7.67660320e-01 -1.05168343e+00
-5.02522767e-01 -3.65312934e-01 1.43327517e-02 2.38697857e-01
-5.80634773e-01 -1.26611501e-01 1.16493218e-01 -5.79072952e-01
-7.18373209e-02 -7.26341724e-01 -2.69710362e-01 3.45491111e-01
-6.25203431e-01 -2.39948809e-01 5.59664726e-01 -7.96781838e-01
1.44163358e+00 -2.04726243e+00 2.85227932e-02 2.00924098e-01
1.73123437e-03 6.24331348e-02 2.28393659e-01 1.00310957e-02
2.09738955e-01 3.59379202e-01 -6.31696343e-01 -5.88731885e-01
2.37469345e-01 1.40912682e-01 -1.05336599e-01 6.23629689e-01
3.79405618e-01 9.25150335e-01 -6.99064195e-01 -5.15955508e-01
2.14049548e-01 7.07786024e-01 -5.24960160e-01 1.70193464e-01
8.94562230e-02 2.34996900e-01 -5.91879904e-01 6.55407250e-01
5.39499640e-01 -3.46357614e-01 -2.95072615e-01 5.80756068e-02
1.14085600e-01 1.46328822e-01 -7.61403501e-01 1.24638760e+00
-8.47158730e-01 6.45462513e-01 -6.33088499e-02 -1.41290951e+00
1.09438896e+00 8.36646557e-02 3.81720036e-01 -1.90194994e-01
1.19560488e-01 3.47689182e-01 -3.58890533e-01 -1.66138917e-01
1.47407249e-01 -3.03208798e-01 1.43046692e-01 7.41088912e-02
-1.86113894e-01 -1.13666743e-01 -1.81257412e-01 -2.89619029e-01
5.33315063e-01 -9.94177759e-02 3.06185037e-01 -3.05655181e-01
5.61480880e-01 -7.51462653e-02 6.29540801e-01 6.88335180e-01
-3.44457209e-01 5.87528408e-01 6.63081527e-01 -5.39166212e-01
-9.54045355e-01 -1.10797822e+00 -6.07459784e-01 6.42340779e-01
7.22599402e-02 5.64028807e-02 -6.68467820e-01 -1.10298479e+00
2.09857747e-01 8.27349305e-01 -5.99824250e-01 -5.79594553e-01
-2.15942651e-01 -1.15704942e+00 4.08323526e-01 4.79510844e-01
5.87135971e-01 -6.81340694e-01 -3.94104689e-01 1.36410326e-01
-1.66031808e-01 -8.34608018e-01 -4.05419618e-01 -5.24778552e-02
-1.15248811e+00 -8.35140705e-01 -8.47288311e-01 -5.34601092e-01
7.21130669e-01 -4.14044976e-01 1.17968321e+00 1.97833125e-02
-2.41292685e-01 1.74846929e-02 2.85740606e-02 -3.65230709e-01
-2.03696966e-01 2.05996931e-01 6.92107901e-02 1.82578132e-01
1.01011589e-01 -2.39199817e-01 -6.91920698e-01 3.10940057e-01
-5.92253387e-01 -4.03474540e-01 4.10868555e-01 1.25274765e+00
9.63498294e-01 9.00894105e-02 7.13223219e-01 -9.24334168e-01
7.92971075e-01 -4.96410310e-01 -7.92797387e-01 4.84889328e-01
-1.18697226e+00 2.57274657e-01 7.79487312e-01 -4.37448680e-01
-9.60927010e-01 -1.09901791e-02 -1.93177551e-01 -6.11661315e-01
3.84198040e-01 5.57779968e-01 -1.28817493e-02 -3.48412514e-01
4.36893493e-01 2.73518026e-01 9.57575291e-02 -5.70862055e-01
-1.09939620e-01 6.75112128e-01 1.62040994e-01 -3.69579375e-01
2.72321016e-01 4.28206623e-01 4.57536988e-02 -4.67666209e-01
-1.16978240e+00 1.55471899e-02 -2.83301204e-01 -9.76361707e-02
4.18811351e-01 -5.79753995e-01 -6.79934323e-01 5.29251277e-01
-9.42871630e-01 -1.62051350e-01 -4.61028695e-01 7.04249859e-01
-6.52395010e-01 6.49814457e-02 -5.65744281e-01 -1.08938789e+00
-4.56517935e-01 -1.10342216e+00 9.63878036e-01 1.90511793e-01
8.75922069e-02 -1.38690734e+00 -3.26700062e-01 1.75193846e-01
4.87760663e-01 4.74916488e-01 6.00831687e-01 -6.98698997e-01
-1.41048059e-01 -7.92734802e-01 -2.08734900e-01 9.79850113e-01
-3.53662111e-02 2.23417506e-01 -8.21968138e-01 -2.24862859e-01
2.69041598e-01 -6.24802828e-01 8.82138014e-01 6.97844923e-01
1.73150313e+00 -5.28784394e-01 -5.94713800e-02 7.96380222e-01
1.45627308e+00 1.06851749e-01 4.51698869e-01 4.08234447e-01
2.92239428e-01 3.19154292e-01 7.19193041e-01 6.10142231e-01
3.14458877e-01 5.36935389e-01 2.83657879e-01 -7.79382437e-02
3.98475200e-01 -1.98571458e-01 3.00405443e-01 5.32569408e-01
1.42391622e-01 -3.74419540e-02 -8.79154146e-01 2.52689481e-01
-1.85184193e+00 -6.84841454e-01 1.86149329e-01 2.36780167e+00
8.79585028e-01 3.03268820e-01 2.66373693e-03 1.48428500e-01
7.79461503e-01 -1.76813692e-01 -8.24552476e-01 -5.26913345e-01
-2.16780737e-01 -5.24912812e-02 5.23318589e-01 6.99606895e-01
-1.13929832e+00 7.06458271e-01 7.55566978e+00 9.91487920e-01
-1.08643281e+00 2.19140261e-01 1.19343698e+00 -2.74306566e-01
-1.81491792e-01 -1.58814996e-01 -7.99456537e-01 5.85954487e-01
5.84839761e-01 -3.28317404e-01 3.04654121e-01 1.13337719e+00
8.11637938e-02 1.25828281e-01 -1.05744195e+00 9.81905580e-01
-1.67441651e-01 -9.92999673e-01 -2.01562166e-01 -8.36950901e-04
7.54669905e-01 -2.61600852e-01 4.18169916e-01 4.37274396e-01
2.07021207e-01 -1.06948030e+00 7.13218033e-01 8.61227036e-01
5.96487582e-01 -9.62660253e-01 1.06162071e+00 4.21379209e-01
-5.94864130e-01 -1.01065397e-01 -5.07901967e-01 -2.90902052e-02
3.31447721e-02 1.01077485e+00 -3.84346068e-01 1.71337157e-01
5.17931461e-01 5.77543497e-01 -2.58782715e-01 1.06982243e+00
-2.21699089e-01 4.69000518e-01 -4.20224607e-01 -3.84864837e-01
3.92153829e-01 -3.95089626e-01 4.66281444e-01 1.06438899e+00
1.93945438e-01 -1.49093226e-01 2.45460898e-01 9.95185733e-01
-3.21364373e-01 1.74963877e-01 -5.63334584e-01 9.40363631e-02
2.68753171e-01 9.14360523e-01 -4.19467509e-01 -2.99687594e-01
-2.78780550e-01 9.67579365e-01 5.42256236e-01 3.87483418e-01
-1.09210253e+00 -6.12759113e-01 5.49596012e-01 -1.10752866e-01
-3.27080749e-02 -1.06012011e-02 -5.43665588e-01 -1.18883204e+00
4.43327725e-01 -4.25582916e-01 3.48544717e-01 -3.74412656e-01
-1.41023910e+00 6.75786436e-01 -1.59562901e-01 -1.04372883e+00
-1.12850532e-01 -6.14719629e-01 -4.53687757e-01 8.37604344e-01
-1.62487626e+00 -6.76420152e-01 -7.89033547e-02 4.47572619e-01
4.07695889e-01 -1.35353550e-01 4.90988284e-01 4.12016392e-01
-9.05690789e-01 1.16910779e+00 6.53280616e-01 -8.24823380e-02
2.80855656e-01 -1.27441621e+00 -1.47351041e-01 3.38255227e-01
-9.96705666e-02 1.82075113e-01 5.88976264e-01 -1.64565071e-01
-7.47830093e-01 -9.91450608e-01 9.04301226e-01 -1.55407861e-01
6.30781770e-01 -1.36601711e-02 -9.88147199e-01 4.40341890e-01
-3.72022063e-01 3.90347928e-01 6.24865949e-01 -2.49418691e-02
-6.46428168e-02 -4.80730891e-01 -1.61159718e+00 2.29427472e-01
6.67219222e-01 -5.11420012e-01 -1.15040198e-01 6.76482201e-01
4.58902478e-01 -2.66454130e-01 -1.06428504e+00 6.67057157e-01
7.30555356e-01 -8.60033035e-01 9.10465360e-01 -9.33604956e-01
4.17838007e-01 3.28369409e-01 -2.46329591e-01 -1.39519298e+00
3.19272801e-02 -2.96722978e-01 -4.31941241e-01 1.08512986e+00
5.60895205e-01 -9.53680813e-01 7.85633206e-01 8.62294734e-01
1.86467990e-01 -1.43927991e+00 -1.09938717e+00 -1.15629661e+00
3.37787598e-01 -3.08073461e-01 4.78829175e-01 8.20678174e-01
-2.01947734e-01 -3.52886356e-02 -4.29318339e-01 -1.48123249e-01
8.97951722e-01 1.72690600e-01 2.38564447e-01 -1.13049591e+00
-1.44315451e-01 -8.54340494e-01 -5.98471820e-01 -6.92160070e-01
5.20588756e-01 -7.79832602e-01 1.29094645e-01 -1.17950380e+00
1.33157164e-01 -6.82730734e-01 -8.07041049e-01 2.38054752e-01
-2.86379486e-01 7.98839182e-02 -5.02448231e-02 3.38193715e-01
-5.18855691e-01 1.05279052e+00 1.00791240e+00 -1.64345130e-01
-2.28644803e-01 5.50611258e-01 -5.38930118e-01 9.14047301e-01
1.20108306e+00 -6.49677098e-01 -1.34811074e-01 -4.15621042e-01
1.15461230e-01 -1.82424802e-02 3.40146840e-01 -8.37999344e-01
-2.34989449e-03 -1.45024329e-01 3.64924103e-01 -1.74596846e-01
2.65622824e-01 -7.05545247e-01 -7.63415173e-02 6.39693260e-01
-7.08750427e-01 1.19971326e-02 -1.10661775e-01 5.34310937e-01
-6.65046632e-01 -4.14397269e-01 1.15349936e+00 7.47618750e-02
-1.31342575e-01 3.91165435e-01 1.37289867e-01 7.12915584e-02
1.00533080e+00 5.39076477e-02 1.93786636e-01 -4.34927076e-01
-9.19707835e-01 4.21860904e-01 3.37314367e-01 1.24556251e-01
8.22447956e-01 -1.57552111e+00 -8.35986733e-01 -7.47547150e-02
-1.14811227e-01 -1.85178846e-01 -1.05200954e-01 1.05453634e+00
-6.17322505e-01 3.45096916e-01 1.43516943e-01 -5.14401436e-01
-1.04580116e+00 3.88444334e-01 6.55016303e-01 -7.04713106e-01
-2.65594065e-01 9.80318129e-01 -1.33680897e-02 -5.27075350e-01
4.28247422e-01 -1.59263328e-01 1.70868620e-01 -5.86930364e-02
2.36999586e-01 6.69493496e-01 5.16242906e-02 -3.98119509e-01
-4.76209611e-01 4.79287028e-01 -9.41887591e-03 -1.09405525e-01
1.24850488e+00 -1.24190539e-01 7.94403814e-03 7.29941487e-01
1.68531644e+00 -5.82221150e-01 -1.36453891e+00 -2.61056274e-01
9.33934227e-02 -6.73747897e-01 4.11025077e-01 -7.00706422e-01
-1.27294981e+00 1.10085738e+00 7.50808477e-01 2.48987302e-01
1.24952030e+00 -3.25196505e-01 7.42468417e-01 5.39233088e-01
1.81884527e-01 -1.28094459e+00 -2.18625829e-01 2.31983364e-01
1.04114938e+00 -1.70666230e+00 -9.69799294e-04 -1.19715072e-01
-6.25462532e-01 9.41360891e-01 5.71417749e-01 -2.27234483e-01
9.37844872e-01 1.01489633e-01 -1.49227619e-01 2.82693088e-01
-7.36823320e-01 2.14899078e-01 3.48122865e-01 1.96484223e-01
4.86003131e-01 1.37733564e-01 -6.66773379e-01 6.58406317e-01
-9.91395712e-02 7.90843442e-02 -9.20874476e-02 7.70210624e-01
-2.73962855e-01 -7.07219005e-01 -1.78829119e-01 6.19097412e-01
-8.14963937e-01 -6.14738613e-02 -3.07159632e-01 6.81385994e-01
-3.27898115e-01 6.59713864e-01 -3.17459814e-02 -3.16756159e-01
2.77378917e-01 1.43322721e-01 2.87998945e-01 -8.95264447e-02
-3.13856155e-01 -4.07669276e-01 -2.94932157e-01 -4.71568733e-01
-9.83962417e-02 -6.24333501e-01 -8.63575578e-01 -3.47804368e-01
-7.51216352e-01 3.15871716e-01 9.27766263e-01 6.63340271e-01
3.55978981e-02 2.68273890e-01 1.17963469e+00 -3.47614616e-01
-1.26121306e+00 -7.15924203e-01 -6.00043058e-01 3.57332528e-01
5.22388339e-01 -7.08276033e-01 -6.31577015e-01 -2.67770231e-01] | [7.9060282707214355, 3.8024120330810547] |
cae67ab3-8a8a-4912-af82-afe4ad37ac15 | label-decoupling-framework-for-salient-object-1 | 2008.11048 | null | https://arxiv.org/abs/2008.11048v1 | https://arxiv.org/pdf/2008.11048v1.pdf | Label Decoupling Framework for Salient Object Detection | To get more accurate saliency maps, recent methods mainly focus on aggregating multi-level features from fully convolutional network (FCN) and introducing edge information as auxiliary supervision. Though remarkable progress has been achieved, we observe that the closer the pixel is to the edge, the more difficult it is to be predicted, because edge pixels have a very imbalance distribution. To address this problem, we propose a label decoupling framework (LDF) which consists of a label decoupling (LD) procedure and a feature interaction network (FIN). LD explicitly decomposes the original saliency map into body map and detail map, where body map concentrates on center areas of objects and detail map focuses on regions around edges. Detail map works better because it involves much more pixels than traditional edge supervision. Different from saliency map, body map discards edge pixels and only pays attention to center areas. This successfully avoids the distraction from edge pixels during training. Therefore, we employ two branches in FIN to deal with body map and detail map respectively. Feature interaction (FI) is designed to fuse the two complementary branches to predict the saliency map, which is then used to refine the two branches again. This iterative refinement is helpful for learning better representations and more precise saliency maps. Comprehensive experiments on six benchmark datasets demonstrate that LDF outperforms state-of-the-art approaches on different evaluation metrics. | ['Jun Wei', 'Chi Su', 'Shuhui Wang', 'Qi Tian', 'Zhe Wu', 'Qingming Huang'] | 2020-08-25 | label-decoupling-framework-for-salient-object | http://openaccess.thecvf.com/content_CVPR_2020/html/Wei_Label_Decoupling_Framework_for_Salient_Object_Detection_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Wei_Label_Decoupling_Framework_for_Salient_Object_Detection_CVPR_2020_paper.pdf | cvpr-2020-6 | ['salient-object-detection'] | ['computer-vision'] | [ 3.38916898e-01 1.54269114e-01 -3.29639494e-01 -3.89573365e-01
-1.38251990e-01 -2.86699142e-02 3.99731696e-01 2.80939966e-01
-2.28616208e-01 5.99173248e-01 2.48171166e-01 2.40591913e-02
1.29510030e-01 -8.29512239e-01 -7.01922536e-01 -7.51359701e-01
2.60044754e-01 -9.02894810e-02 9.75081503e-01 -2.76224583e-01
3.13123137e-01 4.51482460e-02 -1.61149085e+00 3.34535331e-01
1.18002033e+00 1.13334966e+00 7.16173410e-01 -1.63720489e-01
-4.59563226e-01 7.82052636e-01 -3.22826445e-01 -5.73797151e-02
1.05787434e-01 -2.62548387e-01 -6.62181973e-01 7.86863491e-02
1.22047730e-01 -1.50898114e-01 1.08083738e-02 1.44468534e+00
2.01834872e-01 2.99218129e-02 4.52329367e-01 -1.27012730e+00
-5.07952154e-01 6.58115685e-01 -9.93298948e-01 3.23649168e-01
-8.59071612e-02 -4.43660542e-02 1.00394249e+00 -9.66453075e-01
3.39388341e-01 1.00765848e+00 5.41799784e-01 2.13844121e-01
-9.83877420e-01 -7.09297478e-01 5.54100752e-01 2.90321589e-01
-1.33656383e+00 -5.19219972e-02 1.10026252e+00 -1.26200676e-01
3.98669720e-01 9.81339812e-02 5.88694513e-01 6.51272476e-01
4.29783575e-02 1.27077174e+00 1.01599300e+00 -2.18777180e-01
1.30680814e-01 3.42821449e-01 3.28433841e-01 6.76159739e-01
1.94532424e-01 -1.68302506e-02 -4.83246863e-01 2.95821428e-01
8.81400049e-01 2.62549639e-01 -5.66959620e-01 -5.53585827e-01
-1.19906127e+00 6.82173789e-01 1.12769783e+00 4.48214561e-01
-6.10485733e-01 -1.19023547e-01 2.67000973e-01 -1.01491325e-01
2.96690434e-01 2.19647914e-01 -3.85812134e-01 3.57219309e-01
-9.65397537e-01 6.07987344e-02 2.00195476e-01 9.24973965e-01
1.27854800e+00 -2.58095860e-01 -5.01158118e-01 8.69795263e-01
2.51458943e-01 8.45829770e-02 4.98120934e-01 -6.89932525e-01
5.01791656e-01 1.22351646e+00 2.78869402e-02 -1.12809086e+00
-4.66718554e-01 -9.26974535e-01 -8.62893105e-01 1.35150656e-01
2.86954075e-01 2.92744637e-02 -9.93818641e-01 1.56704974e+00
2.31829226e-01 3.34123433e-01 -3.03886861e-01 1.31436181e+00
9.72473621e-01 5.55342376e-01 1.78153098e-01 8.12372342e-02
1.39169455e+00 -1.45663154e+00 -5.73220313e-01 -4.72672969e-01
4.69181597e-01 -5.09457171e-01 1.04620445e+00 1.38046950e-01
-9.01137650e-01 -7.76671290e-01 -1.21809828e+00 -1.08476996e-01
-4.19620067e-01 3.60822499e-01 6.71055317e-01 1.95340306e-01
-1.04534173e+00 5.06256640e-01 -6.57648802e-01 -2.83837300e-02
6.77880585e-01 3.52009714e-01 -1.82036042e-01 -1.55537687e-02
-1.39786923e+00 8.48466337e-01 5.13086081e-01 1.37832001e-01
-5.25289893e-01 -6.68631494e-01 -1.16183543e+00 4.07734066e-01
6.68868124e-01 -3.91484827e-01 1.01773190e+00 -1.30539346e+00
-1.10413229e+00 5.33842862e-01 -2.53853202e-01 -2.52797425e-01
2.37915203e-01 -1.58865094e-01 -3.52033377e-01 9.34623852e-02
2.77894348e-01 1.04062450e+00 9.66294110e-01 -1.40866935e+00
-1.07889712e+00 -2.60750324e-01 1.26818996e-02 3.61936390e-01
-4.83191669e-01 -1.92608967e-01 -7.69394815e-01 -7.11883724e-01
3.91935915e-01 -5.58033049e-01 -1.60506979e-01 -1.62444368e-01
-5.71340084e-01 -3.48640680e-01 1.00434339e+00 -3.96468967e-01
1.52617908e+00 -2.13848615e+00 6.09458834e-02 8.11841860e-02
4.51193541e-01 4.42392468e-01 -1.23205885e-01 -4.48936224e-02
-3.73820215e-01 -1.63668185e-01 -2.01181903e-01 -2.39164233e-01
-1.66920230e-01 -6.82197884e-02 -2.55520821e-01 1.35924714e-02
5.68351984e-01 9.69780982e-01 -1.12020731e+00 -5.54899216e-01
1.82743773e-01 3.45502079e-01 -5.11818171e-01 -1.22888006e-01
-1.32570833e-01 3.85752112e-01 -6.80449486e-01 7.10576534e-01
9.29847658e-01 -5.36417961e-01 -3.30113739e-01 -5.32335997e-01
-4.45041776e-01 3.28568220e-01 -1.19920444e+00 1.64345765e+00
-2.01696098e-01 2.13099003e-01 -2.77917981e-02 -1.03958786e+00
1.16416681e+00 -3.61935534e-02 2.46514574e-01 -7.64991343e-01
3.61993909e-01 9.70268771e-02 -1.46995232e-01 -1.72869429e-01
4.23850238e-01 -2.50850748e-02 5.38806021e-02 2.29566768e-01
-8.15686118e-03 2.88012058e-01 1.16842330e-01 2.14002013e-01
8.43375683e-01 2.41159528e-01 5.94526865e-02 -5.25927067e-01
9.20210898e-01 -1.67206287e-01 1.06605399e+00 3.48809600e-01
-4.31866705e-01 7.43149340e-01 4.73752707e-01 -4.75984335e-01
-5.27074397e-01 -9.67878819e-01 -7.13063776e-02 1.04109347e+00
9.90548968e-01 -4.50625241e-01 -8.82652938e-01 -1.07929158e+00
-1.01815365e-01 6.10983670e-01 -8.04734945e-01 -4.47380722e-01
-4.52472478e-01 -6.53795183e-01 -2.12046385e-01 6.82431638e-01
1.00176024e+00 -1.37802863e+00 -5.77803373e-01 2.34875754e-01
-1.78726867e-01 -7.50121057e-01 -6.84326231e-01 3.06043327e-01
-7.28320837e-01 -8.50824773e-01 -9.38293159e-01 -1.09055424e+00
9.53992069e-01 7.80272543e-01 9.83848333e-01 2.98963517e-01
9.14726481e-02 -4.69705313e-01 -5.40083706e-01 -4.47297871e-01
1.84206128e-01 1.99456424e-01 -3.77731442e-01 2.72082150e-01
5.73714077e-01 -3.91367078e-01 -7.97036886e-01 6.15046680e-01
-8.44964206e-01 5.48116863e-01 9.22093213e-01 9.44131613e-01
6.36163414e-01 2.46136695e-01 6.53093278e-01 -9.50567722e-01
1.59749046e-01 -4.67507839e-01 -4.61295128e-01 2.73384541e-01
-4.51981038e-01 1.46616012e-01 7.12909937e-01 -6.33351654e-02
-1.02696252e+00 1.20065868e-01 2.01146491e-02 -4.93876338e-01
-1.69479877e-01 3.82849932e-01 -2.68371522e-01 8.28667283e-02
2.86292583e-01 3.31087321e-01 -1.48699537e-01 -4.95256692e-01
-9.58098918e-02 6.29179418e-01 5.66097856e-01 -1.61190107e-01
6.30226314e-01 2.92671144e-01 -2.44488969e-01 -2.96751142e-01
-1.15493822e+00 -4.21678066e-01 -5.69913507e-01 -2.39363864e-01
7.62361526e-01 -8.74269485e-01 -3.90774280e-01 5.54893613e-01
-7.93880165e-01 -2.82837570e-01 -1.44728333e-01 2.42512509e-01
-1.36521250e-01 5.61608933e-02 -3.12371552e-01 -4.01383668e-01
-2.04380110e-01 -1.27046716e+00 1.29836655e+00 7.94411480e-01
1.32891059e-01 -8.26629639e-01 -2.89468884e-01 1.45028502e-01
4.86281961e-01 -2.17123772e-03 7.17927516e-01 -3.32490802e-01
-6.27581418e-01 -4.27258238e-02 -8.26311588e-01 2.75042772e-01
4.78060007e-01 -2.03659475e-01 -9.58866894e-01 -2.19785094e-01
2.31845714e-02 -6.05806746e-02 1.38325202e+00 2.95046777e-01
1.24357140e+00 9.23268963e-03 -6.29732788e-01 5.94463646e-01
1.45591450e+00 5.70993545e-03 5.80754936e-01 5.60963571e-01
8.41389894e-01 6.16511047e-01 8.96458805e-01 1.47925079e-01
6.37868881e-01 6.72880054e-01 6.23504162e-01 -5.08551955e-01
-3.43126714e-01 -3.45248342e-01 3.38883638e-01 5.34357846e-01
1.39660358e-01 9.01139230e-02 -6.15489304e-01 6.52228832e-01
-1.90962934e+00 -5.80628872e-01 -1.19525388e-01 2.09500980e+00
8.80582333e-01 4.26601380e-01 2.06412658e-01 2.98052669e-01
1.07231295e+00 2.61018425e-01 -6.75090134e-01 -3.03120655e-03
-1.00415207e-01 -2.78745908e-02 3.46329778e-01 1.48533881e-01
-1.31404865e+00 1.07498550e+00 4.93435574e+00 1.01572764e+00
-1.37987852e+00 7.16845468e-02 8.82810771e-01 1.81782410e-01
-4.16907996e-01 8.52237344e-02 -9.60868955e-01 8.38768244e-01
1.38181284e-01 1.72928840e-01 1.90735176e-01 7.87168384e-01
-5.17415367e-02 -3.91314685e-01 -7.07809508e-01 8.01256835e-01
-6.46156818e-02 -1.19232011e+00 1.64075885e-02 -2.28355467e-01
8.82035136e-01 7.57037997e-02 5.55355975e-04 4.54602540e-01
1.35256201e-01 -7.56811857e-01 7.57362485e-01 4.08051819e-01
4.87724632e-01 -7.48913705e-01 9.82771695e-01 4.91994917e-01
-1.41702902e+00 -2.32201412e-01 -6.02110803e-01 -7.87468976e-04
1.04473703e-01 9.59737003e-01 -4.05513048e-01 6.96505189e-01
7.99295962e-01 1.12770736e+00 -8.46981764e-01 1.20265675e+00
-4.95141089e-01 2.79301822e-01 -1.51786759e-01 1.26587003e-01
4.68876898e-01 -2.87071615e-01 4.22052145e-01 9.49575782e-01
1.05993986e-01 -1.32933501e-02 2.18679577e-01 1.02375126e+00
2.91139949e-02 8.73434991e-02 -1.12883925e-01 5.00394642e-01
4.18900549e-01 1.57597995e+00 -1.07429218e+00 -4.40422744e-01
-5.98210573e-01 9.55960870e-01 4.94739711e-01 2.69572139e-01
-8.95948887e-01 -5.61157703e-01 3.05206388e-01 1.68464720e-01
6.44091547e-01 2.04397514e-01 -5.88783741e-01 -1.12904453e+00
-8.00732803e-03 -2.88866937e-01 2.75960535e-01 -7.19587445e-01
-8.79207492e-01 6.99504673e-01 -3.44950855e-01 -1.39666164e+00
3.64063531e-01 -3.64372462e-01 -9.34329987e-01 9.26639676e-01
-2.11521935e+00 -9.53405023e-01 -6.51413620e-01 4.94465709e-01
5.66280723e-01 2.10483789e-01 2.73299605e-01 2.82922387e-01
-9.08428848e-01 5.81253707e-01 -2.68945694e-01 5.23474216e-02
5.19458294e-01 -1.32694805e+00 1.15528062e-01 8.61989200e-01
-3.35185789e-02 4.33086902e-01 4.15586412e-01 -6.63777173e-01
-7.11347997e-01 -1.38866293e+00 7.07041323e-01 -1.01420522e-01
3.81360173e-01 -1.85105845e-01 -1.13036096e+00 2.94916630e-01
-3.86867225e-02 2.29771525e-01 1.84879377e-01 -1.78151622e-01
-5.75217865e-02 -3.13251913e-01 -8.81903529e-01 4.76655632e-01
9.42713141e-01 -1.90340728e-01 -6.57367826e-01 -1.64936576e-02
7.81832993e-01 -1.70142427e-01 -4.05128866e-01 7.35488176e-01
1.65057853e-01 -1.26939130e+00 7.62591779e-01 5.58081269e-02
4.22555745e-01 -8.65347326e-01 3.41000915e-01 -1.36439943e+00
-7.33278930e-01 -1.29737243e-01 -2.12896559e-02 1.38851082e+00
3.66600543e-01 -4.51039016e-01 8.01361740e-01 1.99250609e-01
-4.82093573e-01 -1.10978436e+00 -3.94968152e-01 -5.95865488e-01
-3.85159999e-01 5.45858704e-02 6.93942249e-01 9.04291630e-01
-4.39065993e-02 3.97978634e-01 -4.54600491e-02 1.52412981e-01
4.06555057e-01 6.12813711e-01 2.98482925e-01 -1.32672155e+00
7.03982040e-02 -7.37183750e-01 -3.48186672e-01 -1.29671228e+00
-3.23987380e-02 -8.54126096e-01 2.79244602e-01 -1.51385283e+00
4.59511876e-01 -6.62846088e-01 -8.16961586e-01 8.50382090e-01
-7.89091110e-01 5.44680119e-01 1.22300260e-01 2.55184650e-01
-7.93993533e-01 7.78290331e-01 1.65440273e+00 -1.25305012e-01
-3.10477048e-01 6.21515363e-02 -1.02909780e+00 8.68346870e-01
6.80161297e-01 -3.10624719e-01 -3.64841014e-01 -3.76935512e-01
-8.74356106e-02 -3.14954132e-01 3.10318679e-01 -1.21157944e+00
4.16310668e-01 6.53741881e-02 5.87053657e-01 -7.64250875e-01
-9.37344283e-02 -7.38278151e-01 -2.13999107e-01 3.38387370e-01
-1.47518799e-01 -4.29881454e-01 6.11284114e-02 2.99471974e-01
-5.55454373e-01 -2.00452864e-01 8.16773772e-01 -1.61060970e-02
-1.04916608e+00 3.79610837e-01 1.24747150e-01 -2.18045518e-01
1.08067453e+00 -3.21042657e-01 -2.73664147e-01 -6.48443624e-02
-6.70101583e-01 6.64934039e-01 6.03029132e-01 5.90265274e-01
5.17120004e-01 -1.49392366e+00 -4.79957461e-01 5.61218798e-01
1.01289399e-01 4.97816920e-01 2.74405658e-01 1.18315327e+00
-2.34938949e-01 3.74967515e-01 -3.03601950e-01 -6.36223793e-01
-7.93226123e-01 7.45216489e-01 1.01305522e-01 -2.73865432e-01
-5.60115039e-01 1.03337026e+00 8.53379369e-01 -1.28575355e-01
2.69580126e-01 -4.40001160e-01 -7.48479187e-01 1.45085856e-01
6.93490386e-01 1.49061456e-01 1.27068935e-02 -7.88767517e-01
-3.60351831e-01 6.83729291e-01 -3.51217002e-01 4.45502609e-01
1.26563299e+00 -1.72624975e-01 -1.56490445e-01 2.81393945e-01
1.06588125e+00 -1.79677814e-01 -1.72169268e+00 -4.26838279e-01
8.10251571e-03 -3.70101243e-01 4.90889817e-01 -7.71611929e-01
-1.56027651e+00 9.29402769e-01 3.69521201e-01 2.61561662e-01
1.47887528e+00 -1.12509746e-02 1.01121473e+00 -2.65503645e-01
3.85847211e-01 -1.03538036e+00 6.79946840e-02 4.93822992e-01
7.45830297e-01 -1.20340621e+00 -1.24538720e-01 -8.31420600e-01
-8.05718362e-01 8.01694334e-01 9.79816437e-01 -3.24241102e-01
6.89145625e-01 -3.17202397e-02 -1.89920112e-01 -2.00326324e-01
-3.61076027e-01 -4.18950915e-01 4.98894930e-01 2.53739119e-01
7.70675838e-02 -2.02049807e-01 -2.26482496e-01 1.03617394e+00
8.55013505e-02 6.34515807e-02 1.73346370e-01 8.26901317e-01
-8.18839967e-01 -9.85282242e-01 -2.87105381e-01 5.50384760e-01
-1.74448818e-01 -2.46834978e-01 -1.15445234e-01 3.87058854e-01
5.58583319e-01 6.94157541e-01 2.56328940e-01 -5.05033135e-01
1.67281017e-01 -2.33686984e-01 -6.96064625e-03 -7.43763506e-01
-5.28991520e-01 2.04944581e-01 -3.83627206e-01 -6.41360581e-01
-4.49568957e-01 -5.66222370e-01 -1.54704511e+00 1.82473764e-01
-7.83509672e-01 3.00909758e-01 4.03194487e-01 9.29866374e-01
3.77406806e-01 8.62156689e-01 9.00888741e-01 -1.06158853e+00
-4.85401489e-02 -9.71270859e-01 -6.17846549e-01 3.53337169e-01
4.16910231e-01 -1.07916594e+00 -3.13988447e-01 -1.75561771e-01] | [9.777902603149414, -0.3974379301071167] |
ab6fe854-9f25-489f-ae05-285f980831da | large-scale-electron-microscopy-image | 1604.00385 | null | http://arxiv.org/abs/1604.00385v1 | http://arxiv.org/pdf/1604.00385v1.pdf | Large-Scale Electron Microscopy Image Segmentation in Spark | The emerging field of connectomics aims to unlock the mysteries of the brain
by understanding the connectivity between neurons. To map this connectivity, we
acquire thousands of electron microscopy (EM) images with nanometer-scale
resolution. After aligning these images, the resulting dataset has the
potential to reveal the shapes of neurons and the synaptic connections between
them. However, imaging the brain of even a tiny organism like the fruit fly
yields terabytes of data. It can take years of manual effort to examine such
image volumes and trace their neuronal connections. One solution is to apply
image segmentation algorithms to help automate the tracing tasks. In this
paper, we propose a novel strategy to apply such segmentation on very large
datasets that exceed the capacity of a single machine. Our solution is robust
to potential segmentation errors which could otherwise severely compromise the
quality of the overall segmentation, for example those due to poor classifier
generalizability or anomalies in the image dataset. We implement our algorithms
in a Spark application which minimizes disk I/O, and apply them to a few large
EM datasets, revealing both their effectiveness and scalability. We hope this
work will encourage external contributions to EM segmentation by providing 1) a
flexible plugin architecture that deploys easily on different cluster
environments and 2) an in-memory representation of segmentation that could be
conducive to new advances. | ['Stephen M. Plaza', 'Stuart E. Berg'] | 2016-04-01 | null | null | null | null | ['electron-microscopy-image-segmentation'] | ['computer-vision'] | [ 9.90554020e-02 -1.60634205e-01 4.83590424e-01 -2.12356195e-01
-3.27692300e-01 -7.72662044e-01 2.81214863e-01 2.52505928e-01
-6.36796296e-01 5.88747144e-01 -4.94930387e-01 -5.66210330e-01
1.05119191e-01 -7.19343245e-01 -7.23024845e-01 -6.04912043e-01
1.21237569e-01 7.98637629e-01 5.70849717e-01 2.71173537e-01
5.22795081e-01 7.76529670e-01 -1.36530566e+00 1.91838309e-01
7.42632985e-01 6.97396934e-01 8.20590496e-01 7.44526327e-01
-2.83360958e-01 2.10680827e-01 -3.81736130e-01 -1.87191874e-01
2.99540550e-01 -1.28846541e-01 -9.71837282e-01 -6.29032776e-02
2.72205323e-01 -2.22700045e-01 1.13027088e-01 1.12123799e+00
2.70333260e-01 -3.09673876e-01 2.63414800e-01 -1.09738135e+00
-1.73815697e-01 2.64483094e-01 -6.37691021e-01 6.61106646e-01
-1.00808583e-01 3.60052139e-01 4.11119670e-01 -3.93326551e-01
9.02805150e-01 7.42531657e-01 6.05906248e-01 3.37225169e-01
-1.50514603e+00 -6.13234997e-01 -2.76076972e-01 4.55141999e-02
-1.19395471e+00 -4.30524260e-01 1.66541085e-01 -6.62430167e-01
1.09278893e+00 4.91510741e-02 7.77526438e-01 6.57368422e-01
3.52984399e-01 1.21500783e-01 1.17784238e+00 -4.95092198e-02
4.61368442e-01 -5.56068346e-02 1.02127157e-01 4.62737799e-01
2.95590103e-01 -5.01269281e-01 -1.49771586e-01 -2.23790675e-01
8.61862540e-01 2.18090773e-01 -1.93039831e-02 -1.56063348e-01
-1.52466822e+00 4.40565586e-01 2.24263683e-01 3.86118263e-01
-1.66956007e-01 8.07367824e-03 3.56079668e-01 8.61084536e-02
1.84341580e-01 4.88125652e-01 -4.59849954e-01 -2.15371445e-01
-1.29187787e+00 8.90139416e-02 6.52025700e-01 6.11654818e-01
9.45835292e-01 -5.07057667e-01 6.01566494e-01 5.65777123e-01
5.27089015e-02 1.58401623e-01 3.93118113e-01 -1.20298183e+00
2.52227858e-02 7.74640501e-01 -9.75871012e-02 -8.69451702e-01
-7.78925121e-01 5.52996434e-02 -6.71880603e-01 3.87489617e-01
7.84538209e-01 -5.76954857e-02 -6.75499022e-01 1.39999819e+00
4.85136926e-01 6.18199110e-02 -5.41399181e-01 8.74740839e-01
3.61525446e-01 2.63704777e-01 -1.42096847e-01 1.27199531e-01
1.29635692e+00 -4.41699028e-01 -1.09694093e-01 -1.48611233e-01
6.96736276e-01 -8.24420154e-01 8.84457767e-01 3.93900841e-01
-1.05842245e+00 -1.95864752e-01 -1.13922358e+00 -3.74501459e-02
-5.52372158e-01 -2.24325925e-01 7.44204462e-01 2.90013701e-01
-1.29111969e+00 9.15444076e-01 -1.42298806e+00 -7.80943930e-01
6.89496636e-01 6.12098336e-01 -5.10016501e-01 8.42382088e-02
-2.96725184e-01 8.08505595e-01 1.80030674e-01 -8.45583901e-02
-5.11318684e-01 -6.81065142e-01 -2.23651037e-01 1.64977796e-02
8.56628418e-02 -7.43099868e-01 8.53247344e-01 -5.31457603e-01
-8.52041304e-01 1.12528217e+00 -2.38700926e-01 -3.69000912e-01
2.50521928e-01 3.97310257e-01 1.39140859e-01 5.07409275e-01
2.49547735e-01 9.73706603e-01 2.23834336e-01 -9.19542551e-01
-4.54702228e-01 -9.50234830e-01 -2.86466271e-01 -3.36414635e-01
-1.81529492e-01 1.74742877e-01 -4.00675595e-01 -8.32248554e-02
2.72443891e-01 -8.62738788e-01 -3.00811142e-01 2.46923603e-02
-1.74569502e-01 2.30590940e-01 8.75520527e-01 -5.58014810e-01
5.71307838e-01 -2.04880404e+00 -5.28456718e-02 1.22314215e-01
6.13914609e-01 1.85556918e-01 1.60877883e-01 2.83534884e-01
1.54573590e-01 1.90391049e-01 -2.02269346e-01 -1.46035552e-01
-3.11272532e-01 1.03971973e-01 2.34633684e-03 7.56463051e-01
-9.15961415e-02 8.37230742e-01 -5.38138330e-01 -4.76184845e-01
1.55088708e-01 4.33783799e-01 -3.42443705e-01 -1.39143355e-02
-1.03228301e-01 6.83954597e-01 -2.49530077e-01 4.55735236e-01
7.34606922e-01 -5.98909974e-01 4.06123012e-01 -1.61554459e-02
-3.40739280e-01 3.54568690e-01 -9.20502841e-01 1.60546303e+00
-7.27435276e-02 7.83100009e-01 5.62746048e-01 -1.21875131e+00
8.18271339e-01 8.44074637e-02 6.60574257e-01 -6.27112925e-01
1.82179436e-01 1.70175746e-01 2.32816234e-01 -2.42024258e-01
6.10347353e-02 -9.22171175e-02 4.50164229e-01 9.04365718e-01
-1.08912900e-01 -7.61586055e-02 3.87884229e-01 3.70929450e-01
1.44459414e+00 3.97699699e-02 -1.25052497e-01 -5.97186029e-01
-1.07173093e-01 3.33773375e-01 3.75072449e-01 4.09884274e-01
-2.76819646e-01 7.06305504e-01 6.13240123e-01 -6.44341052e-01
-1.60836542e+00 -1.08810735e+00 -4.43152428e-01 6.90370679e-01
-9.08318311e-02 -2.10625038e-01 -1.09769070e+00 -1.45193487e-01
-7.08687007e-02 4.31652628e-02 -4.63236392e-01 2.48949870e-01
-3.08489591e-01 -1.07198167e+00 6.22469008e-01 4.37428594e-01
3.43455225e-01 -1.03592145e+00 -1.18947995e+00 1.87411219e-01
5.33820949e-02 -1.21583354e+00 -7.70582911e-03 3.48375440e-01
-1.07295263e+00 -1.18268847e+00 -3.85496974e-01 -6.22919083e-01
9.31269169e-01 4.22336340e-01 9.25005674e-01 5.32290459e-01
-8.52819920e-01 1.71242073e-01 -2.66605932e-02 -2.34408692e-01
-1.68574631e-01 4.85289544e-02 8.19719061e-02 -4.68379647e-01
6.12376392e-01 -1.15752375e+00 -7.39185393e-01 3.75242293e-01
-8.61203074e-01 2.12512597e-01 3.16905946e-01 4.66398031e-01
7.91811168e-01 -7.20381960e-02 4.31721061e-01 -9.33162749e-01
4.98316020e-01 -6.60125136e-01 -8.52086127e-01 2.40211394e-02
-6.66136205e-01 -2.30852127e-01 7.51332104e-01 -1.69074386e-01
-4.15010124e-01 9.86206979e-02 -6.70045018e-02 -1.85852468e-01
-4.15498048e-01 1.66295871e-01 6.58170730e-02 -9.76812840e-02
3.82866859e-01 1.14112295e-01 3.31185579e-01 -4.31014597e-01
-9.25592557e-02 5.79584599e-01 7.09147036e-01 -6.47203803e-01
2.56404191e-01 8.61737192e-01 -3.30802845e-03 -8.47588480e-01
-2.12471873e-01 -4.48772430e-01 -8.41340721e-01 -2.67822258e-02
8.95801961e-01 -4.93878603e-01 -9.77806568e-01 1.39511362e-01
-9.79964852e-01 -5.72752893e-01 1.74263909e-01 2.20747158e-01
-4.60488111e-01 3.62882525e-01 -8.28692138e-01 -3.40215325e-01
-3.81307840e-01 -1.34409320e+00 8.40724289e-01 3.75474542e-01
-3.79033953e-01 -8.05831075e-01 1.64315253e-01 3.99618745e-01
5.19777536e-01 1.36265218e-01 8.51896167e-01 -5.04890084e-01
-6.45257831e-01 1.10903447e-02 -4.41155642e-01 -1.47486478e-01
-1.03787981e-01 3.85530829e-01 -7.95629442e-01 -2.50890523e-01
2.17811819e-02 -2.17287213e-01 6.50466263e-01 4.52655822e-01
1.30477977e+00 1.58123896e-01 -6.90863371e-01 7.71063030e-01
1.27844036e+00 1.12335309e-01 7.11699367e-01 6.00152791e-01
5.65440357e-01 8.28863025e-01 1.06264614e-01 2.18190268e-01
3.27178210e-01 4.60975409e-01 3.65191549e-01 -1.37083709e-01
5.19500077e-02 3.14797968e-01 -1.98750645e-01 7.65369594e-01
-6.93232715e-02 1.84279740e-01 -1.21650743e+00 5.43467045e-01
-1.66863275e+00 -8.40529442e-01 -3.60577911e-01 2.14709163e+00
6.31520152e-01 1.15326770e-01 3.50007862e-01 -8.01171660e-02
7.32692540e-01 -4.77885902e-01 -8.39148343e-01 -4.95659381e-01
2.12266296e-02 2.28723153e-01 4.25887793e-01 1.70641288e-01
-6.30002975e-01 6.76616251e-01 6.64413404e+00 3.01643193e-01
-1.31201518e+00 1.34425431e-01 8.68667126e-01 -3.26936632e-01
-7.87234455e-02 1.98069096e-01 -6.99600816e-01 7.29365110e-01
1.17436171e+00 -4.60400619e-02 9.08131480e-01 5.81549704e-01
3.53099436e-01 -4.85316604e-01 -1.06366241e+00 9.44000125e-01
-4.62250292e-01 -1.60948634e+00 -4.07265514e-01 5.73755145e-01
3.22199702e-01 7.62107790e-01 -2.81589627e-01 -3.64597410e-01
3.22935253e-01 -1.25815117e+00 2.97390968e-01 4.68353271e-01
6.54979646e-01 -7.54896700e-01 5.78789055e-01 4.83231515e-01
-8.55894864e-01 3.17283124e-01 -7.76141644e-01 -2.36012653e-01
6.46996051e-02 7.90534735e-01 -8.65617752e-01 -1.16577327e-01
1.03723049e+00 3.10669839e-01 -8.49015772e-01 1.11051810e+00
5.24401426e-01 3.58528465e-01 -6.36219859e-01 1.08788431e-01
-5.30919060e-02 -5.08442402e-01 1.63844973e-01 1.19432199e+00
3.08710009e-01 -2.81578749e-02 -1.29958943e-01 1.33144522e+00
-1.12160616e-01 2.83059780e-05 -5.53563714e-01 -3.23396206e-01
6.81405127e-01 1.84793890e+00 -1.65795159e+00 -2.56731242e-01
-3.81016344e-01 7.81831563e-01 7.38623679e-01 1.81872286e-02
-3.45815986e-01 -2.54250854e-01 6.26443028e-01 3.52998704e-01
3.81994396e-01 -4.24805492e-01 -7.53899515e-01 -8.71534884e-01
1.14626169e-01 -7.58266628e-01 5.20372726e-02 -7.94488609e-01
-1.07802606e+00 4.35414195e-01 -4.45211083e-01 -4.41751242e-01
8.53903592e-02 -6.08508050e-01 -7.34552562e-01 6.93860710e-01
-8.15884411e-01 -6.13992691e-01 -2.00214297e-01 2.66106874e-01
1.74034014e-03 2.39722982e-01 7.26011932e-01 3.58116776e-01
-6.91766977e-01 1.72150970e-01 2.15800673e-01 -7.03089088e-02
5.08532107e-01 -1.08455741e+00 5.80714345e-01 6.48805320e-01
2.93881029e-01 9.88364518e-01 6.07775033e-01 -5.86122513e-01
-1.39828348e+00 -7.78692782e-01 6.10612988e-01 -5.87007105e-01
7.78956711e-01 -4.69702244e-01 -1.06218803e+00 7.82491982e-01
1.98594667e-02 1.35025397e-01 7.33337760e-01 -3.81326601e-02
-1.63168371e-01 1.26196355e-01 -1.30581117e+00 4.72073555e-01
7.88059890e-01 -3.46604824e-01 -2.81450868e-01 2.46133298e-01
1.99182957e-01 -7.60706840e-03 -1.07779610e+00 -4.10279958e-03
6.59268856e-01 -1.25856614e+00 6.94581211e-01 -3.85439962e-01
4.35494155e-01 -4.33452725e-01 3.85080241e-02 -1.03510153e+00
-1.42369950e-02 -4.88773525e-01 3.18862051e-01 1.21902311e+00
4.66005713e-01 -7.82861948e-01 8.36139381e-01 8.11064541e-01
-8.42567310e-02 -9.31799352e-01 -8.29048634e-01 -5.19284844e-01
3.24278146e-01 -3.67040366e-01 6.00039661e-01 8.42051625e-01
3.69882315e-01 9.55800265e-02 4.65858877e-01 -1.95857082e-02
6.12707138e-01 1.88595533e-01 8.01806748e-01 -1.37585413e+00
-2.92147636e-01 -4.53580648e-01 -6.21556401e-01 -6.33675277e-01
-5.94947748e-02 -1.01210546e+00 -1.41170844e-01 -1.44949114e+00
7.57635355e-01 -4.85077202e-01 4.10560593e-02 5.03217220e-01
1.32474169e-01 6.88463390e-01 9.11232829e-02 3.84576619e-01
-7.70623326e-01 -3.60320598e-01 9.58859563e-01 2.31638581e-01
1.26033828e-01 -4.75324988e-01 -6.90142632e-01 8.35360050e-01
1.05709410e+00 -6.00916445e-01 -4.19711024e-02 -5.84152460e-01
2.23032236e-01 -1.89439252e-01 5.04539490e-01 -1.15547204e+00
5.39814591e-01 2.21298054e-01 6.89178824e-01 -5.54093659e-01
1.04576148e-01 -5.57174861e-01 3.22522014e-01 2.68376023e-01
2.09962279e-02 3.05366784e-01 2.48507261e-01 1.95093811e-01
2.54545480e-01 -1.86917767e-01 9.42697704e-01 -5.52011609e-01
-1.12328276e-01 1.56186372e-01 -5.54197729e-01 5.93623333e-02
1.03406036e+00 -3.05122674e-01 -6.58306897e-01 7.82956704e-02
-5.70595264e-01 2.03489065e-01 1.38997293e+00 -2.29596660e-01
2.68089980e-01 -7.24783480e-01 -2.59603649e-01 1.66291177e-01
-3.25329632e-01 2.77454257e-02 6.76125064e-02 1.13005483e+00
-8.63825917e-01 3.75909626e-01 -6.53613389e-01 -7.55976081e-01
-1.19038510e+00 5.59172153e-01 1.80216223e-01 5.28103486e-02
-9.66683090e-01 6.07651055e-01 6.83625489e-02 -3.76872689e-01
-1.82817549e-01 -6.53622523e-02 1.90029621e-01 -2.28114963e-01
8.04492593e-01 4.30517703e-01 2.32355282e-01 -3.70601654e-01
-4.82326120e-01 4.10226911e-01 -2.94119954e-01 1.63149298e-03
1.56344247e+00 -4.46764320e-01 -6.35518074e-01 4.26200122e-01
9.91507232e-01 -1.95817828e-01 -1.44939673e+00 3.93780380e-01
1.30894974e-01 -2.12428927e-01 3.32780108e-02 -5.09322941e-01
-1.06964600e+00 1.08314788e+00 4.68392491e-01 3.47663015e-01
8.97732198e-01 2.50607848e-01 1.13377500e+00 2.90343374e-01
6.39372170e-01 -1.05425107e+00 -3.43987316e-01 1.57049790e-01
5.64259402e-02 -1.04345548e+00 1.03944816e-01 -5.98513633e-02
-1.03026159e-01 1.28808665e+00 4.73854214e-01 -1.76553980e-01
2.99383879e-01 9.17002976e-01 5.25758155e-02 -5.73867798e-01
-8.09777081e-01 -1.98714528e-02 -3.97730082e-01 6.65516078e-01
4.97256786e-01 3.09257712e-02 -2.62415260e-01 2.81885833e-01
-3.80095810e-01 1.41141042e-01 6.94598973e-01 8.14470708e-01
-6.22112870e-01 -1.08494043e+00 -3.87970030e-01 7.90394604e-01
-7.14769363e-01 -4.95988652e-02 -4.96973872e-01 4.38609451e-01
-6.51729479e-02 8.56003642e-01 4.68758553e-01 -1.75540686e-01
-1.62419796e-01 7.38780424e-02 6.65991664e-01 -5.42421281e-01
-3.25106621e-01 8.38109665e-03 -3.80893856e-01 -6.88730836e-01
-2.60058284e-01 -7.62382567e-01 -1.76791370e+00 -7.98648059e-01
3.31882574e-02 -8.52325335e-02 1.02391624e+00 1.07326591e+00
7.90652692e-01 2.41722852e-01 7.23081529e-02 -1.06902897e+00
4.60459888e-02 -7.55114317e-01 -6.99572742e-01 2.49610052e-01
-1.47067204e-01 -4.72867906e-01 -2.86082655e-01 1.93900242e-01] | [14.208995819091797, -3.071223020553589] |
d2d38125-6a4d-4aaf-befc-46c3cb255e5d | embedding-convolutions-for-short-text-extreme | 2109.07319 | null | https://arxiv.org/abs/2109.07319v2 | https://arxiv.org/pdf/2109.07319v2.pdf | InceptionXML: A Lightweight Framework with Synchronized Negative Sampling for Short Text Extreme Classification | Automatic annotation of short-text data to a large number of target labels, referred to as Short Text Extreme Classification, has found numerous applications including prediction of related searches and product recommendation tasks. In this paper, we propose a convolutional architecture InceptionXML which is light-weight, yet powerful, and robust to the inherent lack of word-order in short-text queries encountered in search and recommendation tasks. We demonstrate the efficacy of applying convolutions by recasting the operation along the embedding dimension instead of the word dimension as applied in conventional CNNs for text classification. Towards scaling our model to datasets with millions of labels, we also propose InceptionXML+ framework which improves upon the shortcomings of the recently proposed dynamic hard-negative mining technique for label shortlisting by synchronizing the label-shortlister and extreme classifier. InceptionXML+ not only reduces the inference time to half but is also an order of magnitude smaller than previous state-of-the-art Astec in terms of model size. Through our proposed models, we outperform all existing approaches on popular benchmark datasets. | ['Devaansh Gupta', 'Rohit Babbar', 'Akash Palrecha', 'Atmadeep Banerjee', 'Siddhant Kharbanda'] | 2021-09-13 | null | null | null | null | ['product-recommendation'] | ['miscellaneous'] | [ 3.31428975e-01 -1.49598375e-01 -4.97075498e-01 -6.65670991e-01
-3.52580905e-01 -8.17734063e-01 7.07740426e-01 2.39417210e-01
-8.68680596e-01 2.67261088e-01 -7.93917477e-02 -8.37579846e-01
-1.88078582e-01 -6.29268944e-01 -4.81549412e-01 -3.28408688e-01
4.09937143e-01 5.83287597e-01 8.61310810e-02 -3.21567319e-02
2.65334904e-01 1.57824069e-01 -1.54225636e+00 6.81168377e-01
5.00487387e-01 1.54866838e+00 1.06428519e-01 5.22473633e-01
-4.75073397e-01 7.45733500e-01 -1.43802688e-01 -6.91203177e-01
3.37717384e-01 4.08730537e-01 -8.39687228e-01 -2.42666885e-01
9.28204238e-01 -2.71189749e-01 -2.19428301e-01 9.70166683e-01
5.85199952e-01 6.58046156e-02 5.37495196e-01 -1.17063320e+00
-7.74149597e-01 8.06643188e-01 -8.20844889e-01 5.88215217e-02
-1.13563426e-01 -1.43444657e-01 1.55693293e+00 -1.06278884e+00
5.08270085e-01 1.12330663e+00 1.05153608e+00 4.80384439e-01
-1.05614364e+00 -1.04236782e+00 3.89204383e-01 2.06833854e-02
-1.16288972e+00 -1.89520702e-01 2.95933425e-01 -3.89501125e-01
1.53174543e+00 3.95388275e-01 2.18059048e-01 1.25482631e+00
6.52726591e-02 8.66122425e-01 7.30223656e-01 -5.07747650e-01
-4.07951325e-02 1.00510120e-01 6.03088975e-01 8.31347227e-01
3.26783121e-01 -2.14969888e-01 -4.64687407e-01 -3.82631570e-01
1.03173457e-01 1.76519230e-01 1.14445865e-01 1.04903415e-01
-1.17356229e+00 9.99135077e-01 1.56999201e-01 1.79385066e-01
-5.03451787e-02 6.33464813e-01 8.96016717e-01 2.88772255e-01
9.70218658e-01 6.38995230e-01 -1.08293486e+00 6.75612837e-02
-9.72396076e-01 1.14121668e-01 8.37720811e-01 1.05010784e+00
4.62284535e-01 -3.20009857e-01 -3.34613353e-01 9.97772336e-01
2.85746068e-01 1.11761361e-01 7.69875646e-01 -8.27897131e-01
6.13777459e-01 8.68077278e-01 4.85349111e-02 -7.45714188e-01
-7.53551960e-01 -8.74227703e-01 -8.04681659e-01 -1.95690632e-01
1.61505535e-01 -1.18853115e-01 -9.51118052e-01 1.46606135e+00
3.43125612e-01 5.57002276e-02 -2.25043610e-01 5.70748985e-01
8.76116812e-01 5.91781914e-01 2.27035508e-01 1.53840706e-01
1.67330897e+00 -1.33650661e+00 -7.06238091e-01 -4.09917384e-01
1.39623785e+00 -8.12232256e-01 1.33799565e+00 7.04709888e-01
-6.01495385e-01 -5.70722938e-01 -1.08628368e+00 -5.06078303e-01
-8.97566199e-01 5.83125412e-01 1.00806320e+00 5.67011833e-01
-9.41364050e-01 6.23627782e-01 -5.13367772e-01 -3.19318444e-01
4.61815566e-01 5.57342052e-01 -2.09037259e-01 1.51298419e-02
-1.27404320e+00 6.86291516e-01 6.03483498e-01 6.65956885e-02
-2.19166309e-01 -7.87553370e-01 -7.25665152e-01 2.70510852e-01
5.40484190e-01 -3.70228201e-01 1.48478734e+00 -5.38954616e-01
-1.19598627e+00 8.47304404e-01 -5.62737882e-02 -4.60557163e-01
1.43125147e-01 -3.93414021e-01 -6.25422120e-01 -2.56452888e-01
4.13307436e-02 8.51984739e-01 6.25314832e-01 -6.09871805e-01
-8.77079785e-01 -3.30492169e-01 6.02601320e-02 -2.90351789e-02
-1.23834682e+00 8.75818823e-03 -6.74949944e-01 -6.61401629e-01
-6.83773160e-02 -1.02170134e+00 -2.27374911e-01 2.09615350e-01
-5.56418538e-01 -9.85823929e-01 8.49988997e-01 -1.90465361e-01
1.66306937e+00 -2.00181055e+00 -3.71821791e-01 6.14967346e-02
6.46896183e-01 4.49101746e-01 -3.23977739e-01 6.25262320e-01
-1.96434438e-01 2.22586229e-01 3.02588284e-01 -5.54061055e-01
4.16029453e-01 2.82047596e-02 -5.19584835e-01 4.51110333e-01
-3.12895365e-02 1.18601739e+00 -9.00596738e-01 -6.33346736e-01
-3.07050031e-02 8.37903619e-02 -3.88342917e-01 -9.47879851e-02
-5.70457041e-01 -4.93126094e-01 -4.34995264e-01 6.36507034e-01
4.80354011e-01 -7.85444200e-01 1.07589588e-01 -2.36915886e-01
-1.04603782e-01 4.80997175e-01 -8.69843066e-01 1.73346102e+00
-8.59851897e-01 6.27193332e-01 -2.47313112e-01 -8.53227377e-01
6.12364590e-01 2.99273014e-01 4.07101274e-01 -6.98093414e-01
2.54658401e-01 4.35002029e-01 -1.71409979e-01 -6.91369057e-01
6.00600183e-01 3.47405881e-01 -1.17748126e-01 8.39803398e-01
2.36012369e-01 4.04197037e-01 3.82446557e-01 1.95194602e-01
1.21335340e+00 1.08511657e-01 7.25027472e-02 -3.56012672e-01
4.27013129e-01 -2.11942136e-01 1.85405523e-01 9.74844575e-01
2.37846747e-01 1.17333189e-01 3.93705130e-01 -8.71492505e-01
-8.51779640e-01 -2.95630246e-01 -2.88983792e-01 1.87607312e+00
-2.33684957e-01 -8.51551533e-01 -4.30440009e-01 -1.13179100e+00
3.64313722e-01 6.75661445e-01 -8.05623174e-01 -8.96927193e-02
-3.46669167e-01 -9.22199368e-01 8.55685651e-01 6.22823894e-01
4.06542897e-01 -9.85003769e-01 -3.25423867e-01 1.30895525e-01
1.86227992e-01 -1.19586062e+00 -6.82794750e-01 7.37177193e-01
-6.65870905e-01 -8.77807617e-01 -3.68942946e-01 -8.95401657e-01
2.90928543e-01 1.16451584e-01 1.30307996e+00 1.23249061e-01
-2.80019343e-01 -9.41961259e-02 -4.49862331e-01 -4.67006832e-01
7.60320649e-02 8.38513851e-01 -1.33464858e-01 4.35644835e-02
9.35604513e-01 -2.57977426e-01 -8.04835200e-01 2.21816555e-01
-1.03354478e+00 6.84598535e-02 5.34497917e-01 9.53593075e-01
3.81246120e-01 -2.83675138e-02 7.13282943e-01 -1.51285458e+00
6.72169864e-01 -6.52328134e-01 -6.37030125e-01 4.04500514e-01
-1.33894241e+00 1.09252952e-01 7.13024497e-01 -7.43744195e-01
-6.19250476e-01 1.50726393e-01 -4.16441709e-01 -7.61177316e-02
9.17798728e-02 6.30463541e-01 4.21635121e-01 8.04752781e-05
6.94145977e-01 -1.45532593e-01 -2.88707018e-01 -6.53208673e-01
6.17571592e-01 1.13755012e+00 1.02221943e-01 -1.48302257e-01
3.29902411e-01 4.38725680e-01 -1.05998060e-02 -3.94673944e-01
-1.35612357e+00 -9.74263787e-01 -3.30187857e-01 1.37515888e-01
7.58899927e-01 -7.71679759e-01 -1.12191987e+00 3.55530322e-01
-1.13947809e+00 -2.19009697e-01 -6.04042038e-03 2.67539561e-01
-5.59304655e-02 2.92732209e-01 -8.97262692e-01 -6.26267910e-01
-9.51348722e-01 -9.54429805e-01 1.35430992e+00 -2.71043420e-01
-3.69104207e-01 -9.81213927e-01 -2.14491248e-01 2.12051794e-01
4.82418418e-01 -3.24603260e-01 1.27213311e+00 -1.21315372e+00
3.92014720e-02 -5.70451498e-01 -4.03331369e-01 2.86685020e-01
-2.88334221e-01 -4.20346469e-01 -1.03801882e+00 -3.73913437e-01
-2.41153523e-01 -7.99154878e-01 1.34534013e+00 3.27385105e-02
1.78655052e+00 -2.22042575e-01 -5.77975869e-01 7.13677168e-01
1.34631801e+00 -5.04349135e-02 1.96703389e-01 5.33432126e-01
6.84391141e-01 5.61623931e-01 6.04093432e-01 4.37504619e-01
9.48030576e-02 7.12764800e-01 6.00974381e-01 -2.95767605e-01
-1.12760505e-02 -1.03222914e-02 2.24726988e-04 7.92432785e-01
6.01691842e-01 -8.10341656e-01 -7.71238089e-01 4.76659894e-01
-1.99102950e+00 -6.22749031e-01 -2.75734305e-01 1.69635904e+00
9.48736489e-01 3.17442566e-01 -3.09358448e-01 8.02114420e-03
3.15282792e-01 1.62313119e-01 -7.81717896e-01 -4.63046640e-01
1.13858454e-01 3.99100661e-01 8.51284564e-01 3.00940305e-01
-1.26664424e+00 9.72864449e-01 5.77090168e+00 1.26099634e+00
-9.72288430e-01 3.86038750e-01 6.46517873e-01 -3.49190354e-01
-1.21774226e-01 -2.19901025e-01 -1.29226875e+00 2.97598720e-01
7.94716120e-01 5.46273351e-01 1.83886707e-01 1.03400624e+00
-2.78750509e-01 2.40410060e-01 -1.27343273e+00 8.64248037e-01
1.56258538e-01 -1.43837488e+00 6.55463710e-02 2.15012375e-02
6.43053949e-01 3.43115449e-01 3.15160453e-01 4.76685166e-01
3.99042219e-01 -1.00937808e+00 5.86488962e-01 -1.85458120e-02
1.12708247e+00 -5.72461486e-01 9.94699717e-01 4.62134600e-01
-1.04236472e+00 -5.11724055e-01 -3.78266811e-01 6.04929253e-02
-1.31430566e-01 8.20074141e-01 -7.69815266e-01 6.84871599e-02
7.28487670e-01 7.91776240e-01 -6.55153453e-01 5.82287848e-01
1.47379950e-01 7.81820059e-01 -3.40983212e-01 -3.86278689e-01
6.71319604e-01 1.12270914e-01 -1.93031579e-01 1.51197171e+00
3.60374212e-01 -3.78632545e-01 2.01645985e-01 5.41005909e-01
-5.09052932e-01 3.07365119e-01 -4.53611970e-01 -3.51169884e-01
4.42139745e-01 1.55034912e+00 -7.01830089e-01 -4.55034167e-01
-6.83226883e-01 9.41592634e-01 4.48245376e-01 2.70554990e-01
-7.39372551e-01 -6.75114214e-01 2.62637198e-01 7.81297162e-02
4.59559530e-01 7.59471506e-02 -2.86870599e-01 -9.43524063e-01
8.65342021e-02 -7.53088772e-01 4.54984635e-01 -5.71457505e-01
-1.41933322e+00 8.06040943e-01 -2.75302708e-01 -1.02101469e+00
-7.80611187e-02 -1.11823058e+00 -1.62951186e-01 7.69903421e-01
-1.63679981e+00 -1.55633056e+00 -1.80229157e-01 3.42098176e-01
8.12719584e-01 -3.10557008e-01 1.01918709e+00 5.94054639e-01
-4.82698113e-01 9.84372556e-01 3.30920964e-01 4.65183929e-02
8.69377017e-01 -1.16424274e+00 7.41295278e-01 2.61003166e-01
2.97717452e-01 7.75376499e-01 2.54834056e-01 -5.64011037e-01
-1.17247343e+00 -1.40635419e+00 1.34464133e+00 -3.20353806e-01
8.98040235e-01 -8.67393553e-01 -5.32242179e-01 7.32235968e-01
3.65262926e-01 1.89575881e-01 9.44926202e-01 4.47719365e-01
-8.30150664e-01 -2.88615912e-01 -5.49898982e-01 5.67141712e-01
1.06607687e+00 -4.91012037e-01 -7.11933300e-02 8.43263209e-01
1.02091157e+00 -5.99198267e-02 -5.93347788e-01 3.52272749e-01
8.50732267e-01 -4.73184854e-01 8.49081397e-01 -8.40577185e-01
4.75048363e-01 1.05680428e-01 -1.91329986e-01 -8.38650823e-01
-5.17556787e-01 -4.38297421e-01 -4.71980453e-01 1.05902100e+00
8.04229200e-01 -6.75465882e-01 9.70961571e-01 1.98415369e-02
-2.71681428e-01 -1.34465706e+00 -8.54430020e-01 -5.71716607e-01
6.15031831e-02 -6.97456598e-01 4.93502319e-01 9.71428216e-01
-6.87875226e-02 5.85726976e-01 -3.67424995e-01 -1.55840516e-01
2.98805475e-01 2.56907612e-01 3.06637108e-01 -1.51353598e+00
-4.67754453e-01 -3.95378560e-01 3.09294723e-02 -1.54523253e+00
4.28496033e-01 -1.37710345e+00 -1.50102884e-01 -1.26654732e+00
1.72768265e-01 -5.89140654e-01 -5.84723532e-01 8.11782777e-01
1.36283748e-02 6.10958576e-01 -2.20884994e-01 1.72897369e-01
-9.84877467e-01 1.70158759e-01 8.47033203e-01 -2.95597941e-01
2.86945462e-01 6.24881759e-02 -7.12172925e-01 6.02711320e-01
6.63178861e-01 -7.41551638e-01 -5.12104094e-01 -7.51582444e-01
1.14890444e+00 -5.51515758e-01 -3.62959281e-02 -6.04546726e-01
3.54858249e-01 1.98037386e-01 2.84709275e-01 -8.25594664e-01
3.06562424e-01 -8.52313221e-01 -3.54797840e-01 3.39355946e-01
-1.02993810e+00 2.39907637e-01 1.96735188e-02 6.03018403e-01
2.07461208e-01 -6.55087054e-01 3.44639897e-01 -1.69952735e-01
-5.26143312e-01 2.76962250e-01 -3.10482055e-01 -1.66857854e-01
7.00410128e-01 3.35492380e-02 -4.79355007e-01 -1.13893487e-01
-2.31005222e-01 4.56989348e-01 -6.57734871e-02 5.10378063e-01
3.53448898e-01 -1.29463935e+00 -3.29106629e-01 1.24612384e-01
4.57291186e-01 -1.96066722e-01 -8.95909965e-02 7.62927055e-01
-4.02995765e-01 9.31671262e-01 4.53794032e-01 -2.77813226e-01
-1.20528471e+00 7.47590601e-01 3.83195020e-02 -7.63510466e-01
-3.69888484e-01 1.15104949e+00 1.00681849e-01 -6.82568312e-01
7.05167115e-01 -4.22124714e-01 -2.85704792e-01 2.63859361e-01
4.85747278e-01 1.68522954e-01 5.70936441e-01 -1.74122304e-01
-1.01772308e-01 4.46059495e-01 -6.68223619e-01 2.16989711e-01
1.27760446e+00 -9.91385430e-02 -2.58996427e-01 4.23274279e-01
1.58878565e+00 -2.44667873e-01 -8.37671280e-01 -4.44936275e-01
3.48769248e-01 -3.82798095e-03 4.44997966e-01 -1.05442500e+00
-1.03088951e+00 1.04282439e+00 6.78132594e-01 2.26576984e-01
9.12394881e-01 9.94471740e-03 1.23748755e+00 9.19558108e-01
-1.45289693e-02 -1.39935923e+00 2.06567690e-01 7.11275518e-01
4.86243695e-01 -1.32538259e+00 6.90802783e-02 -2.38604665e-01
-1.94881961e-01 1.22819710e+00 8.03094149e-01 2.25731477e-01
8.05114925e-01 3.13779294e-01 -2.44985744e-02 -5.42346120e-01
-1.14251113e+00 -9.08848420e-02 3.71441454e-01 1.14095591e-01
6.96791172e-01 -3.96064758e-01 -3.92578095e-01 5.53665221e-01
9.21297446e-02 3.25624309e-02 -6.43075109e-02 8.38123441e-01
-3.16789627e-01 -1.21255457e+00 2.97055334e-01 1.00159907e+00
-8.65527630e-01 -7.83764780e-01 -2.93056428e-01 5.72625756e-01
4.63488072e-01 8.42107117e-01 1.42879814e-01 -7.04063833e-01
6.45739883e-02 3.11302274e-01 -1.74487054e-01 -7.86762297e-01
-9.25412893e-01 1.14903916e-02 1.99561536e-01 -4.46274966e-01
-1.76735535e-01 -1.96149662e-01 -9.28419054e-01 -1.14641868e-01
-9.15748715e-01 3.98409106e-02 8.34577560e-01 7.68843234e-01
6.12482250e-01 2.77149677e-01 3.37198615e-01 -5.33962429e-01
-1.08198535e+00 -1.42994618e+00 -5.04320621e-01 3.38486701e-01
2.24291041e-01 -5.49554586e-01 -2.08499551e-01 -1.77881137e-01] | [9.612666130065918, 4.483583450317383] |
c1dad21c-4db7-4a45-b3fa-718185396129 | q-how-to-specialize-large-vision-language-1 | 2306.03932 | null | https://arxiv.org/abs/2306.03932v1 | https://arxiv.org/pdf/2306.03932v1.pdf | Q: How to Specialize Large Vision-Language Models to Data-Scarce VQA Tasks? A: Self-Train on Unlabeled Images! | Finetuning a large vision language model (VLM) on a target dataset after large scale pretraining is a dominant paradigm in visual question answering (VQA). Datasets for specialized tasks such as knowledge-based VQA or VQA in non natural-image domains are orders of magnitude smaller than those for general-purpose VQA. While collecting additional labels for specialized tasks or domains can be challenging, unlabeled images are often available. We introduce SelTDA (Self-Taught Data Augmentation), a strategy for finetuning large VLMs on small-scale VQA datasets. SelTDA uses the VLM and target dataset to build a teacher model that can generate question-answer pseudolabels directly conditioned on an image alone, allowing us to pseudolabel unlabeled images. SelTDA then finetunes the initial VLM on the original dataset augmented with freshly pseudolabeled images. We describe a series of experiments showing that our self-taught data augmentation increases robustness to adversarially searched questions, counterfactual examples and rephrasings, improves domain generalization, and results in greater retention of numerical reasoning skills. The proposed strategy requires no additional annotations or architectural modifications, and is compatible with any modern encoder-decoder multimodal transformer. Code available at https://github.com/codezakh/SelTDA. | ['Manmohan Chandraker', 'Yun Fu', 'Xiang Yu', 'Samuel Schulter', 'Vijay Kumar BG', 'Zaid Khan'] | 2023-06-06 | q-how-to-specialize-large-vision-language | http://openaccess.thecvf.com//content/CVPR2023/html/Khan_Q_How_To_Specialize_Large_Vision-Language_Models_to_Data-Scarce_VQA_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Khan_Q_How_To_Specialize_Large_Vision-Language_Models_to_Data-Scarce_VQA_CVPR_2023_paper.pdf | cvpr-2023-1 | ['visual-question-answering-1', 'domain-generalization'] | ['computer-vision', 'methodology'] | [ 3.46261382e-01 3.73954713e-01 -5.90845421e-02 -4.61428523e-01
-1.06661701e+00 -1.07158148e+00 5.76582849e-01 -2.93592904e-02
-6.42556965e-01 7.59054661e-01 6.93178177e-02 -5.53565025e-01
3.28411937e-01 -7.43108571e-01 -1.16630936e+00 -3.69999409e-01
6.11425519e-01 6.38989389e-01 2.20162243e-01 -2.99485058e-01
-4.68787737e-02 3.48952383e-01 -1.42025542e+00 5.79140306e-01
8.61225486e-01 1.05617332e+00 3.32836598e-01 9.10021126e-01
-3.63563687e-01 1.07147074e+00 -6.65511131e-01 -8.22231412e-01
3.71967822e-01 -4.83361512e-01 -1.09208357e+00 1.25622407e-01
1.27820790e+00 -6.09516799e-01 -6.27828002e-01 9.49478209e-01
3.22442174e-01 2.81212389e-01 7.01364875e-01 -1.49708402e+00
-1.33673489e+00 3.52429539e-01 -3.33529413e-01 1.97230458e-01
4.10073876e-01 7.07319975e-01 1.00768745e+00 -9.27262604e-01
7.20732152e-01 1.52059305e+00 2.86425322e-01 1.09406412e+00
-1.45474076e+00 -7.15621769e-01 2.01459192e-02 3.86133403e-01
-1.10189450e+00 -3.85852098e-01 7.76340485e-01 -3.67984176e-01
6.85898125e-01 1.46279588e-01 3.52044612e-01 1.46370065e+00
-1.28951579e-01 9.84175563e-01 1.24154818e+00 -3.69129688e-01
3.03219318e-01 4.21893060e-01 6.07480742e-02 7.31383741e-01
-7.63642415e-02 1.29115477e-01 -3.55568647e-01 -6.81873364e-03
8.91918421e-01 -2.69664139e-01 -2.68631101e-01 -7.54554570e-01
-1.32264161e+00 1.16028142e+00 7.33636856e-01 -1.68029681e-01
-2.72270232e-01 2.60718256e-01 3.92676741e-01 7.27761924e-01
-1.01875938e-01 6.60222113e-01 -5.02339959e-01 2.31244460e-01
-6.38336718e-01 3.13388258e-01 4.64658916e-01 1.04941142e+00
9.39584911e-01 3.43697816e-01 -4.74702388e-01 7.23851502e-01
1.54696733e-01 8.60527575e-01 6.58366203e-01 -1.34971797e+00
5.77802062e-01 6.26156151e-01 -8.62538144e-02 -4.61580157e-01
-1.21854603e-01 -7.81522617e-02 -7.40819037e-01 4.82438534e-01
7.86390007e-01 -1.11015424e-01 -1.41333723e+00 1.98745131e+00
1.65709034e-01 -1.86071709e-01 5.52266121e-01 1.00349927e+00
1.27904105e+00 7.29503274e-01 1.67210370e-01 1.04820423e-01
1.37629247e+00 -1.04934621e+00 -4.72382843e-01 -4.89308268e-01
3.39274734e-01 -4.35834348e-01 1.84433246e+00 2.96706885e-01
-1.01617074e+00 -7.46209204e-01 -7.74048805e-01 -5.58963537e-01
-4.70841169e-01 -6.37469068e-03 4.98535216e-01 5.74325383e-01
-1.25457716e+00 -4.97328974e-02 -2.50153303e-01 -1.74279660e-01
8.19413722e-01 3.04089189e-01 -5.54888189e-01 -4.02464241e-01
-1.17714131e+00 9.51784015e-01 4.65647638e-01 -2.59205431e-01
-1.59268451e+00 -7.23946214e-01 -1.26901758e+00 -1.06858753e-01
5.31093657e-01 -7.98061848e-01 1.52907908e+00 -1.32812119e+00
-1.33215427e+00 1.10081697e+00 -3.18656787e-02 -5.96865833e-01
3.83279681e-01 9.07237530e-02 -2.49610424e-01 3.70690167e-01
5.65718859e-02 1.41935623e+00 1.37150586e+00 -1.49303925e+00
-1.77281350e-01 -2.68312275e-01 5.40836096e-01 3.52297723e-01
-1.30409107e-01 -5.64742506e-01 -3.05906057e-01 -6.49588883e-01
-4.31843370e-01 -7.75713623e-01 -1.93805516e-01 2.19515845e-01
-1.50129452e-01 -1.50785714e-01 7.99632609e-01 -5.74367583e-01
5.90717614e-01 -1.98038614e+00 1.87298700e-01 2.67792447e-03
3.34293902e-01 4.48761821e-01 -7.96790898e-01 4.66348641e-02
-9.32000279e-02 -2.45448664e-01 -5.02412617e-01 2.59082429e-02
-2.32175011e-02 5.30738413e-01 -5.90390027e-01 5.54686971e-02
3.72563481e-01 1.52896154e+00 -8.38422716e-01 -5.36438286e-01
2.12872997e-01 1.17872305e-01 -4.86654252e-01 4.07457858e-01
-7.62741506e-01 4.96604562e-01 -8.64946395e-02 6.26289368e-01
5.51962435e-01 -3.96861136e-01 -9.50450376e-02 -2.52244741e-01
4.88608271e-01 -8.59921798e-02 -9.22676802e-01 1.80294323e+00
-5.02379835e-01 8.51424217e-01 -8.07238072e-02 -1.05115426e+00
7.95727909e-01 1.79075152e-01 -2.27493480e-01 -9.57032740e-01
2.57089436e-01 2.97731496e-02 -6.50100559e-02 -6.17999375e-01
3.68046671e-01 -2.46883780e-01 -1.41280934e-01 4.02479142e-01
6.19579017e-01 -6.19517326e-01 2.24297866e-01 5.91631234e-01
8.94549251e-01 8.50852057e-02 2.61145771e-01 1.65768132e-01
6.46078050e-01 4.19594526e-01 8.35405886e-02 7.60099292e-01
-3.64036590e-01 5.26189804e-01 4.14983213e-01 -2.51812488e-01
-1.20189512e+00 -1.34365010e+00 9.84674841e-02 1.36162543e+00
1.15188651e-01 7.89065510e-02 -7.17087984e-01 -1.06899691e+00
2.08390042e-01 9.44500029e-01 -8.46047401e-01 -2.94885248e-01
-3.08319658e-01 -4.39909026e-02 8.53664041e-01 6.76276743e-01
5.50000548e-01 -1.57670808e+00 -5.34170330e-01 -2.54907489e-01
-1.67921498e-01 -1.13387132e+00 -6.20272338e-01 1.56058609e-01
-8.04057002e-01 -1.14831996e+00 -8.23846042e-01 -1.08560753e+00
8.21636677e-01 8.52550864e-02 1.42195463e+00 -3.13983977e-01
-3.20254117e-01 9.20846701e-01 -1.59672722e-01 -3.15985292e-01
-7.62794316e-01 -4.27677274e-01 -7.17322296e-03 -1.12675264e-01
3.77590328e-01 -2.90681630e-01 -4.36900586e-01 1.25695169e-01
-1.21819568e+00 3.80142741e-02 6.76706374e-01 1.12160182e+00
7.30618060e-01 -5.12852371e-01 8.03673208e-01 -9.78503704e-01
7.84492195e-01 -2.07719445e-01 -8.53749573e-01 4.24299747e-01
-3.83423120e-01 3.28842729e-01 6.84268177e-01 -9.10836339e-01
-1.05465305e+00 1.44634098e-01 -5.37627004e-02 -1.03766966e+00
-4.06297296e-01 1.80996299e-01 -3.34140390e-01 -2.36111149e-01
1.09009469e+00 2.81007707e-01 1.39103204e-01 -1.14877045e-01
1.12382925e+00 3.85030359e-01 9.89633679e-01 -5.63758910e-01
1.04649878e+00 2.93521941e-01 -3.65301251e-01 -5.42907059e-01
-6.95282519e-01 -2.39119530e-01 -5.04339218e-01 -6.03903532e-02
1.02521515e+00 -9.70305145e-01 -8.92193556e-01 2.23808765e-01
-1.09816790e+00 -8.25653076e-01 -7.89773524e-01 1.56019464e-01
-6.96149707e-01 1.86622873e-01 -4.05127198e-01 -4.57403213e-01
-2.87558734e-01 -1.12205541e+00 7.80663311e-01 3.89310151e-01
-6.28030822e-02 -1.07145178e+00 2.53424272e-02 9.31662500e-01
2.16022685e-01 6.91298321e-02 1.09232891e+00 -7.05221772e-01
-6.44569635e-01 1.93077058e-01 -3.88718039e-01 7.52976060e-01
-2.75796987e-02 -5.89721620e-01 -1.12560117e+00 -2.68814743e-01
-1.97820961e-01 -1.32442355e+00 9.45605636e-01 1.67368248e-01
1.28191316e+00 -4.60168451e-01 1.82505757e-01 4.84555185e-01
1.33157063e+00 4.59344424e-02 5.62421322e-01 1.76241174e-01
9.04871881e-01 4.78183985e-01 5.86145520e-01 9.06961039e-02
5.39433777e-01 2.47546256e-01 7.16270864e-01 -1.16285414e-01
-4.93202686e-01 -4.26468164e-01 4.80439574e-01 1.62463591e-01
3.12295228e-01 -2.85301179e-01 -1.00429785e+00 8.99298072e-01
-1.43364465e+00 -8.78063142e-01 1.05883211e-01 1.95271242e+00
1.15537024e+00 -1.07328683e-01 6.44474700e-02 -1.95100233e-01
4.16986614e-01 5.76047264e-02 -1.12257981e+00 -4.98546362e-01
-2.89903224e-01 4.68820393e-01 5.21080673e-01 6.28851414e-01
-1.01151168e+00 1.25127637e+00 5.94650745e+00 7.73958564e-01
-8.16280484e-01 2.81099766e-01 4.49729145e-01 -5.59280999e-02
-6.85629427e-01 -2.04999775e-01 -4.36958104e-01 -8.86696801e-02
7.91579247e-01 -1.01699226e-01 6.15217626e-01 9.47588384e-01
-2.82625526e-01 -2.47926060e-02 -1.29643106e+00 1.23965049e+00
4.21876371e-01 -1.51019812e+00 6.06619775e-01 -2.58337677e-01
7.62198389e-01 3.32743954e-03 4.47706223e-01 7.51241148e-01
6.75050020e-01 -1.22658253e+00 5.49992204e-01 1.36038601e-01
1.22829413e+00 -5.37907064e-01 5.35209477e-01 2.63027549e-01
-6.41388655e-01 -2.52862245e-01 -4.83417571e-01 1.82513893e-01
-5.58444783e-02 -2.00929374e-01 -1.06427515e+00 1.18486129e-01
5.65504670e-01 2.28053674e-01 -9.53534782e-01 5.71119428e-01
-4.30324584e-01 5.93927920e-01 6.50121123e-02 1.52475089e-01
3.12712938e-01 5.17686568e-02 3.71352255e-01 8.26116562e-01
-1.76595569e-01 2.00826719e-01 -1.30838528e-03 8.91280413e-01
-5.54690123e-01 5.38684949e-02 -6.87269568e-01 -1.76636249e-01
2.26797044e-01 1.12376010e+00 -2.20335484e-01 -5.92297733e-01
-5.47176242e-01 1.30123687e+00 5.05644381e-01 7.01290250e-01
-7.46935427e-01 -1.51426524e-01 4.96419191e-01 -1.05393216e-01
3.31574827e-01 5.90753034e-02 -2.46348590e-01 -1.33295894e+00
-1.97830409e-01 -1.45959675e+00 7.66927898e-01 -1.31485474e+00
-1.33941686e+00 6.96537137e-01 -6.11385815e-02 -1.01813304e+00
-3.13119441e-01 -8.77363920e-01 -1.74318150e-01 7.80495703e-01
-1.52290332e+00 -1.29078138e+00 -6.07369184e-01 1.33288276e+00
8.45708728e-01 -4.73176122e-01 9.26755786e-01 -1.08589865e-02
-2.26107743e-02 8.68304968e-01 -1.60993069e-01 3.30076635e-01
9.51489866e-01 -1.50928771e+00 3.43137503e-01 7.23984599e-01
4.95493859e-01 2.75594145e-01 5.40484011e-01 -5.58705688e-01
-1.34133387e+00 -1.17755735e+00 2.92896628e-01 -9.12766218e-01
6.89600527e-01 -4.60680097e-01 -1.17496276e+00 8.84022713e-01
4.71061707e-01 1.87323928e-01 5.33626795e-01 -2.65722215e-01
-8.41880262e-01 -9.78289768e-02 -1.34962869e+00 6.44484043e-01
6.90666497e-01 -8.06964695e-01 -1.01867735e+00 3.26119274e-01
9.29136813e-01 -5.54560125e-01 -6.70064688e-01 1.97782055e-01
2.30239272e-01 -5.28420627e-01 1.07074559e+00 -9.24621463e-01
4.56481665e-01 -2.16899663e-01 -3.22752148e-01 -1.26818776e+00
-3.94540131e-02 -3.02165926e-01 -2.38180861e-01 9.53155398e-01
4.38288718e-01 -4.44384605e-01 8.97527695e-01 5.37453294e-01
1.74634591e-01 -3.00796121e-01 -8.50750923e-01 -7.32534647e-01
3.73791099e-01 -4.87358809e-01 3.24212402e-01 9.71039712e-01
-4.01092857e-01 6.34608984e-01 -2.84847140e-01 2.43242040e-01
6.95040464e-01 1.74370319e-01 9.74163234e-01 -9.34138715e-01
-3.06919694e-01 -1.65067554e-01 -3.13115209e-01 -1.04560387e+00
3.43674481e-01 -1.09800327e+00 -5.54900579e-02 -1.61719286e+00
9.95158330e-02 -1.30756959e-01 -1.39868036e-01 8.61863494e-01
-1.25438333e-01 5.67780018e-01 2.96982527e-01 -2.22951323e-02
-5.75557947e-01 6.98183835e-01 1.66213214e+00 -5.83286524e-01
-1.16534062e-01 -3.17106307e-01 -7.34929204e-01 5.59444904e-01
8.25785339e-01 -2.48091802e-01 -1.00454795e+00 -5.36048114e-01
1.11431675e-02 2.05584586e-01 8.79008591e-01 -6.58771455e-01
-1.34686418e-02 -2.30526254e-01 5.58273017e-01 -4.14974958e-01
4.44056958e-01 -7.18121350e-01 -4.19801444e-01 2.73350269e-01
-6.81370914e-01 -7.46942014e-02 6.55749679e-01 5.99092066e-01
-2.74726510e-01 -2.56972790e-01 9.83963549e-01 -2.07031026e-01
-1.17899251e+00 2.95955002e-01 -1.92167610e-01 5.36496937e-01
1.04913020e+00 -1.10484302e-01 -5.69841027e-01 -6.83914781e-01
-8.18265915e-01 5.36573589e-01 3.80095303e-01 6.12066388e-01
1.09028220e+00 -1.37676513e+00 -7.66293406e-01 1.11983381e-01
4.77445811e-01 8.80650356e-02 4.16716278e-01 3.03169757e-01
-5.06524146e-01 4.31248426e-01 -4.92895156e-01 -4.69241977e-01
-1.30712414e+00 1.16246438e+00 3.33968788e-01 -2.31165849e-02
-4.20121878e-01 1.14620936e+00 7.72930384e-01 -7.33328700e-01
2.99345791e-01 -3.08565021e-01 -3.44452411e-01 -6.06022030e-02
5.23862481e-01 -5.23881614e-02 -3.08401138e-01 -5.34219980e-01
-2.23848715e-01 3.00966948e-01 -2.15636820e-01 -3.56985211e-01
8.71577024e-01 -1.31550714e-01 2.41359785e-01 3.70794713e-01
1.03319526e+00 -1.79188013e-01 -1.43323290e+00 -4.73538846e-01
-4.34036762e-01 -2.72813976e-01 -1.98969930e-01 -1.12706625e+00
-9.58822668e-01 1.12025678e+00 6.92075968e-01 -2.73065954e-01
1.28156650e+00 3.62238467e-01 6.77156627e-01 7.91765511e-01
1.25205785e-01 -7.49112725e-01 6.66544199e-01 4.24373895e-01
1.25851500e+00 -1.64252079e+00 -4.09531772e-01 -1.05648125e-02
-1.30068433e+00 8.10339391e-01 1.13080215e+00 -7.63730258e-02
5.26910946e-02 -1.73396781e-01 5.87800622e-01 -1.10030085e-01
-7.59331048e-01 -3.56866598e-01 4.79824960e-01 1.07428896e+00
-1.28209606e-01 -5.98230734e-02 3.08669984e-01 5.35242140e-01
-2.63477117e-01 -1.98835373e-01 6.20066822e-01 5.89924634e-01
-2.48128548e-01 -9.86677468e-01 -5.81649184e-01 3.77997696e-01
1.60971377e-02 -3.66459399e-01 -3.96078855e-01 8.19931209e-01
1.56372011e-01 7.95345068e-01 1.49771599e-02 -1.72905937e-01
3.76198798e-01 2.46005177e-01 7.77007341e-01 -6.64828002e-01
-3.65096211e-01 -4.80767787e-01 -4.40581962e-02 -4.39632833e-01
-1.99178100e-01 -4.31454152e-01 -1.28392982e+00 8.86825919e-02
5.72405197e-02 1.98943843e-03 3.34240794e-01 7.83063233e-01
9.12650079e-02 4.81757462e-01 5.05488403e-02 -3.68802667e-01
-6.28184736e-01 -8.35904479e-01 -2.30578601e-01 6.66919649e-01
6.85181916e-01 -5.94100595e-01 -1.84543371e-01 5.04776537e-01] | [10.762890815734863, 1.7824578285217285] |
a74fe7b7-102d-466e-af96-3eab33602e24 | triangulation-learning-network-from-monocular-1 | 1906.01193 | null | https://arxiv.org/abs/1906.01193v1 | https://arxiv.org/pdf/1906.01193v1.pdf | Triangulation Learning Network: from Monocular to Stereo 3D Object Detection | In this paper, we study the problem of 3D object detection from stereo images, in which the key challenge is how to effectively utilize stereo information. Different from previous methods using pixel-level depth maps, we propose employing 3D anchors to explicitly construct object-level correspondences between the regions of interest in stereo images, from which the deep neural network learns to detect and triangulate the targeted object in 3D space. We also introduce a cost-efficient channel reweighting strategy that enhances representational features and weakens noisy signals to facilitate the learning process. All of these are flexibly integrated into a solid baseline detector that uses monocular images. We demonstrate that both the monocular baseline and the stereo triangulation learning network outperform the prior state-of-the-arts in 3D object detection and localization on the challenging KITTI dataset. | ['Jinglu Wang', 'Zengyi Qin', 'Yan Lu'] | 2019-06-04 | triangulation-learning-network-from-monocular | http://openaccess.thecvf.com/content_CVPR_2019/html/Qin_Triangulation_Learning_Network_From_Monocular_to_Stereo_3D_Object_Detection_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Qin_Triangulation_Learning_Network_From_Monocular_to_Stereo_3D_Object_Detection_CVPR_2019_paper.pdf | cvpr-2019-6 | ['3d-object-detection-from-stereo-images'] | ['computer-vision'] | [ 2.49231204e-01 -7.50774816e-02 8.58717933e-02 -2.73709476e-01
-8.32599044e-01 -6.00399971e-01 5.99847317e-01 -2.53124088e-01
-6.41199887e-01 2.21704528e-01 1.54749369e-02 -7.07127973e-02
3.01296771e-01 -4.83493030e-01 -1.06324279e+00 -5.16352594e-01
1.53702451e-02 4.72156852e-01 3.70668352e-01 1.61006123e-01
4.68466282e-01 7.94775188e-01 -1.55391812e+00 8.78831819e-02
3.17636162e-01 1.19593668e+00 4.06468481e-01 4.78554755e-01
5.08545190e-02 4.97265577e-01 -9.16134343e-02 1.04058109e-01
8.88583004e-01 1.70203805e-01 -5.03094018e-01 4.18003142e-01
1.09296966e+00 -9.06901300e-01 -7.32842028e-01 1.01645648e+00
5.84034383e-01 -2.55482048e-01 7.17428207e-01 -9.18809414e-01
-2.44305015e-01 -7.37173334e-02 -1.01225078e+00 3.05053473e-01
4.84940112e-01 2.82177716e-01 8.92091811e-01 -1.31033075e+00
7.89375424e-01 1.61153281e+00 6.35486484e-01 4.57766622e-01
-1.36656058e+00 -6.97167158e-01 2.81502187e-01 -7.02014863e-02
-1.38731015e+00 -5.55544853e-01 8.88675451e-01 -7.38388121e-01
1.14840341e+00 -4.03894365e-01 6.67146921e-01 8.82969558e-01
1.23283453e-02 1.17876482e+00 1.15809572e+00 -5.37073731e-01
-3.29194628e-02 -2.29566827e-01 7.30299205e-02 8.63110006e-01
2.92617202e-01 7.25538611e-01 -7.39948511e-01 3.18192393e-02
1.28834307e+00 4.82621267e-02 -1.75567701e-01 -1.12906229e+00
-1.23469782e+00 7.63697743e-01 7.60954440e-01 -2.17553392e-01
-3.75214845e-01 4.26981717e-01 1.06891815e-03 7.78660998e-02
4.95499998e-01 3.36191505e-01 -3.24874699e-01 3.70114446e-01
-6.74097776e-01 2.66610980e-01 2.66563207e-01 1.14454961e+00
1.02297950e+00 1.29735414e-02 2.11538561e-02 4.70288634e-01
4.57264185e-01 7.01391160e-01 -1.07970387e-01 -1.32551515e+00
4.84797567e-01 7.13195324e-01 2.02309370e-01 -5.32772005e-01
-4.29357558e-01 -5.65388560e-01 -3.91268462e-01 7.09405363e-01
4.41255152e-01 1.21061042e-01 -1.17434978e+00 1.44182432e+00
6.43207550e-01 2.92089790e-01 -1.45878822e-01 1.31986439e+00
7.92262733e-01 2.47561425e-01 -4.57697421e-01 4.57257897e-01
8.37256551e-01 -6.14944637e-01 -5.87722212e-02 -7.20748365e-01
2.96931833e-01 -7.74060786e-01 5.22255719e-01 1.12553306e-01
-1.21008432e+00 -5.84310532e-01 -1.09005797e+00 -4.60507929e-01
-1.56805709e-01 -8.23030397e-02 7.19605625e-01 2.39155158e-01
-1.11254215e+00 1.17804773e-01 -8.71474862e-01 -3.21178317e-01
5.48511446e-01 4.69102621e-01 -5.09932458e-01 -3.84103626e-01
-8.00860167e-01 1.01216495e+00 3.65771294e-01 -2.33717728e-02
-1.30722058e+00 -6.90008283e-01 -1.22848153e+00 -3.02823246e-01
3.07883412e-01 -1.03442895e+00 1.28689194e+00 -5.72777867e-01
-1.14676154e+00 1.61302567e+00 -2.30059370e-01 -4.14191872e-01
5.82695901e-01 -4.15451646e-01 4.83273506e-01 2.64340848e-01
3.66181254e-01 1.31992197e+00 9.87529874e-01 -1.34838009e+00
-1.09511232e+00 -7.91296184e-01 2.22341880e-01 6.08375728e-01
2.80424356e-01 -2.72854984e-01 -9.48972285e-01 -2.38639250e-01
7.15060949e-01 -8.68925631e-01 -3.06604534e-01 5.13399005e-01
-2.92236865e-01 -4.49709892e-02 7.87603557e-01 -1.95314527e-01
2.39781797e-01 -2.14693451e+00 3.29565793e-01 1.81853518e-01
4.16240633e-01 -2.44268849e-01 -1.17252395e-01 -5.60689829e-02
1.76044956e-01 -5.47957301e-01 6.62576465e-04 -4.26517129e-01
-7.68246427e-02 9.51184854e-02 -2.50174135e-01 9.91371453e-01
2.78162003e-01 1.07553220e+00 -9.05311227e-01 -3.21921170e-01
7.11001575e-01 6.54060364e-01 -7.71055341e-01 3.97209823e-02
-1.81671593e-03 5.20735443e-01 -5.42346120e-01 9.60589051e-01
8.73125434e-01 -1.85758382e-01 -4.80301797e-01 -1.77705973e-01
-2.86135584e-01 4.49450046e-01 -1.16506505e+00 2.09245563e+00
-2.98593462e-01 7.29368150e-01 4.41438347e-01 -6.89504743e-01
8.64745736e-01 -1.67737797e-01 4.03530657e-01 -7.06910908e-01
1.88196540e-01 2.91855276e-01 -4.21582490e-01 -8.39202553e-02
3.28904688e-01 -4.60665859e-02 -3.44818607e-02 2.03723550e-01
9.61423963e-02 -6.22691989e-01 -3.57198030e-01 8.92552435e-02
1.05530226e+00 3.19111198e-01 2.48977125e-01 -1.03545792e-01
3.45531553e-01 2.34688781e-02 4.23916578e-01 9.24960792e-01
-1.80620268e-01 8.18638265e-01 6.69377968e-02 -4.46781844e-01
-1.12513757e+00 -1.41018069e+00 -1.84783772e-01 5.33667624e-01
5.32264829e-01 1.71906993e-01 -3.06421608e-01 -6.26401126e-01
6.73284292e-01 6.38832971e-02 -4.43847239e-01 1.31490484e-01
-5.74877024e-01 -8.59321281e-02 1.36526078e-01 5.88942170e-01
6.28259838e-01 -5.23993194e-01 -9.11516368e-01 1.78781897e-01
-3.86611670e-02 -1.30222642e+00 -4.08613235e-01 5.45427144e-01
-8.69617939e-01 -1.12334073e+00 -6.25789165e-01 -8.40024173e-01
7.25221813e-01 1.07925618e+00 1.03289056e+00 -3.06033194e-01
-3.63021225e-01 7.08410501e-01 -3.89615968e-02 -4.41872239e-01
5.42405248e-02 3.27450372e-02 9.88318920e-02 -3.30959946e-01
6.45729721e-01 -4.37449008e-01 -8.03414226e-01 2.64160782e-01
-4.48345512e-01 1.90413043e-01 6.29991591e-01 8.62606466e-01
7.51077414e-01 -4.07105356e-01 -9.62675512e-02 -3.39952767e-01
-1.66701019e-01 -8.59081075e-02 -1.17562449e+00 -3.69616956e-01
-1.54855996e-01 4.02094834e-02 -1.31854609e-01 -1.48506954e-01
-6.42785192e-01 8.38863790e-01 6.29019886e-02 -8.88243914e-01
-8.60498995e-02 -1.20873447e-03 7.63841644e-02 -6.02518201e-01
6.04603946e-01 1.08636275e-01 -5.64161576e-02 -4.26124066e-01
3.74501884e-01 5.06977201e-01 7.12889791e-01 -4.17634040e-01
1.16313970e+00 1.08927798e+00 2.69496232e-01 -6.39073968e-01
-1.13582802e+00 -8.96982372e-01 -1.04652488e+00 -3.42764616e-01
8.29286098e-01 -1.66090846e+00 -6.32553637e-01 4.71666217e-01
-1.35936248e+00 -1.68380618e-01 -2.29696631e-01 5.81933796e-01
-7.83300400e-01 2.64249504e-01 -5.71874499e-01 -7.54909039e-01
-5.29353991e-02 -1.17411447e+00 1.78855836e+00 -1.26367688e-01
7.03390464e-02 -5.75480998e-01 -2.09816560e-01 2.43595749e-01
3.63832153e-02 2.72179633e-01 5.05593598e-01 1.63817257e-02
-1.26971710e+00 -2.82705992e-01 -5.59003651e-01 1.97607473e-01
-1.66044176e-01 -5.47092557e-01 -1.29876006e+00 -4.91920739e-01
1.06354631e-01 -4.84738827e-01 1.20284474e+00 7.30268121e-01
8.35062206e-01 3.14683706e-01 -4.58008885e-01 1.02621222e+00
1.25305378e+00 -9.28392783e-02 1.98423117e-01 4.91321802e-01
8.17325890e-01 6.36095285e-01 5.90891778e-01 2.01523855e-01
4.29522306e-01 7.26117790e-01 7.05137193e-01 -2.99754590e-01
-2.08480477e-01 -5.07353127e-01 2.49162704e-01 1.34272024e-01
2.82539487e-01 2.63392270e-01 -9.38248217e-01 5.92423737e-01
-1.71334159e+00 -7.42347002e-01 1.42262787e-01 2.13670039e+00
5.33267558e-01 4.39273119e-01 -1.38731927e-01 5.24536707e-02
5.82659304e-01 2.08673045e-01 -8.50588620e-01 3.13684583e-01
-2.53585488e-01 -1.32067308e-01 1.01235139e+00 7.71171749e-01
-1.39462423e+00 1.07110953e+00 6.69106531e+00 3.46283019e-01
-1.02795768e+00 -6.91420734e-02 2.96040982e-01 -3.44777256e-01
2.68307608e-02 -1.03829935e-01 -1.22686899e+00 3.76866162e-02
6.64760694e-02 3.02676827e-01 3.15263510e-01 8.31353664e-01
1.26110852e-01 -1.40116423e-01 -1.57726455e+00 1.40400434e+00
1.46331012e-01 -1.27395236e+00 -1.28926113e-01 2.19668195e-01
8.85267556e-01 7.21929073e-01 2.47605443e-01 -6.53543472e-02
5.43161511e-01 -8.07746053e-01 1.00521100e+00 1.75980404e-01
6.74561083e-01 -4.42514569e-01 3.13666046e-01 4.07392085e-01
-1.12789977e+00 -2.23153219e-01 -4.64994997e-01 -3.31394970e-01
1.11897551e-01 5.79926431e-01 -9.37176883e-01 9.23292637e-02
9.26342130e-01 1.02134609e+00 -3.78274411e-01 1.26084578e+00
-4.40958411e-01 -9.52409506e-02 -6.95544958e-01 3.66778791e-01
4.91625607e-01 7.53094107e-02 7.79637158e-01 9.42577958e-01
2.27467835e-01 4.78924848e-02 2.83844709e-01 9.74330604e-01
-4.94286492e-02 -5.24969697e-01 -1.07822597e+00 6.43624604e-01
5.84090114e-01 9.96603191e-01 -3.86831582e-01 -1.37537748e-01
-5.82416058e-01 9.34721112e-01 3.61387700e-01 3.84291440e-01
-3.02210420e-01 9.09204781e-02 8.59281659e-01 3.86458725e-01
6.92234516e-01 -5.69563329e-01 -3.76236349e-01 -1.27701581e+00
6.36607111e-02 -5.41618466e-01 1.61464944e-01 -9.89023268e-01
-1.35081720e+00 8.89857635e-02 -4.37099412e-02 -1.35402489e+00
-1.72250345e-01 -9.44872260e-01 -1.57220170e-01 8.83273125e-01
-2.02541399e+00 -1.21561539e+00 -4.21316475e-01 6.75987244e-01
4.76207823e-01 5.65318055e-02 1.49100870e-01 2.37411633e-01
7.86240473e-02 2.03220457e-01 -6.28526788e-03 1.06986150e-01
5.30543864e-01 -1.08197331e+00 7.13007152e-01 8.36914957e-01
1.05581976e-01 3.96571785e-01 4.16844189e-01 -4.94268447e-01
-1.66789031e+00 -1.01493311e+00 6.92904472e-01 -7.98098743e-01
4.20130819e-01 -8.55936468e-01 -5.68911135e-01 7.51931131e-01
-2.04125598e-01 1.40782967e-01 -1.34663478e-01 -1.44171104e-01
-5.73556602e-01 -5.24732582e-02 -1.06458843e+00 4.78554934e-01
1.58246922e+00 -8.23517680e-01 -7.60063291e-01 2.72249341e-01
6.08085573e-01 -1.00632262e+00 -2.29160219e-01 4.78816390e-01
6.23200834e-01 -9.23671246e-01 1.58805311e+00 -2.48249859e-01
8.62023309e-02 -5.48678756e-01 -4.87388343e-01 -1.11616910e+00
-2.43152037e-01 -4.10782307e-01 -1.73549596e-02 6.17974818e-01
5.75034656e-02 -4.02962446e-01 1.07309127e+00 3.00273240e-01
-1.89815566e-01 -2.58562744e-01 -1.39188719e+00 -6.24635696e-01
-2.68329624e-02 -4.33560312e-01 2.61371821e-01 5.85302532e-01
-4.36212569e-01 3.93887341e-01 -2.06290841e-01 4.34333175e-01
1.21602499e+00 2.99898833e-01 1.05038393e+00 -1.27879834e+00
-1.62448704e-01 -4.17208582e-01 -8.34127724e-01 -1.88292491e+00
2.54493266e-01 -9.16666448e-01 2.18003377e-01 -1.31201100e+00
2.02901185e-01 -2.36458614e-01 -1.79126337e-01 2.06395090e-01
1.46238089e-01 4.41181123e-01 2.07046866e-01 2.59953946e-01
-6.05930090e-01 4.37996179e-01 1.27806401e+00 -2.81418473e-01
-3.92353125e-02 -1.85994014e-01 -3.93254012e-01 7.61797190e-01
2.90562212e-01 -3.71867448e-01 -1.00798324e-01 -8.95144284e-01
2.57876758e-02 -6.26358613e-02 9.48487043e-01 -8.94024670e-01
4.40448165e-01 1.75583825e-01 8.72986555e-01 -1.25064731e+00
7.44266033e-01 -9.88460779e-01 -2.64605582e-01 4.29002017e-01
-2.90717900e-01 -1.60282224e-01 1.87756211e-01 6.13527656e-01
3.68765220e-02 1.19002327e-01 9.56038892e-01 -3.60741556e-01
-1.05083334e+00 4.84865934e-01 -1.73418775e-01 -6.71886057e-02
9.49079871e-01 -5.23083746e-01 -1.41097546e-01 -2.13961169e-01
-3.85344207e-01 4.76010680e-01 8.34079623e-01 4.22407568e-01
9.76375461e-01 -1.37637377e+00 -6.36037290e-01 5.72822571e-01
3.22356671e-01 4.64843661e-01 -1.76186293e-01 5.36908567e-01
-5.01785636e-01 6.46594524e-01 -2.04024374e-01 -1.30861747e+00
-9.47237074e-01 4.49750125e-01 6.44910455e-01 2.54912466e-01
-8.12675595e-01 1.06157231e+00 7.56334543e-01 -6.57374024e-01
7.20432580e-01 -5.20211875e-01 2.60793358e-01 -2.22214505e-01
2.99281359e-01 5.90503588e-02 -3.88552137e-02 -5.56085169e-01
-4.25578296e-01 1.06936646e+00 -7.28556216e-02 -2.19918504e-01
1.41393876e+00 -5.14775813e-01 1.91507459e-01 2.70445287e-01
1.32477760e+00 -3.98010403e-01 -2.02659869e+00 -8.11360240e-01
-1.57999292e-01 -9.05085206e-01 5.42451084e-01 -5.23079038e-01
-9.08242464e-01 1.09256792e+00 8.14665139e-01 -5.06910026e-01
7.26259649e-01 3.58253568e-01 4.29529935e-01 5.60580969e-01
5.98365784e-01 -8.38628948e-01 9.96406749e-02 8.16156089e-01
7.58637011e-01 -1.70252395e+00 -2.84488406e-03 -3.69231582e-01
-1.04983680e-01 9.01400924e-01 6.90598011e-01 -3.26787353e-01
4.80073631e-01 3.39480788e-01 5.90280294e-02 -5.11946201e-01
-4.35057282e-01 -6.52545750e-01 3.18446040e-01 9.04802561e-01
-1.62901320e-02 -3.36551517e-01 4.85694975e-01 -2.17612863e-01
1.56440347e-01 -2.51962990e-01 1.95045784e-01 1.07094789e+00
-6.82108045e-01 -7.00387239e-01 -5.90115964e-01 2.06610426e-01
-1.06588770e-02 -1.19855121e-01 -5.26691735e-01 8.97352397e-01
6.45447522e-02 6.95086241e-01 3.19167525e-01 -2.09945723e-01
5.29751480e-01 -2.65710771e-01 8.62427771e-01 -7.71738231e-01
-1.16865069e-01 3.19932610e-01 -1.61403909e-01 -8.17580163e-01
-5.90728343e-01 -8.47778499e-01 -9.90263999e-01 -2.75034122e-02
-3.00946921e-01 -4.64791924e-01 7.11091280e-01 7.46874392e-01
1.94951385e-01 -3.63112241e-02 7.05876410e-01 -1.69272816e+00
-4.65229183e-01 -5.49737096e-01 -4.53612059e-01 5.59453554e-02
8.44151318e-01 -1.00182402e+00 -4.48879242e-01 -2.53203392e-01] | [7.837341785430908, -2.6352646350860596] |
2aab104e-0686-4aeb-9322-0093bcf4a252 | flexibert-are-current-transformer | 2205.11656 | null | https://arxiv.org/abs/2205.11656v1 | https://arxiv.org/pdf/2205.11656v1.pdf | FlexiBERT: Are Current Transformer Architectures too Homogeneous and Rigid? | The existence of a plethora of language models makes the problem of selecting the best one for a custom task challenging. Most state-of-the-art methods leverage transformer-based models (e.g., BERT) or their variants. Training such models and exploring their hyperparameter space, however, is computationally expensive. Prior work proposes several neural architecture search (NAS) methods that employ performance predictors (e.g., surrogate models) to address this issue; however, analysis has been limited to homogeneous models that use fixed dimensionality throughout the network. This leads to sub-optimal architectures. To address this limitation, we propose a suite of heterogeneous and flexible models, namely FlexiBERT, that have varied encoder layers with a diverse set of possible operations and different hidden dimensions. For better-posed surrogate modeling in this expanded design space, we propose a new graph-similarity-based embedding scheme. We also propose a novel NAS policy, called BOSHNAS, that leverages this new scheme, Bayesian modeling, and second-order optimization, to quickly train and use a neural surrogate model to converge to the optimal architecture. A comprehensive set of experiments shows that the proposed policy, when applied to the FlexiBERT design space, pushes the performance frontier upwards compared to traditional models. FlexiBERT-Mini, one of our proposed models, has 3% fewer parameters than BERT-Mini and achieves 8.9% higher GLUE score. A FlexiBERT model with equivalent performance as the best homogeneous model achieves 2.6x smaller size. FlexiBERT-Large, another proposed model, achieves state-of-the-art results, outperforming the baseline models by at least 5.7% on the GLUE benchmark. | ['Niraj K. Jha', 'Shreshth Tuli', 'Bhishma Dedhia', 'Shikhar Tuli'] | 2022-05-23 | null | null | null | null | ['graph-similarity'] | ['graphs'] | [-2.62139380e-01 -1.80678740e-02 -4.72372770e-01 -1.31801754e-01
-7.80578792e-01 -4.33913261e-01 4.71713781e-01 -4.30144906e-01
-1.09892823e-01 3.19082230e-01 1.67862132e-01 -5.94935119e-01
-3.53310615e-01 -6.96689844e-01 -6.62383974e-01 -4.72045928e-01
-1.38341263e-01 5.11152089e-01 1.04836628e-01 -3.94491017e-01
1.36569858e-01 1.48355469e-01 -1.01729500e+00 1.18321352e-01
7.33748376e-01 1.11421466e+00 4.23386455e-01 2.97272563e-01
-9.88460630e-02 3.09373051e-01 -3.62684190e-01 -5.81055284e-01
4.87666607e-01 -3.08532231e-02 -5.37338138e-01 -5.14976680e-01
3.09649616e-01 -4.74481672e-01 -5.65829813e-01 6.93323612e-01
5.58716655e-01 -1.15858316e-01 4.77995157e-01 -1.21684968e+00
-8.25757921e-01 1.10761726e+00 -4.76887256e-01 5.13937771e-02
-2.73261160e-01 6.95493221e-02 1.39930606e+00 -1.06886232e+00
1.24181926e-01 1.47146809e+00 8.08538854e-01 6.10783458e-01
-1.46579528e+00 -1.07677507e+00 4.54893082e-01 8.82482752e-02
-1.63542080e+00 -5.12700319e-01 1.01660419e+00 -1.38399467e-01
1.25056648e+00 6.98898882e-02 5.15132368e-01 1.37564290e+00
2.73828894e-01 8.64864647e-01 1.04664695e+00 -4.62014854e-01
3.29994977e-01 1.13497607e-01 1.25403613e-01 8.44499469e-01
3.88410240e-01 2.62339681e-01 -5.39996326e-01 -3.85190338e-01
6.21737838e-01 -1.39095942e-02 -1.73438624e-01 -5.12001514e-01
-1.08978045e+00 9.29747820e-01 4.66633707e-01 2.81385422e-01
-2.03911260e-01 4.91261125e-01 2.03512892e-01 2.23652184e-01
2.59923071e-01 8.16543877e-01 -6.57519996e-01 -1.80902973e-01
-1.08065796e+00 1.43144131e-01 8.00627291e-01 1.03358495e+00
5.51990509e-01 4.58853960e-01 -2.07192793e-01 9.10672784e-01
5.85013688e-01 1.43731132e-01 5.18329322e-01 -7.17968464e-01
7.56395042e-01 6.63006306e-01 -4.03452694e-01 -8.61536741e-01
-2.71097928e-01 -9.23367143e-01 -9.36891913e-01 -2.29185715e-01
-5.71462959e-02 -8.89844447e-02 -9.21195328e-01 1.98235214e+00
5.88372797e-02 3.68288666e-01 -1.80631019e-02 6.95050597e-01
4.69049305e-01 7.04558432e-01 -9.32999104e-02 2.74842352e-01
1.25521076e+00 -1.22369611e+00 -3.30309659e-01 -3.96859497e-01
5.55887759e-01 -5.43370783e-01 1.32633710e+00 4.23792064e-01
-8.95045757e-01 -4.32632148e-01 -1.43728173e+00 3.07037503e-01
-2.63488382e-01 2.81533301e-01 6.84217572e-01 7.91884661e-01
-1.38002837e+00 7.17150331e-01 -1.01201987e+00 -2.47027591e-01
1.67650267e-01 6.39641464e-01 4.69928719e-02 1.72878280e-02
-1.41381502e+00 8.77463460e-01 4.92808551e-01 -8.27899959e-04
-1.09980023e+00 -7.70618618e-01 -6.31598234e-01 3.77862573e-01
5.25257289e-01 -8.96691799e-01 1.09832323e+00 -4.37010854e-01
-1.84152019e+00 2.48733014e-02 1.59192294e-01 -5.75079560e-01
5.04897796e-02 -2.21943721e-01 -3.87290269e-01 -1.39778033e-01
-4.36375409e-01 6.19777620e-01 8.40385020e-01 -1.22171056e+00
-1.23276785e-01 -1.53296620e-01 3.92760962e-01 -4.15289961e-03
-1.08293641e+00 -1.16076857e-01 -7.96639144e-01 -8.25673282e-01
-6.09451048e-02 -1.25808370e+00 -3.70885998e-01 -1.66219100e-01
-7.69109249e-01 -1.65976405e-01 7.22216606e-01 -4.75996643e-01
1.94817364e+00 -1.91125154e+00 2.89833188e-01 3.45870197e-01
4.03994679e-01 3.15361798e-01 -3.84182692e-01 5.95833302e-01
-2.49094050e-02 6.15917802e-01 8.55739415e-02 -6.27992272e-01
3.07714850e-01 2.58301228e-01 -1.55754685e-01 8.31989497e-02
1.85970381e-01 8.67385209e-01 -4.68737334e-01 -2.43488804e-01
4.23084088e-02 5.92779994e-01 -8.96261454e-01 2.46850893e-01
-1.08822353e-01 -1.71363037e-02 -6.71465814e-01 6.26736283e-01
4.74405229e-01 -6.25756145e-01 3.89401942e-01 -4.49038088e-01
8.43717977e-02 3.43297660e-01 -1.08057153e+00 1.76217079e+00
-7.60107338e-01 4.48685884e-01 -9.60473865e-02 -1.03112578e+00
1.15476537e+00 4.55659777e-02 2.80808091e-01 -5.56317985e-01
1.21495105e-01 2.77001500e-01 1.23186871e-01 -7.80151412e-02
4.71794099e-01 4.33380157e-01 -4.70518582e-02 4.72973853e-01
6.33565485e-02 3.73798788e-01 -5.63982166e-02 1.45222887e-01
1.43058515e+00 9.44769979e-02 8.59332457e-02 -3.70982051e-01
1.57721341e-01 -5.23569345e-01 5.37761629e-01 7.30779171e-01
1.22517854e-01 3.93693894e-01 3.79901081e-01 -3.49641919e-01
-1.15099490e+00 -9.75223064e-01 -5.15294597e-02 9.74219143e-01
-9.30957589e-03 -9.56524968e-01 -6.86590672e-01 -5.22942066e-01
5.54624796e-02 8.55421424e-01 -4.37966257e-01 -6.06736958e-01
-7.19819009e-01 -9.27073419e-01 6.69499099e-01 5.18996775e-01
6.59172237e-01 -3.59915018e-01 -4.68711913e-01 1.61061019e-01
8.11661407e-02 -9.03543890e-01 -4.90022928e-01 2.16470882e-01
-9.34300721e-01 -6.16975427e-01 -6.59411430e-01 -5.19426823e-01
3.56331468e-01 -2.42401902e-02 1.20024908e+00 6.11969866e-02
1.82127595e-01 1.04705699e-01 -3.13467741e-01 3.57846357e-02
-4.36465770e-01 5.86697936e-01 1.59807727e-01 2.29243301e-02
1.12452336e-01 -7.54967451e-01 -8.48150671e-01 4.13525701e-01
-8.32854271e-01 1.96330577e-01 1.14575386e+00 1.14499891e+00
4.34432715e-01 -8.74806121e-02 5.40817976e-01 -6.15286469e-01
8.72088909e-01 -5.93842924e-01 -4.92005169e-01 5.53532898e-01
-1.40652752e+00 5.99649012e-01 8.37902009e-01 -6.84649110e-01
-6.41950190e-01 -3.51018965e-01 2.52860263e-02 -9.69378293e-01
4.20119554e-01 7.42012441e-01 -9.31950361e-02 -9.38794017e-02
4.94974434e-01 1.21437535e-01 1.26171659e-03 -7.80942202e-01
2.53221303e-01 5.75009823e-01 2.97964667e-03 -8.10239255e-01
8.80316854e-01 -1.73514754e-01 -1.74312875e-01 -4.98028517e-01
-4.61282104e-01 -1.27607971e-01 -1.48844108e-01 1.69203639e-01
2.71902740e-01 -9.65577483e-01 -4.47960347e-01 1.48175746e-01
-9.06822383e-01 -3.28214407e-01 2.60685086e-01 4.41267997e-01
-2.23027572e-01 1.37165517e-01 -5.17723203e-01 -7.46123314e-01
-7.14228153e-01 -1.46826446e+00 1.06069314e+00 -1.31871432e-01
-2.47864157e-01 -8.84441495e-01 -4.51778211e-02 1.89580441e-01
8.55893910e-01 3.22841480e-02 1.33597577e+00 -8.07403147e-01
-7.46467412e-01 -3.92673649e-02 2.59508360e-02 4.68074322e-01
-1.35646343e-01 -5.62712625e-02 -6.92260504e-01 -6.51950002e-01
-2.17369273e-01 -1.92219213e-01 8.11876297e-01 2.94960052e-01
1.36452150e+00 -5.31742573e-01 -4.94521141e-01 8.51056457e-01
1.50993741e+00 2.89681286e-01 4.04970884e-01 4.82354105e-01
5.80665231e-01 9.53681320e-02 2.21211284e-01 3.79788369e-01
5.63795626e-01 1.12244296e+00 4.78525370e-01 4.85021546e-02
-1.44125268e-01 -4.50322151e-01 6.21799231e-01 1.39531732e+00
2.22495750e-01 -3.74230593e-01 -1.07887566e+00 3.80668402e-01
-1.75179780e+00 -5.08671880e-01 5.83197951e-01 2.02630901e+00
7.66653955e-01 3.99947196e-01 -3.74748223e-02 1.02923267e-01
5.44539630e-01 1.12808138e-01 -7.94027269e-01 -4.00369853e-01
1.06117673e-01 2.65066445e-01 4.94147092e-01 1.07773766e-01
-7.43235767e-01 7.99677551e-01 6.42168474e+00 1.11063194e+00
-1.20751119e+00 1.49459571e-01 7.98240840e-01 -2.48216733e-01
-5.47630310e-01 2.00824380e-01 -1.11178446e+00 5.17706752e-01
1.44238985e+00 -1.57651544e-01 6.68015242e-01 9.84387755e-01
1.28047705e-01 5.30697227e-01 -1.20054138e+00 9.59669471e-01
1.43444210e-01 -1.48473418e+00 2.04735875e-01 1.67706817e-01
6.07371926e-01 1.30430952e-01 3.06038022e-01 7.58018494e-01
2.77941763e-01 -9.62737501e-01 7.94107020e-01 3.09888333e-01
7.76541710e-01 -8.53188217e-01 4.20114517e-01 1.63844332e-01
-1.21881461e+00 -5.06630421e-01 -2.83164680e-01 2.47523710e-01
1.69786103e-02 4.83633995e-01 -8.60929489e-01 5.47453940e-01
7.64856398e-01 4.46143657e-01 -7.43427277e-01 9.39250767e-01
9.47929546e-02 9.89981592e-01 -4.62682694e-01 -3.61765116e-01
3.34576666e-01 8.56527090e-02 5.80832005e-01 9.65389371e-01
6.15932941e-01 -5.20702958e-01 2.61736095e-01 1.11013198e+00
-1.90565288e-01 -4.34670448e-02 -4.78084058e-01 -1.18844062e-01
1.10753655e+00 1.09825814e+00 -2.02847362e-01 -6.47922829e-02
-3.50447506e-01 4.59985942e-01 5.67800701e-01 4.47759598e-01
-1.07522082e+00 -3.02769780e-01 6.13203406e-01 1.31100833e-01
4.22116607e-01 -3.31179649e-01 -1.24628894e-01 -1.12123358e+00
9.86963958e-02 -1.16376066e+00 2.16347337e-01 -5.38678765e-01
-1.18890464e+00 1.03740680e+00 3.93637151e-01 -1.07323027e+00
-4.73170280e-01 -6.16645515e-01 -4.70074415e-01 8.26026618e-01
-1.54122722e+00 -1.20451415e+00 2.46049669e-02 4.05437648e-01
5.28748453e-01 -5.88684261e-01 8.85389984e-01 4.25455213e-01
-1.00340569e+00 1.25640082e+00 3.35327923e-01 -1.36528283e-01
4.37412918e-01 -8.92562449e-01 5.77823699e-01 7.25177944e-01
2.25519657e-01 1.09421384e+00 4.92704898e-01 -2.75213808e-01
-1.87344873e+00 -1.18482161e+00 4.75622356e-01 -1.42789930e-01
8.10811043e-01 -6.67817891e-01 -6.97864294e-01 4.85906124e-01
2.91346282e-01 -3.29528511e-01 6.10515893e-01 3.10450286e-01
-5.51679611e-01 -4.42913502e-01 -6.95708454e-01 8.88083637e-01
1.06932354e+00 -4.59174663e-01 -2.92348921e-01 5.38676120e-02
1.15489721e+00 -2.42332980e-01 -1.08822978e+00 5.39320111e-01
6.23940825e-01 -7.65395701e-01 1.14773476e+00 -4.97427821e-01
2.23979041e-01 -3.19064707e-02 -3.72386515e-01 -1.44983613e+00
-5.07005394e-01 -9.96024132e-01 -7.30866611e-01 1.21042538e+00
7.15267301e-01 -7.90635169e-01 7.89404154e-01 5.82986355e-01
-3.50299567e-01 -1.56237721e+00 -1.00705183e+00 -1.16271710e+00
2.38209277e-01 -4.50828999e-01 8.87238860e-01 7.13316798e-01
-2.61226714e-01 4.90224332e-01 -4.94846195e-01 4.42122063e-03
4.63940978e-01 1.01008855e-01 6.02016270e-01 -1.07248795e+00
-6.58468664e-01 -6.76962435e-01 -1.81439489e-01 -1.34157920e+00
2.27042034e-01 -9.86196280e-01 -3.44381154e-01 -1.23987067e+00
1.25186950e-01 -8.41729283e-01 -5.82032740e-01 6.75423503e-01
8.06521550e-02 -2.01004162e-01 2.55609035e-01 3.06937963e-01
-2.34869272e-01 7.76696682e-01 9.16204154e-01 -2.91315585e-01
-1.28830701e-01 -1.98996231e-01 -9.43698287e-01 2.93742448e-01
9.56184387e-01 -3.55427206e-01 -8.41892719e-01 -7.49988437e-01
2.85571158e-01 -2.28908196e-01 1.72249392e-01 -1.07052004e+00
2.80449033e-01 8.93455520e-02 1.53208394e-02 -3.82470161e-01
5.81861734e-01 -8.40324342e-01 3.97538066e-01 4.35394913e-01
-5.16105056e-01 7.15264261e-01 2.76712716e-01 5.63697934e-01
-1.32847242e-02 -2.01619819e-01 4.80443537e-01 1.95360929e-01
-5.45916378e-01 4.95374411e-01 1.10884756e-02 -5.89866526e-02
7.07866967e-01 -1.85661510e-01 -3.47466946e-01 -2.84221917e-01
-3.16648722e-01 2.79076308e-01 9.32306200e-02 6.90752447e-01
7.25903869e-01 -1.66498709e+00 -5.46059430e-01 9.42469910e-02
-1.37065062e-02 -1.98954016e-01 1.44842006e-02 7.79372752e-01
-1.17593654e-01 5.86393058e-01 2.03484386e-01 -5.64238966e-01
-8.64611268e-01 4.77273375e-01 3.09911966e-01 -7.32891977e-01
-4.25800741e-01 7.06354558e-01 3.54178287e-02 -4.94595081e-01
1.93697840e-01 -3.55888128e-01 -1.51800908e-04 -3.74357283e-01
1.44994989e-01 3.05007905e-01 1.56329870e-01 -3.13698977e-01
-2.66243368e-01 4.03091311e-01 -2.28502393e-01 1.14928678e-01
1.26985216e+00 -1.28523475e-02 8.90252739e-02 2.50470489e-01
1.44326270e+00 -2.81733215e-01 -1.02789891e+00 -3.69796008e-01
3.59875299e-02 -3.22328180e-01 5.36899984e-01 -8.98844957e-01
-1.35606873e+00 7.91433096e-01 7.08027363e-01 6.43259957e-02
1.11046243e+00 -2.53942579e-01 8.70193779e-01 5.78794718e-01
5.94891787e-01 -9.81336176e-01 2.62008220e-01 6.14632487e-01
9.07381713e-01 -7.67785013e-01 -1.87809944e-01 -5.11132367e-02
-3.90231162e-01 1.07613039e+00 8.72783244e-01 4.86238711e-02
7.66084075e-01 3.64925385e-01 -2.77159035e-01 -1.07378177e-01
-1.36740017e+00 3.33488554e-01 4.62550461e-01 2.54115760e-01
3.19652736e-01 -1.15942666e-02 7.32377097e-02 8.35159004e-01
-3.09389979e-01 -3.42340738e-01 1.22807622e-02 7.20758379e-01
-5.27479388e-02 -1.33231318e+00 -2.03745544e-01 7.30746627e-01
-2.30581552e-01 -4.36771482e-01 -1.02768034e-01 9.07519400e-01
-2.47777984e-01 8.76198590e-01 -1.56367704e-01 -1.03890502e+00
3.86601985e-01 2.44620442e-02 1.54507145e-01 -3.01208764e-01
-6.47407889e-01 1.19912155e-01 2.37315357e-01 -6.71080947e-01
-8.38029832e-02 -2.15130612e-01 -7.98810005e-01 -4.46397066e-01
-7.03617394e-01 3.98503840e-02 7.04141736e-01 5.81419647e-01
9.26862299e-01 6.82837427e-01 6.47713244e-01 -7.36833453e-01
-1.23224759e+00 -9.35501158e-01 -2.38931388e-01 -1.13346681e-01
7.73293078e-02 -9.34401035e-01 -1.93773001e-01 -6.87939525e-01] | [8.714158058166504, 3.4158711433410645] |
28822391-5dba-436f-80fe-b3c9ba89c7a4 | uda-cope-unsupervised-domain-adaptation-for | 2111.12580 | null | https://arxiv.org/abs/2111.12580v2 | https://arxiv.org/pdf/2111.12580v2.pdf | UDA-COPE: Unsupervised Domain Adaptation for Category-level Object Pose Estimation | Learning to estimate object pose often requires ground-truth (GT) labels, such as CAD model and absolute-scale object pose, which is expensive and laborious to obtain in the real world. To tackle this problem, we propose an unsupervised domain adaptation (UDA) for category-level object pose estimation, called UDA-COPE. Inspired by recent multi-modal UDA techniques, the proposed method exploits a teacher-student self-supervised learning scheme to train a pose estimation network without using target domain pose labels. We also introduce a bidirectional filtering method between the predicted normalized object coordinate space (NOCS) map and observed point cloud, to not only make our teacher network more robust to the target domain but also to provide more reliable pseudo labels for the student network training. Extensive experimental results demonstrate the effectiveness of our proposed method both quantitatively and qualitatively. Notably, without leveraging target-domain GT labels, our proposed method achieved comparable or sometimes superior performance to existing methods that depend on the GT labels. | ['Kuk-Jin Yoon', 'In So Kweon', 'Ukcheol Shin', 'Jaesung Choe', 'Inkyu Shin', 'Byeong-Uk Lee', 'Taeyeop Lee'] | 2021-11-24 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Lee_UDA-COPE_Unsupervised_Domain_Adaptation_for_Category-Level_Object_Pose_Estimation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Lee_UDA-COPE_Unsupervised_Domain_Adaptation_for_Category-Level_Object_Pose_Estimation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['6d-pose-estimation-using-rgbd'] | ['computer-vision'] | [ 4.18967977e-02 1.27148584e-01 -1.11723594e-01 -5.74735343e-01
-1.04587281e+00 -5.82847834e-01 4.71890748e-01 5.30766835e-03
-2.11659044e-01 4.37462032e-01 -2.47397542e-01 7.92488605e-02
-1.80735633e-01 -8.02459359e-01 -1.11936855e+00 -5.87222219e-01
3.87972236e-01 9.95413065e-01 5.49946249e-01 1.62147358e-01
1.54677629e-01 7.71184087e-01 -1.42992139e+00 -2.47931942e-01
8.30636024e-01 1.33962858e+00 5.06496012e-01 7.18690157e-02
5.27277738e-02 4.71737921e-01 -4.37627733e-01 -1.14701517e-01
4.30502832e-01 4.20636423e-02 -7.32051909e-01 3.99238020e-01
7.21135974e-01 -3.90520006e-01 -8.42182264e-02 1.02786326e+00
2.87050843e-01 2.10482195e-01 8.75007033e-01 -1.31190586e+00
-4.20588046e-01 2.79397309e-01 -6.52993679e-01 -3.88601005e-01
1.23485789e-01 1.11629769e-01 8.21519852e-01 -1.07791197e+00
6.82823181e-01 1.25137067e+00 7.88793266e-01 3.00194323e-01
-1.24223888e+00 -9.97995377e-01 3.58486980e-01 9.32715461e-02
-1.54635894e+00 -1.96032003e-02 1.23085380e+00 -5.05120516e-01
3.26643765e-01 -2.30433252e-02 4.45535034e-01 1.00316131e+00
-2.89585948e-01 7.69316196e-01 1.16855276e+00 -2.41876513e-01
2.02043742e-01 3.06745827e-01 -1.75194651e-01 7.13882387e-01
1.92413419e-01 -1.25086308e-01 -3.97857815e-01 -2.75027454e-02
1.03916776e+00 -2.80918423e-02 -6.93950728e-02 -1.21614110e+00
-1.37785625e+00 6.27586424e-01 8.63783181e-01 -8.45881552e-02
-3.36069018e-01 9.99146104e-02 8.75140280e-02 -8.25365558e-02
4.74319220e-01 2.76280433e-01 -6.44236565e-01 2.19135553e-01
-5.64814746e-01 1.06887504e-01 4.48139369e-01 1.31533873e+00
1.12043691e+00 -4.23209183e-02 6.06531575e-02 7.21602440e-01
7.01824367e-01 7.39341974e-01 2.19797224e-01 -8.62813532e-01
4.72620368e-01 7.65969992e-01 3.58392090e-01 -1.10056674e+00
-3.43383610e-01 -5.95429122e-01 -6.37678862e-01 1.59000039e-01
6.09250724e-01 2.45030060e-01 -9.03933227e-01 1.73969841e+00
8.72519612e-01 3.93705308e-01 -1.51146039e-01 1.09509921e+00
7.24803507e-01 3.02833587e-01 -3.55901234e-02 7.60253798e-03
1.05067861e+00 -8.10251653e-01 -2.05333978e-01 -2.20256239e-01
5.17105639e-01 -8.40399265e-01 1.09284735e+00 2.45647907e-01
-7.87402928e-01 -8.20213556e-01 -9.28935885e-01 5.95687367e-02
-1.95035771e-01 5.71555793e-01 4.77259487e-01 3.77415448e-01
-5.63478827e-01 5.47709227e-01 -1.05644953e+00 -3.17276597e-01
5.78420043e-01 5.23888111e-01 -3.45038891e-01 -8.95242766e-02
-6.50151253e-01 8.79123509e-01 6.11642003e-01 5.28070666e-02
-1.02069736e+00 -8.02242935e-01 -8.88146877e-01 -2.72352338e-01
5.68769336e-01 -5.03830194e-01 1.35418522e+00 -5.90159059e-01
-1.60055542e+00 6.72852993e-01 1.02544688e-01 -1.36413172e-01
7.18193591e-01 -4.18419927e-01 1.01241775e-01 1.26983881e-01
1.93879575e-01 9.22843575e-01 1.00181043e+00 -1.69203699e+00
-5.35246730e-01 -6.29587054e-01 -6.16992526e-02 2.09069997e-01
-4.23570693e-01 -4.63975877e-01 -5.07617712e-01 -6.03374600e-01
6.86130106e-01 -1.07319176e+00 -1.31519765e-01 4.97640640e-01
-4.33121115e-01 -5.94159484e-01 1.24130106e+00 -3.13558638e-01
4.24363583e-01 -1.98838282e+00 6.71315119e-02 3.71134847e-01
1.26803339e-01 3.33070159e-02 -1.09929793e-01 2.92441081e-02
6.46117656e-03 -4.90493625e-01 -6.30836561e-02 -4.54864591e-01
6.90021664e-02 3.16626519e-01 -2.77945191e-01 8.36319208e-01
2.59274334e-01 7.69790113e-01 -9.39211071e-01 -6.74568474e-01
4.49998528e-01 4.24673587e-01 -5.39431930e-01 4.27704394e-01
-5.88749290e-01 9.21334445e-01 -7.22137094e-01 6.95067585e-01
9.63911831e-01 -4.63608176e-01 -9.35911313e-02 -6.91537321e-01
5.09884953e-02 1.78195864e-01 -1.51649487e+00 1.90361202e+00
-5.49435914e-01 2.23475277e-01 -8.77406225e-02 -1.01480520e+00
1.16386378e+00 8.39978755e-02 5.79258204e-01 -2.94487476e-01
3.41426551e-01 2.84460545e-01 -1.85073301e-01 -2.23898336e-01
2.23572001e-01 -1.02781795e-01 -4.52297516e-02 2.68797398e-01
4.27264124e-01 -4.55119669e-01 -1.38537556e-01 -4.67563607e-02
5.39928913e-01 6.04675233e-01 1.27199814e-01 -1.34033293e-01
6.16155982e-01 9.68657359e-02 7.30186939e-01 5.40393114e-01
-1.40474871e-01 7.02331841e-01 -4.67298441e-02 -3.08508456e-01
-8.82270575e-01 -1.19707584e+00 -2.95439929e-01 9.19192195e-01
6.08161926e-01 -6.12359755e-02 -5.26414812e-01 -9.67144787e-01
2.32013628e-01 4.51207817e-01 -3.66134197e-01 -7.72455856e-02
-6.26488626e-01 -2.90422112e-01 7.81900138e-02 9.19586182e-01
5.99954545e-01 -7.94657350e-01 -2.69200474e-01 1.99190989e-01
-1.34414226e-01 -1.33023691e+00 -4.78676140e-01 8.19711387e-02
-9.60902572e-01 -1.05126894e+00 -7.09231853e-01 -8.64673853e-01
9.55284119e-01 1.91249415e-01 8.50409091e-01 -2.13198557e-01
-1.76572632e-02 4.64009047e-01 -2.05105811e-01 -5.56698024e-01
-1.74527422e-01 7.52970055e-02 3.85254711e-01 1.20615818e-01
3.10565621e-01 -7.85129309e-01 -6.15187943e-01 7.26169109e-01
-5.97279608e-01 1.34375036e-01 8.27311456e-01 6.11734748e-01
8.81917894e-01 -9.24265459e-02 6.02785528e-01 -7.73692191e-01
-8.84273276e-02 -1.51327178e-01 -9.63428080e-01 1.09417848e-01
-5.50717771e-01 -6.65851403e-03 4.04253453e-01 -7.18235731e-01
-1.09796929e+00 6.70538783e-01 -6.56807646e-02 -8.31693769e-01
-4.15083259e-01 2.35207781e-01 -3.80386204e-01 -5.90779066e-01
6.37005031e-01 1.83958903e-01 -8.83361325e-02 -7.24433899e-01
2.91670710e-01 5.49305141e-01 8.12047362e-01 -7.83357382e-01
1.49662888e+00 6.08068168e-01 9.29833129e-02 -3.19177181e-01
-1.07627261e+00 -6.12457454e-01 -1.14733171e+00 -3.41533005e-01
7.84666479e-01 -1.15838861e+00 -8.46478224e-01 4.45520997e-01
-1.26650190e+00 -1.03739962e-01 -1.30747169e-01 6.75699055e-01
-6.18305027e-01 2.13106558e-01 -2.41087258e-01 -5.35327435e-01
4.44164313e-02 -1.13259566e+00 1.44059575e+00 1.82941362e-01
1.15436733e-01 -7.38583982e-01 -1.95299283e-01 5.09017646e-01
8.30275565e-02 3.19186628e-01 5.92257321e-01 -6.72653735e-01
-9.63100493e-01 -2.82140821e-01 -4.74287927e-01 2.31464013e-01
2.72980332e-01 -4.03650463e-01 -9.16536033e-01 -4.29961711e-01
7.35609839e-03 -5.87726116e-01 2.68957049e-01 9.77229849e-02
1.33723378e+00 3.91363725e-02 -4.15316224e-01 4.91744190e-01
1.27804005e+00 -2.53643841e-01 -4.41707969e-02 3.09837371e-01
1.06326234e+00 5.40451407e-01 1.13251805e+00 4.48874533e-01
5.71336925e-01 8.63412380e-01 7.48165071e-01 -5.04423752e-02
-1.47479996e-01 -5.34527600e-01 1.60846382e-03 9.44889784e-01
2.78486349e-02 1.89656034e-01 -9.50634360e-01 4.80184913e-01
-1.71081638e+00 -1.94466144e-01 -9.09872353e-02 2.09171081e+00
8.29311550e-01 2.37455800e-01 1.03655748e-01 1.45004243e-02
6.18955255e-01 -1.23017989e-01 -8.45978260e-01 5.76308072e-01
2.36690894e-01 -3.49560380e-02 5.51826060e-01 1.06410496e-01
-1.25773501e+00 9.27209616e-01 5.06150961e+00 9.64435458e-01
-1.17562985e+00 2.42373034e-01 8.74731019e-02 3.22777838e-01
5.79263680e-02 -2.02818543e-01 -8.16973925e-01 2.43885890e-01
3.90869886e-01 1.11045271e-01 -3.96425836e-02 1.49384725e+00
-6.28960878e-02 1.97465415e-03 -1.32218122e+00 9.82524097e-01
-4.03882004e-02 -1.18801486e+00 -5.72967343e-02 -5.28632570e-03
8.33011568e-01 5.71325049e-02 -7.37140551e-02 3.97074401e-01
4.37010169e-01 -5.31102419e-01 9.36940730e-01 1.27773434e-01
7.45899975e-01 -7.35391498e-01 6.30441785e-01 6.50272250e-01
-1.49862480e+00 9.87544507e-02 -4.45529789e-01 1.68815091e-01
-7.21230879e-02 2.92451888e-01 -1.10260594e+00 6.95328832e-01
7.97683716e-01 7.99752891e-01 -7.86857247e-01 1.07364559e+00
-3.31551641e-01 4.58796084e-01 -6.30423427e-01 1.70663670e-01
1.09997861e-01 -8.37030709e-02 6.72381878e-01 6.68811142e-01
2.26528332e-01 -1.08810253e-01 6.49467289e-01 8.96728992e-01
-1.62838593e-01 -6.74855113e-02 -3.66502225e-01 2.54406422e-01
6.55464113e-01 1.32245755e+00 -9.92696643e-01 -1.51463330e-01
-2.97518224e-01 7.35685289e-01 3.11426044e-01 1.22736409e-01
-8.02572250e-01 -9.75625291e-02 4.23997790e-01 9.76101160e-02
4.78111953e-01 -3.31180453e-01 -3.36240679e-01 -1.05516517e+00
1.82630807e-01 -5.53187013e-01 1.90382645e-01 -8.08321416e-01
-1.42235374e+00 4.44081187e-01 2.81654119e-01 -1.84038579e+00
-2.04652235e-01 -6.84838951e-01 -3.10515761e-01 6.75020933e-01
-1.44021285e+00 -1.71763349e+00 -5.95033348e-01 5.31723559e-01
4.46174055e-01 -9.32348706e-03 5.21884799e-01 3.62549812e-01
-2.74596632e-01 6.07898593e-01 -2.23103520e-02 2.95864493e-01
8.50768387e-01 -1.22819018e+00 3.63422669e-02 4.47782904e-01
2.69788146e-01 4.66608852e-01 7.51945198e-01 -6.22861028e-01
-1.45069277e+00 -1.48387432e+00 2.16358706e-01 -7.07048535e-01
6.48789048e-01 -5.55804312e-01 -9.87754285e-01 8.75603676e-01
-4.98215467e-01 3.67049694e-01 1.60754248e-01 1.16433077e-01
-4.24892098e-01 -3.07283223e-01 -1.26680517e+00 1.79977715e-01
9.67022598e-01 -5.22409856e-01 -7.15322912e-01 6.42148912e-01
7.38558471e-01 -9.89098072e-01 -9.52629447e-01 8.09246540e-01
4.41634566e-01 -4.68528777e-01 1.12864971e+00 -2.44588763e-01
8.63162875e-02 -6.82067692e-01 -8.45514685e-02 -1.21317255e+00
-1.22112766e-01 3.23964506e-02 -7.49509186e-02 1.37138295e+00
4.29344177e-02 -5.43578804e-01 1.20221281e+00 4.29051906e-01
-2.47492313e-01 -7.09103286e-01 -9.61103261e-01 -1.18518829e+00
5.21394722e-02 -6.26863241e-01 6.82814777e-01 9.62177098e-01
-7.31733859e-01 1.11317627e-01 -1.02220938e-01 8.50585043e-01
9.07263577e-01 4.82379496e-01 1.12410438e+00 -1.60070586e+00
-1.27814516e-01 -2.27465574e-02 -6.90423369e-01 -1.33029532e+00
3.06870639e-01 -8.22789133e-01 4.04059827e-01 -1.24186194e+00
3.52318659e-02 -7.11845219e-01 -1.75088391e-01 5.09547174e-01
4.74591032e-02 6.11725211e-01 1.07883520e-01 4.08574253e-01
-7.96276152e-01 7.95409381e-01 1.39275885e+00 -8.70592967e-02
-4.30010743e-02 1.84697896e-01 -2.30884463e-01 1.07868421e+00
3.45078349e-01 -7.16736436e-01 -3.98999512e-01 -4.20523167e-01
-2.49842420e-01 -1.16578586e-01 6.47281051e-01 -1.19488311e+00
3.55093032e-01 -2.63772368e-01 4.87267911e-01 -1.16689503e+00
4.83040988e-01 -1.23770082e+00 3.32745127e-02 3.02092671e-01
-5.73089086e-02 -3.91962677e-01 8.12503546e-02 7.84850597e-01
-1.68503001e-01 -3.20291482e-02 9.67620611e-01 1.83211789e-01
-7.70930052e-01 5.29663503e-01 2.41286546e-01 -2.39232168e-01
1.19339895e+00 -1.52258918e-01 -2.75072847e-02 -2.09271476e-01
-6.62337482e-01 4.29692060e-01 4.12745774e-01 4.78084236e-01
4.39716220e-01 -1.53317988e+00 -3.76495808e-01 1.06874488e-01
4.83413398e-01 8.30221355e-01 1.02061376e-01 8.46515596e-01
-2.86150008e-01 4.00691152e-01 -1.47325853e-02 -1.20070767e+00
-1.14209497e+00 4.49571729e-01 1.43778026e-01 5.89321516e-02
-6.15819275e-01 9.71516073e-01 3.74841779e-01 -1.09702384e+00
4.67031837e-01 -4.64543492e-01 8.06555301e-02 -7.03698695e-02
1.42569691e-01 1.22299924e-01 1.53194070e-01 -8.60551119e-01
-4.41734999e-01 8.49389017e-01 -7.73869827e-02 2.11302936e-01
1.38675439e+00 -1.66223094e-01 2.41870746e-01 4.64588404e-01
1.08344555e+00 -2.51019448e-01 -1.57985413e+00 -7.47444212e-01
1.36186704e-01 -6.13131881e-01 1.36994701e-02 -5.86566210e-01
-1.06221926e+00 7.31128275e-01 7.31513619e-01 -2.26767316e-01
8.82345080e-01 3.34261745e-01 7.10770845e-01 6.42587721e-01
7.23246038e-01 -9.65629816e-01 5.07095993e-01 2.39190891e-01
9.43349719e-01 -1.60129750e+00 3.18797119e-02 -7.59808600e-01
-3.48261982e-01 9.75677013e-01 1.13768017e+00 -2.69731045e-01
6.01367354e-01 -4.68413569e-02 2.94514507e-01 -9.69570652e-02
-2.10340217e-01 -1.02773741e-01 6.01414084e-01 7.84061253e-01
1.00671984e-02 -1.94614887e-01 3.25914979e-01 5.41454494e-01
-2.51675308e-01 -1.10589959e-01 -6.89201728e-02 9.71590757e-01
-3.89706373e-01 -1.07255292e+00 -7.26921737e-01 2.90380538e-01
8.79167020e-02 4.75446105e-01 -2.93571502e-01 9.69793260e-01
2.29061469e-01 4.87305969e-01 1.22176602e-01 -2.99371421e-01
4.86049980e-01 -1.50323510e-01 6.18700743e-01 -7.86365032e-01
-2.74014413e-01 1.79647237e-01 -4.46338534e-01 -4.67822701e-01
-6.08240485e-01 -6.30652070e-01 -1.29425263e+00 3.89702842e-02
-6.33700907e-01 -5.52692600e-02 6.95418835e-01 1.16550410e+00
1.74212396e-01 4.35912043e-01 7.06460536e-01 -1.39223850e+00
-6.14771426e-01 -1.13445663e+00 -4.14353102e-01 4.03525472e-01
2.91506767e-01 -1.28057563e+00 -1.65641218e-01 7.96861053e-02] | [7.6106743812561035, -2.653414249420166] |
34ea8fef-4869-41c7-afb7-d64e53035cd9 | particle-transformer-for-jet-tagging | 2202.03772 | null | https://arxiv.org/abs/2202.03772v2 | https://arxiv.org/pdf/2202.03772v2.pdf | Particle Transformer for Jet Tagging | Jet tagging is a critical yet challenging classification task in particle physics. While deep learning has transformed jet tagging and significantly improved performance, the lack of a large-scale public dataset impedes further enhancement. In this work, we present JetClass, a new comprehensive dataset for jet tagging. The JetClass dataset consists of 100 M jets, about two orders of magnitude larger than existing public datasets. A total of 10 types of jets are simulated, including several types unexplored for tagging so far. Based on the large dataset, we propose a new Transformer-based architecture for jet tagging, called Particle Transformer (ParT). By incorporating pairwise particle interactions in the attention mechanism, ParT achieves higher tagging performance than a plain Transformer and surpasses the previous state-of-the-art, ParticleNet, by a large margin. The pre-trained ParT models, once fine-tuned, also substantially enhance the performance on two widely adopted jet tagging benchmarks. The dataset, code and models are publicly available at https://github.com/jet-universe/particle_transformer. | ['Sitian Qian', 'Congqiao Li', 'Huilin Qu'] | 2022-02-08 | null | null | null | null | ['jet-tagging'] | ['graphs'] | [-4.38098520e-01 -9.14568752e-02 -3.45645308e-01 -2.51400530e-01
-1.15764403e+00 -1.03648651e+00 9.26107168e-01 1.54778650e-02
-3.31547439e-01 9.36540663e-01 5.82612574e-01 -4.36447948e-01
3.55274417e-02 -7.53840864e-01 -6.78211153e-01 -6.79054201e-01
4.01079208e-02 1.01531124e+00 1.01533496e+00 -1.11696661e-01
-1.32936150e-01 3.52485597e-01 -1.05172777e+00 4.65650916e-01
4.90614504e-01 9.90851879e-01 1.38661489e-01 7.63128579e-01
-2.51176387e-01 6.59111917e-01 -4.52360600e-01 -9.23477829e-01
2.95443624e-01 -3.63379806e-01 -8.70918036e-01 -9.07009244e-01
3.41058165e-01 2.32272029e-01 -6.25109613e-01 1.07095504e+00
8.90717626e-01 3.42836440e-01 2.32576102e-01 -8.72501194e-01
-9.02362347e-01 9.68859255e-01 3.82444322e-01 8.79630804e-01
-1.29526109e-01 -1.34999916e-01 1.51570916e+00 -6.24888837e-01
6.96980000e-01 1.27669632e+00 8.86077762e-01 4.18487161e-01
-7.96077192e-01 -8.88256133e-01 -1.26614437e-01 1.36986738e-02
-7.37485826e-01 -2.60493100e-01 6.69282854e-01 -5.40962338e-01
1.20955694e+00 4.02258426e-01 8.41761708e-01 1.07089531e+00
7.04133928e-01 3.53809714e-01 9.98135924e-01 -6.67043179e-02
-1.33899003e-01 3.40272076e-02 8.18308741e-02 6.64072573e-01
2.44889513e-01 5.65521002e-01 -3.35906208e-01 -4.18359488e-01
6.04579210e-01 -1.80272445e-01 1.35634080e-01 -2.66115487e-01
-1.49215221e+00 8.92545164e-01 6.76054239e-01 4.94530767e-01
2.68225074e-01 1.87025562e-01 8.21059346e-01 2.99038552e-02
9.99944270e-01 8.40412974e-01 -7.88715482e-01 -3.60799372e-01
-5.16675711e-01 4.83100623e-01 7.60225296e-01 8.59935403e-01
7.38167688e-02 1.06717154e-01 -7.46347129e-01 6.50911033e-01
2.33156681e-01 6.40480995e-01 2.62156487e-01 -6.33979857e-01
5.53651392e-01 4.23980534e-01 1.86126500e-01 -3.35289270e-01
-5.73877394e-01 -8.76312137e-01 -6.46251142e-01 5.25940172e-02
1.96586028e-01 -2.70237010e-02 -8.96269143e-01 1.40414453e+00
4.96469826e-01 2.09022030e-01 6.70304224e-02 9.94609237e-01
1.38605189e+00 6.07313573e-01 4.59097594e-01 3.94569337e-01
1.57536864e+00 -1.58871996e+00 -5.67418694e-01 1.33742867e-02
2.51589596e-01 -1.19537222e+00 5.61663091e-01 5.27899079e-02
-1.06338441e+00 -9.10306454e-01 -7.54309952e-01 -4.81746286e-01
-7.40765095e-01 1.98923856e-01 1.24748158e+00 6.13584340e-01
-7.87741303e-01 7.95624197e-01 -9.73595142e-01 -2.82034755e-01
2.56453633e-01 3.87636483e-01 -2.20365942e-01 3.41195762e-01
-1.35096502e+00 5.66266358e-01 4.89451736e-01 -2.66141564e-01
-8.44921768e-01 -1.14519584e+00 -6.02363706e-01 1.04158282e-01
-2.72084046e-02 -8.91021192e-01 1.70984530e+00 1.21877581e-01
-1.44599736e+00 6.19667292e-01 5.99104576e-02 -5.43478727e-01
5.08562744e-01 -3.03770691e-01 -9.03615952e-01 -1.38560161e-01
3.40494931e-01 6.70552552e-01 4.63895380e-01 -9.68078494e-01
-6.76688254e-01 4.51012015e-01 4.08039331e-01 -2.52716839e-01
9.73661710e-03 5.94870210e-01 -5.88959038e-01 -8.31130147e-01
-4.09660190e-01 -1.01006413e+00 -3.18547159e-01 -5.67727506e-01
-6.11962199e-01 -3.44709367e-01 4.86017704e-01 -5.27746856e-01
7.56078660e-01 -1.88939309e+00 -6.67644441e-02 -5.59028804e-01
3.41827393e-01 5.69473922e-01 -1.81767032e-01 5.33832312e-01
-2.27776200e-01 2.63900489e-01 6.91287369e-02 -6.02251530e-01
5.30079365e-01 1.21317916e-01 -4.78804916e-01 6.63840845e-02
1.07581735e-01 1.23174596e+00 -1.12362492e+00 -1.63106754e-01
2.51267850e-01 4.75915194e-01 -7.76493013e-01 2.35809326e-01
-4.21615154e-01 9.17057514e-01 -6.16907239e-01 6.32548034e-01
7.65733778e-01 -3.23496759e-01 -1.29293218e-01 -4.84236211e-01
-3.85299355e-01 1.00310588e+00 -4.16704595e-01 1.71736634e+00
-2.16341808e-01 4.38690454e-01 -2.02754408e-01 -7.02189147e-01
8.41450274e-01 4.90060776e-01 6.87838972e-01 -8.61561894e-01
2.50371426e-01 7.61854723e-02 5.24374664e-01 -1.27505764e-01
8.60546172e-01 -1.75013706e-01 -3.89362931e-01 2.31857091e-01
3.56633633e-01 1.14968568e-01 4.68744248e-01 2.66763598e-01
1.37641048e+00 7.48726875e-02 -2.16930047e-01 -3.69103104e-01
1.42375112e-01 2.39663675e-01 9.50517237e-01 9.38971698e-01
-1.06858395e-01 3.84508908e-01 2.75333881e-01 -6.43386841e-01
-1.25422359e+00 -1.19529307e+00 -4.07384396e-01 1.14470506e+00
1.33506000e-01 -9.18863952e-01 -1.48470089e-01 -9.53409016e-01
6.90175742e-02 3.85376662e-01 -3.62195313e-01 -2.66982466e-02
-5.83876073e-01 -1.06202781e+00 6.62631273e-01 8.82734120e-01
4.38752264e-01 -1.27648258e+00 1.47113770e-01 3.07875603e-01
1.67026728e-01 -1.39982176e+00 -3.22943032e-01 2.25509331e-01
-4.42847967e-01 -9.91545022e-01 -2.17729986e-01 -6.28300607e-01
1.96473166e-01 -8.29879865e-02 1.66702104e+00 -1.64856732e-01
-3.19799781e-01 1.13038667e-01 -6.50241792e-01 -5.19628167e-01
-5.04023492e-01 2.81995356e-01 1.19966246e-01 -7.02019572e-01
2.42328420e-02 -2.89121270e-01 -5.30031681e-01 4.21808288e-02
-3.85229051e-01 -2.52060760e-02 7.44457543e-01 8.57932508e-01
5.46014726e-01 3.64497602e-01 3.60422969e-01 -1.03950691e+00
1.80286109e-01 -5.37761390e-01 -8.43431413e-01 -1.72787741e-01
-1.55823797e-01 -1.22377977e-01 7.70381808e-01 -1.66524947e-01
-1.22058010e+00 -3.95055652e-01 -8.09478104e-01 -1.76003799e-01
-8.29773862e-03 7.54318088e-02 5.96612468e-02 -2.99877405e-01
-1.04948781e-01 -2.25492179e-01 -8.82903814e-01 -1.11722505e+00
3.25660348e-01 1.56745508e-01 5.72023511e-01 -7.62605906e-01
1.09896553e+00 3.45381647e-01 -6.05579168e-02 -9.38561484e-02
-1.36485207e+00 -5.90655386e-01 -7.01067567e-01 2.32874259e-01
1.17102027e+00 -1.03482234e+00 -5.62028170e-01 -7.67173469e-02
-1.18887508e+00 -2.63371438e-01 -3.91619980e-01 8.28333914e-01
-6.59651086e-02 4.01250236e-02 -8.33368123e-01 -4.57091302e-01
-5.11336744e-01 -1.28616989e+00 1.23638582e+00 -7.25359693e-02
2.87634432e-01 -1.08689559e+00 5.09996116e-01 4.66327995e-01
4.38182443e-01 -1.40114725e-01 9.48707044e-01 -1.09387386e+00
-6.97483420e-01 -8.53311270e-02 -5.33319235e-01 2.15873241e-01
-5.34675941e-02 -3.87018502e-01 -8.22337151e-01 -3.56814951e-01
-5.34154356e-01 -1.96561486e-01 1.27247965e+00 1.08743478e-02
1.03360438e+00 1.64689943e-01 -6.81705832e-01 5.78881264e-01
1.04408002e+00 3.46053809e-01 1.45004958e-01 2.35001042e-01
6.34837747e-01 -3.63559604e-01 5.24406075e-01 1.31013095e-01
3.45448673e-01 8.88705254e-01 2.42773652e-01 -3.44975948e-01
-6.96037233e-01 -3.28763455e-01 3.17885309e-01 1.52008808e+00
-2.68380076e-01 -6.50466025e-01 -7.18032718e-01 3.35870177e-01
-1.64272797e+00 -1.04037285e+00 -1.46764562e-01 1.76341963e+00
6.08682811e-01 5.87834895e-01 -1.23728566e-01 -3.51772189e-01
4.14740682e-01 2.61689067e-01 -2.28655294e-01 -2.69258827e-01
2.65513808e-01 7.75357842e-01 5.78961134e-01 4.71162260e-01
-1.77475536e+00 1.19663763e+00 6.45950413e+00 8.74606073e-01
-7.28635848e-01 8.34915996e-01 1.22407831e-01 6.29275888e-02
-3.76373321e-01 2.31229693e-01 -1.39914310e+00 6.75845802e-01
9.74709868e-01 1.36042774e-01 9.72397253e-02 7.64698267e-01
-3.05302292e-01 6.05804384e-01 -8.52374196e-01 3.36554915e-01
-1.79699361e-01 -1.78591812e+00 -7.30148256e-02 -1.05769977e-01
8.25582087e-01 1.01280808e+00 -2.04163119e-01 1.06443691e+00
7.84597397e-01 -5.83934546e-01 8.57619405e-01 5.11155069e-01
5.65755665e-01 -5.57844400e-01 1.02528107e+00 -2.84501538e-03
-1.79551160e+00 1.46453872e-01 -9.34181571e-01 2.51508672e-02
5.13079882e-01 9.61067677e-01 -5.88455856e-01 1.04331028e+00
9.89000261e-01 7.75059223e-01 -6.23270154e-01 1.40407348e+00
-1.04001299e-01 1.00713527e+00 -2.88767666e-01 2.90266797e-02
3.47903877e-01 -1.42436758e-01 9.52976882e-01 1.60323191e+00
4.81020331e-01 -1.63524911e-01 7.51643896e-01 7.69629240e-01
-3.63515019e-01 -2.25299031e-01 -4.13870156e-01 -5.53669631e-01
4.35761690e-01 1.73743176e+00 -9.06569481e-01 -4.45349276e-01
-6.20453835e-01 5.74166477e-01 4.16575789e-01 9.48959813e-02
-1.26915157e+00 -1.05419122e-01 9.94429827e-01 -3.25645328e-01
6.67988002e-01 -2.89761037e-01 4.63523597e-01 -9.79309022e-01
-3.53122532e-01 -2.40883261e-01 4.44879979e-01 -5.68940222e-01
-1.66086650e+00 9.60435510e-01 -1.17793694e-01 -1.16274679e+00
2.92573571e-01 -9.25006568e-01 -8.24537575e-01 5.31602502e-01
-1.77573109e+00 -1.32872844e+00 -2.97789555e-02 -9.14005488e-02
4.43148494e-01 -5.15967429e-01 8.71070504e-01 8.07880759e-01
-4.73137885e-01 2.71382064e-01 2.38484353e-01 3.05797338e-01
7.60672688e-01 -1.52579939e+00 1.12982666e+00 4.43015665e-01
4.05248493e-01 3.78190756e-01 4.56313252e-01 -1.01163232e+00
-1.36742437e+00 -1.64003682e+00 1.03831756e+00 -9.70852494e-01
1.09982598e+00 -9.40589666e-01 -4.65185404e-01 8.21841657e-01
5.36479115e-01 4.98088866e-01 6.29280150e-01 2.11091667e-01
-2.19991878e-01 9.84375626e-02 -7.68804133e-01 1.23331182e-01
1.49943924e+00 -4.34294194e-01 -5.46893120e-01 1.01967883e+00
1.13026083e+00 -6.68846905e-01 -1.44563031e+00 6.63319468e-01
1.34484380e-01 -8.36028695e-01 1.44500649e+00 -9.21973944e-01
7.24177137e-02 -2.41285294e-01 -1.88716725e-02 -1.36252761e+00
-1.04448700e+00 -4.87422973e-01 1.18408620e-01 1.53625190e+00
5.20414770e-01 -7.55110145e-01 5.83484650e-01 -1.08720183e-01
-8.75860631e-01 -6.44234359e-01 -1.20226490e+00 -1.04906774e+00
3.20691198e-01 -4.81344163e-01 7.78731883e-01 7.09968984e-01
-3.35545868e-01 3.39837849e-01 -3.05014431e-01 4.77297455e-01
4.01370466e-01 5.31382263e-01 4.02428746e-01 -1.41864204e+00
-6.09406352e-01 -3.89023602e-01 -3.51069123e-01 -8.30223501e-01
3.36794436e-01 -1.82830477e+00 -2.73681041e-02 -1.63319027e+00
3.69611263e-01 -9.49121714e-01 -6.88338399e-01 5.99579215e-01
-5.03413007e-02 7.29608536e-01 2.49610320e-01 -4.68977056e-02
-1.02268720e+00 8.55234265e-01 1.43611217e+00 -3.81063193e-01
5.04724562e-01 -1.73604593e-01 -3.16551566e-01 7.62604296e-01
1.04544842e+00 -8.73263299e-01 1.77207738e-02 -4.78417426e-01
3.30098391e-01 -3.29365790e-01 4.94653761e-01 -1.40052557e+00
1.11450881e-01 3.49816531e-01 7.24585176e-01 -7.16781676e-01
8.39548647e-01 -3.48965734e-01 1.98507011e-01 2.99451143e-01
-2.22952574e-01 2.89941609e-01 4.98810500e-01 4.30967569e-01
-1.45674899e-01 -2.88209677e-01 5.11819661e-01 -3.20408523e-01
-6.94729865e-01 7.67486274e-01 -3.83247994e-02 4.79686409e-01
5.59011221e-01 8.82521629e-01 -8.71484756e-01 3.81906986e-01
-6.67703748e-01 7.36486679e-03 -3.53863761e-02 6.96721971e-01
-3.55023518e-02 -1.67488980e+00 -5.37910998e-01 1.65338919e-01
-6.10556453e-02 -4.51009214e-01 6.79257438e-02 7.12676764e-01
-1.26224250e-01 1.00854707e+00 4.56659272e-02 -2.22828150e-01
-9.71009970e-01 4.96751875e-01 1.78727712e-02 -6.84000432e-01
-1.18834507e+00 1.04709911e+00 1.09363444e-01 -9.69919860e-01
-1.82641581e-01 -7.32119203e-01 -4.39495891e-02 -3.97254586e-01
2.65290290e-01 3.08257192e-02 3.12121540e-01 -5.06031454e-01
-2.49954507e-01 5.07420838e-01 -2.28838082e-02 3.43112588e-01
1.20800626e+00 5.18715978e-01 -2.55062878e-01 1.40229240e-01
8.52372587e-01 7.06013322e-01 -8.98614287e-01 2.74585765e-02
-2.53718704e-01 -2.92142391e-01 9.15531348e-03 -1.19258857e+00
-1.14455616e+00 6.54522836e-01 3.70504200e-01 4.00718123e-01
3.40441287e-01 6.40675545e-01 1.42256021e+00 4.93477941e-01
2.36466929e-01 -6.04640663e-01 -2.01705754e-01 8.67105663e-01
6.52158976e-01 -1.23711538e+00 -1.95072144e-01 -6.36758745e-01
-2.97096401e-01 7.13444233e-01 6.43319428e-01 -1.60329148e-01
7.98094511e-01 6.65717602e-01 -9.92712192e-03 -4.17185515e-01
-9.30043936e-01 -4.34498638e-01 7.69419968e-01 2.92054981e-01
7.01140523e-01 2.02770695e-01 -1.49980858e-01 1.03225386e+00
-7.13386774e-01 -3.53456855e-01 -2.54326731e-01 7.03038812e-01
-5.25534630e-01 -1.62870765e+00 3.71438148e-03 5.22340417e-01
-5.74706316e-01 -2.80925274e-01 -2.66730309e-01 7.73546576e-01
6.50858998e-01 5.97783029e-01 1.12718575e-01 -2.60867745e-01
1.56929836e-01 1.21028565e-01 5.45111477e-01 -7.09933221e-01
-8.11963022e-01 -2.31064886e-01 5.56475103e-01 -5.73201954e-01
-3.93461108e-01 -3.94910067e-01 -1.31281221e+00 -1.29835056e-02
-9.23292711e-02 8.36529016e-01 5.22088945e-01 8.19743872e-01
3.95650953e-01 1.18101764e+00 -4.68911976e-02 -1.15618801e+00
-2.95597047e-01 -9.20600474e-01 -3.52759272e-01 3.52598339e-01
6.75611347e-02 -1.27877247e+00 -1.20632134e-01 -3.96772802e-01] | [15.701129913330078, 2.9201831817626953] |
90ecc2ab-48f1-4c20-a631-32a7a09b5791 | clio-role-interactive-multi-event-head | null | null | https://aclanthology.org/2022.coling-1.221 | https://aclanthology.org/2022.coling-1.221.pdf | CLIO: Role-interactive Multi-event Head Attention Network for Document-level Event Extraction | Transforming the large amounts of unstructured text on the Internet into structured event knowledge is a critical, yet unsolved goal of NLP, especially when addressing document-level text. Existing methods struggle in Document-level Event Extraction (DEE) due to its two intrinsic challenges: (a) Nested arguments, which means one argument is the sub-string of another one. (b) Multiple events, which indicates we should identify multiple events and assemble the arguments for them. In this paper, we propose a role-interactive multi-event head attention network (CLIO) to solve these two challenges jointly. The key idea is to map different events to multiple subspaces (i.e. multi-event head). In each event subspace, we draw the semantic representation of each role closer to its corresponding arguments, then we determine whether the current event exists. To further optimize event representation, we propose an event representation enhancing strategy to regularize pre-trained embedding space to be more isotropic. Our experiments on two widely used DEE datasets show that CLIO achieves consistent improvements over previous methods. | ['Yi Liu', 'Wei Ma', 'Zheng Lin', 'Ping Guo', 'Fang Fang', 'Yanan Cao', 'Yubing Ren'] | null | null | null | null | coling-2022-10 | ['event-extraction', 'document-level-event-extraction'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.63586491e-01 2.31759809e-03 -2.02484593e-01 -1.74642593e-01
-9.48036849e-01 -6.66174173e-01 6.80839300e-01 4.14038956e-01
-4.81959552e-01 7.34755933e-01 1.06278813e+00 -1.18118688e-01
-2.57237583e-01 -8.64288390e-01 -7.62534857e-01 -5.18723845e-01
5.74408881e-02 5.30250967e-01 2.57883102e-01 -2.95636263e-02
-1.70180313e-02 4.50174451e-01 -1.34965849e+00 6.42514348e-01
6.56370699e-01 7.32214153e-01 1.79685712e-01 3.70247513e-01
-5.81918120e-01 1.02402329e+00 -7.53384650e-01 -1.52146176e-01
-7.31302053e-02 -5.15788555e-01 -1.02404845e+00 2.72311661e-02
-2.21613571e-01 -1.69796765e-01 -3.53658229e-01 8.71042013e-01
4.03066695e-01 3.34911376e-01 7.45266020e-01 -1.39412916e+00
-5.29504597e-01 9.87342298e-01 -5.28227687e-01 5.07985532e-01
3.08414429e-01 -4.72576022e-01 1.40633345e+00 -1.16741359e+00
8.90703380e-01 1.32501304e+00 3.76187623e-01 1.69768825e-01
-9.16175187e-01 -6.13442719e-01 3.86144191e-01 4.67922658e-01
-1.27436268e+00 -1.81682721e-01 8.25488687e-01 -3.47968370e-01
1.02492893e+00 2.12726891e-01 4.85627979e-01 1.11448812e+00
1.18972585e-01 9.96999502e-01 3.35160643e-01 -2.99248487e-01
5.10854647e-02 -1.03323735e-01 3.43265355e-01 2.15517059e-01
2.14117527e-01 -5.01191199e-01 -5.76579690e-01 -1.61260337e-01
6.11416578e-01 1.89143762e-01 -1.25513539e-01 -4.85533178e-02
-1.49923718e+00 9.09354031e-01 1.67283952e-01 5.20719826e-01
-8.04560125e-01 6.92773461e-02 6.21258616e-01 -1.39527485e-01
2.11988688e-01 2.25263417e-01 -5.30273080e-01 -1.38362408e-01
-5.47250092e-01 4.47440237e-01 6.76794231e-01 7.87816465e-01
5.12069583e-01 -2.99018532e-01 -5.06459832e-01 8.03095877e-01
2.89483666e-01 -1.16401173e-01 4.63496834e-01 -3.96364897e-01
1.04375494e+00 9.68871534e-01 3.47965837e-01 -9.75721896e-01
-3.72479975e-01 -3.54677737e-01 -7.20916212e-01 -3.43960345e-01
1.66855782e-01 -2.91840285e-01 -5.28696418e-01 1.75905883e+00
5.92318833e-01 2.09994391e-01 2.85185814e-01 7.47415900e-01
9.66790557e-01 1.31683099e+00 4.26997602e-01 -2.14608312e-01
1.84133685e+00 -9.08577025e-01 -1.15058625e+00 -4.39888448e-01
3.54070842e-01 -6.23019993e-01 9.10363078e-01 1.14855722e-01
-8.88472855e-01 -3.72900873e-01 -1.01851916e+00 -3.78297865e-01
-5.84981263e-01 4.48013872e-01 4.84061599e-01 -4.45987619e-02
-2.61999577e-01 1.78280771e-01 -6.62771881e-01 -1.87531322e-01
3.73543054e-01 1.08265486e-02 -3.73232424e-01 1.92097977e-01
-1.69947267e+00 5.03037274e-01 1.02576482e+00 -1.28518268e-01
-5.88744223e-01 -6.40726030e-01 -1.03229475e+00 4.65388566e-01
8.32440853e-01 -3.77185792e-01 1.07894647e+00 -4.10919845e-01
-9.85895097e-01 5.84124684e-01 -5.05691409e-01 -4.47305202e-01
1.07590603e-02 -4.92952704e-01 -6.49265826e-01 2.17654943e-01
5.84323943e-01 3.35174531e-01 5.38817465e-01 -1.21986818e+00
-9.27045822e-01 -3.88105124e-01 -7.80079886e-02 3.90242249e-01
-6.26358151e-01 5.34719288e-01 -7.87445784e-01 -9.68415976e-01
1.15247786e-01 -4.93614137e-01 -7.10867718e-02 -4.47550863e-01
-5.51138103e-01 -7.78491855e-01 7.89066136e-01 -8.81937146e-01
1.81130493e+00 -2.12126851e+00 3.58741164e-01 -8.60980824e-02
2.08047241e-01 4.75196876e-02 1.02315292e-01 7.34094858e-01
-4.61310893e-01 1.45813927e-01 -1.16356798e-02 -2.50376433e-01
1.90080851e-01 2.55985856e-01 -8.07017207e-01 7.87644759e-02
4.46179390e-01 8.90069783e-01 -9.18998659e-01 -7.30961382e-01
6.63011596e-02 3.50815028e-01 -3.22867870e-01 3.92805755e-01
-2.01092780e-01 2.51789719e-01 -6.86634362e-01 3.14410627e-01
5.43401539e-01 -4.94199395e-01 1.09474629e-01 -6.36170745e-01
-1.79096565e-01 6.01684809e-01 -1.66732728e+00 1.42085826e+00
-3.33700746e-01 3.79559994e-01 -1.55385450e-01 -1.07810199e+00
6.56078875e-01 6.66905463e-01 7.31809974e-01 -3.45916122e-01
1.20286137e-01 -1.21688008e-01 -2.24750802e-01 -4.63137925e-01
5.35086036e-01 -1.01964802e-01 -4.50921357e-01 6.34432197e-01
1.34597763e-01 4.98027772e-01 5.36986113e-01 5.19925237e-01
1.08775806e+00 -1.28576577e-01 6.78459644e-01 1.09178051e-01
5.49501240e-01 -1.27817884e-01 9.44296658e-01 6.12497747e-01
1.53531078e-02 6.14895463e-01 8.06796551e-01 -4.11406457e-01
-8.79216552e-01 -1.11450994e+00 6.77434914e-03 1.11496365e+00
2.61996686e-01 -8.02433729e-01 -4.57476348e-01 -9.87855375e-01
-2.74920434e-01 8.51736665e-01 -6.36056721e-01 5.71738258e-02
-9.28099632e-01 -8.67399931e-01 3.80195469e-01 8.66596758e-01
3.63263577e-01 -1.33848178e+00 -6.30453110e-01 8.79993200e-01
-6.89480782e-01 -1.22585893e+00 -6.06027424e-01 2.37592518e-01
-2.61918575e-01 -1.01528108e+00 -4.67135936e-01 -7.33381152e-01
3.37626874e-01 -2.97773350e-02 1.08306897e+00 -3.81595761e-01
-1.66012615e-01 2.70850514e-03 -6.19500935e-01 -6.06372356e-01
-9.59270969e-02 5.29568754e-02 -3.83433372e-01 3.62718284e-01
5.19588709e-01 -3.89450639e-01 -3.93259943e-01 1.17469527e-01
-1.19272470e+00 3.60299237e-02 5.00187755e-01 6.62001491e-01
8.35531235e-01 4.94238377e-01 7.81510949e-01 -7.95756936e-01
7.06106067e-01 -9.56575394e-01 -2.22918838e-01 3.77662480e-01
-1.31043524e-01 1.29702896e-01 8.27837467e-01 -4.80543852e-01
-1.26064742e+00 -2.65901219e-02 -2.27800786e-01 -1.34759322e-01
-2.14232713e-01 8.14035416e-01 -6.37123466e-01 9.68690634e-01
2.59037644e-01 3.03225338e-01 -7.33110666e-01 -4.58823353e-01
3.15424085e-01 6.22029066e-01 5.45658052e-01 -6.53359354e-01
5.22531629e-01 5.58144331e-01 -3.36594194e-01 -5.30376673e-01
-1.16028583e+00 -6.73990786e-01 -2.84190774e-01 -2.05407999e-02
1.15191257e+00 -8.54183555e-01 -3.43982369e-01 5.58157451e-02
-1.51671159e+00 8.87899026e-02 -3.26136947e-01 5.28392017e-01
-1.78037047e-01 2.52044320e-01 -5.71651638e-01 -5.54533601e-01
-3.49835455e-01 -9.33569551e-01 1.29419935e+00 2.36430109e-01
-4.20408696e-01 -7.85914779e-01 9.81022343e-02 1.56263053e-01
-1.70304731e-01 2.57444948e-01 1.11065626e+00 -9.18865502e-01
-3.48061204e-01 -8.96421820e-02 -3.61459285e-01 -1.11457773e-01
1.74313888e-01 -2.11942196e-01 -4.43415582e-01 4.42452990e-02
3.28464732e-02 -7.77771473e-02 8.03587317e-01 1.68496400e-01
1.40560079e+00 -3.26850712e-01 -5.23608923e-01 3.64820212e-01
1.19364190e+00 4.56174910e-01 6.48274302e-01 4.66740251e-01
7.36714780e-01 5.49471378e-01 6.36237204e-01 6.98881924e-01
5.46103954e-01 6.40392184e-01 1.18303582e-01 -9.43058729e-02
-1.58008412e-01 -4.60191995e-01 3.61132234e-01 6.46390855e-01
2.83287913e-01 -7.49513090e-01 -7.70716369e-01 8.78924608e-01
-2.07395411e+00 -1.08202505e+00 -1.89034984e-01 1.73631454e+00
9.41281915e-01 1.08213022e-01 -7.31492639e-02 2.75338382e-01
8.59723151e-01 3.46166551e-01 -4.06227171e-01 1.66343935e-02
-1.22391112e-01 1.45625263e-01 4.22253311e-02 7.02577978e-02
-1.38612258e+00 8.09659898e-01 4.98973513e+00 1.01797998e+00
-6.26059294e-01 1.85574889e-01 2.68405557e-01 2.77435072e-02
-4.24789220e-01 4.80242595e-02 -1.22250760e+00 6.86509311e-01
7.93590605e-01 -3.34661931e-01 1.05323724e-01 5.89476407e-01
1.39569417e-01 1.26618922e-01 -1.24037409e+00 8.76354277e-01
1.10027172e-01 -1.51459289e+00 3.92814517e-01 -3.49643938e-02
4.43778634e-01 -4.70786005e-01 -4.36057627e-01 6.05865419e-01
2.15107650e-01 -7.36649394e-01 9.00018692e-01 1.45598039e-01
3.66045833e-01 -8.23769271e-01 5.24797440e-01 3.84075999e-01
-1.82645273e+00 -1.16980318e-02 -1.45833403e-01 3.86399686e-01
6.37606442e-01 6.35769486e-01 -5.17149210e-01 7.94825137e-01
7.18149722e-01 8.27782571e-01 -2.77132064e-01 6.85903013e-01
-5.20429611e-01 7.11254835e-01 -3.40375364e-01 8.82017687e-02
2.79080361e-01 -7.98409954e-02 7.93794334e-01 1.31092370e+00
3.75231117e-01 2.40604579e-01 2.72659063e-01 1.12241220e+00
-3.42482507e-01 2.37737015e-01 -3.23169738e-01 -2.62097955e-01
6.93852007e-01 1.15997589e+00 -8.67339671e-01 -5.41422069e-01
-6.42913759e-01 1.09636772e+00 4.57556605e-01 3.49257827e-01
-1.14589095e+00 -8.03319037e-01 4.15831834e-01 -1.66076183e-01
6.22973263e-01 4.81201075e-02 -4.58248854e-02 -1.40123582e+00
2.94381380e-01 -7.47641206e-01 9.61573601e-01 -7.60302901e-01
-1.16749883e+00 5.03600419e-01 2.77027693e-02 -1.15840244e+00
-1.77474841e-01 -2.85355568e-01 -7.93791056e-01 7.08295345e-01
-1.14236486e+00 -1.13122165e+00 -1.72664691e-02 6.69824958e-01
8.61437023e-01 4.81059439e-02 6.67135835e-01 6.73761070e-01
-8.32976103e-01 3.86956900e-01 -1.01248331e-01 4.29410994e-01
5.75296044e-01 -1.23929560e+00 2.67400801e-01 1.11381304e+00
4.48735297e-01 5.82387507e-01 4.18805897e-01 -9.48600113e-01
-1.27123034e+00 -1.25825965e+00 1.34388685e+00 -4.18270469e-01
7.95388460e-01 -4.46419746e-01 -1.06242561e+00 1.05222297e+00
1.22957267e-01 -1.63180873e-01 7.09018052e-01 1.51361942e-01
-2.45146528e-01 2.14015343e-03 -6.53532803e-01 7.29074895e-01
8.08566868e-01 -5.22582889e-01 -1.18427730e+00 3.71372461e-01
9.81435537e-01 -1.97561860e-01 -7.78412402e-01 2.64633417e-01
1.32749245e-01 -5.09128928e-01 1.20179152e+00 -7.14003146e-01
5.90885758e-01 -5.15820920e-01 -5.70612203e-04 -1.08131862e+00
-2.25149274e-01 -4.63410497e-01 -5.22311628e-01 1.64978898e+00
3.67948353e-01 -2.53001988e-01 5.27355671e-01 3.31170738e-01
-1.78514048e-01 -7.31270313e-01 -8.87030780e-01 -5.49264967e-01
-2.26835430e-01 -4.92494017e-01 8.29329967e-01 8.87743890e-01
-1.23328120e-01 8.88666689e-01 -3.96088094e-01 4.89472836e-01
3.93645167e-01 3.88546586e-01 4.54095721e-01 -1.13684320e+00
-2.34794542e-01 -2.93499857e-01 4.93975654e-02 -1.00922418e+00
2.60500550e-01 -9.56797957e-01 1.00578070e-02 -1.96862936e+00
4.58764941e-01 -1.05699331e-01 -4.03323054e-01 5.63790143e-01
-5.51051080e-01 -3.75472695e-01 -1.06504127e-01 8.32198635e-02
-8.50043178e-01 7.63319433e-01 7.94766188e-01 -5.38534634e-02
-3.37711096e-01 -2.96626449e-01 -7.41717458e-01 7.13391781e-01
6.41235232e-01 -7.62468100e-01 -3.57722789e-01 -3.05174708e-01
4.35794711e-01 1.93926722e-01 3.15648913e-01 -7.91779101e-01
3.02209556e-01 -1.24118723e-01 4.16959405e-01 -1.13387024e+00
1.24849528e-01 -8.16641271e-01 1.40344828e-01 5.18203713e-02
-4.39041823e-01 -6.38404191e-02 2.61398375e-01 7.35109925e-01
-5.51408291e-01 -4.87609297e-01 3.56314123e-01 1.36011727e-02
-8.84119868e-01 2.95952678e-01 -4.24263835e-01 4.44691718e-01
1.07533336e+00 1.51590839e-01 -1.23440281e-01 -2.02956215e-01
-7.61594415e-01 3.75228912e-01 -2.72505879e-01 5.10187507e-01
6.23839796e-01 -1.52306569e+00 -8.45727205e-01 2.42247153e-02
3.09980541e-01 1.80344000e-01 3.59303117e-01 4.20146704e-01
-8.26691091e-02 2.70736277e-01 2.26487577e-01 -1.07661709e-01
-1.16781557e+00 5.11433303e-01 -1.45091370e-01 -8.48852217e-01
-8.91641557e-01 7.14060247e-01 3.71567696e-01 -2.95269400e-01
2.84391791e-01 -3.04800689e-01 -6.39634252e-01 4.71649200e-01
7.51685023e-01 2.79502690e-01 9.08540487e-02 -5.88853478e-01
-4.16596293e-01 3.15797180e-01 -3.09359491e-01 -1.86843693e-01
1.47636461e+00 2.30679773e-02 -2.98982263e-01 2.86444485e-01
1.28465486e+00 7.84503669e-02 -1.08360684e+00 -2.70147592e-01
1.99862763e-01 -1.38987929e-01 1.69726983e-02 -5.69796622e-01
-6.36653781e-01 6.45911098e-01 1.47679001e-01 2.39712432e-01
1.01017618e+00 4.24524277e-01 1.14878464e+00 3.73190373e-01
-8.25850442e-02 -1.11714840e+00 2.11069122e-01 6.64619386e-01
1.13223803e+00 -9.10803080e-01 -9.83390734e-02 -4.67723250e-01
-7.54958689e-01 1.01668453e+00 4.68497187e-01 1.12365410e-01
5.61736882e-01 1.72239721e-01 -4.86640304e-01 -4.87352282e-01
-6.67564750e-01 -2.44036302e-01 4.07785952e-01 -1.09818224e-02
2.73974389e-01 -1.13143340e-01 -5.14840841e-01 1.31197500e+00
8.97380412e-02 -1.15526110e-01 1.79885387e-01 1.01720071e+00
-2.72755742e-01 -1.19800198e+00 -4.49291050e-01 4.60841179e-01
-7.65624583e-01 -1.19607188e-01 -2.46690854e-01 7.07288682e-01
3.45057279e-01 9.02539194e-01 2.06474885e-01 -4.86470945e-02
5.79087138e-01 3.11843395e-01 -1.16451876e-02 -8.17186952e-01
-5.82060695e-01 3.28075826e-01 1.71841964e-01 -4.83856559e-01
-1.61232606e-01 -8.77815604e-01 -1.77213144e+00 3.50787640e-01
-2.28042603e-02 2.68550038e-01 4.17890072e-01 9.87342358e-01
4.72508043e-01 9.19258356e-01 3.91881973e-01 -4.02661622e-01
-3.13919604e-01 -7.39447832e-01 -4.70641166e-01 8.72630060e-01
8.70788768e-02 -6.79374695e-01 -3.44690055e-01 3.08916628e-01] | [9.058799743652344, 9.163472175598145] |
db91ab91-c1db-4a2e-affc-56e1acba95e7 | relational-model-for-parameter-description-in | 2005.05046 | null | https://arxiv.org/abs/2005.05046v1 | https://arxiv.org/pdf/2005.05046v1.pdf | Relational Model for Parameter Description in Automatic Semantic Web Service Composition | Automatic Service Composition is a research direction aimed at facilitating the usage of atomic web services. Particularly, the goal is to build workflows of services that solve specific queries, which cannot be resolved by any single service from a known repository. Each of these services is described independently by their providers that can have no interaction with each other, therefore some common standards have been developed, such as WSDL, BPEL, OWL-S. Our proposal is to use such standards together with JSON-LD to model a next level of semantics, mainly based on binary relations between parameters of services. Services relate to a public ontology to describe their functionality. Binary relations can be specified between input and/or output parameters in service definition. The ontology includes some relations and inference rules that help to deduce new relations between parameters of services. To our knowledge, it is for the first time that parameters are matched not only based on their type, but on a more meaningful semantic context considering such type of relations. This enables the automation of a large part of the reasoning that a human person would do when manually building a composition. Moreover, the proposed model and the composition algorithm can work with multiple objects of the same type, a fundamental feature that was not possible before. We believe that the poor model expressiveness is what is keeping service composition from reaching large-scale application in practice. | ['Andrei Netedu', 'Liana Ţucăr', 'Paul Diac'] | 2020-05-08 | null | null | null | null | ['service-composition'] | ['miscellaneous'] | [-1.93906296e-02 2.80952960e-01 1.29988059e-01 -5.53880632e-01
9.94096920e-02 -6.73086524e-01 9.43746090e-01 1.92552865e-01
-3.19427341e-01 6.92624748e-01 1.76904798e-02 -2.91996509e-01
-6.97741330e-01 -1.26560462e+00 -1.82643726e-01 -4.96630162e-01
7.58755952e-02 9.02844191e-01 9.19426739e-01 -6.87334895e-01
-2.17221007e-01 8.00464749e-01 -2.29255772e+00 4.69217896e-01
6.87329352e-01 1.07357454e+00 3.38736147e-01 4.88913283e-02
-9.24120665e-01 5.04642487e-01 -4.44860369e-01 -5.04811108e-01
6.99260039e-03 -3.16086978e-01 -1.06166661e+00 9.11984518e-02
-3.91246885e-01 -5.14930785e-02 5.71886837e-01 1.05896533e+00
1.49305603e-02 1.40164532e-02 4.39909101e-01 -1.64433193e+00
-1.14765577e-01 8.05081546e-01 4.59917456e-01 -4.80073363e-01
8.21947098e-01 -5.51881157e-02 1.02942097e+00 -1.75176576e-01
1.03337193e+00 1.02328837e+00 1.25870436e-01 6.64984524e-01
-9.07192349e-01 -2.49197125e-01 2.44917750e-01 6.79516315e-01
-1.30703712e+00 -3.11930209e-01 4.28499967e-01 -1.99408680e-01
9.00815606e-01 9.39175010e-01 7.76238739e-01 7.51837730e-01
-3.96820694e-01 7.25998282e-02 9.52780783e-01 -5.73039353e-01
4.33333516e-01 6.02450192e-01 1.84312224e-01 2.34142944e-01
5.80929279e-01 -6.33407056e-01 -1.91995557e-02 -2.88368553e-01
1.58517629e-01 9.65525284e-02 -3.93112868e-01 -6.04780793e-01
-7.28535831e-01 4.44754213e-01 -7.70806149e-02 1.11781466e+00
-1.45699754e-01 -3.01381230e-01 3.20288748e-01 3.61049920e-01
-4.05911833e-01 3.04051489e-01 -7.32234478e-01 -1.97174206e-01
-2.74175644e-01 2.43858218e-01 1.75741553e+00 1.28068829e+00
8.67045760e-01 -2.61772424e-01 5.83960414e-01 5.14622808e-01
5.34353316e-01 3.45271349e-01 4.62905943e-01 -7.98023164e-01
-1.98398188e-01 1.31663895e+00 4.90207851e-01 -6.36734784e-01
-2.73954034e-01 -1.34682313e-01 -2.20742092e-01 4.17387366e-01
4.73842353e-01 4.16153342e-01 -2.51034498e-01 1.46007514e+00
5.97990632e-01 -1.79032519e-01 3.01200360e-01 8.99341822e-01
7.07959771e-01 2.72809416e-01 1.31187350e-01 -4.38538492e-01
1.69829631e+00 -4.79040593e-01 -9.79309738e-01 2.85599828e-01
4.68083769e-01 -6.38536930e-01 8.53349447e-01 4.72070575e-01
-8.20716977e-01 5.72898202e-02 -8.14917862e-01 2.19301701e-01
-9.53515351e-01 -5.70439041e-01 7.49608099e-01 5.67425549e-01
-7.75877714e-01 4.74718601e-01 -4.02606130e-01 -8.88849914e-01
-2.03559816e-01 4.06571984e-01 -4.61612642e-01 1.36125967e-01
-1.32922316e+00 1.05962789e+00 5.62455654e-01 -1.21260900e-02
-5.73058367e-01 -1.54803500e-01 -5.70471346e-01 4.53405142e-01
9.90860105e-01 -5.51465571e-01 1.29833496e+00 -1.16923404e+00
-1.32214367e+00 6.71289980e-01 -1.66917533e-01 -4.57887381e-01
4.97636467e-01 4.61698741e-01 -1.31490111e+00 -8.12375546e-02
2.26653703e-02 -2.26869419e-01 2.13872522e-01 -1.67286420e+00
-1.11608696e+00 -3.07802081e-01 6.01127982e-01 -1.36055246e-01
-2.83972025e-01 3.40322822e-01 -5.65630078e-01 4.26096678e-01
2.44865984e-01 -4.49257374e-01 -2.69497186e-01 -2.20026046e-01
3.30789164e-02 -1.97296843e-01 5.08180737e-01 -3.30224872e-01
1.21602130e+00 -1.66711736e+00 -1.82664208e-02 6.09525681e-01
-4.86465916e-03 2.37896472e-01 2.51773000e-01 1.01900768e+00
2.34325975e-01 2.42571995e-01 -2.96532005e-01 4.01719719e-01
6.96830750e-01 7.80316651e-01 -1.27075100e-02 -1.70370787e-01
-3.93031448e-01 1.66203409e-01 -7.56302655e-01 -4.34677124e-01
7.61372745e-02 5.15088022e-01 -3.23551029e-01 9.61385593e-02
-7.93192804e-01 3.10229659e-01 -7.07344651e-01 5.17734885e-01
7.06485450e-01 -1.65552914e-01 9.74146247e-01 -2.41092965e-01
-3.19211572e-01 3.39449257e-01 -1.73937702e+00 1.50296175e+00
-6.25372171e-01 -5.36899984e-01 2.48979703e-01 -9.19917703e-01
1.04900157e+00 8.98948669e-01 7.11248517e-01 -4.80163574e-01
3.04274838e-02 7.27272511e-01 2.29990467e-01 -7.87421703e-01
-2.28986833e-02 5.13220727e-02 1.37218580e-01 5.31942129e-01
-3.64016108e-02 -3.30270790e-02 9.02517080e-01 -1.94182262e-01
9.52044964e-01 5.64087152e-01 5.35614192e-01 -1.32741645e-01
1.22984982e+00 -1.07228458e-01 6.48394287e-01 4.67566937e-01
4.45505381e-01 -1.32696629e-01 7.25919843e-01 -6.43093586e-01
-9.92582440e-01 -8.82912159e-01 -1.14587493e-01 9.30264473e-01
2.68044859e-01 -7.88265169e-01 -5.66740036e-01 -6.17610455e-01
-1.29045352e-01 6.76443398e-01 5.04357107e-02 3.91569674e-01
-3.07653993e-01 -2.91981936e-01 3.10425907e-01 -1.89428806e-01
1.98809713e-01 -1.15707743e+00 -9.11262155e-01 4.74319011e-01
-5.46345413e-02 -1.46473670e+00 3.73033762e-01 -9.88151059e-02
-6.57708466e-01 -1.43454278e+00 1.39608130e-01 -2.79287219e-01
4.71679807e-01 -1.11234598e-01 1.18737781e+00 6.07561231e-01
1.51506644e-02 2.74409026e-01 -6.51537359e-01 -2.86202580e-01
-6.57582700e-01 -2.23089848e-03 -1.88234925e-01 4.11767483e-01
6.49361432e-01 -1.02525198e+00 -2.09556490e-01 8.33675444e-01
-1.39561594e+00 1.40286162e-01 3.46553981e-01 3.48455347e-02
2.82789916e-01 3.44986916e-01 3.67382884e-01 -9.81919706e-01
4.49545413e-01 -4.24051017e-01 -7.13203192e-01 4.59999263e-01
-9.51173127e-01 1.99743629e-01 7.69583225e-01 -5.29855154e-02
-1.21330118e+00 -8.63184854e-02 -3.92273664e-01 4.52743590e-01
-6.75963879e-01 5.11334419e-01 -1.09331691e+00 1.31547257e-01
2.42566302e-01 2.16653034e-01 -4.58033793e-02 -1.01343131e+00
3.14947933e-01 6.49183035e-01 1.39397457e-01 -8.85045350e-01
7.96611011e-01 6.63277030e-01 2.42735580e-01 -4.33616489e-01
-7.19563812e-02 -4.92188781e-01 -4.71263766e-01 -1.30448490e-01
5.91723919e-01 -7.98626393e-02 -1.24201727e+00 -4.37608883e-02
-1.12869060e+00 1.44670218e-01 -3.45817745e-01 3.23432654e-01
-4.79153693e-01 2.54931778e-01 -2.61982054e-01 -9.19543982e-01
-3.76327373e-02 -1.22612858e+00 4.90670085e-01 4.87636477e-02
-1.39867082e-01 -7.07504153e-01 -2.95794547e-01 1.75032958e-01
7.10877240e-01 -7.50425607e-02 8.69696558e-01 -1.22273362e+00
-6.86094642e-01 -3.03430319e-01 -9.72931683e-02 1.60452217e-01
2.58030444e-01 4.55352627e-02 -5.25017738e-01 1.98051706e-01
-7.06174746e-02 7.07380831e-01 -1.84484392e-01 -5.86386979e-01
7.11058795e-01 -3.62109959e-01 -3.97277236e-01 2.05235824e-01
1.70312059e+00 6.16074741e-01 7.44611681e-01 5.71508288e-01
2.41966859e-01 7.05083311e-01 3.92316550e-01 4.06325191e-01
4.34116274e-01 1.31231380e+00 6.40384793e-01 2.30031446e-01
3.14469938e-03 1.33134738e-01 -4.44010496e-02 5.49636602e-01
-6.38829708e-01 9.16261971e-02 -9.99540985e-01 6.92167431e-02
-2.05756283e+00 -1.10741067e+00 -5.74700534e-01 2.19584799e+00
6.85744762e-01 5.48959561e-02 2.13808492e-02 4.26216334e-01
5.63703895e-01 -2.73350209e-01 -4.26744446e-02 -5.26552558e-01
-1.15414783e-01 8.22474360e-02 2.22231001e-01 6.78736091e-01
-3.52259815e-01 6.92573130e-01 4.57319212e+00 3.32265079e-01
-8.01902592e-01 3.90093386e-01 -5.23798704e-01 2.91597664e-01
-8.17550778e-01 5.80923557e-01 -1.04367173e+00 5.05896389e-01
9.46350753e-01 -5.03497541e-01 8.69274855e-01 7.61447668e-01
4.58496809e-01 -1.18903004e-01 -1.08824599e+00 4.22527552e-01
-2.15709042e-02 -1.05814517e+00 1.44388229e-01 2.07478791e-01
2.32767701e-01 -3.87555808e-01 -1.05590355e+00 -1.23914309e-01
2.27225408e-01 -4.59241360e-01 6.83363199e-01 9.04923201e-01
2.06712142e-01 -4.55496699e-01 7.86091506e-01 2.98808336e-01
-1.15516448e+00 -2.52447367e-01 -3.46065462e-02 5.20866290e-02
4.40946072e-01 5.92878103e-01 -6.87513828e-01 1.17283678e+00
7.79584408e-01 5.82973808e-02 -1.34693712e-01 1.27403498e+00
-3.46197873e-01 5.68625890e-02 -4.64807004e-01 -2.25672439e-01
-1.36114642e-01 -6.83307648e-01 4.65990186e-01 1.10106015e+00
5.90160191e-01 -4.75053079e-02 -5.47353663e-02 7.46323645e-01
3.23300987e-01 7.18801975e-01 -4.41789061e-01 1.22295924e-01
3.71179700e-01 1.23029888e+00 -7.35668123e-01 -5.69435596e-01
-7.52886653e-01 5.89216709e-01 -2.90805876e-01 5.00178695e-01
-6.53710067e-01 -2.86566675e-01 1.01866543e+00 5.32190382e-01
2.40910694e-01 -1.44240260e-01 2.55446553e-01 -1.24805880e+00
2.25047931e-01 -1.06349540e+00 4.18307692e-01 -7.30026603e-01
-1.15949225e+00 8.52262795e-01 2.01355591e-01 -1.15683699e+00
-4.19717580e-01 -4.06867445e-01 -2.02080812e-02 9.24966335e-01
-1.53905392e+00 -1.24250317e+00 -3.34721118e-01 8.57686102e-01
-8.50286260e-02 3.78380343e-02 1.43026924e+00 6.29782438e-01
-3.93231288e-02 -2.91910410e-01 -3.76406491e-01 -2.99656808e-01
5.01584888e-01 -9.53571439e-01 -1.67120859e-01 8.11033607e-01
2.49592699e-02 6.66547775e-01 9.92965698e-01 -5.16473651e-01
-1.37884808e+00 -2.12808609e-01 1.43659854e+00 -8.80898791e-04
8.58026385e-01 -2.65307844e-01 -9.68120098e-01 7.57871151e-01
-1.33635430e-03 -2.95386463e-02 6.28182113e-01 3.50184850e-02
-3.92561078e-01 -7.40317643e-01 -1.29114318e+00 5.52198291e-01
1.28428221e+00 -3.18490922e-01 -5.73478043e-01 3.46128076e-01
5.67834675e-01 1.17034100e-01 -1.16336036e+00 2.82994002e-01
6.62309408e-01 -1.28092492e+00 8.41169655e-01 -8.02587986e-01
-2.22271502e-01 -8.99237216e-01 -2.52931863e-01 -7.64317155e-01
2.48157367e-01 -4.37989414e-01 -3.30264010e-02 1.55785334e+00
5.02344787e-01 -1.08075440e+00 3.50730568e-01 1.03782809e+00
-1.80547044e-01 -2.14677930e-01 -8.58552873e-01 -1.00557983e+00
-8.31246674e-01 -6.65865660e-01 1.70537925e+00 8.82550299e-01
1.77387297e-01 -1.67974979e-01 1.76861398e-02 2.79843211e-01
3.70633543e-01 4.10246670e-01 7.08440244e-01 -1.96911979e+00
-4.99025345e-01 -5.89604378e-01 -5.80612421e-01 -2.70039380e-01
1.35002267e-02 -8.01681757e-01 -4.06005144e-01 -2.12688351e+00
-4.50872451e-01 -7.43133724e-01 -1.39721990e-01 6.58982217e-01
8.34016681e-01 -1.25188261e-01 2.16543034e-01 3.11025947e-01
-5.17293930e-01 -1.20623350e-01 1.17722988e+00 1.16291463e-01
-4.30399738e-02 2.23506331e-01 -5.48063517e-01 7.17621386e-01
6.23924911e-01 -5.89976966e-01 -3.81328762e-01 -1.74147353e-01
6.44457638e-01 9.42177176e-02 3.73322815e-02 -9.39212739e-01
3.77822995e-01 -7.57316709e-01 -4.70552176e-01 3.00569862e-01
4.21084374e-01 -1.97128904e+00 1.23803091e+00 2.94330359e-01
-6.65896609e-02 -3.55516493e-01 -5.57757318e-01 -7.18872920e-02
-2.01109916e-01 -8.75550330e-01 2.17790321e-01 -4.74863052e-01
-9.97016668e-01 7.65996352e-02 -3.95137578e-01 -5.48267484e-01
1.10062981e+00 -7.76502192e-02 -1.15431063e-01 -2.17204377e-01
-1.11658454e+00 -1.11466855e-01 7.27814198e-01 4.20576036e-01
-2.64932122e-03 -1.13242877e+00 -2.50541955e-01 7.59421512e-02
3.34266484e-01 -3.51301640e-01 -1.23228647e-01 7.35643029e-01
-4.42301303e-01 3.48859549e-01 -3.98114830e-01 -1.77378461e-01
-1.22931063e+00 8.42096150e-01 4.53570366e-01 -3.82941142e-02
-7.10072994e-01 -2.02921331e-01 -5.03263891e-01 -3.84529591e-01
7.78230056e-02 -2.61182815e-01 -6.53142810e-01 2.23958850e-01
8.20877433e-01 9.66666564e-02 4.74762022e-01 -8.22638035e-01
-5.39334893e-01 4.89109725e-01 6.79139256e-01 -3.35989118e-01
1.46101224e+00 -1.60446227e-01 -7.66233087e-01 2.93676585e-01
5.75384617e-01 3.05015713e-01 -6.11622393e-01 -1.65533558e-01
5.48721969e-01 -4.72992510e-01 -6.15198493e-01 -9.97546256e-01
-8.94369602e-01 2.13363454e-01 2.47148871e-01 9.97109890e-01
1.15472639e+00 1.85479924e-01 4.90047693e-01 4.17496711e-01
9.26130772e-01 -1.02581203e+00 -8.98895264e-01 1.48365587e-01
8.03744614e-01 -5.79581738e-01 -2.26396710e-01 -8.53458405e-01
-3.70175540e-01 1.43057859e+00 1.19303256e-01 4.22388077e-01
4.94400412e-01 7.02393055e-01 3.34247313e-02 -2.56387711e-01
-8.12915146e-01 -9.47169006e-01 -1.28167957e-01 7.66098499e-01
1.92874372e-01 1.03248149e-01 -1.39351046e+00 8.80237460e-01
1.03501320e-01 4.36834216e-01 4.56845820e-01 9.09686208e-01
-5.75976670e-01 -2.09858394e+00 -4.23028648e-01 3.46150547e-02
-2.40202680e-01 2.39375949e-01 -4.66501713e-02 9.14115906e-01
6.84712589e-01 1.11693811e+00 -1.09096460e-01 2.50974856e-02
8.35488677e-01 4.32824671e-01 3.75438631e-01 -4.62469220e-01
-6.92119718e-01 -8.66945088e-02 5.75689316e-01 -6.97475016e-01
-5.21922827e-01 -6.42809868e-01 -1.70799768e+00 -5.11017069e-02
1.55219855e-02 7.93271840e-01 1.29499078e+00 1.17842090e+00
2.73732301e-02 4.90901262e-01 3.21156740e-01 -3.05270135e-01
-3.75660568e-01 -4.61050361e-01 -6.40136957e-01 9.00617898e-01
-4.56365854e-01 -5.68234146e-01 -4.68627810e-01 2.30532169e-01] | [8.672204971313477, 7.010528087615967] |
579ea15c-e3cc-49b1-b7b2-4daa50651273 | generating-compressed-combinatory-proof | 2209.12592 | null | https://arxiv.org/abs/2209.12592v1 | https://arxiv.org/pdf/2209.12592v1.pdf | Generating Compressed Combinatory Proof Structures -- An Approach to Automated First-Order Theorem Proving | Representing a proof tree by a combinator term that reduces to the tree lets subtle forms of duplication within the tree materialize as duplicated subterms of the combinator term. In a DAG representation of the combinator term these straightforwardly factor into shared subgraphs. To search for proofs, combinator terms can be enumerated, like clausal tableaux, interwoven with unification of formulas that are associated with nodes of the enumerated structures. To restrict the search space, the enumeration can be based on proof schemas defined as parameterized combinator terms. We introduce here this "combinator term as proof structure" approach to automated first-order proving, present an implementation and first experimental results. The approach builds on a term view of proof structures rooted in condensed detachment and the connection method. It realizes features known from the connection structure calculus, which has not been implemented so far. | ['Christoph Wernhard'] | 2022-09-26 | null | null | null | null | ['automated-theorem-proving', 'automated-theorem-proving'] | ['miscellaneous', 'reasoning'] | [ 4.67397928e-01 8.82810414e-01 -1.82806045e-01 7.02661127e-02
-1.90620705e-01 -1.01738763e+00 9.91130590e-01 3.57282519e-01
2.87415832e-01 8.58433843e-01 -6.64255843e-02 -1.18315208e+00
-5.65075278e-01 -1.19643450e+00 -5.90570450e-01 -2.53651410e-01
-5.21787107e-01 5.77920496e-01 7.19098210e-01 -3.07849765e-01
3.60636175e-01 3.14913601e-01 -1.45993423e+00 5.43703735e-01
7.44181156e-01 3.99158865e-01 1.87011987e-01 8.01583290e-01
-7.12402582e-01 8.25318813e-01 -5.72573841e-01 -7.24248409e-01
5.86633198e-02 -4.27689075e-01 -1.13022089e+00 -1.13487341e-01
3.81807268e-01 -2.54129469e-01 9.94578153e-02 1.17384589e+00
-5.93969166e-01 -5.52027643e-01 5.20099163e-01 -1.72947991e+00
3.87781449e-02 1.14016914e+00 -2.69941688e-02 -1.71611026e-01
6.68059409e-01 -4.47239846e-01 1.21490264e+00 -2.14359090e-01
1.21081829e+00 1.46558928e+00 6.31463885e-01 4.03684914e-01
-1.75925732e+00 -1.99327230e-01 8.85233730e-02 3.26495230e-01
-1.14392781e+00 -2.23410591e-01 5.56236982e-01 -1.97336301e-01
1.20302188e+00 7.65419006e-01 6.61701918e-01 4.64779466e-01
6.53904438e-01 4.52762395e-01 9.13989902e-01 -9.23219681e-01
1.97711855e-01 1.57835886e-01 6.02886260e-01 1.20332885e+00
8.57594430e-01 -2.08033398e-01 -2.12145746e-01 -5.07575631e-01
7.19633102e-01 -4.19221073e-01 -9.23822895e-02 -8.54657471e-01
-9.42593038e-01 8.08255970e-01 -1.77930266e-01 4.35197145e-01
2.94538289e-01 4.31790173e-01 6.17552817e-01 4.65593040e-01
-2.43601322e-01 2.94490546e-01 -5.20451665e-01 1.33753285e-01
-5.60030699e-01 6.92592859e-01 1.34333396e+00 1.23234212e+00
6.87794507e-01 -5.24377227e-01 3.22096124e-02 2.87201613e-01
3.65136415e-01 2.26477236e-01 -2.56608278e-01 -1.15528715e+00
3.30960721e-01 9.60255444e-01 4.59764116e-02 -6.21327281e-01
-1.60637379e-01 -1.18571632e-01 -2.54561037e-01 3.44284683e-01
7.10060894e-01 2.36758575e-01 -2.63252139e-01 1.65098238e+00
4.25672650e-01 -2.04450693e-02 1.30419627e-01 1.47638842e-01
3.90411913e-01 2.77720541e-01 -2.09830493e-01 -4.48249549e-01
1.46535909e+00 -6.11388862e-01 -6.44566834e-01 6.37614131e-01
1.09482062e+00 -3.50742042e-01 2.35901013e-01 8.32443595e-01
-1.45256078e+00 2.30657205e-01 -1.09884632e+00 -2.03197807e-01
-6.31652772e-01 -3.80709082e-01 9.71126497e-01 5.51039815e-01
-1.14576626e+00 6.93010330e-01 -4.95872617e-01 -1.60755664e-01
-4.74740118e-02 3.71011019e-01 -2.60657132e-01 -1.29460663e-01
-1.07357001e+00 7.97712982e-01 5.74184239e-01 3.72548215e-03
-9.87178862e-01 -5.84528387e-01 -8.92499924e-01 1.31015629e-01
7.07613289e-01 -8.95688295e-01 1.55822861e+00 -4.32699978e-01
-1.07184780e+00 9.24360693e-01 -2.09938854e-01 -6.63762093e-01
2.21217141e-01 6.27069831e-01 -3.61898869e-01 1.49343669e-01
3.44140559e-01 -3.17942421e-03 5.60059488e-01 -1.14493430e+00
-7.26762831e-01 -2.45315164e-01 9.95092273e-01 -4.36366290e-01
4.31783468e-01 3.15076530e-01 6.84890226e-02 -4.37742770e-02
2.48785570e-01 -6.41909540e-01 -4.69347872e-02 -2.37744913e-01
-4.84353602e-01 -5.58441520e-01 7.24895775e-01 -5.97793274e-02
1.55452704e+00 -1.70746112e+00 3.60153407e-01 4.49815154e-01
5.23991346e-01 1.68986563e-02 2.95922518e-01 8.52650464e-01
-1.14414223e-01 4.03370202e-01 5.41981682e-02 2.80040562e-01
6.52007341e-01 7.14211822e-01 -3.87597740e-01 2.24043623e-01
1.11399256e-02 7.30529010e-01 -1.02219713e+00 -6.67774677e-01
-6.05543097e-03 -4.09224182e-01 -7.35417962e-01 -1.51519790e-01
-9.49334502e-01 -6.76552057e-01 -6.24643385e-01 4.32283521e-01
8.99627686e-01 -8.10569972e-02 8.80326986e-01 2.11289614e-01
-5.98230362e-01 8.77695262e-01 -1.50467372e+00 1.51140606e+00
-2.30413809e-01 6.36438504e-02 6.17596745e-01 -8.66627991e-01
5.92137456e-01 1.90090820e-01 -1.58648640e-01 -1.56026825e-01
9.46341306e-02 2.60331690e-01 1.33120745e-01 -3.19003284e-01
3.51259530e-01 -8.03605095e-02 -1.36032954e-01 4.24304008e-01
-3.95756820e-03 -2.70556122e-01 9.20832336e-01 7.37134218e-01
1.37742054e+00 3.94251466e-01 4.11731064e-01 -6.10077798e-01
1.02213502e+00 3.61473382e-01 3.32028627e-01 8.59031439e-01
5.29945970e-01 -4.19149369e-01 1.28728414e+00 -4.43939239e-01
-8.90273809e-01 -1.24441516e+00 -2.98229694e-01 6.26702249e-01
6.02785312e-02 -1.54703844e+00 -9.28265631e-01 -7.63512433e-01
2.65542120e-02 7.21250951e-01 -3.02406460e-01 2.48344243e-01
-5.41502833e-01 1.02292404e-01 7.63856053e-01 8.46237466e-02
2.35191658e-01 -6.17611051e-01 -9.45609093e-01 3.17761958e-01
1.36369929e-01 -9.26982284e-01 3.23063910e-01 3.68439943e-01
-1.02998042e+00 -1.61712158e+00 2.60981381e-01 -5.41402340e-01
6.87442124e-01 2.14678962e-02 1.03300369e+00 6.71113789e-01
-3.50593716e-01 3.38655084e-01 -2.50000030e-01 -5.10776818e-01
-8.42071176e-01 -2.09531430e-02 -3.45977485e-01 -5.65468788e-01
5.96624799e-02 -7.59058833e-01 1.38062716e-01 -6.63831383e-02
-8.49736452e-01 2.46482939e-01 5.62018380e-02 7.37955928e-01
1.83955461e-01 2.78304875e-01 -7.47991577e-02 -8.69929016e-01
7.01106250e-01 -2.71635175e-01 -1.20002234e+00 5.15687764e-01
-5.52218676e-01 5.47354400e-01 7.17088699e-01 -2.93349829e-02
-7.89561391e-01 -4.00109470e-01 5.78759730e-01 -1.64092496e-01
-9.66574103e-02 7.81304538e-01 -1.30337432e-01 5.23348637e-02
3.37459743e-01 -1.32486805e-01 1.24542765e-01 -3.30166817e-01
4.50326890e-01 1.15461133e-01 5.63419342e-01 -1.34402585e+00
7.30048895e-01 1.30652636e-01 1.18956459e+00 -3.00582409e-01
-7.18764961e-02 9.02897641e-02 -2.99054027e-01 -1.33190304e-01
3.70021075e-01 -2.06835628e-01 -8.89306784e-01 -6.69492111e-02
-1.64489186e+00 -3.07875007e-01 -4.10027474e-01 -4.59996983e-03
-6.52320027e-01 6.23798072e-01 -5.69712520e-01 -1.10898435e+00
2.29371265e-01 -9.17648554e-01 9.30684209e-01 -2.21926123e-01
-5.26285708e-01 -9.68398988e-01 1.88900083e-01 -2.27761716e-01
4.19032387e-02 2.23004594e-01 1.75150990e+00 -7.01323390e-01
-1.06705225e+00 -6.39313906e-02 -2.11917222e-01 -2.24878922e-01
-3.09356481e-01 2.08476633e-01 -3.33370179e-01 2.31556952e-01
-2.77001560e-01 1.76374286e-01 9.68846083e-02 -1.33362979e-01
6.09876990e-01 -4.76139992e-01 -6.52949691e-01 9.56933722e-02
1.58083129e+00 2.14107465e-02 7.78732538e-01 5.09703100e-01
-9.77622066e-03 3.80751729e-01 2.31137171e-01 5.48111275e-03
3.44748139e-01 5.83595276e-01 4.91324246e-01 6.62429154e-01
-2.39612591e-02 -4.57156360e-01 3.00815374e-01 3.16921890e-01
-1.88313797e-01 1.56538144e-01 -8.99308383e-01 3.62911403e-01
-1.83431625e+00 -1.28747189e+00 -9.87400472e-01 2.08576012e+00
8.39899600e-01 3.07115525e-01 4.81626345e-03 6.11153662e-01
4.94906157e-01 -2.74405956e-01 3.89300734e-01 -1.25223160e+00
3.14842850e-01 4.73580033e-01 3.47917944e-01 1.08720183e+00
-4.26231772e-01 6.67500615e-01 7.16444683e+00 5.82279503e-01
-3.94961923e-01 1.08953848e-01 -4.76782918e-01 2.42619038e-01
-7.92326212e-01 8.56174946e-01 -7.95031667e-01 1.71523690e-02
8.00297976e-01 -3.59192580e-01 7.93112218e-01 6.99391007e-01
-1.27253473e-01 -4.34710622e-01 -1.54215813e+00 1.20056912e-01
-3.78908455e-01 -1.59532046e+00 4.60095257e-02 1.87736735e-01
3.21793914e-01 -5.34486949e-01 -6.77377522e-01 3.08671653e-01
3.43290329e-01 -7.17711627e-01 9.90646601e-01 2.70652264e-01
5.57710171e-01 -5.73592246e-01 3.82208556e-01 4.69979852e-01
-1.17154014e+00 -2.45093331e-01 -9.01858807e-02 -6.15326464e-01
5.74632734e-02 2.38738284e-01 -7.27649629e-01 1.07402229e+00
8.60795900e-02 3.24922390e-02 -1.32181212e-01 8.67359221e-01
-5.73308289e-01 3.21910471e-01 -4.27823603e-01 -3.04149240e-01
2.04518855e-01 -8.18448365e-01 8.55845034e-01 1.28602707e+00
3.40246893e-02 1.98921576e-01 -5.40331826e-02 1.22491896e+00
4.06676710e-01 -2.32836336e-01 -8.62155616e-01 2.67601848e-01
3.27041000e-01 1.03406906e+00 -7.81064987e-01 -5.73725820e-01
-3.66780996e-01 2.27498680e-01 1.02716707e-01 2.25080952e-01
-8.90224159e-01 -5.06933689e-01 2.77649641e-01 4.38876003e-01
3.40313822e-01 -3.94403309e-01 -1.88548058e-01 -9.96809185e-01
1.29831880e-01 -8.01123500e-01 2.84147710e-01 -9.79721129e-01
-6.21975005e-01 1.10146515e-01 9.83363509e-01 -4.95380700e-01
-6.08936906e-01 -8.17896545e-01 -8.12439144e-01 1.05048609e+00
-1.17003155e+00 -1.13782871e+00 2.79244393e-01 5.94228923e-01
4.02501784e-02 2.30736434e-01 1.31207466e+00 -1.64793491e-01
-2.85108179e-01 3.71523857e-01 -3.45125616e-01 -4.34877843e-01
-2.37577021e-01 -1.29269266e+00 -7.08532780e-02 9.46861267e-01
-6.23245463e-02 1.25151134e+00 8.67102981e-01 -5.57651103e-01
-1.85892034e+00 -5.03507912e-01 1.61325383e+00 -2.67746925e-01
1.30536413e+00 -6.13335431e-01 -6.02743983e-01 8.27596545e-01
3.40375662e-01 -5.70744991e-01 1.04680195e-01 4.44813341e-01
-6.78621948e-01 -1.39338568e-01 -1.17902720e+00 8.57040584e-01
1.35317874e+00 -5.26713729e-01 -9.22810018e-01 3.40390027e-01
3.58933270e-01 -3.46929699e-01 -7.53611445e-01 1.67965397e-01
6.22708142e-01 -9.49082017e-01 5.17167926e-01 -6.69527352e-01
4.80285197e-01 -6.53531551e-01 -1.00905888e-01 -6.90503597e-01
-2.34358847e-01 -1.10680068e+00 -3.67008388e-01 1.30837047e+00
5.72208285e-01 -8.12918901e-01 3.78677011e-01 6.44223273e-01
-1.60404667e-01 -7.00413167e-01 -1.21121740e+00 -9.11533773e-01
1.41254947e-01 -4.51726377e-01 8.13829064e-01 6.72870457e-01
1.14965427e+00 4.78129774e-01 4.53993171e-01 -6.62321523e-02
6.41350031e-01 3.99382502e-01 7.07736552e-01 -1.56885850e+00
-4.88059878e-01 -7.86692739e-01 -3.98415655e-01 -6.38423085e-01
4.06936556e-01 -1.54680943e+00 -4.12982017e-01 -1.64434743e+00
7.88671300e-02 -5.45660853e-01 3.44306737e-01 7.56007731e-01
7.78169811e-01 -4.23962027e-01 1.42465025e-01 -3.28455925e-01
-5.60552955e-01 -1.81300312e-01 9.18792784e-01 -6.27414137e-02
2.78335690e-01 -4.97073412e-01 -3.56038600e-01 8.65381420e-01
6.32052600e-01 -5.88491082e-01 -6.44376338e-01 3.30160442e-03
7.42042959e-01 5.81852853e-01 7.06163585e-01 -5.66812634e-01
2.22692057e-01 -3.27879906e-01 -6.33419871e-01 -3.32515389e-01
-2.29293212e-01 -9.16151047e-01 7.76892543e-01 5.44162273e-01
-2.59347439e-01 1.42018765e-01 1.82555467e-01 2.18204767e-01
2.13916302e-01 -1.07131112e+00 1.11860164e-01 -1.82943076e-01
-2.89743215e-01 -4.67539638e-01 -6.69242442e-01 -3.08974475e-01
9.17817593e-01 -2.59786576e-01 -4.38592404e-01 1.56300396e-01
-6.09020889e-01 2.32123405e-01 6.19397283e-01 -1.10428467e-01
2.49805972e-01 -1.05013645e+00 -2.14252561e-01 4.42633778e-02
-1.68490663e-01 1.50952563e-01 -3.62215012e-01 9.25156355e-01
-5.68447769e-01 6.09497488e-01 -9.84932333e-02 -3.14567029e-01
-1.56304753e+00 8.47446740e-01 3.75099361e-01 -6.59509480e-01
-5.84795177e-01 2.60863185e-01 -9.05735344e-02 -4.46758904e-02
1.80071499e-02 -8.55455041e-01 1.84671536e-01 -2.86000699e-01
5.22731543e-01 3.90064448e-01 2.30719452e-03 -8.36677290e-03
-4.95643318e-01 4.64653552e-01 1.36485174e-01 -1.52365237e-01
9.72561359e-01 9.70622972e-02 -1.14049327e+00 2.15369031e-01
7.02437758e-01 3.85601968e-01 -2.58353531e-01 2.99658269e-01
1.46364495e-01 -6.40498772e-02 -4.72845227e-01 -7.60148942e-01
-1.15890928e-01 3.83572102e-01 -3.91420305e-01 9.75387573e-01
7.99046159e-01 1.97042361e-01 1.54416800e-01 5.80775797e-01
8.81706238e-01 -6.86392665e-01 -8.11883688e-01 6.84831262e-01
7.52528906e-01 6.76544681e-02 1.11474171e-01 -9.85583961e-01
3.25666636e-01 1.57457888e+00 -1.33376583e-01 -3.75250399e-01
2.55370140e-01 9.18399334e-01 -6.39799476e-01 -3.64610672e-01
-1.07792366e+00 -1.11694753e-01 -3.41153234e-01 6.03404403e-01
7.69879222e-02 4.32092361e-02 -1.25170898e+00 7.76238978e-01
-2.37306222e-01 5.47346771e-01 9.07557786e-01 1.09115517e+00
-3.57434720e-01 -1.88778734e+00 -6.74147487e-01 -7.20071932e-03
-4.56966192e-01 -1.80220693e-01 -4.29048568e-01 1.25343740e+00
6.76534474e-01 8.90221179e-01 -6.25828952e-02 -1.16852066e-02
2.49073550e-01 4.64172542e-01 1.47852981e+00 -5.79508841e-01
-6.33457959e-01 -2.34752223e-01 9.92939234e-01 -3.22884381e-01
-3.50994378e-01 -8.40072572e-01 -1.47336888e+00 -6.84474349e-01
-6.03258491e-01 5.94374120e-01 5.30666292e-01 9.05026674e-01
-1.10460743e-01 4.88944560e-01 2.15103060e-01 -4.13390249e-01
-6.86323583e-01 -3.98012966e-01 -6.45540237e-01 -1.46302328e-01
4.01665531e-02 -6.32364094e-01 -2.25460857e-01 1.04877884e-02] | [8.77391242980957, 6.851644992828369] |
8e4e4c6c-ed1f-4a55-a088-d1299c3187fc | anomaly-detection-inspired-few-shot-medical | 2203.02048 | null | https://arxiv.org/abs/2203.02048v1 | https://arxiv.org/pdf/2203.02048v1.pdf | Anomaly Detection-Inspired Few-Shot Medical Image Segmentation Through Self-Supervision With Supervoxels | Recent work has shown that label-efficient few-shot learning through self-supervision can achieve promising medical image segmentation results. However, few-shot segmentation models typically rely on prototype representations of the semantic classes, resulting in a loss of local information that can degrade performance. This is particularly problematic for the typically large and highly heterogeneous background class in medical image segmentation problems. Previous works have attempted to address this issue by learning additional prototypes for each class, but since the prototypes are based on a limited number of slices, we argue that this ad-hoc solution is insufficient to capture the background properties. Motivated by this, and the observation that the foreground class (e.g., one organ) is relatively homogeneous, we propose a novel anomaly detection-inspired approach to few-shot medical image segmentation in which we refrain from modeling the background explicitly. Instead, we rely solely on a single foreground prototype to compute anomaly scores for all query pixels. The segmentation is then performed by thresholding these anomaly scores using a learned threshold. Assisted by a novel self-supervision task that exploits the 3D structure of medical images through supervoxels, our proposed anomaly detection-inspired few-shot medical image segmentation model outperforms previous state-of-the-art approaches on two representative MRI datasets for the tasks of abdominal organ segmentation and cardiac segmentation. | ['Michael Kampffmeyer', 'Robert Jenssen', 'Srishti Gautam', 'Stine Hansen'] | 2022-03-03 | null | null | null | null | ['cardiac-segmentation'] | ['medical'] | [ 5.48877537e-01 2.94302255e-01 -1.84908912e-01 -4.98637468e-01
-7.75735199e-01 -1.03236102e-01 4.32407767e-01 7.97049224e-01
-4.52278048e-01 3.77103060e-01 -1.04828291e-01 1.26801595e-01
1.02134332e-01 -6.59948707e-01 -5.92331767e-01 -9.81643856e-01
5.12219071e-02 7.33561277e-01 8.35134149e-01 -4.86269733e-03
6.30578101e-02 2.72622734e-01 -1.56181347e+00 2.59007961e-01
1.03052759e+00 9.36165631e-01 2.08983906e-02 5.84093690e-01
-4.36244339e-01 6.57556057e-01 -5.44689000e-01 -3.80807705e-02
2.35144213e-01 -8.61768603e-01 -8.54094505e-01 6.81542754e-01
3.33518624e-01 -2.69380450e-01 8.59565884e-02 1.50491905e+00
3.04868639e-01 3.64738613e-01 8.42754424e-01 -9.07189488e-01
-6.53347820e-02 4.18639332e-01 -6.91625059e-01 6.47606015e-01
-1.10856876e-01 1.72959104e-01 9.06051219e-01 -6.36995435e-01
7.85962880e-01 6.74585760e-01 4.73437339e-01 6.57135487e-01
-1.33629370e+00 -7.77660757e-02 3.02688241e-01 -3.74619471e-04
-1.25457394e+00 -2.44672462e-01 8.50125611e-01 -5.59461474e-01
4.67291564e-01 2.40298986e-01 7.86087692e-01 7.05121994e-01
5.24544716e-02 9.31701124e-01 8.76807332e-01 -4.67143089e-01
7.84808934e-01 8.70113745e-02 5.94427824e-01 8.35388958e-01
3.44663739e-01 -4.15395379e-01 -7.82352760e-02 -2.38682091e-01
3.46206218e-01 3.43857974e-01 -4.59333099e-02 -9.30171788e-01
-1.03153729e+00 8.79260361e-01 4.64059055e-01 7.00976670e-01
-6.39443576e-01 -8.92281085e-02 6.51446640e-01 5.24558425e-02
6.48110688e-01 1.87143639e-01 -1.69148803e-01 2.50065356e-01
-1.42394233e+00 -5.18613271e-02 7.18328297e-01 5.64766347e-01
9.19254065e-01 -6.22728616e-02 -3.03372651e-01 7.53932595e-01
1.50163695e-01 -9.54694077e-02 5.95117211e-01 -7.80530214e-01
-2.03901067e-01 6.69983327e-01 -1.49751142e-01 -7.46599257e-01
-3.53232384e-01 -2.63209492e-01 -7.54850686e-01 1.43151641e-01
5.75556576e-01 7.14576989e-02 -1.65966630e+00 1.37884986e+00
7.63189554e-01 3.45271677e-01 -9.25688744e-02 9.57737446e-01
7.58289039e-01 2.13312596e-01 2.60791868e-01 -4.31122005e-01
1.27930534e+00 -1.10776818e+00 -8.74078691e-01 -2.39149407e-01
7.99194574e-01 -4.18410569e-01 9.77583885e-01 1.99681103e-01
-1.03329802e+00 -2.12755173e-01 -1.17719412e+00 3.87365311e-01
-2.70056695e-01 -5.79522073e-01 5.36308289e-01 8.00395131e-01
-7.13631928e-01 6.37188435e-01 -1.32461560e+00 -5.99900961e-01
8.99116516e-01 2.24901706e-01 -8.62648040e-02 -2.51253545e-01
-8.24855864e-01 6.40776575e-01 5.59934676e-01 -2.38237053e-01
-8.17996979e-01 -6.73839986e-01 -1.05756557e+00 -9.96655300e-02
7.18100369e-01 -6.37638450e-01 1.03753376e+00 -1.22454357e+00
-1.15056217e+00 1.08447850e+00 -3.99232581e-02 -6.49759412e-01
5.34244478e-01 1.10567793e-01 -1.85954332e-01 7.54350245e-01
2.82003790e-01 6.51604712e-01 9.87894773e-01 -1.29564369e+00
-4.35540974e-01 -5.14680147e-01 -1.99843079e-01 1.14881814e-01
-2.15441823e-01 1.12076998e-02 -5.17745137e-01 -6.94436312e-01
4.38538432e-01 -8.05828691e-01 -8.20623577e-01 8.02506879e-02
-3.65337372e-01 1.62983820e-01 7.27398574e-01 -3.78570557e-01
1.14836240e+00 -2.24466658e+00 -6.66733831e-02 3.37702632e-01
2.45423883e-01 2.60140687e-01 3.34200710e-01 -3.56027097e-01
1.03124291e-01 -3.85082886e-02 -8.35222185e-01 -2.28195876e-01
-3.32614720e-01 6.31721139e-01 2.50218391e-01 7.09840655e-01
1.04140475e-01 7.83185005e-01 -1.07662106e+00 -1.12899101e+00
5.28136790e-01 2.00395435e-01 -5.95906615e-01 1.43261515e-02
-4.13038313e-01 6.13960445e-01 -4.53867346e-01 8.83438647e-01
4.73336786e-01 -4.28792387e-01 -9.72641259e-02 1.21847475e-02
2.42199183e-01 -2.38026708e-01 -1.18976092e+00 2.05811095e+00
3.89085971e-02 -7.25165457e-02 7.34819844e-02 -1.33632779e+00
5.12939513e-01 3.24157178e-01 1.08836806e+00 -3.90661448e-01
3.29195797e-01 3.57596546e-01 1.42534792e-01 -4.17059094e-01
1.40298181e-03 -5.31089723e-01 3.62814702e-02 4.01605964e-01
2.39028648e-01 -1.41214624e-01 3.86433959e-01 3.93884271e-01
1.27363598e+00 -9.38046128e-02 5.00416636e-01 -2.86014795e-01
4.79935527e-01 3.24779004e-01 8.27817738e-01 9.83592153e-01
-7.94153154e-01 1.01454341e+00 3.61108989e-01 -4.46249098e-01
-8.52083862e-01 -1.08479309e+00 -2.38338366e-01 1.01672482e+00
3.89275610e-01 -2.22315490e-01 -1.08321190e+00 -9.84978437e-01
-3.56447428e-01 7.23941207e-01 -7.83923030e-01 -1.73146561e-01
-5.29251635e-01 -1.03454840e+00 2.46515676e-01 3.78489822e-01
1.81899309e-01 -1.05448520e+00 -1.18094206e+00 4.53901768e-01
-6.74327761e-02 -1.06529558e+00 -3.59119028e-01 4.28398788e-01
-1.18311095e+00 -1.15927637e+00 -9.61899996e-01 -5.74899852e-01
1.03703201e+00 2.12350145e-01 1.09477377e+00 3.95325214e-01
-8.77845287e-01 5.51517367e-01 -3.24128509e-01 -2.81596810e-01
-4.74875867e-01 -1.92405596e-01 -2.55190879e-01 2.27431819e-01
4.23049986e-01 -4.37515438e-01 -8.48757982e-01 -6.01840997e-03
-1.01620591e+00 -2.26240039e-01 5.21900296e-01 9.70294297e-01
9.42135811e-01 -8.47990513e-02 2.76403338e-01 -1.33602679e+00
-1.25861928e-01 -4.56073046e-01 -2.83602834e-01 2.49145910e-01
-4.13834035e-01 -7.50493631e-02 3.22194397e-01 -4.35793757e-01
-1.02683914e+00 3.86938125e-01 -5.26269004e-02 -7.49598742e-01
-6.01096213e-01 2.27505371e-01 2.08514690e-01 4.74024005e-02
6.84041023e-01 1.22973688e-01 9.32062939e-02 -2.66464800e-01
2.47412801e-01 3.36235583e-01 5.63520133e-01 -2.67242014e-01
4.17020619e-01 8.68711591e-01 4.09472808e-02 -1.14114618e+00
-1.11927700e+00 -1.06808507e+00 -9.47989702e-01 -1.71259895e-01
1.33386314e+00 -6.19982660e-01 1.32895619e-01 2.37344384e-01
-7.78878570e-01 -2.04839706e-01 -7.53283203e-01 4.79637265e-01
-6.36740148e-01 6.35166287e-01 -7.49776006e-01 -6.92285955e-01
-3.59266341e-01 -1.33133578e+00 1.09169018e+00 9.23811942e-02
-4.29506630e-01 -9.90440726e-01 4.25187461e-02 3.08855116e-01
2.02593192e-01 4.99730170e-01 1.01040781e+00 -1.17770767e+00
-4.03180182e-01 -2.90957004e-01 1.82113368e-02 1.98893890e-01
2.22755715e-01 -1.89719871e-01 -9.22051072e-01 -1.21678032e-01
3.22682500e-01 -2.30386078e-01 1.10335112e+00 7.48332560e-01
1.18373382e+00 1.63474634e-01 -4.33650553e-01 5.15849352e-01
1.29003072e+00 1.87187158e-02 3.45617980e-01 1.11827560e-01
7.81773627e-01 5.05405188e-01 7.80825973e-01 3.80796224e-01
3.33587863e-02 3.67125928e-01 4.35274929e-01 -3.23289871e-01
-5.51584996e-02 2.99811006e-01 -1.67197034e-01 6.40872598e-01
1.15555719e-01 1.48769826e-01 -1.08705211e+00 8.29998553e-01
-1.89510584e+00 -8.05167973e-01 4.58315462e-02 2.13906837e+00
7.11017609e-01 3.70169878e-01 3.20781946e-01 1.32171601e-01
7.50229716e-01 2.64302552e-01 -6.65697753e-01 -1.64780989e-01
2.14889854e-01 2.04496324e-01 3.92164916e-01 2.56501108e-01
-1.41509211e+00 8.30768049e-01 5.84193897e+00 5.76671958e-01
-9.37693954e-01 4.01396155e-01 8.01105917e-01 -7.95622468e-02
-2.65952814e-02 5.95069714e-02 -4.19699013e-01 4.50494438e-01
6.91214740e-01 1.73907112e-02 -1.75162837e-01 9.19234395e-01
-1.09473810e-01 -3.92932326e-01 -1.11366832e+00 7.72886515e-01
3.30005527e-01 -1.19978309e+00 5.36268987e-02 -1.66377917e-01
8.56654704e-01 -1.48193794e-03 -2.45488122e-01 1.18827224e-01
-2.98351180e-02 -6.84040368e-01 2.67961860e-01 3.26414168e-01
3.24989438e-01 -5.43512881e-01 7.62490869e-01 3.92685652e-01
-9.08951759e-01 6.70762956e-02 -3.92071545e-01 3.07280511e-01
1.65493280e-01 6.99450612e-01 -7.43996978e-01 3.96618158e-01
6.20893002e-01 5.43546259e-01 -5.51832259e-01 1.30938637e+00
1.33080482e-01 5.99055588e-01 -3.25146079e-01 4.20733005e-01
5.11001229e-01 -9.45311189e-02 7.24806011e-01 1.09566569e+00
5.44007495e-03 2.25824758e-01 6.70071006e-01 8.52270365e-01
1.90192744e-01 3.26369524e-01 -4.65109617e-01 1.43724471e-01
-1.00263499e-01 1.29543734e+00 -1.71545255e+00 -7.83892751e-01
-6.14243507e-01 1.17117500e+00 -4.36902381e-02 1.00641921e-01
-6.64585233e-01 -3.68810110e-02 3.01279843e-01 1.61669835e-01
5.32002449e-01 1.00880921e-01 -3.55195701e-01 -1.17049479e+00
-2.30191350e-01 -7.05816627e-01 8.64376426e-01 -1.38346404e-01
-1.14059520e+00 3.54498744e-01 3.65352929e-02 -1.11149716e+00
-2.07637623e-01 -1.80296689e-01 -7.83904016e-01 2.29635626e-01
-1.34828162e+00 -1.05075943e+00 -2.54358292e-01 6.04833901e-01
8.39611113e-01 1.28891632e-01 8.65144730e-01 2.57304639e-01
-4.58838791e-01 2.25939393e-01 -7.56494105e-02 1.49802268e-01
5.93098938e-01 -1.59092879e+00 -3.18544582e-02 9.48648572e-01
2.98312664e-01 4.26487476e-01 8.52607548e-01 -8.11697543e-01
-8.21717203e-01 -9.73367691e-01 3.07023078e-01 -2.70198315e-01
5.33874094e-01 -1.97805896e-01 -1.36988890e+00 4.61860329e-01
-1.22286469e-01 9.45288539e-01 8.84028733e-01 -2.53258377e-01
2.48351991e-01 3.23484153e-01 -1.34440660e+00 4.89944875e-01
8.34810436e-01 -7.96567090e-03 -8.15944076e-01 4.81578082e-01
5.41337907e-01 -3.85452867e-01 -7.15914905e-01 4.04138923e-01
7.30424002e-02 -9.48379099e-01 9.14153576e-01 -6.07678235e-01
1.76640317e-01 -2.82448828e-01 -2.46652327e-02 -9.98334527e-01
1.56456620e-01 -3.33128184e-01 -2.87154764e-01 7.14845955e-01
5.53562157e-02 -1.64886266e-01 1.07013631e+00 7.27925301e-01
-2.54432589e-01 -7.14961648e-01 -9.31116998e-01 -6.01320386e-01
-1.06033452e-01 -2.98326552e-01 1.16816752e-01 9.49030936e-01
-2.31801391e-01 4.41754572e-02 -3.90072241e-02 1.67725623e-01
1.00822437e+00 2.36710429e-01 4.50324297e-01 -1.40209365e+00
-1.77943647e-01 -2.98702478e-01 -7.80406833e-01 -3.79305214e-01
6.07036501e-02 -8.04802537e-01 4.08904105e-01 -1.44489837e+00
3.82950336e-01 -3.18617374e-01 -7.16060996e-01 2.95543879e-01
-4.21055198e-01 5.24521530e-01 -1.09500021e-01 1.38242662e-01
-1.15875888e+00 2.80649394e-01 1.09406912e+00 -2.94238448e-01
-3.15009177e-01 1.48386985e-01 -3.16599756e-01 1.15156794e+00
5.63007534e-01 -6.62607789e-01 -2.73446023e-01 1.98951542e-01
-3.49402279e-01 -1.56204611e-01 3.28996032e-01 -1.08468616e+00
3.81087214e-01 -3.28968428e-02 3.38826895e-01 -4.89536166e-01
1.58253625e-01 -7.84608006e-01 -3.65185440e-01 7.22333610e-01
-2.32702255e-01 -5.76320410e-01 -1.23122431e-01 8.44675243e-01
-2.97272146e-01 -5.32746315e-01 1.23772478e+00 -6.52328730e-01
-9.41205323e-01 4.10600871e-01 -4.74641323e-01 3.30029577e-01
1.36882138e+00 -4.89203215e-01 2.70072430e-01 -1.95206273e-02
-1.28257203e+00 1.60579756e-01 6.48610473e-01 9.82231647e-03
5.25672793e-01 -9.17660832e-01 -3.76403600e-01 2.39825040e-01
3.22183102e-01 3.54122758e-01 3.89007628e-01 1.22162735e+00
-4.74614143e-01 -1.55661047e-01 -1.98118672e-01 -1.06909108e+00
-1.38290787e+00 7.53218234e-01 4.45716202e-01 -3.26403350e-01
-1.21034360e+00 7.12251604e-01 3.41564864e-01 -5.77148870e-02
2.55597800e-01 -3.34291548e-01 -1.54519215e-01 1.61831692e-01
4.75605488e-01 2.16066360e-01 1.28160819e-01 -6.70038044e-01
-5.16747594e-01 4.84957486e-01 -4.58902478e-01 6.01355582e-02
1.23067188e+00 -1.54744923e-01 1.35855051e-03 8.13983738e-01
9.51665640e-01 -3.25447619e-01 -1.04088175e+00 -5.43708205e-01
3.87462974e-01 -4.60314184e-01 2.14357167e-01 -3.99862021e-01
-1.03312719e+00 8.95512223e-01 7.30961382e-01 6.60124421e-02
9.66673076e-01 1.88851357e-01 8.66524935e-01 2.15741605e-01
1.51311189e-01 -1.39331400e+00 3.67154330e-01 2.14416906e-01
3.44243869e-02 -1.68691289e+00 1.50610328e-01 -4.75842893e-01
-7.60108948e-01 8.35903287e-01 5.70545673e-01 -1.19433619e-01
6.92343235e-01 1.96514070e-01 2.32965618e-01 -5.01978278e-01
-3.61301750e-01 -5.01553357e-01 3.33220541e-01 3.61803740e-01
2.11992547e-01 -1.48932517e-01 -3.31750475e-02 2.91045964e-01
4.77360070e-01 -1.17818743e-01 4.25683349e-01 1.39872515e+00
-7.17965364e-01 -8.12871158e-01 -3.09426874e-01 7.73519099e-01
-9.05958533e-01 1.04118578e-01 -1.37853146e-01 6.61187232e-01
6.06797263e-02 5.77218890e-01 2.20114723e-01 3.82770002e-01
2.12507062e-02 3.82831603e-01 4.78896618e-01 -1.10626543e+00
-3.99790466e-01 4.57557350e-01 -4.53112662e-01 -6.14738703e-01
-5.74015081e-01 -6.15421355e-01 -1.56051338e+00 4.32509959e-01
-4.06823933e-01 6.56638443e-02 4.04662132e-01 1.10615861e+00
-1.22171953e-01 6.66720092e-01 3.40313494e-01 -6.71686709e-01
-4.37897086e-01 -6.26847446e-01 -7.74099410e-01 8.88240993e-01
4.62857485e-01 -8.02546680e-01 -4.67152774e-01 4.98533435e-02] | [14.63752269744873, -2.19032883644104] |
56d98fbd-73e9-4097-a6b1-92986c1ea873 | continuous-decomposition-of-granularity-for | 2209.01765 | null | https://arxiv.org/abs/2209.01765v2 | https://arxiv.org/pdf/2209.01765v2.pdf | Continuous Decomposition of Granularity for Neural Paraphrase Generation | While Transformers have had significant success in paragraph generation, they treat sentences as linear sequences of tokens and often neglect their hierarchical information. Prior work has shown that decomposing the levels of granularity~(e.g., word, phrase, or sentence) for input tokens has produced substantial improvements, suggesting the possibility of enhancing Transformers via more fine-grained modeling of granularity. In this work, we propose a continuous decomposition of granularity for neural paraphrase generation (C-DNPG). In order to efficiently incorporate granularity into sentence encoding, C-DNPG introduces a granularity-aware attention (GA-Attention) mechanism which extends the multi-head self-attention with: 1) a granularity head that automatically infers the hierarchical structure of a sentence by neurally estimating the granularity level of each input token; and 2) two novel attention masks, namely, granularity resonance and granularity scope, to efficiently encode granularity into attention. Experiments on two benchmarks, including Quora question pairs and Twitter URLs have shown that C-DNPG outperforms baseline models by a remarkable margin and achieves state-of-the-art results in terms of many metrics. Qualitative analysis reveals that C-DNPG indeed captures fine-grained levels of granularity with effectiveness. | ['Jung-Woo Ha', 'Kang Min Yoo', 'Sang-Woo Lee', 'Zhaowei Zhang', 'Xiaodong Gu'] | 2022-09-05 | null | https://aclanthology.org/2022.coling-1.554 | https://aclanthology.org/2022.coling-1.554.pdf | coling-2022-10 | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 1.12491407e-01 4.23545420e-01 -2.60780126e-01 -1.56809092e-01
-1.21459079e+00 -6.51208758e-01 8.88414383e-01 3.22566181e-01
-7.61220083e-02 7.64330983e-01 1.06894565e+00 -4.77569818e-01
4.83556002e-01 -1.22057307e+00 -1.03580403e+00 -2.94532239e-01
3.27752411e-01 4.21463370e-01 1.36706069e-01 -2.65622348e-01
4.77719903e-01 3.60372290e-02 -1.50379908e+00 1.02015758e+00
1.09022498e+00 9.38265920e-01 1.13629401e-01 7.50879765e-01
-7.52426147e-01 9.62518096e-01 -9.60352838e-01 -4.93651956e-01
6.97014183e-02 -5.16240239e-01 -1.17496741e+00 1.10794358e-01
6.82699025e-01 -4.49043542e-01 -1.82567850e-01 7.60868490e-01
2.67547905e-01 -1.36227250e-01 6.29043698e-01 -8.24399173e-01
-1.03108490e+00 1.23693669e+00 -6.57540858e-01 4.13162112e-01
3.01956981e-01 3.27460647e-01 1.65229380e+00 -6.38778031e-01
3.53026152e-01 1.42884755e+00 5.94696105e-01 3.44380230e-01
-1.44942403e+00 -3.21419269e-01 1.87524870e-01 -2.81557953e-03
-1.04931068e+00 -2.49355972e-01 5.99140644e-01 -4.24458802e-01
1.34600151e+00 1.90773651e-01 6.23916924e-01 1.02135146e+00
4.21424299e-01 8.19178581e-01 1.00459909e+00 -5.38821995e-01
2.62518346e-01 -1.85637176e-01 2.81392932e-01 5.86680770e-01
3.21246952e-01 -3.89351994e-01 -5.86455822e-01 -2.48282149e-01
7.75824189e-01 -2.78474361e-01 -6.22544251e-02 4.22499388e-01
-1.10833693e+00 9.74473059e-01 6.02572322e-01 2.57025242e-01
-5.40378690e-01 3.76046360e-01 5.98125041e-01 1.07255422e-01
6.15291357e-01 8.64618301e-01 -1.90642059e-01 -2.64162302e-01
-1.05242682e+00 5.46576619e-01 8.20079327e-01 1.19099939e+00
6.38061404e-01 -5.68940379e-02 -1.10161686e+00 6.43887758e-01
-1.07337102e-01 9.27051082e-02 6.98276699e-01 -8.77153158e-01
8.57468426e-01 8.28383505e-01 2.40386218e-01 -5.41360080e-01
-1.47509873e-01 -4.11159873e-01 -6.25470042e-01 -4.36436445e-01
2.04167619e-01 -1.36171594e-01 -9.58744764e-01 1.78742433e+00
4.43175882e-02 7.21809687e-03 -2.94748425e-01 6.17189944e-01
8.04805934e-01 8.66029620e-01 2.13969246e-01 1.84070572e-01
1.84661496e+00 -1.18004227e+00 -4.72555727e-01 -2.27930382e-01
5.20042598e-01 -3.75335336e-01 1.58720648e+00 3.75975221e-02
-1.57657242e+00 -6.34946167e-01 -1.00101590e+00 -6.60733163e-01
-2.63003707e-01 -5.56211174e-02 5.23426175e-01 5.91004252e-01
-1.17309475e+00 5.22475421e-01 -5.70502162e-01 -1.13124199e-01
4.01230127e-01 -4.05963190e-04 2.22778961e-01 1.73837602e-01
-1.41606081e+00 6.53948426e-01 4.20212954e-01 -2.47350201e-01
-6.43502295e-01 -9.45379257e-01 -7.79922724e-01 5.59661627e-01
5.42739443e-02 -1.21088696e+00 1.46116376e+00 -5.43851554e-01
-1.47812009e+00 6.13056898e-01 -4.27173823e-01 -7.45337009e-01
1.09485365e-01 -1.14050530e-01 2.19209209e-01 2.23288402e-01
3.99914503e-01 9.78846371e-01 7.67075002e-01 -9.30868566e-01
-7.33407497e-01 -1.11082934e-01 6.01386726e-01 3.42381567e-01
-5.91448426e-01 -1.68759525e-01 -2.55533546e-01 -7.62954533e-01
-2.68698245e-01 -5.31169713e-01 6.67663962e-02 -5.51230431e-01
-5.01411796e-01 -6.50569737e-01 3.43253255e-01 -8.68704438e-01
1.74581409e+00 -1.65846467e+00 2.61639237e-01 -4.88521725e-01
2.85733968e-01 -7.87071288e-02 -2.51709729e-01 7.54869938e-01
2.71527231e-01 5.72974324e-01 -1.23797923e-01 -6.92632318e-01
3.20006818e-01 1.07958643e-02 -6.95855856e-01 -3.02316368e-01
3.96600246e-01 1.30749929e+00 -9.68364060e-01 -4.52946961e-01
-1.99176684e-01 2.42311358e-01 -8.76507163e-01 2.17434615e-01
-7.76773036e-01 -4.84727286e-02 -2.97110498e-01 5.01190782e-01
2.52772242e-01 -6.70681059e-01 -8.29378739e-02 -1.80743158e-01
-1.02127105e-01 1.07115138e+00 -4.38553035e-01 1.60387206e+00
-8.27791691e-01 4.66072053e-01 -1.79096445e-01 -3.81859928e-01
4.71151263e-01 4.13300842e-01 -7.49197900e-02 -6.85847700e-01
-6.18791804e-02 -8.25573131e-02 -1.03226915e-01 -2.35964745e-01
1.29259324e+00 -2.05889374e-01 -5.47313094e-01 6.96374357e-01
-1.41304225e-01 -2.08968684e-01 4.32584345e-01 5.17704189e-01
1.26288009e+00 -1.79026574e-01 3.36083435e-02 -2.71379143e-01
2.34014362e-01 -1.07143521e-01 2.13932455e-01 9.49804783e-01
1.46550938e-01 7.09458053e-01 8.84087324e-01 -6.09526858e-02
-1.37816477e+00 -9.23101068e-01 1.45364657e-01 1.43307137e+00
-2.39147842e-01 -6.21129394e-01 -1.21995580e+00 -6.42885029e-01
1.80284724e-01 1.02610266e+00 -9.55144286e-01 4.92633805e-02
-6.75432146e-01 -6.29884243e-01 7.55829573e-01 8.99460256e-01
5.33542693e-01 -1.29750931e+00 -5.27224541e-01 4.67762828e-01
-6.66275561e-01 -1.15623248e+00 -8.70838344e-01 -8.11980888e-02
-7.06712782e-01 -4.00295734e-01 -4.40206766e-01 -4.72846299e-01
4.39190686e-01 2.06326827e-01 1.56165159e+00 9.18377712e-02
4.58746627e-02 -1.22718744e-01 -5.31205118e-01 -2.49378294e-01
-2.54002303e-01 6.43201590e-01 -5.94392598e-01 -2.14924917e-01
2.28303790e-01 -5.31437218e-01 -7.90356696e-01 -1.88010857e-01
-1.06136346e+00 2.40601793e-01 7.19786286e-01 9.04445529e-01
2.37357572e-01 -2.36839905e-01 8.17591786e-01 -1.07572627e+00
1.12904847e+00 -6.12756670e-01 -1.72496811e-01 6.88496307e-02
-3.14343005e-01 2.72401661e-01 9.16619420e-01 -1.33095011e-01
-9.52883303e-01 -7.80128241e-01 -3.05895060e-01 -7.23425597e-02
-9.34488233e-03 5.49811840e-01 -2.07052380e-01 5.09941638e-01
5.51685095e-01 4.05827135e-01 -3.72070074e-01 -3.53067338e-01
5.85609913e-01 7.98126936e-01 3.30613345e-01 -1.03136206e+00
3.62744153e-01 2.78227717e-01 -3.90008807e-01 -6.44840777e-01
-1.24650669e+00 -2.98191279e-01 -2.58871377e-01 3.06075126e-01
9.11695361e-01 -9.52542245e-01 -3.69845808e-01 3.52499187e-01
-1.37750292e+00 -5.66104949e-01 -5.16611636e-01 -3.10405523e-01
-4.57742035e-01 3.19404840e-01 -1.22443271e+00 -3.37690711e-01
-7.40485668e-01 -1.01499307e+00 1.54682982e+00 7.94994980e-02
-5.06924033e-01 -1.06030738e+00 -1.69286534e-01 6.89940214e-01
6.24918759e-01 9.41459090e-02 1.31145465e+00 -3.81909281e-01
-8.42572033e-01 1.20581180e-01 -5.75321138e-01 1.98530883e-01
5.91308549e-02 -2.52516359e-01 -8.09658170e-01 -9.43964049e-02
-1.51880875e-01 -2.45784774e-01 1.08259153e+00 3.12669963e-01
1.29380023e+00 -8.35434139e-01 -2.82149520e-02 3.44220698e-01
1.30840182e+00 -2.07702100e-01 8.00915837e-01 3.53761703e-01
7.19162405e-01 3.74319941e-01 8.82295221e-02 4.92989391e-01
8.63848209e-01 5.15469372e-01 3.71968567e-01 1.89918667e-01
-3.79094630e-01 -5.07984638e-01 2.46794969e-01 7.70681798e-01
2.38461658e-01 -4.30322051e-01 -7.18798578e-01 9.60821986e-01
-1.65160847e+00 -1.09749973e+00 6.29789978e-02 1.94553447e+00
1.18766475e+00 4.67351794e-01 3.19622397e-01 -3.59930806e-02
5.49907386e-01 4.28123921e-01 -3.21306348e-01 -1.02572393e+00
1.44090727e-01 2.60978520e-01 2.75165439e-01 5.43322563e-01
-6.63297594e-01 1.05022562e+00 6.14462090e+00 9.35297012e-01
-9.30785894e-01 1.07986771e-01 7.66389370e-01 -1.13014840e-01
-8.30259264e-01 -1.15242891e-01 -1.18916857e+00 9.62854505e-01
1.03945744e+00 -2.61084020e-01 4.01253372e-01 4.50275630e-01
2.42348477e-01 1.37342617e-01 -1.14993930e+00 1.77697971e-01
9.98445135e-03 -1.70468104e+00 6.41732991e-01 1.96770817e-01
9.63691533e-01 -1.48517534e-01 -1.88327543e-02 5.70695102e-01
4.36575085e-01 -9.65037167e-01 9.92486775e-01 3.35282624e-01
7.52932847e-01 -6.88853323e-01 5.72592974e-01 4.53663141e-01
-1.25665963e+00 -1.98553205e-01 -4.68400270e-01 -3.99592847e-01
8.74683335e-02 6.12661839e-01 -8.62394094e-01 4.78573114e-01
3.67948800e-01 2.94955015e-01 -6.10728681e-01 7.16904879e-01
-3.81118447e-01 9.08764601e-01 -1.13141509e-02 -2.38459915e-01
6.19777739e-01 1.85642511e-01 2.10426137e-01 1.53938448e+00
1.67718008e-01 7.51211122e-02 -9.43915844e-02 1.12820208e+00
-5.46166241e-01 -1.43315047e-01 -7.11611435e-02 -2.26823941e-01
8.59877288e-01 1.10462832e+00 -5.66198707e-01 -7.22460330e-01
-3.62180680e-01 7.96414137e-01 6.71328485e-01 2.07009047e-01
-8.66032064e-01 -5.87845862e-01 5.89641809e-01 3.86937737e-01
6.19360328e-01 -6.41679391e-02 -9.01006877e-01 -1.09819829e+00
1.39596984e-02 -8.63189816e-01 2.19865680e-01 -7.96223700e-01
-1.27115798e+00 7.00603485e-01 -1.12914920e-01 -7.72754848e-01
-2.88235605e-01 -1.45099625e-01 -8.13809633e-01 1.18384075e+00
-1.33959234e+00 -1.37823093e+00 -2.24783123e-01 9.47391689e-02
9.87795174e-01 4.00396883e-01 5.13402998e-01 -1.02925077e-01
-4.46819514e-01 7.79775441e-01 -2.03168899e-01 1.94164783e-01
3.87546122e-01 -1.61752748e+00 1.04956400e+00 7.04032421e-01
-8.69884938e-02 9.00408030e-01 6.63702011e-01 -5.70481360e-01
-1.23460352e+00 -1.24507928e+00 1.36846924e+00 -6.95092380e-01
9.15534914e-01 -5.65518022e-01 -9.05563772e-01 6.22029006e-01
6.44306660e-01 -6.08777881e-01 5.26437223e-01 2.19723195e-01
-6.11086667e-01 6.71787411e-02 -9.91918325e-01 5.92236400e-01
8.57236624e-01 -7.46504962e-01 -9.42423522e-01 3.24115723e-01
1.38850093e+00 -4.11619961e-01 -1.01721740e+00 -4.23744470e-02
4.69509423e-01 -9.80100095e-01 7.21194565e-01 -4.10280377e-01
1.14530790e+00 -3.63735259e-02 1.86121408e-02 -1.27337396e+00
-4.98779953e-01 -4.14757580e-01 -4.42707211e-01 1.38436162e+00
4.21913207e-01 -4.72623855e-01 8.40436101e-01 5.00990272e-01
-4.87197340e-01 -1.17897868e+00 -6.13712192e-01 -5.32880008e-01
5.20414650e-01 -8.38459507e-02 1.16317248e+00 5.90435326e-01
3.17704529e-01 7.13327289e-01 9.54575185e-03 1.48821637e-04
4.60621923e-01 3.73576701e-01 4.18668091e-01 -7.15759873e-01
-5.09762645e-01 -8.76786053e-01 9.92182568e-02 -1.48281765e+00
1.54277638e-01 -7.09536791e-01 5.33840731e-02 -2.17571330e+00
3.73420298e-01 -7.73781687e-02 -3.50871030e-03 3.92836779e-01
-5.57047665e-01 2.59048969e-01 1.16010614e-01 1.33210063e-01
-5.37726879e-01 4.41851437e-01 1.33343065e+00 -1.46271974e-01
1.11390697e-02 -1.50932193e-01 -1.15296078e+00 3.17643851e-01
7.12686956e-01 -9.76803452e-02 -2.84157783e-01 -7.73282290e-01
3.47824961e-01 2.92987049e-01 1.94516793e-01 -7.19796360e-01
1.42688826e-01 3.98535021e-02 1.05257981e-01 -7.23290861e-01
1.50350630e-01 -1.08066678e-01 -4.37533885e-01 2.40545914e-01
-8.00122499e-01 1.80431157e-01 2.49592841e-01 4.81795102e-01
-1.99368626e-01 -2.20164627e-01 4.23828840e-01 -4.49783176e-01
-2.07090139e-01 1.64486587e-01 -3.84117544e-01 3.95716339e-01
2.58481264e-01 -2.15372637e-01 -7.75619984e-01 -2.53593385e-01
-3.06993961e-01 3.19855601e-01 6.20940089e-01 3.65419298e-01
2.46111467e-01 -1.18287718e+00 -7.72780716e-01 6.85417578e-02
-2.11037211e-02 1.13732405e-01 1.45690948e-01 4.31835413e-01
-3.64029050e-01 8.13540757e-01 2.68868282e-02 -3.45195949e-01
-7.70786941e-01 3.14734846e-01 6.68251701e-03 -9.43174183e-01
-4.72140431e-01 1.03524256e+00 4.10814852e-01 -7.94185847e-02
-8.28458741e-03 -9.79997098e-01 6.92880107e-03 2.33558536e-01
5.57933271e-01 2.94268578e-01 1.72412977e-01 -2.75094390e-01
1.44513249e-01 2.61687607e-01 -4.16971147e-01 -1.44961968e-01
9.35521364e-01 -1.90613762e-01 -3.29250067e-01 3.15284163e-01
1.12363160e+00 -5.07989489e-02 -1.35207486e+00 -1.67049572e-01
7.10094394e-03 -2.32064843e-01 -1.17425315e-01 -9.22071457e-01
-4.57038343e-01 1.05025506e+00 -4.29672062e-01 7.33569980e-01
9.96952116e-01 -1.51094003e-02 1.25060797e+00 -1.86051782e-02
3.93368065e-01 -7.32325256e-01 3.92003894e-01 8.25366199e-01
9.81708288e-01 -6.97443187e-01 -1.26437023e-01 -1.93042830e-01
-5.12680948e-01 7.75884748e-01 7.06719995e-01 -1.77359074e-01
5.40138595e-02 2.36088336e-01 -4.65632468e-01 6.39298558e-02
-1.24270380e+00 -6.63422719e-02 3.10033202e-01 4.04015034e-02
6.85559750e-01 1.14702582e-01 -3.49264145e-01 6.87798679e-01
-6.19674802e-01 -1.25125840e-01 7.54216790e-01 7.26500988e-01
-8.23194385e-01 -9.50461388e-01 -2.35396773e-01 8.17963958e-01
-6.12484813e-01 -6.51390433e-01 -2.67470747e-01 4.97772157e-01
4.25932556e-02 8.41324270e-01 3.74259651e-01 -2.41018504e-01
2.14743778e-01 1.40822783e-01 5.66607654e-01 -1.00617731e+00
-1.22505808e+00 -2.13975787e-01 2.26005137e-01 -2.97950476e-01
2.52679646e-01 -5.69558620e-01 -1.08857477e+00 -5.47004819e-01
-1.85742229e-01 4.28932220e-01 4.18936729e-01 8.73980999e-01
6.80301368e-01 8.93363297e-01 5.23202121e-01 -8.22935760e-01
-7.48249650e-01 -1.25390005e+00 -3.72607738e-01 2.48424187e-01
3.33624482e-01 -1.10241145e-01 -4.55148876e-01 -3.39892097e-02] | [11.73184585571289, 8.961880683898926] |
b3b7d160-1d1a-4e6c-b223-c95c55ec6c47 | learning-customized-visual-models-with | 2301.07094 | null | https://arxiv.org/abs/2301.07094v1 | https://arxiv.org/pdf/2301.07094v1.pdf | Learning Customized Visual Models with Retrieval-Augmented Knowledge | Image-text contrastive learning models such as CLIP have demonstrated strong task transfer ability. The high generality and usability of these visual models is achieved via a web-scale data collection process to ensure broad concept coverage, followed by expensive pre-training to feed all the knowledge into model weights. Alternatively, we propose REACT, REtrieval-Augmented CusTomization, a framework to acquire the relevant web knowledge to build customized visual models for target domains. We retrieve the most relevant image-text pairs (~3% of CLIP pre-training data) from the web-scale database as external knowledge, and propose to customize the model by only training new modualized blocks while freezing all the original weights. The effectiveness of REACT is demonstrated via extensive experiments on classification, retrieval, detection and segmentation tasks, including zero, few, and full-shot settings. Particularly, on the zero-shot classification task, compared with CLIP, it achieves up to 5.4% improvement on ImageNet and 3.7% on the ELEVATER benchmark (20 datasets). | ['Chunyuan Li', 'Yong Jae Lee', 'Jianfeng Gao', 'Ce Liu', 'Jianwei Yang', 'Kilho Son', 'Haotian Liu'] | 2023-01-17 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Liu_Learning_Customized_Visual_Models_With_Retrieval-Augmented_Knowledge_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Liu_Learning_Customized_Visual_Models_With_Retrieval-Augmented_Knowledge_CVPR_2023_paper.pdf | cvpr-2023-1 | ['zero-shot-transfer-image-classification', 'semi-supervised-image-classification'] | ['computer-vision', 'computer-vision'] | [ 3.23948205e-01 -2.26963863e-01 -3.62857342e-01 -3.67185265e-01
-1.08539510e+00 -6.08655751e-01 4.09611821e-01 4.51387800e-02
-6.68609321e-01 2.92153388e-01 -7.33224601e-02 3.25584337e-02
8.21591392e-02 -4.46435422e-01 -8.94790411e-01 -4.38197523e-01
1.25590414e-01 4.82776523e-01 7.19957471e-01 -2.02619866e-01
2.25285143e-01 2.25507114e-02 -1.91412628e+00 5.84007084e-01
7.71043003e-01 1.29399204e+00 5.73349953e-01 6.11809850e-01
-1.75255343e-01 5.98642707e-01 -5.39807796e-01 -4.79338229e-01
2.70871639e-01 -1.43784642e-01 -6.33935452e-01 3.99869502e-01
9.33958650e-01 -4.78219748e-01 -1.79878846e-01 8.31427276e-01
6.35445595e-01 2.93825090e-01 5.45899332e-01 -1.28169823e+00
-8.55454385e-01 5.58552027e-01 -7.73276508e-01 2.53976494e-01
1.97530925e-01 2.97480375e-01 1.12765372e+00 -1.36199284e+00
9.63989973e-01 1.01668143e+00 4.80843633e-01 5.93914211e-01
-1.19763052e+00 -8.84469271e-01 2.17310101e-01 5.49035251e-01
-1.53568614e+00 -5.58302701e-01 5.23191929e-01 -4.21130329e-01
1.01493442e+00 1.51384950e-01 8.32377195e-01 1.20309973e+00
-1.47010177e-01 1.13867402e+00 6.93088114e-01 -4.22301203e-01
2.11437210e-01 2.95216084e-01 1.98289946e-01 6.84696436e-01
9.75929126e-02 -3.37725610e-01 -9.38294470e-01 1.50180116e-01
4.40296441e-01 1.00740999e-01 -2.67598242e-01 -6.79918289e-01
-8.81730378e-01 6.23242795e-01 5.71542919e-01 -4.61720116e-02
-1.43376082e-01 1.86666191e-01 5.13685882e-01 1.99821830e-01
3.39345455e-01 4.34224248e-01 -4.35554653e-01 4.46602516e-03
-1.03221881e+00 2.00132057e-01 4.38432157e-01 1.47420084e+00
9.86109436e-01 -9.22326669e-02 -5.49893141e-01 1.15073192e+00
-1.60406549e-02 7.01519430e-01 5.96874416e-01 -9.45244730e-01
6.39278114e-01 4.13443267e-01 -1.12108856e-01 -5.66925347e-01
-1.02965767e-02 -2.95218259e-01 -4.36840475e-01 1.21292481e-02
-1.01094015e-01 5.60266636e-02 -1.55165815e+00 1.52852392e+00
1.57249346e-01 2.54498333e-01 -5.10873161e-02 8.87999356e-01
1.00115430e+00 7.27767169e-01 4.10224855e-01 9.07574221e-02
1.38312435e+00 -1.20371044e+00 -4.13085610e-01 -5.57939768e-01
3.68288279e-01 -6.56673610e-01 1.48365998e+00 2.42759719e-01
-1.20488513e+00 -6.29955292e-01 -1.20539665e+00 -1.94533050e-01
-5.46865523e-01 1.26721077e-02 2.78749496e-01 2.70702899e-01
-9.41976845e-01 3.78353089e-01 -5.38809836e-01 -4.51420039e-01
6.93390608e-01 -1.51965925e-02 -2.83965558e-01 -6.13226950e-01
-1.05664253e+00 5.91497421e-01 5.94488442e-01 -3.99427801e-01
-1.31607234e+00 -1.19676471e+00 -8.29455018e-01 2.33650729e-01
6.14692450e-01 -6.77635968e-01 1.28850842e+00 -9.96817827e-01
-1.06159055e+00 9.14632380e-01 1.01441078e-01 -2.80607671e-01
5.49228013e-01 -2.99826235e-01 -2.11617783e-01 5.17353773e-01
3.04843336e-01 1.16401076e+00 1.12011671e+00 -1.43838441e+00
-8.94569933e-01 -1.67248145e-01 7.36299157e-02 3.68472278e-01
-6.92053914e-01 -1.04666546e-01 -1.42495477e+00 -4.87112731e-01
-1.74636424e-01 -8.63744497e-01 -3.74486204e-03 2.39149913e-01
-1.52894169e-01 -9.37048197e-02 8.36886942e-01 -6.52557731e-01
1.13370609e+00 -2.19609737e+00 -1.31707424e-02 1.32966250e-01
2.60119259e-01 3.62960130e-01 -4.84386295e-01 3.33321482e-01
1.17933370e-01 7.87201077e-02 -9.23585221e-02 -4.35212731e-01
-4.89134416e-02 1.50930300e-01 -4.16980349e-02 -5.60290441e-02
2.27951914e-01 1.15611219e+00 -7.94462204e-01 -8.74954581e-01
3.11459512e-01 2.25774333e-01 -5.89751482e-01 1.58340931e-01
-3.48015368e-01 -3.13503027e-01 -3.30079585e-01 7.54052341e-01
5.07210910e-01 -5.90809166e-01 6.76331967e-02 -3.02492470e-01
2.80053169e-01 -3.15523922e-01 -9.73747849e-01 1.95154035e+00
-3.76857013e-01 8.13298762e-01 -1.36075988e-01 -8.25135648e-01
6.08778834e-01 1.25387594e-01 4.02341366e-01 -1.12438893e+00
-4.24993895e-02 -3.25320661e-01 -3.98411602e-01 -8.13726366e-01
5.85372448e-01 3.10505629e-01 -1.24722898e-01 1.66817293e-01
3.71786624e-01 -8.41011927e-02 5.29454648e-01 6.21520400e-01
1.02589297e+00 3.66800018e-02 -4.33489941e-02 2.92734355e-02
6.42659813e-02 3.02783102e-01 4.03680086e-01 9.39233422e-01
-1.52505338e-01 8.78453016e-01 2.05239952e-01 -3.43080685e-02
-1.18276083e+00 -1.24768364e+00 -1.00054979e-01 1.59529579e+00
4.11766231e-01 -6.55754328e-01 -6.14423215e-01 -7.48131752e-01
7.55545795e-02 6.97938025e-01 -6.76841199e-01 -3.45959187e-01
-2.90289998e-01 -5.30543506e-01 3.05495858e-01 8.69931161e-01
6.59675777e-01 -1.00458324e+00 -6.61123931e-01 5.26608936e-02
-3.96782041e-01 -1.12108767e+00 -7.68491268e-01 1.29976511e-01
-5.23228407e-01 -1.11398935e+00 -8.24068606e-01 -9.49932218e-01
5.80464125e-01 6.91795886e-01 1.19444013e+00 2.94961017e-02
-7.74364769e-01 6.41512930e-01 -5.85197508e-01 -2.92082995e-01
4.07719910e-02 3.27838701e-03 -4.21932936e-01 -2.07915500e-01
3.06113482e-01 -2.21525833e-01 -8.02269220e-01 3.53499204e-01
-9.92192566e-01 1.80248424e-01 6.98815823e-01 9.04689968e-01
6.85033262e-01 -2.91986555e-01 4.65465665e-01 -8.11809659e-01
4.72360343e-01 -3.75044435e-01 -4.02476460e-01 6.70640469e-01
-7.72579551e-01 -2.72913784e-01 2.40125641e-01 -7.16733694e-01
-1.22283447e+00 2.13317759e-02 2.29661405e-01 -9.29985583e-01
1.14125580e-01 6.03321314e-01 3.74758691e-02 1.84817091e-02
9.53905642e-01 2.97740370e-01 -3.23311061e-01 -3.80041331e-01
6.33243561e-01 6.63281381e-01 5.27002990e-01 -6.32844329e-01
7.33749092e-01 3.45194042e-01 -6.88835382e-01 -8.87523174e-01
-9.60167468e-01 -8.43449950e-01 -5.48269510e-01 -4.08418655e-01
9.79025245e-01 -1.29942226e+00 -1.96046129e-01 3.46572936e-01
-7.62521029e-01 -4.96856123e-01 -2.36934066e-01 1.40230104e-01
-4.54998404e-01 2.79454768e-01 -5.54668188e-01 -4.16630089e-01
-6.14736199e-01 -1.02347755e+00 1.14579427e+00 2.83060014e-01
1.77609876e-01 -5.16598940e-01 -2.69548476e-01 5.22451997e-01
3.45351368e-01 -5.94128445e-02 9.07884419e-01 -4.73060995e-01
-6.82551980e-01 -2.56341755e-01 -6.01618528e-01 4.00458485e-01
-3.63104194e-01 -1.48896854e-02 -1.00219274e+00 -4.94734406e-01
-6.56364322e-01 -8.39208305e-01 1.10730672e+00 2.18449265e-01
1.34111166e+00 1.33824959e-01 -4.97438997e-01 6.22123837e-01
1.64083469e+00 1.84483781e-01 5.03377914e-01 2.65921056e-01
6.90475821e-01 3.68752480e-01 8.52885604e-01 5.95924079e-01
3.92063171e-01 6.63152635e-01 1.96258709e-01 -8.92796442e-02
-4.28476244e-01 -4.13653195e-01 1.31256267e-01 6.75730348e-01
7.71510750e-02 -1.32774830e-01 -9.43637133e-01 7.61939287e-01
-1.96947575e+00 -8.93919170e-01 2.80510336e-01 2.13420177e+00
9.20209527e-01 3.02034408e-01 -3.52695920e-02 -4.03893322e-01
6.59146130e-01 9.94929671e-02 -9.04473603e-01 2.16628551e-01
4.08335589e-02 2.54223764e-01 6.20899379e-01 -1.20271053e-02
-1.05487072e+00 1.28220463e+00 6.60894203e+00 1.13325596e+00
-9.82221901e-01 3.21566433e-01 4.28273290e-01 -4.27068859e-01
-7.14945644e-02 -9.71514061e-02 -9.01263475e-01 3.24079514e-01
6.26245499e-01 -4.90763962e-01 3.05934817e-01 1.17601883e+00
-9.22454447e-02 -8.36521983e-02 -9.79307353e-01 1.12955892e+00
5.67482531e-01 -1.40333188e+00 2.82809764e-01 -3.08478922e-01
1.01730931e+00 2.54874289e-01 1.58294559e-01 7.39963710e-01
3.19737732e-01 -5.49294353e-01 8.52925003e-01 4.02199537e-01
1.07597375e+00 -5.69706142e-01 3.46937448e-01 2.96737254e-02
-1.24160540e+00 -1.51064038e-01 -5.37136376e-01 5.72504401e-01
-3.33676790e-03 1.33150280e-01 -8.03899229e-01 3.32351118e-01
1.25295794e+00 8.32257628e-01 -1.03321350e+00 1.18789577e+00
-5.85786887e-02 5.57836533e-01 -1.72712967e-01 5.63155413e-02
2.12707341e-01 2.90414959e-01 1.57145575e-01 1.26825738e+00
6.42977804e-02 -3.65928151e-02 4.05956328e-01 6.75566733e-01
-3.60863924e-01 1.98501930e-01 -2.85563022e-01 7.43965432e-02
5.16662896e-01 1.34122694e+00 -5.59522092e-01 -8.03386807e-01
-5.33942759e-01 1.05952024e+00 4.75869060e-01 7.12034523e-01
-1.02739108e+00 -7.32529759e-01 4.57128733e-01 6.91726105e-03
7.66648471e-01 2.83972323e-02 4.78902608e-02 -1.19412768e+00
7.61156604e-02 -7.28012025e-01 5.82081139e-01 -1.19752240e+00
-1.28548241e+00 3.00199389e-01 3.68909210e-01 -1.21104538e+00
-4.65442613e-02 -5.57613552e-01 -4.82446730e-01 5.15794575e-01
-1.44073808e+00 -1.32094276e+00 -7.38780797e-01 8.41648042e-01
1.02822661e+00 -9.79921594e-02 4.12517428e-01 4.50763613e-01
-5.99724710e-01 8.22609425e-01 1.40737712e-01 1.25490800e-01
1.04443300e+00 -1.23776543e+00 3.37693721e-01 6.35610938e-01
8.15981105e-02 3.50492716e-01 5.02478778e-01 -5.70562541e-01
-1.31945252e+00 -1.24606049e+00 3.57667029e-01 -4.43627149e-01
6.75338268e-01 -4.97858524e-01 -9.10395622e-01 7.34682620e-01
3.57095689e-01 1.69383865e-02 5.82841933e-01 1.01838568e-02
-6.57012045e-01 -2.77905554e-01 -7.95785487e-01 6.61999881e-01
1.23119080e+00 -3.96623522e-01 -7.17988789e-01 4.21369493e-01
9.89844024e-01 -2.97022432e-01 -8.62557650e-01 1.85978770e-01
7.16532052e-01 -6.16534770e-01 1.07753348e+00 -6.40252888e-01
6.54730201e-01 -1.36565134e-01 -1.44260079e-01 -1.06727970e+00
-3.77369821e-01 -7.21328482e-02 -6.54478297e-02 1.20185030e+00
5.97679377e-01 -6.32431880e-02 8.05018425e-01 6.09525204e-01
-1.62965223e-01 -3.94677967e-01 -5.84207296e-01 -7.51118660e-01
-2.37140507e-01 -5.14267683e-01 1.90366209e-01 9.02561903e-01
3.82976085e-02 4.74319309e-01 -4.76320684e-01 -1.55111104e-01
7.45153606e-01 1.25907153e-01 9.38304543e-01 -9.33637440e-01
-3.07271481e-01 -2.32340634e-01 -2.56310374e-01 -1.08615732e+00
1.18983565e-02 -9.57617998e-01 2.82597244e-01 -1.63693571e+00
5.46792865e-01 -2.32875898e-01 -6.48423910e-01 6.88383400e-01
-3.43091667e-01 4.81205553e-01 5.88553429e-01 2.87307739e-01
-1.30244136e+00 4.65428144e-01 1.30186021e+00 -5.72585344e-01
-2.14800403e-01 -2.79558748e-01 -5.64444602e-01 3.44549447e-01
5.30298769e-01 -2.94626057e-01 -6.88579321e-01 -5.60162306e-01
-2.21911613e-02 -1.98200032e-01 3.10633391e-01 -1.21111691e+00
3.23087066e-01 -2.22505257e-01 6.98101521e-01 -7.43468344e-01
5.33308089e-01 -7.90524185e-01 -1.04940914e-01 1.26129717e-01
-5.38417101e-01 -1.67222157e-01 2.89217561e-01 8.06848109e-01
-1.35574251e-01 -3.10584545e-01 8.02970350e-01 -1.81281388e-01
-1.48701048e+00 4.69422907e-01 -1.32077605e-01 3.29030693e-01
1.24076617e+00 -1.60431638e-01 -5.49820542e-01 -2.93028712e-01
-8.49819839e-01 6.87453270e-01 3.64363045e-01 8.00950766e-01
8.42308760e-01 -1.20404196e+00 -4.47703421e-01 1.63457438e-01
9.51881826e-01 -1.35143682e-01 6.02689564e-01 5.71786761e-01
-3.54847401e-01 1.98075756e-01 -3.27976763e-01 -8.16519201e-01
-1.33956289e+00 8.33355486e-01 5.70341907e-02 7.95037672e-02
-7.76549160e-01 7.83293545e-01 3.09345543e-01 7.39480779e-02
4.70187277e-01 -7.07315058e-02 -1.72510874e-02 1.91155270e-01
6.20124996e-01 1.80789217e-01 4.39548753e-02 -1.90478653e-01
-1.72565654e-01 6.43098652e-01 -5.35282135e-01 -1.75823778e-01
1.08216834e+00 -1.88070774e-01 4.87915993e-01 3.04175854e-01
1.18375528e+00 -5.58087528e-01 -1.48854077e+00 -4.88909513e-01
-6.90107197e-02 -4.30420607e-01 9.99128297e-02 -9.84928131e-01
-1.08921254e+00 8.41226399e-01 8.04526687e-01 -1.82660416e-01
1.04737723e+00 1.76260635e-01 6.46405697e-01 7.04932451e-01
2.17468157e-01 -1.64890385e+00 4.99787956e-01 3.01317900e-01
7.97326028e-01 -1.43489182e+00 -5.30437417e-02 -1.58449829e-01
-9.75490451e-01 7.25265980e-01 1.08849025e+00 -8.24303627e-02
4.92455661e-01 -1.57515134e-03 2.00449973e-02 -1.15723178e-01
-1.02904153e+00 -6.15665674e-01 3.26931417e-01 6.30662501e-01
-1.85944941e-02 -2.07301855e-01 -2.66972184e-02 4.32965636e-01
2.46348739e-01 1.93095565e-01 1.93003252e-01 1.11526990e+00
-7.88498998e-01 -6.19281769e-01 -4.07956392e-02 6.96568668e-01
-1.05959885e-01 -3.21332246e-01 -1.60637006e-01 9.33719218e-01
-6.96977898e-02 7.95327425e-01 1.52720138e-01 -4.27017868e-01
6.56748414e-01 2.72221506e-01 2.38093957e-01 -6.85864210e-01
-4.14202571e-01 2.06090882e-01 3.21538001e-02 -6.63207114e-01
-2.27403462e-01 -3.91225666e-01 -1.00768232e+00 -3.86231542e-02
-3.15557539e-01 -1.17071435e-01 7.59494722e-01 6.81345880e-01
5.19596100e-01 6.89101934e-01 3.00967544e-01 -9.20740664e-01
-4.69481021e-01 -9.37458336e-01 -4.64287519e-01 8.28773797e-01
-1.44716008e-02 -7.40865171e-01 -2.29971856e-01 4.64108795e-01] | [10.257715225219727, 1.7929413318634033] |
cf3e9b25-1c65-4384-8c80-35a82858438b | on-decoder-only-architecture-for-speech-to | 2307.03917 | null | https://arxiv.org/abs/2307.03917v1 | https://arxiv.org/pdf/2307.03917v1.pdf | On decoder-only architecture for speech-to-text and large language model integration | Large language models (LLMs) have achieved remarkable success in the field of natural language processing, enabling better human-computer interaction using natural language. However, the seamless integration of speech signals into LLMs has not been explored well. The "decoder-only" architecture has also not been well studied for speech processing tasks. In this research, we introduce Speech-LLaMA, a novel approach that effectively incorporates acoustic information into text-based large language models. Our method leverages Connectionist Temporal Classification and a simple audio encoder to map the compressed acoustic features to the continuous semantic space of the LLM. In addition, we further probe the decoder-only architecture for speech-to-text tasks by training a smaller scale randomly initialized speech-LLaMA model from speech-text paired data alone. We conduct experiments on multilingual speech-to-text translation tasks and demonstrate a significant improvement over strong baselines, highlighting the potential advantages of decoder-only models for speech-to-text conversion. | ['Yu Wu', 'Linquan Liu', 'Bo Ren', 'Shujie Liu', 'Jinyu Li', 'Tianrui Wang', 'Yimeng Zhu', 'Long Zhou', 'Zhuo Chen', 'Yashesh Gaur', 'Jian Wu'] | 2023-07-08 | null | null | null | null | ['speech-to-text-translation', 'language-modelling'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.45592916e-01 2.47685134e-01 -5.24793044e-02 -5.99419653e-01
-1.54134154e+00 -4.39054638e-01 8.46591353e-01 -1.40922427e-01
-3.60759586e-01 3.26959610e-01 5.87876618e-01 -6.95681393e-01
6.41976953e-01 -2.22265765e-01 -9.43912864e-01 -3.96983325e-01
2.78276712e-01 6.45706952e-01 -1.38459161e-01 -2.16051146e-01
-2.87981898e-01 -1.61533013e-01 -9.14167464e-01 6.14458263e-01
6.52372003e-01 7.09330142e-01 7.24923790e-01 7.65555561e-01
-2.80858964e-01 6.38314724e-01 -5.74342072e-01 -2.14216694e-01
1.17262572e-01 -6.56203330e-01 -8.22004080e-01 -1.68923721e-01
3.01957667e-01 -1.84345469e-01 -5.03844380e-01 8.13911259e-01
6.04792416e-01 2.03619763e-01 3.54177028e-01 -8.14984262e-01
-7.53958344e-01 8.75998080e-01 -2.41163105e-01 2.17736974e-01
5.46727598e-01 -5.22076972e-02 1.14524639e+00 -1.19610262e+00
5.60636342e-01 1.75979018e+00 4.46680605e-01 4.39456552e-01
-1.46088350e+00 -5.36414564e-01 1.94701076e-01 1.93533078e-01
-1.23461175e+00 -1.19942772e+00 5.68382502e-01 -1.19489118e-01
1.69097662e+00 1.20487407e-01 3.79480362e-01 1.52890837e+00
2.48600960e-01 9.20493305e-01 1.01476634e+00 -7.84233868e-01
7.71937296e-02 2.16780797e-01 -3.58092666e-01 4.65361565e-01
-5.57858646e-01 1.60670236e-01 -9.94497836e-01 2.50927389e-01
3.97131354e-01 -5.01290977e-01 9.82000381e-02 1.22054197e-01
-1.60776258e+00 7.33141184e-01 1.47528678e-01 3.34330916e-01
-1.42807782e-01 2.79297799e-01 6.20512903e-01 7.72289515e-01
6.10496521e-01 2.84893394e-01 -5.80053389e-01 -4.26450878e-01
-1.05745113e+00 -1.79716736e-01 6.91590965e-01 1.03652394e+00
3.41381609e-01 3.18967819e-01 -1.91287294e-01 1.23883295e+00
4.88037944e-01 9.84309375e-01 7.50178933e-01 -6.50006711e-01
9.37419415e-01 -1.27504885e-01 -4.34902698e-01 -4.38303530e-01
-1.76540256e-01 -1.97514921e-01 -6.50495350e-01 -4.25451279e-01
1.03640435e-02 -1.16272427e-01 -7.71902561e-01 1.99182773e+00
-7.00534461e-03 2.10303128e-01 4.54435319e-01 6.32980287e-01
5.15518188e-01 1.27211607e+00 -1.55294351e-02 -2.10800052e-01
1.21901536e+00 -1.33657837e+00 -8.40406477e-01 -5.89678288e-01
6.91932976e-01 -1.00314951e+00 1.35405576e+00 1.38559014e-01
-1.22278225e+00 -5.98405242e-01 -8.78081203e-01 -4.06669080e-01
-3.09573680e-01 3.90091091e-01 2.99602211e-01 3.88695210e-01
-1.29099655e+00 1.68670103e-01 -1.16519690e+00 -5.94460309e-01
1.87797025e-02 2.19823271e-01 -3.34063500e-01 1.40635639e-01
-1.39438868e+00 9.93730724e-01 1.64476737e-01 1.13403328e-01
-9.43506181e-01 -3.72004539e-01 -1.05985487e+00 -8.57327692e-03
1.38974190e-01 -6.99172795e-01 1.79111338e+00 -7.52853334e-01
-2.00958681e+00 6.35886550e-01 -7.30132043e-01 -8.29661131e-01
1.11177512e-01 -1.76798284e-01 -6.25578821e-01 9.29538757e-02
1.61902383e-01 9.09605265e-01 1.05492961e+00 -7.96441138e-01
-5.31423628e-01 -1.70992091e-01 -3.41992408e-01 4.96502161e-01
-3.93964082e-01 3.12423408e-01 -3.39813083e-01 -9.93072689e-01
3.28021869e-02 -1.09596789e+00 -8.90348256e-02 -2.97243536e-01
-3.29746842e-01 -2.00504765e-01 5.64974844e-01 -9.17109907e-01
1.23684430e+00 -2.15347600e+00 2.28548884e-01 -1.28295243e-01
-4.09899622e-01 1.98455110e-01 -6.14268363e-01 6.71034575e-01
8.40764269e-02 -7.03605488e-02 -2.06802189e-01 -9.85525608e-01
1.92268252e-01 1.92717746e-01 -8.32680881e-01 6.40255287e-02
4.40865159e-01 1.22910237e+00 -6.37022257e-01 -2.47408301e-01
3.33816439e-01 5.62422156e-01 -4.93169874e-01 2.32641876e-01
-2.96027064e-01 4.66608912e-01 -1.82028040e-01 3.72179449e-01
2.06967279e-01 9.47793201e-02 1.42895758e-01 8.81043598e-02
-2.41870452e-02 1.22997713e+00 -4.02819455e-01 2.30387950e+00
-9.94355321e-01 9.81459141e-01 2.68089414e-01 -1.01495266e+00
6.65768385e-01 7.92877018e-01 1.38284802e-01 -1.15177536e+00
5.85095994e-02 4.37192261e-01 -5.13565727e-02 -3.11572343e-01
4.33353812e-01 -3.69758844e-01 -2.79257894e-01 5.72830498e-01
3.07769358e-01 -5.12572765e-01 -2.54304707e-01 1.98607519e-01
9.37779248e-01 1.33320987e-01 1.33115426e-01 -1.32529140e-02
3.13134044e-01 -1.13369301e-01 1.89186819e-02 6.05436563e-01
-2.81027611e-02 7.38771081e-01 -8.36157426e-02 2.56068110e-01
-9.92846131e-01 -1.17727995e+00 -5.62779792e-02 1.41375935e+00
-2.45907366e-01 -5.30349553e-01 -9.18535054e-01 -3.39832217e-01
-2.18455404e-01 8.42268348e-01 -1.82863533e-01 -3.02652687e-01
-6.60855114e-01 -5.17468572e-01 1.02723372e+00 3.96148533e-01
7.52365962e-02 -9.73454297e-01 1.10237740e-01 5.00032187e-01
-6.37357891e-01 -1.61008656e+00 -8.97507012e-01 3.00253332e-01
-8.17013979e-01 -1.62123069e-01 -6.58487678e-01 -1.15716350e+00
2.45277211e-01 2.74118364e-01 7.73079395e-01 -6.52743995e-01
1.03547424e-01 1.94315761e-01 -2.62532145e-01 -3.28051299e-01
-1.08901179e+00 3.58580470e-01 3.96495163e-01 1.44117018e-02
3.62363994e-01 -5.62585592e-01 1.65566858e-02 2.27157205e-01
-4.86971676e-01 3.01282197e-01 8.73639345e-01 7.09678471e-01
4.23890501e-01 -5.59154332e-01 9.88301873e-01 -3.49104613e-01
8.17827821e-01 -3.24169576e-01 -2.82248497e-01 2.00046197e-01
-5.21508396e-01 3.91369015e-01 9.18339014e-01 -5.16301513e-01
-1.20793355e+00 -1.45861506e-02 -5.62666714e-01 -3.90045315e-01
-1.71587646e-01 6.48735762e-01 -2.27870166e-01 3.96441966e-01
3.75694782e-01 5.03980100e-01 7.85158053e-02 -6.90988541e-01
6.50942981e-01 1.24380541e+00 8.46527457e-01 -5.81084728e-01
5.98833263e-01 1.74428135e-01 -4.66148645e-01 -9.11506653e-01
-5.72380722e-01 -4.66160744e-01 -5.75305343e-01 2.99941659e-01
8.52478743e-01 -1.29524899e+00 -1.76951945e-01 3.47574323e-01
-1.45342541e+00 -4.56540495e-01 -2.33632386e-01 8.02364171e-01
-7.68568873e-01 3.27132046e-01 -8.73518229e-01 -6.47427499e-01
-3.82891625e-01 -1.23192692e+00 1.47538757e+00 -3.95377964e-01
-3.67885590e-01 -1.07995927e+00 1.63797840e-01 5.53145349e-01
6.38743401e-01 -7.60109663e-01 8.28880310e-01 -7.40934432e-01
-4.95207667e-01 -2.59788372e-02 9.75974202e-02 3.99497598e-01
1.91575997e-02 -4.45472866e-01 -1.34242797e+00 -4.66519386e-01
2.63069868e-02 -4.40626740e-01 8.48621011e-01 3.91126096e-01
5.82252860e-01 -3.49210978e-01 -2.94786364e-01 5.16399682e-01
6.66273594e-01 1.26295954e-01 4.09052879e-01 -1.47578714e-04
5.32687545e-01 3.96939784e-01 3.78232509e-01 7.50450194e-02
7.15175271e-01 8.65676045e-01 -1.75298989e-01 -2.04162791e-01
-6.29720569e-01 -6.37663424e-01 1.16269505e+00 1.95785892e+00
7.04540074e-01 -3.58778149e-01 -7.77717113e-01 6.06071115e-01
-1.52467930e+00 -6.03902757e-01 2.91769087e-01 1.98491442e+00
1.03798962e+00 1.60714611e-01 -2.49330670e-01 -2.00079650e-01
5.24894536e-01 2.50763237e-01 -3.78270924e-01 -6.33385837e-01
-1.94425538e-01 2.48542368e-01 3.39500234e-02 7.90064573e-01
-1.02246082e+00 1.33380830e+00 6.36104250e+00 1.00248194e+00
-1.27520943e+00 4.93147343e-01 3.55908841e-01 -2.46196181e-01
-3.32755119e-01 1.25672638e-01 -7.61757970e-01 2.35686034e-01
1.81673157e+00 -2.42872313e-01 8.43729079e-01 3.56999964e-01
6.33478105e-01 2.27401122e-01 -1.56221163e+00 1.14774859e+00
1.50321335e-01 -1.04767728e+00 2.84641415e-01 -6.42049089e-02
5.80371559e-01 5.74187160e-01 1.30556703e-01 6.17813528e-01
9.93911773e-02 -1.04036748e+00 1.11570406e+00 1.06766433e-01
1.18251455e+00 -3.07915092e-01 2.62686759e-01 4.59374607e-01
-1.40520108e+00 6.51486441e-02 -1.35926664e-01 7.32539669e-02
5.27887821e-01 1.53877735e-01 -1.23301280e+00 4.09425050e-01
4.02290881e-01 8.40605617e-01 -3.41507345e-01 3.12760919e-01
1.86874270e-02 1.08646822e+00 -4.30904001e-01 1.37199312e-01
3.11621100e-01 -9.62360948e-02 7.38501310e-01 1.54732251e+00
4.09146726e-01 -3.25943530e-01 1.91324428e-01 7.79014289e-01
-3.38188738e-01 3.40126336e-01 -6.97413027e-01 -5.50236881e-01
5.19456089e-01 7.89682209e-01 -3.76982450e-01 -2.62522250e-01
-6.10244691e-01 1.28742373e+00 1.77792475e-01 3.87911797e-01
-5.25208354e-01 -3.70525450e-01 6.57228410e-01 -1.20510519e-01
4.01958376e-02 -6.16658568e-01 -1.77847475e-01 -1.33727384e+00
2.96592981e-01 -9.95757520e-01 -7.88268447e-02 -8.31095099e-01
-1.14258349e+00 8.90861690e-01 -2.65451163e-01 -1.12881374e+00
-7.26574421e-01 -4.12192851e-01 -2.72304982e-01 1.01879573e+00
-1.48564136e+00 -1.32322049e+00 4.56535578e-01 4.31247026e-01
1.22240269e+00 -4.77099240e-01 1.06126845e+00 3.44791174e-01
-3.39713812e-01 7.42295384e-01 3.99057984e-01 1.27152503e-01
9.74578083e-01 -8.92252505e-01 1.25712359e+00 9.43158329e-01
6.44631028e-01 6.22181118e-01 4.12017941e-01 -4.84370500e-01
-1.63287175e+00 -1.11367369e+00 1.24773991e+00 -7.23505378e-01
8.44186306e-01 -1.00935590e+00 -8.14758301e-01 8.68314028e-01
6.20867312e-01 -3.49524260e-01 5.00986695e-01 2.57604960e-02
-2.84677207e-01 -1.42449990e-01 -5.02931535e-01 7.24307775e-01
9.11566257e-01 -1.37144160e+00 -7.67869473e-01 2.28054240e-01
1.36846006e+00 -2.85570264e-01 -6.36666059e-01 4.65948582e-02
4.88467485e-01 -2.47150436e-01 9.97407913e-01 -4.99579400e-01
2.20365182e-01 -4.94999141e-02 -5.43803036e-01 -1.73103976e+00
2.53741384e-01 -9.53248501e-01 5.07935472e-02 1.18773699e+00
7.55492270e-01 -6.32514119e-01 9.56675634e-02 5.04027586e-04
-5.35537541e-01 -3.25677186e-01 -1.27379453e+00 -8.64367306e-01
3.42534751e-01 -8.26270044e-01 4.24591750e-01 7.77646422e-01
4.65687096e-01 8.08860183e-01 -3.82336646e-01 9.60201249e-02
1.36061102e-01 -1.17018312e-01 4.78632063e-01 -8.80222023e-01
-6.28204942e-01 -5.00888228e-01 -4.12723428e-04 -1.71403491e+00
5.08650482e-01 -1.51983702e+00 4.66068029e-01 -1.55665433e+00
5.66728860e-02 -2.78185129e-01 -4.52161208e-02 4.09084558e-01
1.64814323e-01 5.85121363e-02 2.68415838e-01 1.85662955e-01
-4.71664786e-01 9.57598448e-01 9.84545588e-01 -2.51046002e-01
-2.14320004e-01 -1.72345772e-01 -4.18199241e-01 2.77262330e-01
6.06444359e-01 -4.56040740e-01 -5.24752378e-01 -9.45637584e-01
2.20238930e-03 5.61510265e-01 -2.53213830e-02 -9.20772254e-01
3.44753653e-01 2.59481817e-02 -1.87502027e-01 -2.96166301e-01
7.50386357e-01 -5.54984152e-01 -2.40273356e-01 -4.55814824e-02
-8.26001108e-01 1.26858562e-01 3.12758505e-01 5.95936239e-01
-5.71784735e-01 2.09670678e-01 5.79591215e-01 7.87343532e-02
-3.59779119e-01 -4.62999521e-03 -8.89874399e-01 9.75611284e-02
4.93507773e-01 1.93046272e-01 -8.49084482e-02 -6.89903021e-01
-9.22245860e-01 5.54051697e-02 1.53101608e-02 1.02433336e+00
5.37451506e-01 -1.27864814e+00 -8.27516019e-01 5.83049059e-01
1.24306582e-01 -2.43295595e-01 -2.43068010e-01 9.85952079e-01
-1.83747839e-02 1.11809397e+00 3.79912019e-01 -6.94258630e-01
-1.02630877e+00 2.19191283e-01 3.38639289e-01 6.72212765e-02
-6.42754674e-01 7.69777119e-01 1.63165629e-01 -8.94080937e-01
3.80256116e-01 -7.20999002e-01 3.44982684e-01 -1.81289792e-01
3.33687395e-01 4.85023893e-02 4.12708402e-01 -9.30030644e-01
-3.16979468e-01 3.03719103e-01 -5.37646823e-02 -8.83793235e-01
1.14662945e+00 -7.48018503e-01 5.26753031e-02 8.66629422e-01
1.38465428e+00 4.79331017e-02 -8.87383699e-01 -6.08979702e-01
1.66158855e-01 1.03924774e-01 9.51294899e-02 -7.84147620e-01
-5.30988514e-01 1.55852830e+00 3.70173037e-01 -1.56442061e-01
9.29698110e-01 1.38100445e-01 1.24662125e+00 8.31153393e-01
3.70124549e-01 -1.41476643e+00 8.89866799e-02 1.01597595e+00
8.92499685e-01 -1.27282071e+00 -8.41355443e-01 -1.30868126e-02
-7.88992941e-01 1.03899884e+00 1.84761196e-01 4.41121072e-01
5.97586572e-01 6.12777352e-01 3.73470843e-01 3.08994561e-01
-1.28797245e+00 -4.70652394e-02 2.19561458e-01 4.10279930e-01
6.18087053e-01 2.54574627e-01 2.45418891e-01 6.77197099e-01
-5.50538480e-01 -1.43227696e-01 2.84935176e-01 6.25735223e-01
-3.95868450e-01 -1.26623261e+00 -3.87802035e-01 2.08712786e-01
-3.27341318e-01 -6.92429364e-01 -3.79520804e-01 6.57350570e-02
-5.35663307e-01 1.29884493e+00 1.34329200e-01 -3.94690096e-01
1.90369040e-01 5.95887423e-01 2.74747401e-01 -7.85955429e-01
-4.19591218e-01 6.20842218e-01 2.14801401e-01 -5.17806530e-01
-3.70567851e-02 -7.67011404e-01 -1.65026402e+00 -3.70610058e-02
-2.39019394e-01 1.39398098e-01 1.03098822e+00 1.04944468e+00
5.42695403e-01 4.92073148e-01 7.00631559e-01 -8.21447730e-01
-8.18046570e-01 -1.16062272e+00 -2.77473897e-01 2.92435884e-02
7.60749817e-01 -9.60020423e-02 -2.00281084e-01 3.13113958e-01] | [14.47337532043457, 7.045506954193115] |
6c24bfe7-d7f6-41a6-87d4-cde2bbdbeb26 | query-based-industrial-analytics-over | 2209.11089 | null | https://arxiv.org/abs/2209.11089v1 | https://arxiv.org/pdf/2209.11089v1.pdf | Query-based Industrial Analytics over Knowledge Graphs with Ontology Reshaping | Industrial analytics that includes among others equipment diagnosis and anomaly detection heavily relies on integration of heterogeneous production data. Knowledge Graphs (KGs) as the data format and ontologies as the unified data schemata are a prominent solution that offers high quality data integration and a convenient and standardised way to exchange data and to layer analytical applications over it. However, poor design of ontologies of high degree of mismatch between them and industrial data naturally lead to KGs of low quality that impede the adoption and scalability of industrial analytics. Indeed, such KGs substantially increase the training time of writing queries for users, consume high volume of storage for redundant information, and are hard to maintain and update. To address this problem we propose an ontology reshaping approach to transform ontologies into KG schemata that better reflect the underlying data and thus help to construct better KGs. In this poster we present a preliminary discussion of our on-going research, evaluate our approach with a rich set of SPARQL queries on real-world industry data at Bosch and discuss our findings. | ['Evgeny Kharlamo', 'Ahmet Soylu', 'Ernesto Jiménez-Ruiz', 'Gong Cheng', 'Dongzhuoran Zhou', 'Baifan Zhou', 'Zhuoxun Zheng'] | 2022-09-22 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [-2.42919043e-01 1.87038839e-01 -1.15503632e-01 -3.87185872e-01
-1.21101163e-01 -5.80263197e-01 9.52678397e-02 1.03550971e+00
1.51964411e-01 6.75721526e-01 -4.55280572e-01 -2.34109581e-01
-8.74619067e-01 -1.46515036e+00 -5.70471644e-01 7.84014538e-03
7.29890242e-02 7.61321783e-01 5.70937514e-01 -4.68411595e-01
1.54161349e-01 7.57178068e-01 -2.17202163e+00 2.57394552e-01
1.11717141e+00 1.30076313e+00 2.30306163e-01 5.91229983e-02
-7.42422104e-01 1.02804792e+00 -8.45343828e-01 -4.65242803e-01
3.26476246e-01 2.60135233e-01 -8.25938404e-01 -5.76126166e-02
1.63452953e-01 2.62910038e-01 1.80960804e-01 1.29181814e+00
1.59035534e-01 -1.90841600e-01 4.85526696e-02 -1.70433462e+00
-5.26849806e-01 4.13160294e-01 1.76539674e-01 -1.36468023e-01
5.01050830e-01 -1.63010731e-01 6.75892055e-01 -3.51269454e-01
1.13779032e+00 9.01311338e-01 6.01347566e-01 -2.15751790e-02
-1.00851858e+00 -4.42823827e-01 -5.62851653e-02 5.23407519e-01
-1.32712245e+00 -1.34478420e-01 6.79593325e-01 -3.27151716e-01
1.10210693e+00 6.26158357e-01 9.02559996e-01 4.54749882e-01
2.48227462e-01 4.47072014e-02 5.80281973e-01 -3.27564806e-01
3.18370581e-01 7.21001863e-01 2.97119528e-01 5.62956870e-01
9.90585983e-01 -4.41455960e-01 -6.92242503e-01 -2.00431824e-01
4.84757125e-01 9.98752117e-02 9.46465284e-02 -8.77273321e-01
-6.69867635e-01 4.31336403e-01 1.51078135e-01 4.41569686e-01
-5.33405304e-01 -3.54442686e-01 6.05845273e-01 7.11161196e-01
-7.38883838e-02 6.18614435e-01 -6.68861210e-01 -5.32313250e-02
-2.26918265e-01 4.53194588e-01 1.12031925e+00 1.44954884e+00
1.04558587e+00 -7.31735155e-02 4.41962063e-01 6.31357312e-01
3.56100917e-01 2.49720633e-01 2.84407377e-01 -6.06827557e-01
5.89677811e-01 1.56879973e+00 8.92543271e-02 -1.06890595e+00
-4.25759971e-01 -1.93158641e-01 -3.02715361e-01 3.52098674e-01
1.85765639e-01 4.75047946e-01 -5.37949383e-01 1.03878736e+00
4.93038267e-01 -4.02680218e-01 5.11897564e-01 5.41234553e-01
6.10429943e-01 2.04167366e-01 2.44679824e-01 -1.70329530e-02
1.49939883e+00 -3.31776947e-01 -1.16897023e+00 2.79328793e-01
7.62428403e-01 -7.23179877e-01 8.54016900e-01 6.15728855e-01
-9.47432399e-01 -6.13716602e-01 -1.35649347e+00 8.30472447e-03
-1.33685863e+00 -3.67599994e-01 7.08230972e-01 6.10964835e-01
-5.29871225e-01 7.79965043e-01 -5.16741753e-01 -6.93561733e-01
2.31287360e-01 2.83590078e-01 -4.79926378e-01 -1.54356107e-01
-1.34899879e+00 1.11099946e+00 1.07208192e+00 -1.78485677e-01
-1.33937627e-01 -9.72888052e-01 -7.55640209e-01 -8.35125372e-02
7.54083097e-01 -5.37327826e-01 8.29093099e-01 -2.99215704e-01
-1.01152074e+00 5.36501884e-01 4.94516909e-01 -5.36275625e-01
2.50427634e-01 -1.00441255e-01 -1.45423508e+00 -2.15949461e-01
1.01025999e-01 -2.91278660e-01 2.81228036e-01 -1.14814019e+00
-9.20006931e-01 -7.00958490e-01 -2.11442187e-02 -3.46868306e-01
-2.03763381e-01 -1.73682362e-01 -1.80995509e-01 -5.51016778e-02
2.14910597e-01 -3.68061095e-01 1.41611043e-02 -4.13177431e-01
-1.50897101e-01 -2.95821458e-01 1.05222809e+00 -6.57889187e-01
1.52372754e+00 -1.99146855e+00 -6.44331351e-02 5.36067843e-01
1.08564988e-01 1.92293063e-01 4.01708931e-01 9.27441001e-01
-1.86868027e-01 1.08369634e-01 3.50533500e-02 3.75731826e-01
3.05355102e-01 6.76055193e-01 -1.54506296e-01 -1.50609612e-01
1.97587952e-01 4.75840002e-01 -7.03899622e-01 -4.81383771e-01
4.57618237e-01 7.52332434e-02 -3.07138950e-01 3.11995953e-01
-5.81077456e-01 2.48291921e-02 -6.29089057e-01 9.86108601e-01
6.00931406e-01 -8.86293650e-02 6.68091178e-01 -5.87283731e-01
-2.97669142e-01 2.06323802e-01 -1.67716992e+00 1.77109814e+00
-3.06218684e-01 -8.79856795e-02 -3.21607850e-02 -8.92160952e-01
1.30999112e+00 4.06427652e-01 9.03553963e-01 -8.18194330e-01
-7.05050454e-02 4.38366473e-01 -4.52709287e-01 -1.05875063e+00
7.16898203e-01 7.39092231e-02 -2.08560243e-01 -7.52465352e-02
9.00378376e-02 -1.82233676e-01 7.71700501e-01 -9.68147721e-03
1.05300915e+00 2.79918641e-01 3.74157459e-01 -3.33835334e-01
7.72105217e-01 3.56567115e-01 6.91892326e-01 2.10499302e-01
3.04171294e-01 -4.60158177e-02 6.96430027e-01 -7.96717167e-01
-9.58605230e-01 -1.02278590e+00 -9.76391807e-02 5.41657269e-01
1.55551881e-01 -9.19550776e-01 -4.69722688e-01 -5.51366031e-01
5.28354824e-01 9.43832338e-01 -6.65518716e-02 -1.44424736e-01
-2.00979322e-01 -3.90969902e-01 3.42038959e-01 3.16388935e-01
2.39319548e-01 -9.55917299e-01 -8.01976264e-01 4.95428830e-01
1.39583260e-01 -1.25604010e+00 6.76309586e-01 1.37825549e-01
-9.13510799e-01 -1.68919706e+00 4.55006897e-01 -3.81346822e-01
3.60367179e-01 -3.54161322e-01 1.52898967e+00 1.67890519e-01
-5.46665967e-01 6.11962862e-02 -4.95524317e-01 -1.10360861e+00
-7.82434940e-01 1.75220072e-01 3.87817807e-02 -2.72353351e-01
4.81868476e-01 -6.41542494e-01 1.60387885e-02 3.15991729e-01
-1.46416569e+00 -2.98119426e-01 3.84507537e-01 -1.80333257e-02
6.25760972e-01 6.87811792e-01 8.19542825e-01 -1.22966313e+00
7.24529207e-01 -5.75585186e-01 -1.20496893e+00 7.64747798e-01
-1.31660450e+00 1.89410597e-01 5.34696579e-01 2.31803745e-01
-9.17654216e-01 -7.94735029e-02 5.71052507e-02 -3.85815859e-01
-2.49585897e-01 7.27862656e-01 -5.31543493e-01 1.41930714e-01
3.41094345e-01 -2.97959268e-01 1.22194424e-01 -1.02799332e+00
2.21591756e-01 6.70783281e-01 5.48403919e-01 -6.92125916e-01
6.91550493e-01 2.06010774e-01 1.65160209e-01 -5.92153490e-01
-2.19969720e-01 -4.08858895e-01 -7.06204772e-01 -1.67808682e-01
8.07202935e-01 -5.45533240e-01 -7.94228852e-01 5.81385158e-02
-1.10539448e+00 3.13908875e-01 -8.19335163e-01 1.81972355e-01
-4.86553907e-01 2.62718022e-01 -1.57579016e-02 -7.78841734e-01
-6.25965744e-02 -8.91191423e-01 7.47036874e-01 -1.60722986e-01
-1.95450231e-01 -7.68532693e-01 -7.59734213e-02 4.62332278e-01
4.10736442e-01 6.69607520e-01 1.33694565e+00 -7.38350034e-01
-9.58007812e-01 -5.54075718e-01 -1.37293369e-01 3.42449695e-01
4.90058273e-01 1.42716998e-02 -5.90850055e-01 4.77267709e-03
-2.28086740e-01 1.47712648e-01 -2.56490767e-01 -6.66055024e-01
1.04007864e+00 8.75924751e-02 -3.26217949e-01 2.20836639e-01
1.87556374e+00 4.03024048e-01 8.27175677e-01 5.97328365e-01
6.18807435e-01 8.35070014e-01 9.78091896e-01 4.58795786e-01
3.29760671e-01 8.20705295e-01 7.76658475e-01 3.13820034e-01
7.91116580e-02 -2.06996784e-01 -4.41407673e-02 9.58728552e-01
-2.65647531e-01 -9.14501101e-02 -8.07495236e-01 3.14622372e-01
-2.08424306e+00 -7.13944256e-01 -4.59382594e-01 2.37400579e+00
6.04496837e-01 1.20936781e-01 2.68987492e-02 4.43732828e-01
4.80919510e-01 -3.99859041e-01 -3.63580644e-01 -3.27591121e-01
-2.36378442e-02 2.81939358e-01 7.42685854e-01 1.75490201e-01
-5.77049315e-01 5.94923198e-01 5.78150558e+00 2.37824753e-01
-8.30264807e-01 8.39787498e-02 -6.31554484e-01 9.48677808e-02
-2.14347214e-01 -1.56242559e-02 -6.28525674e-01 4.70711142e-01
1.37061846e+00 -4.79176491e-01 5.11607349e-01 1.09861732e+00
-1.87100977e-01 1.75845966e-01 -1.11600232e+00 1.01965797e+00
-3.48941922e-01 -1.59900796e+00 1.09281532e-01 2.54989117e-01
3.45430791e-01 -7.35000288e-03 -7.43185222e-01 1.92482043e-02
3.05125654e-01 -6.48652017e-01 8.97340894e-01 8.47681642e-01
2.36313298e-01 -7.41588950e-01 1.04961681e+00 -1.25369921e-01
-1.31453288e+00 -2.82143325e-01 -5.10826647e-01 1.70146272e-01
2.13152602e-01 7.15678871e-01 -7.33075678e-01 1.77325678e+00
9.68013465e-01 3.89641434e-01 -6.89829051e-01 7.33182788e-01
3.24547440e-01 -3.00074637e-01 -3.82700861e-01 3.34346503e-01
-4.59793448e-01 -3.00518632e-01 3.25865775e-01 6.26611292e-01
5.79850793e-01 -4.57657874e-01 1.52150646e-01 7.82988250e-01
1.64859533e-01 4.48648542e-01 -8.58282804e-01 -4.45018411e-01
5.14948845e-01 1.00897872e+00 -3.58741730e-01 -3.67393821e-01
-6.15180075e-01 3.71110618e-01 2.76384115e-01 1.14076331e-01
-6.09109819e-01 -5.60156465e-01 1.05558753e+00 5.32914758e-01
1.93967685e-01 -1.26152009e-01 -9.75357890e-02 -8.94461513e-01
3.37478638e-01 -7.99241185e-01 8.43972087e-01 -8.22418272e-01
-1.41927564e+00 6.50746047e-01 1.23078655e-02 -1.29038322e+00
-3.88007611e-01 -7.70699263e-01 7.72872269e-02 7.42400408e-01
-1.46460783e+00 -1.09024525e+00 -6.85261250e-01 6.26227558e-01
9.49787870e-02 -1.76902384e-01 1.19274044e+00 8.32020819e-01
-4.92076516e-01 -2.92539597e-02 -1.46016687e-01 -4.04112458e-01
5.16007364e-01 -1.25000000e+00 -7.41036907e-02 4.50902313e-01
-6.68498129e-02 6.22428954e-01 7.41525590e-01 -9.93997693e-01
-1.72961771e+00 -1.06299436e+00 7.04649448e-01 -5.91956675e-01
8.70267153e-01 -9.21773240e-02 -1.16775203e+00 7.28035808e-01
1.08342789e-01 -2.10960787e-02 7.67214417e-01 1.67433340e-02
-3.04419845e-01 -8.35333645e-01 -1.43177176e+00 -3.33217788e-03
9.64632034e-01 -4.16169316e-01 -6.79537833e-01 3.71562034e-01
6.43223226e-01 -1.06888197e-01 -1.98148704e+00 5.17443538e-01
4.03493851e-01 -1.22901046e+00 8.05372775e-01 -8.91550720e-01
-2.70619363e-01 -6.93721473e-01 -3.89740825e-01 -8.55093241e-01
7.55811110e-02 -4.60424006e-01 -1.84955791e-01 1.42838347e+00
3.02986354e-01 -1.02348351e+00 5.06816447e-01 7.75763690e-01
-3.54101747e-01 -4.60454464e-01 -6.93717837e-01 -1.14173174e+00
-6.60610080e-01 -6.13975883e-01 1.59904587e+00 1.09471285e+00
2.49930680e-01 -1.07123874e-01 2.04657584e-01 4.67598319e-01
5.15521824e-01 1.28369376e-01 8.85465622e-01 -2.10495019e+00
1.72207147e-01 7.38317668e-02 -1.01490128e+00 3.34328979e-01
-3.23406756e-01 -8.54420841e-01 -5.36957324e-01 -1.77380514e+00
-6.15260661e-01 -8.71059299e-01 -5.73835671e-01 3.30836236e-01
6.85745299e-01 -7.27931336e-02 1.42584324e-01 2.91709065e-01
-4.15868670e-01 5.88077530e-02 8.15213680e-01 1.18773833e-01
-1.41725749e-01 -2.83979684e-01 -4.58424181e-01 4.98598307e-01
7.53788173e-01 -4.87649500e-01 -6.28715754e-01 -1.69526756e-01
7.70317674e-01 -2.26665735e-01 1.68605343e-01 -1.38350117e+00
2.55154669e-01 -3.47516477e-01 -2.52926908e-02 -6.18971527e-01
2.18209624e-01 -1.65357530e+00 1.25541961e+00 3.77278805e-01
3.09073269e-01 4.66146737e-01 1.65360093e-01 4.07725930e-01
-3.53576064e-01 -5.06489754e-01 2.63962358e-01 -2.08770394e-01
-9.21748936e-01 9.28736851e-02 1.15175448e-01 -5.37954807e-01
1.33251667e+00 -1.99049622e-01 -4.27198201e-01 3.51944298e-01
-9.29697156e-01 1.55525193e-01 8.60624611e-01 8.04776847e-01
7.47382268e-02 -1.31040323e+00 -3.18822712e-02 6.71644747e-01
5.88344336e-01 2.16427207e-01 9.04183984e-02 3.89282435e-01
-8.87660205e-01 6.17457628e-01 -6.20388687e-01 -1.71351492e-01
-1.10126424e+00 9.99642134e-01 1.86922520e-01 -4.40773875e-01
-6.48676455e-01 -1.48909822e-01 -6.63595378e-01 -5.55880904e-01
7.15325400e-02 -5.90484262e-01 -2.86760539e-01 2.43855596e-01
2.88119197e-01 6.52542353e-01 7.14127123e-01 -3.62538904e-01
-4.71916586e-01 2.96117783e-01 2.94264585e-01 4.39559758e-01
1.46329856e+00 -1.96272343e-01 -5.58151364e-01 6.40692770e-01
6.22766137e-01 1.58895060e-01 -5.42723358e-01 9.44236144e-02
6.78483605e-01 -5.12755513e-01 -2.93184221e-01 -8.65783632e-01
-1.03681695e+00 1.99064776e-01 6.65972173e-01 1.27044034e+00
9.74633813e-01 1.47387534e-01 6.24523282e-01 5.49566329e-01
6.96188688e-01 -1.46340632e+00 -5.09937346e-01 -1.41900197e-01
8.26075494e-01 -7.47946799e-01 2.98757643e-01 -1.11842728e+00
-3.74953955e-01 1.29720485e+00 5.97422063e-01 3.49821240e-01
7.45420694e-01 3.23116899e-01 2.34505609e-01 -8.68060410e-01
-5.22630751e-01 -3.39088172e-01 -8.93070027e-02 9.08153176e-01
3.69530506e-02 -3.78588513e-02 -2.56253034e-01 6.94211602e-01
-2.08823517e-01 5.61718941e-01 2.66125411e-01 1.40363562e+00
-2.52117932e-01 -1.78408611e+00 -2.81332552e-01 4.80670750e-01
-3.31026733e-01 5.86804092e-01 -1.86441958e-01 1.16868770e+00
5.61169982e-01 7.77310312e-01 2.19604552e-01 -3.32125992e-01
1.25178111e+00 5.45407653e-01 4.46488857e-01 -4.67428327e-01
-5.04018664e-01 -6.69524908e-01 5.06264329e-01 -8.10416281e-01
-2.44276643e-01 -5.55750012e-01 -1.32101905e+00 -3.05070698e-01
-1.89329997e-01 4.33562636e-01 1.17230463e+00 6.99852824e-01
7.29846418e-01 9.51371610e-01 1.48152322e-01 1.36872858e-01
-4.19574440e-01 -7.28849769e-01 -1.06333423e+00 8.61048579e-01
-3.54956120e-01 -9.19084072e-01 3.05417001e-01 2.66723573e-01] | [9.120279312133789, 7.7033538818359375] |
c1b03033-11ab-4295-b48a-d998083c7a3f | transferring-from-formal-newswire-domain-with | null | null | https://aclanthology.org/D18-1275 | https://aclanthology.org/D18-1275.pdf | Transferring from Formal Newswire Domain with Hypernet for Twitter POS Tagging | Part-of-Speech (POS) tagging for Twitter has received considerable attention in recent years. Because most POS tagging methods are based on supervised models, they usually require a large amount of labeled data for training. However, the existing labeled datasets for Twitter are much smaller than those for newswire text. Hence, to help POS tagging for Twitter, most domain adaptation methods try to leverage newswire datasets by learning the shared features between the two domains. However, from a linguistic perspective, Twitter users not only tend to mimic the formal expressions of traditional media, like news, but they also appear to be developing linguistically informal styles. Therefore, POS tagging for the formal Twitter context can be learned together with the newswire dataset, while POS tagging for the informal Twitter context should be learned separately. To achieve this task, in this work, we propose a hypernetwork-based method to generate different parameters to separately model contexts with different expression styles. Experimental results on three different datasets show that our approach achieves better performance than state-of-the-art methods in most cases. | ['Xuanjing Huang', 'Keyu Ding', 'Di Liang', 'Tao Gui', 'Minlong Peng', 'Qi Zhang', 'Jingjing Gong'] | 2018-10-01 | null | null | null | emnlp-2018-10 | ['stock-prediction'] | ['time-series'] | [-2.49452055e-01 1.31468788e-01 -6.97222412e-01 -6.91337407e-01
-3.29894304e-01 -6.50630772e-01 6.37680829e-01 3.62001926e-01
-4.14740622e-01 6.71306968e-01 4.32760894e-01 -8.46361443e-02
4.73953664e-01 -7.04340398e-01 -3.17639589e-01 -5.30182838e-01
3.73321891e-01 4.84615654e-01 5.09384573e-01 -3.34520280e-01
-1.61836281e-01 1.22663509e-02 -1.15200555e+00 2.38458380e-01
7.30529010e-01 7.74610639e-01 1.07866377e-01 -3.22530530e-02
-8.53089511e-01 8.03737581e-01 -5.71240425e-01 -6.19044840e-01
-6.77211434e-02 -1.67975664e-01 -6.37577951e-01 -1.70285907e-02
9.33261812e-02 2.66913980e-01 -2.12490007e-01 1.19455934e+00
4.83842939e-01 9.88903418e-02 6.10246539e-01 -1.11688173e+00
-4.94756162e-01 1.14134479e+00 -5.09004533e-01 5.51936962e-02
1.23174638e-01 -3.20701361e-01 1.13807285e+00 -5.11793733e-01
5.16584516e-01 9.90892112e-01 7.21426904e-01 6.31499946e-01
-9.06467497e-01 -8.16768765e-01 4.95665461e-01 -2.59617627e-01
-1.33895040e+00 -1.66302308e-01 9.32819605e-01 -3.99802774e-01
3.87477815e-01 -9.00424495e-02 6.07090890e-01 1.34849060e+00
-3.02097291e-01 9.43844676e-01 1.15688360e+00 -3.68317574e-01
1.66101858e-01 6.31656647e-01 1.68080106e-01 3.04350436e-01
5.20221740e-02 -4.44641352e-01 -2.67310232e-01 -2.04593077e-01
3.31544667e-01 2.54881024e-01 1.41357616e-01 3.97268943e-02
-9.42928433e-01 7.60527074e-01 1.93776295e-01 8.85643125e-01
-1.35332435e-01 -3.31823736e-01 3.86224508e-01 -1.09115884e-01
8.48069429e-01 3.25085163e-01 -8.17963123e-01 -1.95866123e-01
-7.32742786e-01 -8.95714313e-02 1.08795118e+00 1.10041571e+00
1.01910985e+00 -2.95819730e-01 -2.01480135e-01 1.17058873e+00
5.11095166e-01 5.49454391e-01 7.43910193e-01 -3.91444027e-01
5.84815800e-01 7.81302929e-01 1.28982767e-01 -9.68302310e-01
-5.12504578e-01 -2.71091163e-01 -7.09445953e-01 -8.50786924e-01
4.06265050e-01 -7.64097691e-01 -6.47886038e-01 1.72007859e+00
3.53550196e-01 4.22716022e-01 1.68034524e-01 3.97142678e-01
8.95369887e-01 9.42536533e-01 6.31667078e-01 -2.25185275e-01
1.45200014e+00 -8.87268484e-01 -7.66407251e-01 -6.56269968e-01
8.76517355e-01 -7.33555019e-01 1.18525577e+00 -8.26136395e-02
-6.19485378e-01 -2.86963820e-01 -6.34510577e-01 3.54984999e-01
-6.58399105e-01 7.50318319e-02 5.84459722e-01 6.11081898e-01
-5.95227361e-01 2.00585559e-01 -6.47149146e-01 -6.45558357e-01
1.81731224e-01 1.02762498e-01 -1.51599824e-01 1.23296238e-01
-1.62760305e+00 4.22078192e-01 5.67226350e-01 -2.38626987e-01
-6.91332072e-02 -4.06081229e-01 -8.00159395e-01 1.60079613e-01
2.13408276e-01 -1.64202422e-01 1.49575341e+00 -1.42199910e+00
-1.54207730e+00 8.16920638e-01 -1.73180267e-01 -1.17795683e-01
5.19780405e-02 1.21271603e-01 -7.28962541e-01 -7.97306150e-02
1.71314955e-01 5.93653619e-01 5.48007727e-01 -1.34391463e+00
-8.77120376e-01 6.74626976e-02 2.57236540e-01 2.93294061e-02
-9.75883424e-01 2.61837393e-01 -5.84547758e-01 -7.38280535e-01
-3.45658302e-01 -1.07708013e+00 -2.90450871e-01 -6.07240915e-01
-3.62663329e-01 -7.20213473e-01 7.18312621e-01 -2.58757204e-01
1.62286687e+00 -2.43919992e+00 -5.37768483e-01 2.47058958e-01
5.41373640e-02 5.67366898e-01 1.99472755e-02 5.27257502e-01
1.48727775e-01 3.76281798e-01 7.02110454e-02 -5.51218867e-01
9.51708034e-02 5.90841651e-01 -2.75247991e-01 4.38366309e-02
-3.56048644e-02 5.89361489e-01 -1.24090147e+00 -9.40672576e-01
-3.10740471e-01 3.01064789e-01 -5.07263958e-01 2.61986554e-01
-4.41811562e-01 5.43975115e-01 -8.69205773e-01 5.21292746e-01
5.53290665e-01 -4.48274910e-01 6.57684267e-01 7.41555169e-02
-2.53823966e-01 2.70426869e-01 -6.95399642e-01 1.29046655e+00
-9.05363142e-01 3.95985246e-01 -1.85367361e-01 -8.18794787e-01
9.83653009e-01 5.17690897e-01 6.59623027e-01 -4.41076636e-01
3.37683737e-01 1.57789946e-01 -1.66922569e-01 -5.59128761e-01
4.47148383e-01 -4.75522637e-01 -4.78098243e-01 3.24379772e-01
-8.73581097e-02 2.21730411e-01 2.43207783e-01 -9.26964357e-02
7.97081113e-01 -1.79261282e-01 2.77114928e-01 5.84555343e-02
4.12679881e-01 -1.40742719e-01 1.05948615e+00 4.41320717e-01
-3.02253664e-01 2.66951501e-01 5.07652342e-01 -2.72045821e-01
-6.47581756e-01 -4.75466847e-01 -1.89201877e-01 1.47442758e+00
2.07128406e-01 -6.54717565e-01 -8.20412934e-01 -1.16921949e+00
-3.52431625e-01 5.51473975e-01 -4.55191493e-01 2.69728005e-01
-6.35396719e-01 -7.86906540e-01 7.00942934e-01 5.94532669e-01
6.22752070e-01 -1.16771245e+00 2.00896189e-01 3.70334238e-01
-3.78028452e-01 -1.49761486e+00 -5.65023065e-01 1.23967424e-01
-5.20976245e-01 -7.23427355e-01 -7.06230283e-01 -1.18586206e+00
6.77945077e-01 2.56561726e-01 9.81562912e-01 7.19197914e-02
7.48693287e-01 7.61915147e-02 -8.30677330e-01 -6.29827619e-01
-4.07165587e-01 5.83161533e-01 -4.84436825e-02 4.02502596e-01
7.72091746e-01 -6.16892397e-01 -2.95177430e-01 6.30880892e-01
-1.14958930e+00 -2.13745937e-01 7.10114777e-01 5.96182346e-01
3.58796120e-01 2.04403475e-01 5.75086713e-01 -1.42380416e+00
3.83373022e-01 -8.28769565e-01 -2.39401355e-01 2.91958988e-01
-2.84417719e-01 -1.13998458e-01 8.88449669e-01 -9.11037564e-01
-1.39509308e+00 9.88081843e-02 -3.01788926e-01 -9.39093009e-02
-5.05390227e-01 9.35071766e-01 -3.53264660e-01 3.07745993e-01
4.62110311e-01 5.33957407e-03 -5.10133088e-01 -5.61610401e-01
1.89116283e-03 1.16184604e+00 -1.42222255e-01 -6.34288847e-01
7.39967942e-01 3.63929540e-01 -3.36180300e-01 -9.28144872e-01
-1.58385479e+00 -8.90550613e-01 -5.18522263e-01 -5.82652213e-03
8.46551955e-01 -1.03634131e+00 -2.51789361e-01 4.39104229e-01
-9.83665586e-01 -5.18135965e-01 -1.38830468e-01 4.14040208e-01
-5.75764221e-04 2.91611910e-01 -4.44459736e-01 -5.38181186e-01
3.42307277e-02 -7.26713181e-01 1.00542068e+00 5.35517335e-01
-2.61940897e-01 -1.65812004e+00 3.16292763e-01 1.00410655e-01
3.85939479e-01 -8.77419934e-02 8.40806186e-01 -1.37772822e+00
1.82664201e-01 -3.51321965e-01 -1.68985516e-01 2.80574888e-01
2.70440638e-01 6.77034408e-02 -9.01133478e-01 1.94173768e-01
-3.08195382e-01 -2.25071535e-01 3.16382438e-01 -5.80451600e-02
1.00589275e+00 -6.53700531e-01 -5.66508293e-01 2.20513210e-01
1.08838391e+00 7.73257762e-02 3.25856835e-01 4.49952453e-01
7.05327988e-01 7.93202698e-01 8.31402242e-01 5.59610426e-01
7.39070475e-01 6.32370114e-01 -2.77161207e-02 -8.17686394e-02
1.99465290e-01 -6.76352262e-01 5.55988550e-01 1.20190287e+00
2.71943569e-01 -3.35728496e-01 -1.16901219e+00 5.84098160e-01
-1.97206628e+00 -9.69717979e-01 -3.34803574e-02 1.80561328e+00
1.29726565e+00 2.67139494e-01 2.79457778e-01 -2.56195694e-01
1.02705133e+00 3.54914248e-01 -1.36362553e-01 -7.60805532e-02
-1.18767425e-01 -7.23290667e-02 4.14204925e-01 2.75817990e-01
-1.33755863e+00 1.19986284e+00 5.62000227e+00 9.75103855e-01
-1.45071292e+00 2.19902575e-01 4.67867464e-01 3.50172073e-01
-3.68767112e-01 2.66245782e-01 -1.27811468e+00 8.82232189e-01
9.98084068e-01 -8.51299763e-02 -1.97200596e-01 9.84499395e-01
1.43623099e-01 2.79653996e-01 -7.18488991e-01 8.75883222e-01
-1.52797043e-01 -8.41097176e-01 -3.29110436e-02 -6.90447018e-02
9.39123809e-01 6.32746890e-02 -1.46040589e-01 7.19677210e-01
3.57514977e-01 -5.18851578e-01 5.77399373e-01 3.45049769e-01
4.97472733e-01 -6.07840180e-01 1.09465921e+00 6.81154490e-01
-1.17932522e+00 5.18593639e-02 -2.95238912e-01 -2.95674242e-02
1.36563137e-01 8.24364781e-01 -8.68054569e-01 1.80777982e-01
5.22817612e-01 9.77649212e-01 -5.26783586e-01 8.92499208e-01
-6.04134440e-01 1.10592365e+00 -4.09483790e-01 -4.87501144e-01
4.98646826e-01 6.35761470e-02 3.04083049e-01 1.65154207e+00
2.88372666e-01 8.55654385e-03 6.51894569e-01 1.52210668e-01
-3.41486484e-01 5.20164549e-01 -5.87990284e-01 -3.30371439e-01
7.19956934e-01 1.39707470e+00 -7.21022367e-01 -3.71217102e-01
-8.30769956e-01 3.63409907e-01 1.95889011e-01 2.04022363e-01
-7.69413710e-01 -4.27811265e-01 5.34638703e-01 3.56114805e-01
2.68808633e-01 -1.56669781e-01 8.56773928e-02 -1.35980499e+00
-1.41753480e-01 -5.91525614e-01 4.16402102e-01 -4.16510165e-01
-1.72647643e+00 6.48442745e-01 1.21048637e-01 -1.60772407e+00
-2.11916938e-01 -3.49547774e-01 -7.40314007e-01 4.49152678e-01
-1.77779698e+00 -1.31658804e+00 -1.66195869e-01 4.58299518e-01
3.56796831e-01 -5.67329600e-02 6.38016343e-01 5.04382491e-01
-6.39434516e-01 7.62734711e-01 1.40516773e-01 7.25507557e-01
1.05561280e+00 -1.10034049e+00 1.10299490e-01 4.67858106e-01
1.43341228e-01 6.87190354e-01 6.49588645e-01 -5.47322273e-01
-7.92888701e-01 -1.36446905e+00 1.49248731e+00 -2.01539919e-01
1.07210386e+00 -3.12756240e-01 -8.67588103e-01 8.22581291e-01
3.84211093e-01 1.78707644e-01 1.38389683e+00 4.20390785e-01
-3.27500284e-01 -1.92836270e-01 -9.62076426e-01 7.68257201e-01
8.15818727e-01 -3.22720021e-01 -6.44827783e-01 4.79746431e-01
6.14686906e-01 -1.36457279e-01 -8.43973994e-01 1.17992558e-01
2.69168466e-01 -6.07251763e-01 3.73741180e-01 -5.21885395e-01
1.90853238e-01 -2.88274616e-01 -1.03446757e-02 -1.48863542e+00
-1.66532487e-01 -6.50382698e-01 1.27344489e-01 2.07094789e+00
6.70543432e-01 -7.24581122e-01 8.03496480e-01 4.60170716e-01
-1.88251827e-02 -4.96062249e-01 -6.32965147e-01 -7.34282196e-01
8.79405811e-02 -2.66894370e-01 6.84452236e-01 1.42256129e+00
3.98631454e-01 8.40223134e-01 -2.57681102e-01 1.33189037e-01
4.01366875e-02 7.70265609e-02 5.79190433e-01 -1.55484700e+00
-1.55414179e-01 -4.23873186e-01 -2.23821655e-01 -1.16657579e+00
6.93998098e-01 -8.51546168e-01 1.79084212e-01 -1.24430013e+00
2.96204593e-02 -1.12650812e+00 -2.30058402e-01 8.91962707e-01
-3.10776472e-01 4.38850857e-02 6.21733442e-03 3.30527872e-01
-9.15802300e-01 4.02423590e-01 9.92494643e-01 -1.98128924e-01
-3.35984051e-01 1.91407561e-01 -8.68029952e-01 1.13521576e+00
9.03907299e-01 -7.98294008e-01 -3.26451302e-01 -3.67316782e-01
7.82407939e-01 -2.03308731e-01 -2.66091883e-01 -9.19907868e-01
1.32590637e-01 -5.35737514e-01 -6.03965744e-02 -1.42958909e-01
5.36445938e-02 -1.01088095e+00 -2.85566866e-01 -1.76537335e-01
-3.17522138e-01 -1.78724632e-01 2.46760696e-02 5.11938035e-01
-5.65335512e-01 -4.86420929e-01 7.63731480e-01 -6.19004630e-02
-6.10004425e-01 5.07540822e-01 -5.33824563e-01 4.11684811e-01
9.27593231e-01 1.45005405e-01 -2.16530159e-01 -3.77385199e-01
-7.13462055e-01 2.91559964e-01 5.33999145e-01 3.32610488e-01
-1.94286574e-02 -1.32678020e+00 -3.23901683e-01 -1.69587657e-01
4.97071922e-01 2.01255232e-01 9.30771902e-02 7.72350907e-01
-7.60555863e-02 2.22986713e-01 1.85416311e-01 -3.16709012e-01
-1.01182497e+00 4.29036081e-01 7.52795907e-03 -5.88670373e-01
-9.26383138e-02 6.26862347e-01 2.23691300e-01 -6.45570099e-01
9.53010097e-02 -1.83389604e-01 -7.01621115e-01 5.21116793e-01
5.07194638e-01 -1.96428418e-01 -7.45453611e-02 -8.04178655e-01
-1.75885260e-01 3.75591040e-01 -1.24967851e-01 7.05536678e-02
1.34871030e+00 -1.49596989e-01 3.49217132e-02 5.13883710e-01
1.19159245e+00 5.90718806e-01 -1.06438029e+00 -6.75126731e-01
2.82728255e-01 -1.86297059e-01 -2.06170693e-01 -4.47271019e-01
-1.04050910e+00 7.99631774e-01 2.01908816e-02 7.07611501e-01
8.70786250e-01 2.74128467e-01 1.08818352e+00 4.55085218e-01
3.24533075e-01 -1.29025948e+00 3.03209145e-02 1.04775202e+00
1.92794397e-01 -1.17939389e+00 -4.86003816e-01 -5.17380178e-01
-9.79738593e-01 9.75203276e-01 7.17650473e-01 1.70953870e-01
1.14898729e+00 2.45221049e-01 4.67112064e-01 6.29817843e-02
-5.42082489e-01 -3.86366874e-01 4.82561067e-02 6.07342303e-01
8.36322010e-01 2.72598881e-02 -3.33240479e-01 1.01223886e+00
-2.23366812e-01 -2.00046659e-01 3.71704340e-01 1.19667995e+00
-4.05202955e-01 -1.67625308e+00 -5.51360734e-02 1.66053951e-01
-7.88827956e-01 -1.62955895e-01 -5.00324786e-01 5.66034436e-01
2.84771681e-01 1.04077721e+00 8.04556832e-02 -3.70186448e-01
3.65543395e-01 2.59309858e-01 -5.60884774e-02 -1.02528918e+00
-7.22366214e-01 7.93396831e-02 3.43112439e-01 5.51416166e-02
-8.94148707e-01 -7.64846683e-01 -1.33414006e+00 -2.38542885e-01
-2.13637277e-01 6.06188416e-01 5.81899881e-01 1.28146410e+00
3.01297277e-01 2.27329656e-02 8.44615281e-01 -7.51076639e-01
-2.17292041e-01 -1.13067639e+00 -7.03936219e-01 4.65552658e-01
-3.99225652e-02 -6.86504006e-01 -2.52458423e-01 5.83799034e-02] | [9.74415111541748, 9.473616600036621] |
6e884b01-1470-487b-8e6e-77f74d4d44d2 | off-beat-multi-agent-reinforcement-learning | 2205.13718 | null | https://arxiv.org/abs/2205.13718v2 | https://arxiv.org/pdf/2205.13718v2.pdf | Off-Beat Multi-Agent Reinforcement Learning | We investigate model-free multi-agent reinforcement learning (MARL) in environments where off-beat actions are prevalent, i.e., all actions have pre-set execution durations. During execution durations, the environment changes are influenced by, but not synchronised with, action execution. Such a setting is ubiquitous in many real-world problems. However, most MARL methods assume actions are executed immediately after inference, which is often unrealistic and can lead to catastrophic failure for multi-agent coordination with off-beat actions. In order to fill this gap, we develop an algorithmic framework for MARL with off-beat actions. We then propose a novel episodic memory, LeGEM, for model-free MARL algorithms. LeGEM builds agents' episodic memories by utilizing agents' individual experiences. It boosts multi-agent learning by addressing the challenging temporal credit assignment problem raised by the off-beat actions via our novel reward redistribution scheme, alleviating the issue of non-Markovian reward. We evaluate LeGEM on various multi-agent scenarios with off-beat actions, including Stag-Hunter Game, Quarry Game, Afforestation Game, and StarCraft II micromanagement tasks. Empirical results show that LeGEM significantly boosts multi-agent coordination and achieves leading performance and improved sample efficiency. | ['Changjie Fan', 'Yingfeng Chen', 'Jianye Hao', 'Zinovi Rabinovich', 'Svetlana Obraztsova', 'Yujing Hu', 'Bo An', 'Rundong Wang', 'Weixun Wang', 'Wei Qiu'] | 2022-05-27 | null | null | null | null | ['starcraft-ii'] | ['playing-games'] | [-1.16594970e-01 -2.11886212e-01 -3.99537623e-01 1.06905058e-01
-5.31382084e-01 -3.54534686e-01 6.72508597e-01 2.68838316e-01
-8.18685174e-01 1.31252742e+00 6.12045564e-02 8.92970040e-02
-4.70160216e-01 -9.09010351e-01 -8.30967903e-01 -9.06488299e-01
-5.77562988e-01 9.92333651e-01 2.61373669e-01 -3.22977811e-01
2.18928993e-01 -3.61415707e-02 -1.37967455e+00 -9.95668396e-03
7.96064019e-01 4.56187606e-01 3.66764694e-01 5.94582558e-01
4.62699354e-01 1.42738152e+00 -7.55502462e-01 1.17536979e-02
1.53136685e-01 -4.48655248e-01 -7.34372437e-01 1.96742743e-01
-4.72009212e-01 -4.03677881e-01 -4.33066785e-01 7.15321600e-01
5.36271036e-01 6.64369881e-01 3.94573063e-01 -1.65962541e+00
-3.80068630e-01 9.55412507e-01 -7.82277405e-01 2.89530516e-01
1.95375353e-01 2.96604991e-01 1.04720080e+00 -2.59459049e-01
4.95170623e-01 1.24556565e+00 4.62180376e-01 5.29504597e-01
-9.94556367e-01 -5.60689569e-01 6.35924041e-01 5.35924911e-01
-8.56509686e-01 -1.80967078e-01 5.22096992e-01 -7.05724657e-02
1.28875983e+00 -1.14475675e-01 7.23454297e-01 1.26720965e+00
5.45249522e-01 1.06605089e+00 1.21418726e+00 -1.92715362e-01
6.72436655e-01 -7.24106789e-01 -2.39567190e-01 4.88876045e-01
6.75383210e-02 4.91766870e-01 -8.63904476e-01 -2.98687130e-01
7.23516643e-01 1.86803445e-01 2.76452184e-01 -1.44395903e-01
-1.32671106e+00 8.47819805e-01 -4.59714681e-02 -1.25999227e-01
-8.84272993e-01 4.70858604e-01 4.39663768e-01 6.90471590e-01
2.09186599e-01 3.87542486e-01 -3.60226035e-01 -5.16439080e-01
-5.71005702e-01 5.95600724e-01 7.55723238e-01 5.90704024e-01
6.11844540e-01 2.03991815e-01 -1.88578442e-01 7.19289541e-01
6.96413964e-02 5.43664217e-01 6.32733047e-01 -1.33532584e+00
3.03687125e-01 1.06152602e-01 4.92194742e-01 -6.32181346e-01
-7.41385877e-01 -2.92534679e-01 -8.24738264e-01 4.00812358e-01
4.29044783e-01 -3.27328563e-01 -5.15362084e-01 2.10369754e+00
4.03150797e-01 6.15266979e-01 3.61705601e-01 8.62311006e-01
1.42967314e-01 5.72553515e-01 7.78476819e-02 -6.31453395e-01
1.07993472e+00 -1.23031628e+00 -5.01765132e-01 -4.84721988e-01
4.79104847e-01 -2.29466856e-01 7.43622661e-01 4.64715719e-01
-1.10624444e+00 -8.47278237e-02 -6.97228193e-01 7.84961522e-01
1.13521479e-01 -5.58894277e-01 8.73694479e-01 1.16237037e-01
-6.40305400e-01 5.54998815e-01 -1.22677326e+00 3.36922035e-02
2.24535584e-01 3.75858784e-01 -9.11889374e-02 6.53670058e-02
-1.26162481e+00 9.73091304e-01 2.99871743e-01 -1.77659020e-01
-1.46504962e+00 -4.61471945e-01 -5.50035477e-01 -1.31721467e-01
1.06029987e+00 -5.34669816e-01 1.66179109e+00 -9.52561140e-01
-1.66590858e+00 2.60472357e-01 2.39743561e-01 -7.91884065e-01
6.15795135e-01 -1.75410151e-01 -3.24838668e-01 3.99199724e-02
3.28039765e-01 2.44132385e-01 7.82549560e-01 -9.68542039e-01
-7.91024745e-01 -7.10863546e-02 3.26960266e-01 6.30927682e-01
2.81944033e-02 -3.28408211e-01 6.77861348e-02 -6.73066199e-01
-5.12329340e-01 -1.18556118e+00 -5.09054124e-01 -7.13144541e-01
-1.50100957e-03 -3.89409214e-01 4.26207274e-01 -1.29161820e-01
1.03447306e+00 -1.95380795e+00 2.96767354e-01 7.47575760e-02
1.59556270e-01 -1.17127616e-02 -5.73623121e-01 9.04033601e-01
5.31475902e-01 -4.50183988e-01 -1.16522051e-01 -4.34544176e-01
3.93840283e-01 8.09107006e-01 -3.65732402e-01 4.01556343e-01
-2.80804336e-01 8.30160797e-01 -1.36250877e+00 -4.27950501e-01
2.86044851e-02 -2.34588627e-02 -6.71176672e-01 2.84590304e-01
-6.71854198e-01 5.10637581e-01 -5.35983086e-01 7.51117349e-01
2.57321745e-01 -3.10416222e-01 5.70226192e-01 6.89195454e-01
-4.18844745e-02 2.33259082e-01 -1.31392431e+00 1.57295239e+00
-4.96974856e-01 1.27078861e-01 1.25042070e-02 -9.77213144e-01
4.56695497e-01 3.39824528e-01 8.14281166e-01 -1.26979065e+00
-1.53970718e-01 -3.15322131e-02 1.74068004e-01 -2.79046357e-01
7.23324120e-01 -1.77371383e-01 -2.98445135e-01 1.05034041e+00
-2.54387110e-01 1.97118357e-01 5.39239645e-01 2.74129719e-01
1.51075566e+00 1.92493618e-01 4.69382048e-01 1.44778579e-01
-1.62939593e-01 -6.31552935e-02 1.16807747e+00 1.29700816e+00
-3.80861461e-01 -3.95504273e-02 5.54954350e-01 -4.27889466e-01
-7.44145751e-01 -1.04740059e+00 5.23325384e-01 1.61329460e+00
4.25607413e-01 -1.87493950e-01 -4.25399333e-01 -5.41127443e-01
8.52839053e-02 3.87697548e-01 -7.28435397e-01 -2.79381990e-01
-7.81761825e-01 -1.25876009e+00 3.39777172e-01 4.36728537e-01
6.35809422e-01 -1.78205955e+00 -1.35391295e+00 8.26736271e-01
-2.45032653e-01 -7.77930140e-01 -6.27477765e-01 3.47279787e-01
-5.31326234e-01 -1.12429833e+00 -3.53161395e-01 -4.90527391e-01
2.68296570e-01 2.28651240e-01 1.38393939e+00 2.53929406e-01
-1.32981245e-03 5.53268731e-01 -5.52910447e-01 -1.03067517e-01
-2.23748654e-01 1.52870923e-01 4.38351512e-01 -1.18335046e-01
1.01622567e-01 -9.16002095e-01 -5.39941490e-01 4.10537928e-01
-7.71866441e-01 -2.44823098e-02 9.07939494e-01 1.26245749e+00
5.86683869e-01 2.77050436e-01 1.23721051e+00 -6.73804283e-01
6.42130256e-01 -7.83631325e-01 -6.01607800e-01 2.72286654e-01
-5.92619658e-01 4.69315350e-02 6.07034266e-01 -9.49013770e-01
-9.10587728e-01 -3.98359537e-01 3.20306510e-01 -2.14358225e-01
8.37417468e-02 6.51657164e-01 1.02942199e-01 3.12446117e-01
3.22385460e-01 3.96364480e-01 6.74151257e-02 8.93190317e-03
1.20039783e-01 2.33705521e-01 4.51428294e-01 -1.11421227e+00
4.54607695e-01 2.99842507e-01 7.55853504e-02 -4.90836769e-01
-7.28473365e-01 1.29156671e-02 -4.88371663e-02 -4.05052602e-01
3.25420052e-01 -9.50867712e-01 -1.28446436e+00 8.47759247e-01
-8.72404575e-01 -1.21724463e+00 -3.60221893e-01 4.83789861e-01
-1.02137101e+00 1.72475412e-01 -8.81970227e-01 -8.57426703e-01
6.14386126e-02 -1.05211294e+00 4.79387611e-01 4.03780222e-01
-4.12588716e-02 -1.07581151e+00 7.09318936e-01 1.89009160e-01
2.66946822e-01 2.95475096e-01 6.69501424e-01 -5.35748661e-01
-6.40690804e-01 4.79578227e-01 3.59983593e-01 -2.84076750e-01
1.90149993e-02 -5.25971591e-01 -4.55724776e-01 -6.83522224e-01
-3.26804638e-01 -7.63100326e-01 6.63953304e-01 3.90623271e-01
5.83306491e-01 -5.98793566e-01 -2.12304801e-01 6.71794489e-02
1.11890197e+00 4.41279739e-01 3.61258775e-01 8.24421346e-01
2.10899234e-01 2.26615712e-01 1.13356400e+00 1.27652335e+00
8.50775838e-01 7.60879576e-01 7.86621571e-01 3.22116286e-01
1.87896907e-01 -1.66433275e-01 8.11332941e-01 6.56458676e-01
-2.21011952e-01 -3.88381839e-01 -7.00962186e-01 6.98030889e-01
-2.56767130e+00 -1.37249410e+00 2.61842102e-01 2.03433919e+00
1.09078300e+00 1.67858139e-01 4.97106135e-01 -2.29855463e-01
4.40726936e-01 3.94788027e-01 -1.02973545e+00 -2.34543249e-01
-2.43699506e-01 5.97282834e-02 4.05224591e-01 5.88534713e-01
-1.02444696e+00 9.50811863e-01 5.53325129e+00 7.56958008e-01
-7.32289433e-01 4.74053085e-01 1.46921590e-01 -4.66605812e-01
7.36725852e-02 -2.98159365e-02 -4.73746598e-01 6.91702306e-01
9.62009370e-01 -1.90873131e-01 9.17424381e-01 6.26187682e-01
3.38649094e-01 -4.52828735e-01 -1.04245448e+00 7.03414381e-01
-1.86670512e-01 -1.10976040e+00 -3.33409518e-01 1.54575557e-01
9.18234766e-01 2.87874907e-01 6.28578812e-02 6.80122137e-01
1.20524693e+00 -9.45907712e-01 8.21850479e-01 5.11412859e-01
2.41995245e-01 -1.10389054e+00 4.68302876e-01 5.40319383e-01
-1.25751901e+00 -4.37938750e-01 -2.03061432e-01 -6.55747294e-01
3.04363787e-01 3.64965826e-01 -4.26242858e-01 5.43542743e-01
4.61307526e-01 8.67824614e-01 -4.87350523e-02 8.54174674e-01
-1.68180570e-01 4.23792899e-01 -1.22700647e-01 -8.79155248e-02
4.80511129e-01 -2.06598967e-01 5.83589315e-01 6.76185131e-01
7.38482550e-02 2.45464861e-01 7.76422918e-01 4.00412112e-01
6.24940991e-02 -4.85960543e-01 -4.42794532e-01 3.89732420e-02
9.61376488e-01 9.50732112e-01 -6.81462407e-01 -2.51210868e-01
-3.14961582e-01 8.96699965e-01 5.09042501e-01 4.06599671e-01
-1.07932222e+00 -9.90391597e-02 8.94490361e-01 -2.67595470e-01
2.50247806e-01 -3.11383367e-01 2.02079698e-01 -1.06660020e+00
-9.66826752e-02 -1.26705694e+00 5.61250925e-01 -3.61950606e-01
-1.47353625e+00 2.76413143e-01 -2.30243608e-01 -1.15171540e+00
-6.89863920e-01 -3.41746211e-03 -7.23186791e-01 4.91161719e-02
-1.64288795e+00 -1.04621100e+00 1.44681647e-01 7.41742849e-01
7.00052261e-01 -3.17219973e-01 7.63914168e-01 3.14247847e-01
-7.22317398e-01 3.69146198e-01 2.94976234e-01 -2.17911601e-01
6.33711219e-01 -1.24226415e+00 -2.04936154e-02 7.19033122e-01
6.75370023e-02 6.90244734e-02 7.26344943e-01 -8.41165423e-01
-1.42877626e+00 -1.18313670e+00 3.66389662e-01 -2.62973011e-01
8.28848541e-01 8.48046038e-03 -7.41121292e-01 8.59008908e-01
3.22151840e-01 -9.04115513e-02 5.80558598e-01 1.40291423e-01
-1.19617276e-01 7.38701895e-02 -7.90486991e-01 6.59652472e-01
9.59706545e-01 -2.55326271e-01 -6.68269038e-01 3.30751598e-01
3.65853578e-01 -3.55572939e-01 -7.67428577e-01 -1.65771623e-03
3.72485101e-01 -8.15669656e-01 8.89714539e-01 -7.62376428e-01
3.72883588e-01 -2.96609730e-01 -2.23599210e-01 -1.74697125e+00
-2.83075809e-01 -9.30028319e-01 -5.60665667e-01 1.02442849e+00
4.42062207e-02 -6.50995076e-01 5.33024251e-01 3.10383558e-01
-2.24744320e-01 -6.12933159e-01 -1.15727770e+00 -1.21050239e+00
3.63553651e-02 -1.08160749e-01 6.24681175e-01 9.81405020e-01
1.32800892e-01 4.90980931e-02 -1.00147605e+00 2.95414478e-01
9.45096850e-01 4.41486448e-01 6.96520329e-01 -9.32459712e-01
-9.28729355e-01 -3.95334721e-01 1.73084661e-01 -8.27783763e-01
6.02100849e-01 -5.39682388e-01 2.95107692e-01 -1.27964878e+00
4.64845985e-01 -7.60626197e-01 -4.80886549e-01 7.71728754e-01
-7.98352510e-02 2.95662172e-02 3.68942678e-01 3.41508448e-01
-1.47014713e+00 9.09364939e-01 1.04955578e+00 -1.22804545e-01
-4.11830634e-01 4.03432269e-03 -2.36709937e-01 5.95774174e-01
9.34898555e-01 -7.07290351e-01 -4.13527519e-01 -3.58917922e-01
4.21081483e-01 4.62440968e-01 3.95952642e-01 -9.32788432e-01
3.44791293e-01 -9.88158941e-01 -2.67684817e-01 -2.18931571e-01
2.87330657e-01 -4.11430389e-01 2.64077991e-01 7.88710535e-01
-3.89048457e-01 4.08943206e-01 -1.42867506e-01 8.70653808e-01
-5.17777279e-02 -9.27023366e-02 5.57682991e-01 -2.94872284e-01
-8.53859365e-01 3.35567594e-01 -8.26607287e-01 4.07511473e-01
1.27691770e+00 2.47152969e-01 -5.81899643e-01 -4.23223287e-01
-7.39318132e-01 9.43650782e-01 3.33431005e-01 2.52166003e-01
5.10209799e-01 -1.25878787e+00 -6.60427928e-01 -2.24677235e-01
-1.49737060e-01 -6.11892976e-02 5.83143234e-01 1.01919746e+00
1.03653021e-01 -7.32497722e-02 -5.80001354e-01 -1.98674649e-01
-9.92627442e-01 5.30792415e-01 3.90084058e-01 -8.43193054e-01
-6.80538774e-01 3.54592144e-01 2.44799517e-02 -3.39876294e-01
2.41691157e-01 1.45621791e-01 -1.69714447e-02 1.85410067e-01
5.34139395e-01 6.31615698e-01 -2.62215942e-01 -1.76088139e-01
-2.62517005e-01 7.14731514e-02 -2.49786377e-01 -4.24092263e-01
1.67446268e+00 -2.22442210e-01 7.50987530e-02 6.05272651e-01
1.05638437e-01 -3.23417515e-01 -1.89270186e+00 -3.93023610e-01
-1.33670811e-02 -2.03528464e-01 -9.14329216e-02 -9.68237102e-01
-9.81296301e-01 3.85211080e-01 1.58787400e-01 1.49571717e-01
9.58723724e-01 -1.02344207e-01 9.44198132e-01 6.55537486e-01
1.03392100e+00 -1.64064002e+00 6.82472110e-01 8.15276980e-01
5.86008847e-01 -1.15910089e+00 -1.50097653e-01 4.84156817e-01
-9.30761993e-01 6.04546785e-01 8.60314488e-01 -1.20060094e-01
2.71032184e-01 2.27075085e-01 -2.70660698e-01 -1.90285072e-02
-1.45274186e+00 -3.24690908e-01 -6.25963449e-01 7.10702837e-01
-3.94324362e-01 4.65204149e-01 -8.97851735e-02 6.87950671e-01
1.68855339e-01 -3.37785929e-02 9.40993190e-01 1.37527120e+00
-4.89847779e-01 -1.26013970e+00 -2.39354402e-01 3.84135723e-01
-2.50153005e-01 6.54945821e-02 3.11412588e-02 7.44287014e-01
-4.13319543e-02 9.40138340e-01 2.31197387e-01 -6.93260599e-03
4.65056449e-02 -1.82492301e-01 5.64853072e-01 -3.69623542e-01
-7.18339384e-01 -5.46766259e-03 1.75415669e-02 -6.31404281e-01
-6.03430986e-01 -9.26667869e-01 -1.58057880e+00 -6.85754299e-01
2.08991200e-01 2.94775128e-01 -9.71258655e-02 1.15210652e+00
5.35968781e-01 7.52890527e-01 8.05863738e-01 -8.84752452e-01
-8.01054537e-01 -8.22068989e-01 -8.02921653e-01 2.59282798e-01
5.49311817e-01 -1.19516444e+00 -1.94838002e-01 -2.81324148e-01] | [3.7826409339904785, 2.007352352142334] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.